Sunday, December 31, 2023

Electrostatic speakers and Bass EQ

This is two separate topics I was thinking about yesterday.

It seems like the good channel of my newest Acoustat 2+2's sounds louder than my older 2+2's, and this is true regardless of interface used.  Substituting a plain vanilla C mod interface makes it sound more like my older C modded 2+2, but still not identical.

(Note: I have not tested speakers in identical or equivalent positions yet.  The speakers are very hard to move and I may not even get around to such rigor until I've removed the two 'bad' speakers from the room, one from the older pair and one from the newer pair.  Then I will have a pair of speakers, one old and one new, in equivalent positions ("left" and "right") though not actually identical, and I will be forced to level match and EQ match them for good stereo, and only THEN will I be able to state with better authority how the old and new speakers differ.)

I'm thinking electrostatics can age in various ways:

1) The HV can droop.  I wouldn't have thought this to be a problem because all my interfaces (except my original 1+1's which may be the best sounding of all) have been refurbished by knowledgeable people.

2) The membrane can stretch.  I was previously thinking this would only be a problem if there were actual 'slapping,' otherwise the panel would just move to where it was supposed to regardless of stretching.  BUT that was wrong, in fact the resistance of the panel to movement is a critical part of its operation.  Stretched membranes can cause bad resonances and intermodulation.

3) The insulation on the stators (if applicable, as with Acoustats) can weaken and break.  In serious cases you get periodic arcing (as with the bad new speaker) and then the speaker develops level dependent distortion.  But there could be distortion from mere 'weak' spots that weren't arcing as such.

Every panel is also going to vary from every other panel in minor construction details that result in slight resonances at different slightly different points.  (Though, by and large, I'd expect most brand new speakers to be almost the same, but diverging with aging.)

ALL speaker units differ, even if made by the same manufacturer on the same day in sequential serial numbers.  However, you'd expect that the closer you get to being on the same day in sequential numbers, the closer they are likely to be.  So units made years apart in different factories are likely to be different than their mates made at the same time.

I'm thinking because of how large electrostatic speakers effectively have been hand-made they probably differ more in their tiny resonances and such more than factory made dynamic drivers.   OTOH the electrostatics generally have (less so nowadays) less distortion and resonances to begin with.

Funny I've never read anything from Linkwitz about using electrostatic speakers.  He made dipolar speakers using dynamic drivers.   So I don't know what he thought about electrostats.  But his methodology for picking out the better drivers was ultimately quite simple.  You have to listen to them.  Only then can you tell if the particular set of defects every speaker has is problematic or not.

For rigorous testing, playing one speaker by itself on a mono signal is the most revealing.  (Linkwitz also may have recommended female voice, which I hardly ever listen to.)

And so I've been doing, with the good (non-arcing) unit for a few days now.  Playing just that one speaker.  And this convinces me that although it may sound different from my older pair, it's still very good if not significantly better.  It was the arcing speaker that made the new pair fatiguing.

The new good speaker just seems to have remarkable clarity and musicality.  I'm afraid I might have to 'dumb it down' with EQ to make it match the other one, but it will probably still sound better, more headroom, etc.

I'm thinking this may be because it has a tighter membrane.  Or maybe other construction details varied, as this newer speaker likely was made in a different factory.  However it doesn't have the '5 wire' connection that would indicate the new kind improved wire insulation, if I'm understanding that correctly.

*****

I've just made another tweak to my 'background' EQ.  (I lost my original 'background' EQ when my chairside EQ unit died, so perhaps I'm just 'recovering' what I had previously done, which I had to guess at first.)

This proved necessary when I was listening to some bass records including 'Bass Erotica' which was almost unlistenable before making them, and afterwards could (for the first time) be cranked way up and enjoyed much more, with the bass lines becoming easy to follow.

I'm now cutting 50 Hz by 5dB, and the flanking frequencies of 40 and 63 by 4.5.

I believe is needed because of the room modes around 45 and 52 Hz.  In the center listening position, these modes don't get amplified (in fact, they get attenuated) so I don't cut them in the primary EQ (as when I last did it, I was running the oscillator while sitting in the listening chair..previously I had often done tuning while listening near the DEQ, which is more like the room boundary).  So they have to be cut in the background EQ now.

Which brings to mind how many EQ systems deliberately have multiple presets.  I am now thinking it is a requirement that if you use EQ at all, and if you listen to background music on your main system at all, you must have separate EQ's for serious listening at the listening position and casual listening everywhere else.

I need this in particular because my listening position is anti-modal.  Near the center of the room, the larger modes cause cancellations instead of augmentations (they augment at the peripheries).

I choose to listen in the center of the room anyway so I can be as close to the speakers as required for the widest possible stereo separation.  That opens everything up in an incredible way.  A friend clued me into this.  I was previously a back-of-the-room listener, with speakers in front, and that felt 'natural.'

I could in principle achieve similar separation with the speakers in the middle of the room, and listening in the back.  From the standpoint of getting good bass (and especially getting the best possible bass out of electrostatics) that is far better.  Then I would not be listening from an anti-modal position, and I might even get away with no EQ and no subwoofer too.  (Though any time a subwoofer is being used, you are almost certain to need EQ anyway simply because of that, subwoofers excite room modes in ways that panel speakers avoid.  But if the panel speakers were in the center of the room, you might get away with only one EQ for both serious and background listening.)

But that just doesn't fit my home and my lifestyle.  I do not have a 'dedicated' listening room.  I have a living room I walk through about 100 times a day, and have movie parties in 1-3 times a month.  Speakers in the middle of the room would mess everything up, and Acoustat 2+2's are not very easy to move around.

Then people tell me I could simply "enlarge" my living room.  Geez.  I have enough trouble keeping up with home issues as it is.  Last year I spent $21,000 on foundation repair and related issues.  It was because of the fear of issues like that I built an entirely separate climate controlled storage building with its own very heavy duty foundation ten years ago.  I didn't want to tack more rooms onto my already struggling house.  (I had bigger plans for the storage building, but in the end it became just a storage building, because I needed that.)  As serious of an audiophile as I am, the changes needed to enlarge my living room are basically unthinkable.  And I have other major home improvements already long in the queue, such as a new patio cover and carport.

Ultimately I think it's not bad at all for me to figure out how to fix these sorts of problems either.

Anyway, with these changes, it seems I can play anything at any level and enjoy it much more in the background now.

 

Monday, December 25, 2023

The Grinch Lost

 Since October, the Grinch has visited me many times.

Upon returning from vacation on October 31, I discovered the DEQ for the supertweeters had become dysfunctional.  Attempting to bypass, I then discovered the entire right channel of the supertweeter had died--they were shorted.  Some kind of catastrophic failure I hadn't noticed.  It appears the amplifier is still OK.

I managed to get a replacement for the nearly unobtainable D21AF right away.  Suprisingly, it has been much harder to replace the Vifa tweeter in back.

But then, I discovered that merely moving the supertweeter towers out of the center made the imaging far better.  It turns out there's a "central" rule: don't put stuff in between dipolar speakers.

So, the grinch lost that one.  While it may take a year (I've determined not to hurry it) to get a new supertweeter system in place again, it will be far different.  It will not be on a tower in between the speakers.  My current plan is to have the supertweeters behind the Acoustats, just above "interface" level.  More recent testing has revealed the Acoustats are nearly acoustically transparent, so little harm in putting the supertweeter behind and then compensating for the time difference digitally, as I always do.

The grinch has really determined to mess up my audio life in the past couple days.

In the previous week I figured out to hook up my modified QSC ABX switch box to switch speakers instead of amplifiers.  That seems to work fine (but I need to inspect if the amplifier is ever "shorted").  Only then was I able to nail down the difference in sound between my newest 2+2's and my original pair, with instantanous switching.  Then I learned a bunch of things:

1) Instantaneous switching is near essential for making audio comparisons

2) The new speakers sounded way louder, but measured softer.  They could sound more dynamic without any level compensation (and which way???) but had a glare that became objectionable quickly.  Possibly lowering HF level and changing to C mod would help.

I then started working on lowering the level.  It took sanding the variable resistance bar in the interface in the left channel.  I had disconnected the speaker wires during this time.  But the first time I plugged the speaker wires in it didn't work, and it took several more attempts of sanding to get it to work.  I also plugged the speaker in while it was connected to the amplifier (which I now think could be a no-no).

I had the channels reversed and messed with fixing that.  Somehow in the process of modifying the speaker and reversing the (complicated extended and automatically switched) connections I managed to kill my Hafler 9300.  One channel first became increasingly noisy (which at first I thought was the modified speaker, or the connections) and ultimately it went out completely.  It took some testing before I realized it was the 9300 that had failed, it was not a matter of the (now very complicated) speaker connections.

I'm now worried that it was actually the new 2+2 on the right side, which makes a very soft arcing sound about once per second, that had killed the amplifier.  Perhaps especially when I plugged in the speaker.

I can't blame the new speaker for certain, but I now think it's a risk not worth taking.  I'm not going to play that speaker with a non-expendable amplifier until the arcing has been fixed.

So the grinch blew up my favorite amplifier, killed my new pair of speakers, and killed my new speaker testing plans and even my imagined post-testing scenario: how I first planned to use the new Acoustats in bedroom if I couldn't get them to sound as good as the older ones (which of course had been the original hope).

Pretty bad.

But it gave a renewed purpose in life to my Aragon 8008 BB amplifier that's been waiting there in the living room all this time, as my backup and "alternative" amplifier to test to see how audible amplifier differences are (try as I might, in blind A/B testing, I've never been able to consistently identify which is which...which right now comes in handy...because I know I'm not 'missing' anything...I don't have to pine over the now dysfunctional Hafler).

In fact, I had spent the last month or more pondering the sale of the Aragon amplifier.  Though it's clearly the 'highest end' working amplifier I own, I just didn't need it.  If anything, I thought the Hafler sounded slightly better overall (though not being able to prove I could hear this difference).  My sighted listening impression is that the Hafler sounds 'sweeter' (less distortion) whereas the Aragon sounds 'punchier.'  

I didn't need a backup amplifier, I was thinking, because the Hafler (unlike the Krell FPB 300 before it) seemed totally reliable.  I'd never had one of my two Hafler 9300's fail.  I thought they were virtually indistructable.

And now that I'm leaving the Aragon on all day, I'm thinking I might like it better.  It's got at least twice as much power and current handling capability (it has two 1100VA toroidal transformers, that's a total of 2200VA--too much for one Insteon switch to handle btw).  I'm feeling the extra bass punch and solidity, exactly what I'd been hoping to get from the newer speakers (and sort of did, but also with listening fatigue).

Both amplifiers have impressively wide bandwidth, I know it's 300kHz for the Hafler, I thought the Aragon was much less than that but in fact Secrets of Hi Fi and Home Theater reported 550kHz for the Aragon.  There is no indication of any early rolloff in the frequency response shown by Stereophile for the 8008 (regular model), though there is a slow rise in the distortion floor above 1kHz, nevertheless the amp exceeds it's 0.07% THD spec.  The Hafler does have lower distortion, but it hardly matters at these levels.  The Aragon gets high praise in all the High End Audiophile magazines, even The Absolute Sound, meanwhile the Hafler got high praise in Audio magazine (it SHOULD have been noticed more IMO, it was a sleeper).  The Aragon has the basic "good" design standard by the 1990's for bipolar amplifiers: complementary symmetry DC coupled with servo, not unlike the GAS Ampzilla of 1973, but with much better parts down to the teflon coated wire and high end circuit boards.  And massive heatsinks so it doesn't need a fan.  And the high bandwidth plastic transistors were a special limited edition from Toshiba before with extra fat die for elevated ruggedness on top of high linearity.  They ran out of replacements in the 90's, all you can get now are the more conventional variety.  Mine was in the earliest series that was contract built by a manufacturer in Connecticut that really knew their stuff, the circuit boards and everything look instrumentation grade.  Later Aragon had their own factory in California.  Then they were absorbed by Klipsch, then spun off a separate company again today.  Always in the niche right below the insanely high end.  Back in the 80's they were distributed by the same people as Krell, perhaps that's how the rumor got stated that Dan D'Agostino himself designed the basic 8008 series of amplifiers.  If there's any truth to that, it would have to be the cocktail napkin kind of design, where D'Agostino sketched out the basic design, and an engineer named Robbii did the detailed design work, his name is on the circuit boards.  I think it's more or less textbook as far as the basic amplification circuit, except the bias is probably simpler than the textbook kind which I think would be more regulated.  I don't think it's as finely tuned as something by D'Agostino, Curl, Pass, or another of the amplifier designer legends might have created.  But if you're not cutting corners, by the 90's it was fairly easy to do a top notch Class AB+ amplifier that would blow everything from the 70's away.  So this is the "value" high end amplifier with over-the-top high end parts and build but without the fancy designer label (though nowadays Aragon is plenty fancy).  And that's good enough, it doesn't take a one in 10 million genius to design a good audio amplifier, pretty much all you have to do is copy the one of the transistor manufacturer design examples, with a bit of fine tuning.  And if you're not pushing things to the limit, it's probably going to last longer too.

I met the Mondial (Acurus and Aragon) co-founder named Tony.  He showed me the Aragon in a back room of an audio meeting (after the Meitner amps had failed spectacularly in an audio meeting playing Apogees...they weren't supposed to be doing that and the store owner was pissed).  Tony bragged about how could stack stuff on the amp without issues and he even "warmed it up" the sound by putting a blanket over it for 15 minutes.  He was absolutely fearless about it, but I was scared to leave it covered that long so I took the blanket off a bit earlier, but indeed the sound was already warmer.  I've discovered the bias supply is very "flexible."  As the amp warms up, the bias goes even higher.  But if you put the amp on tall feet, as I did, you then give it "too much" airflow and the amp cools down and the bias goes down too.  So with tall feet I had to re-adjust the bias higher.  I've set it so that with the tall feet and nothing on top it runs at a perfect 125F after one hour where it measures and sounds great.  Balancing the heat of the two channels is essential too, otherwise they "cycle" pushing each other up and down.  There is incredible thermal mass and it takes a long time to heat up or change.  And of course if it gets too hot, it shuts down, or at least is supposed to, well before anything breaks.

I guess you could say this is "old school biasing" before Nelson Pass patented 'optical bias' went mainstream (btw, it's what made most later era Krell amps with plateau biasing possible, and I'm pretty sure it's used in Levinson and most other high end class AB amps).  Old school biasing has it's "advantages" in being kind of self-limiting, but it really needs to be adjusted for your exact situation to perform perfectly.  But then it does really well.  It's fairly easy to access the Aragon controls but not easy to get them set just right, to avoid cycling, etc, and depending on the height of the feet under the unit.

The Mondial founders were not themselves engineers, they were audio marketing guys who felt there was a missing niche just below the very high end they could sell along with other brands (which originally included Krell, which was sold by the same distribution chain) and not compete with them, by having more basic designs.

I've owned the Aragon 8008 BB longer than any other amplifier without any failures, and the way it's conservatively designed and built you'd expect that (but often don't get as much durability from the more prestige brands, which are often more like race cars needing pit stops after every lap).  So on top of everything else, the Aragon is a survivor.  At this point in my life, my system is all survivors, and I honor survivors.  (Some things survive because it's easy for me to get them fixed, others like the Aragon just keep on working, and that's good because fixing it could be somewhere between very expensive and impossible.)

I know really don't have to continue the speaker test in earnest, which in a way is a kind of relief.  The plans I had for moving everything in the house around to make way for another major pair of speakers are for the time unimportant.  Now I merely need to plan for the somewhat simpler storage of all the new speakers, and keeping the arcing pair out in an accessible way so I can work on it.

So here are my wins:

1) My most high end working amplifier has renewed its purpose, and I'm liking it better than ever.

2) I don't have to sweat all the speaker moves I was planning.

3) I get to tear down an electrostat speaker with no worries since it's already broken (unlike the clocks I took apart as a kid and into my teens).

4) I've learned to love my original pair of Acoustat 2+2's better than ever before.  They've stood high even against another pair of 2+2's.  And it's clear my slightly attenuated HF helps.  The default position sounds way too bright and becomes fatiguing.

5) Failure after failure, and my living room stereo keeps working and sounding better all the time despite them.

6) I'm finding that indeed the higher power of the Aragon 8008 BB makes it possible to play louder than I felt like doing before, and then it's magic.  To really appreciate this at first it may help to be drunk, as I was on the evening of December 25.  I had to move the Mapleshade damper back on to the Hafler to stop the rattling.  And the right channel of even my older 2+2's has a rattle at louder levels, even with my 8th order linear phase crossover.  Eventually I'm going to have to rebuild that too...  (I was hoping to THIS time around, but the new pair aren't a suitable replacement yet.)

I plan to continue to test the one new 2+2 that is not arcing, to see how similar I can get it to sound to my originals, etc, so I can use at least that one as a backup unit for now.

I not only fear that an arcing speaker could destroy the amplifier driving it (by sending spikes of HV back to the amplifier) it probably cannot sound as good, because the uniform charge of the panels is disrupted, causing distortion.  So not really worth further sonic testing.  Once it's arcing, it's over.  I should have known that.




Friday, December 22, 2023

My B mod and C mod Acoustats

I am still struggling to define the difference between my two sets of Acoustat 2+2's.

It's certainly not just a matter of B mod vs C mod.  It's also that the level control was adjusted downward on my first 2+2's.  I adjusted it that way thinking it was flat, but it was actually 2-3dB down perhaps.  And I also liked it that way, in fact EQ'ing down a bit more.  Though it might be better flattening out above 12kHz rather than continuing on a downward slope.

But I haven't adjusted the newer-to-me B mod speakers to that same level, because it's a bit harder to do, you have to open the interface case and fool with a strap (which requires deoxit cleaning).  I'm now very certain I like the old knob level control better, but there should be terminals you could measure for exact settings.  (Actually, in the older interface style, like my first speakers, with the original A mod in them, you could measure the DC resistance across the input terminals to set the HF level repeatably.  But since mine have C mod, I can't do that, and have to break the case apart--difficult in the older interface style--to make measurements, which I'd been planning to do since forever.)

Also, I wasn't even sure whether I liked the greater HF output or not.  It looks like it might even be objectively an improvement.  And sometimes I've thought it was.  But I'm leaning the other way right now.

But there could be still other factors.

What's really mind boggling is that the newer-to-me speakers (B mod, higher HF level) sound a lot louder.  So I adjusted the level downward, finding the matching point to be about 2.5dB.  After making that change, a lot of the added "punch" of the newer-to-me speakers went away.  But they seemed more similar if still different.

Anyway, objective level measurements don't back this up at all.  In fact the C-weighted pink noise of the newer-to-me speakers doesn't measure louder...it measures 2-3dB softer!  Changing the weighting to pink noise brought the two speakers much closer, but with the newer-to-me speakers still about 0.5dB softer (about the margin of error).

Note that this is still with the newer-to-me speakers in front of the others, which seems to make surprisingly little difference.

The 'louder sound' of the newer to me speakers was especially apparent to me with pink noise at a high level.  At levels I used to think OK the newer speakers sound overwhelming, whereas the older ones sound slighly distant and unfrightening.  (And this is with the C weighted noise of the more distant sounding speakers actually being 2-3dB louder!)

I'm still finding the 'harder' sound of the newer speakers to be sometimes bothersome.

So I don't know what the issue is.  Is it that the B mod has more distortion, making it 'sound louder' ?  Or is it that some of the old-to-me speaker panels have weakened?  Or is it just the HF level difference?

Should I swap my old modified C mod interfaces in, and with or without adjusting HF level to -2dB?

Should I just reset the levels in the newer interfaces to -2dB (which isolates the level change)?



Sunday, December 10, 2023

Acoustat "C" mod

 The Acoustat "C" mod  (as found in the MK 121-C interfaces) is always strongly recommended by Acoustat gurus such as Andy Szabo (who has answered questions about Acoustat for over 10 years at DIYAudio) and Roy Esposito (who rebuilds transformer interfaces).  They are both former Acoustat employees (the Acoustat company does not exist anymore).  "C" mod can easily be applied to any previous Acoustat transformer interface.  Though the factory "C" version included the Medallion transformers also included with "B" version, the Medallion transformers are not necessary to perform the mod.


Acoustat A and B vs C mod

(Ignore the green arrows, which are intended to show the backward secondary coupling that is alleged by one DIYAudio poster to cause saturation in the HF transformer of the original design.  Others dispute that.  I wonder why saturation would not also be caused directly by the direct coupled primary circuit of the original design.)

"C" mod was the last iteration of interface designs before the significantly different Spectra series of Acoustat speakers were introduced.  So "C" mod was the last factory design for Acoustat models such as 1+1 and 2+2.

One of the reasons it is recommended is that it attenuates the low frequencies being applied to the HF transformer, reducing distortion. 

At first glance, it seems very strange that in the original MK-121 interface there is no actual "crossover" on the primary side of the two Acoustat transformers.   With both the "A" and "B" mod interface designs, the HF transformer is DC coupled to the input signal, with only a capacitor bypassed series resister in line (whose purpose is tuning the high frequency peak around 10kHz, though it seems weird such a large capacitor and small resistor would have that effect, when their own RC constant suggests they work in the midbass, but their effect is just enough to shift the ultimate HF pole).

The actual "crossover" for the HF transformer is only on the secondary side, a 0.01uF capacitor coupling the HF transformer output to the panel drive signal.  The LF transformer has no "crossover" at all except it's own inductance rolling off the high frequency response, which was the reason why the Acoustat dual transformer interface was invented in the first place.

Is there a difference between crossing over the transformer on the primary or secondary side?  I'm not sure, I would have never thought about crossing over a coupling transformer on the secondary side, that it even works OK wouldn't have been obvious to me.  It would have seemed to me that even with the secondary wide open, the transformer primary would operate like a choke, and would still be affected by potential bass saturation.  (However, chokes do less and less at low frequencies anyway.)

Comparing that capacitor with the 50k resistor coupling the LF transformer, a first approximation of the highpass cutoff of the HF circuit is (2pi/RC) 318 Hz.

I can see a one pole at 318 Hz might be a bit inadequate for fully isolating the HF transformer from low frequencies.  If the C mod HF control were turned all the way up, it would introduce a second cutoff caused by the 57uF capacitor being loaded by 16 ohms, around 174 Hz.  Additional loading from the transformer itself and turning the control to midpoint, push this cutoff somewhat lower than that.

Further protecting the HF transformer from low frequencies seems like a good idea (though it's not clear how much of a problem it was).

But at what cost?

People smarter than me, or at least more familiar with SPICE modeling, have analyzed the B and C mod circuits.  And the incredibly curious thing is despite how they look entirely different, and wouldn't work the same at all, in fact they work very much the same.  There is relatively small difference between the interface outputs with B or C mod in place.  Which is as it should be, presuming B mod was reasonably well designed in the first place.


But there is some difference, the two very different circuits obviously do not work exactly the same.  I would have expected more rolloff in the deepest bass, but in fact the C mod might have slightly higher output, like 0.1dB or so, at 20-40 Hz.  But it's in the midbass through the midrange, 100-1000 Hz, that the C mod has up to 2dB less output.  The absolute advertised gain for the C mod in the extreme highs around 10kHz and above is very very slight, it's only relative to the 2dB attenuated output at 1kHz that it's "relatively" significant.

If you played these two versions in a simple test without adjusting the levels to compensate for the greater loudness of the B mod,  I'd bet that almost all of the time audiophiles would prefer the louder seeming "B" mod.  It might still win in a properly level matched (best to use C-weighted pink noise) comparison because 100-1000 Hz is basically the heart of the music.  The relative and tiny absolute gain in the very extreme highs above 10kHz is debateable, I've had an ear tuned notches tamping 2200, 5200 and 9100 for C mod actually.  It certainly has more 'punch' if you don't adjust the level downwards (when switching from C mod to B mod) to compensate (I've found a 2.5dB downward adjustment in gain for B mod (vs previous C mod) for the Acoustats to work best by ear, eliminating the 'listening fatigue' I get when it's set higher, relative to the subs, but after you compensate for the difference in output with a 2.5dB downward adjustment...I don't yet if the 'punch' is different, I think it is slightly but not as much as I'd hoped when I first heard it, the increase in punch was shocking, but I hadn't realized I was playing 2.5dB louder 100-1000 Hz either.  (Plus, I'd always set the HF level down before by accident, not realizing "flat" was 3 o'clock instead of 12 o'clock.  So it's louder in the highs not, despite not being a difference because of the change to B mod in the my most recent 2+2's but instead because so far I've left the newer speaker at it's factory "0dB" setting, which is tricker to change then with my first 2+2's where I could just turn a knob, but now I'm thinking I like more highs anyway, so I'm not planning to adjust the new but rather to measure how the HF level control on the older one's works, and how much and whether I still need the EQ's with the newer speaker for it to sound best.  So, added highs is an additional factor possibly adding punch.)

In fact, a number of Acoustat users and even some self-appointed Acoustat gurus do prefer the "B" mod, a fact I only discovered yesterday (note: the figures in this post come from that thread).

I'd generally found the C mod midbass through midrange to be somewhat weak, but I've never broadly EQ'd that upwards,  I've only notched out a few peculiar resonances to make it more smooth, and curiously added two tiny 1/3 octave 3dB boosts at 850 and 1kHz, where there was a curious suckout, I determined when I had the chairside EQ (in the repair queue for about a year now).  But I had been disturbed by the general 'weakness' of that region.  Well, going back to the B mod fixes that general 'weakness' from 100-1000 Hz, or at least the new speakers do.

The significant loudness increase with my newest 2+2's (which have B mod) might be entirely due to B mod!  Which I had previously assumed was a step backwards.  But there might be other reasons too.

With more midrange, midbass, and deep bass, the B mod might have "punchier" sound as well.

The only noticeable lack in the B mod from the measurements is a very slight extension in ultimate high frequency response, which looks pretty small and fairly unimportant to me.

I'm not going to assume C mod is necessarily better any more.  I now have the ability to swap B and C mod interfaces into my newest 2+2's and I will measure them and listen to them and see.

Note that while the B mod will play as much as 1.5dB louder than C mod with the same input signal, the ultimate loudness limit (provided sufficient amplifier power is available, < 200 watts into 4 ohms) will likely be about the same.  The transformer and panel limitations will remain the same.

DIYAudio poster Bolserst posted some more illuminating simulations and measurements to the aforementioned thread.

First, he shows B mod mixer drives the HF transformer with the full bass signal, but the C mod mixer rolls off the bass to the HF transformer (this doesn't seem to show the effect of the capacitor in the secondary circuit however).

In later graphs, he shows these spice models do fit the measurements almost exactly.  Including the curious lump in the LF transformer response, which goes away in the C mod.  It's actually that "lump" which causes the boost in B mod vs C mod output in the midrange.  The HF midrange response is identical!

There is no change to the LF circuit at all, so what explains this lump and how it goes away with C mod?  There are no changes in the LF circuit at all!  That brings us to Bolserst's theory that the C mod reduces the low frequency backward coupling of the LF transformer into the HF circuit, which somehow pulls the LF response higher.

Greater output in the midrange might make it sound better, on the other hand, if there's more distortion added, that could make it sound worse.  In my previous measurements the Acoustat (with C mod, all I had ever used before) is very low distortion, but perhaps distortion is slightly higher in B mod.

Since I've readjusted the level to -2.5dB from before, I haven't had any more episodes of listening fatigue, but I wouldn't be surprised if C mod was slightly cleaner.  So perhaps it will win in the end?  I'm going with my newer speakers anyway because I can switch from the current B mod to my original 1+1 speakers hot rodded (external Solen cap) interfaces, because my latest 2+2's have the compact interfaces which are quickly swappable.

That feature alone is important,  but for now I'm thinking they sound at least as good with B mod as my originals do with C mod, or possibly better.

 *****

Update December 23

No question anymore, the 'hardness' of the new-to-me 2+2 with a B mod interface is intolerable.  I don't know yet what the problem is, perhaps I just need to turn down the HF level comparable to my other 2+2's.  Perhaps the C mod does reduce audible distortion, and once you've gotten used to it nothing else will do.  I can make all these changes with equipment on hand, I would love to swap out the interfacees in the new-to-me 2+2's with my hot rodded 1+1 interfaces from back in the day, with huge external polypropylene cap and everything.  But to be fair, I'm going to take the first step and changed the new-to-me 2+2's to their lowest HF level and see what that does.

It occurs to me that the way the HF level control "works" is that it actually creates HF ringing.  The "flat" position is with a little ringing, and the +2 is with more ringing.  -2 is probably what you need for no ringing.

And the ringing probably is more audible with B mod having in effect higher Q.

Still, I imagine many would prefer the 'louder' sounding sound.

Update December 28

I should not have conflated the differences between my two sets of Acoustats with the difference between B and C mod.

It now appears the fatiguing quality may have been mostly that the arcing in the right speaker, though impossible to hear except with ear at speaker, was making it more distorted.  The electric charge was non-uniform.

Also the difference in HF level.  But now, very indirectly so far, it does seem like B mod may sound somewhat louder irrespective of HF differences.  (And also that in my room and to my taste, the HF level needs to be turned down at least 2dB.)

I will shortly be able to swap in my C mod interface, which will be the closest I've come to doing a fair test.  I still have the B mod speakers in front of the C mod, giving the B mod an artificial advantage one would think.

I've disconnected the arcing speaker, because it might have damaged my amplifier, and I can't risk my next amplifier, the amplifier I'm using now, which I'm liking more than anything ever right now, the Aragon 8008 BB, with Mapleshade carpet feet and my superlative bias adjustment.

Until I get that arcing fixed, I'm only going to play it with "expendable" amplifiers.  For now that speaker going back to the repair pool.

I might be first repairing my original right channel with C mod.  It also seems to have some distortion playing You Think Too Much About Flying Saucers, but, playing the same right channel signal on the new B mod 2+2, there is no distortion.

So it's looking very much like when and if I get them matched, I'm going to be mixing my new and old 2+2 pairs until I get at least one of the defective ones fixed.

And to really do that correctly, I'm going to pull my old modified C mod interface out, so both will have C mod, at least for starters.



Thursday, December 7, 2023

Acoustat 2+2 (updated Dec 8)

I've complicated my life by buying (at very low price) two more stereo sets of Acoustat speakers, 2+2's and 1+1's.  This means I now have 2 sets of each variety.

I don't actually intend to use these speakers in any current room, except that I might place 1 set of 2+2's (whichever I decide is least good) into the Laboratory, replacing the 1+1's already there.  After all, the 2+2's are better, etc.  Ultimately, the 2+2's will be the best speaker for the new Gym, after I expand my existing Gym to the full width of the garage after I build a carport.  (I'm not sure I will live to see all these imagined improvements anymore, however, because I haven't done anything in more than a decade of thinking about them, and I have less money now.)

My main hope was that the new 2+2's would be better than my pre-existing set, and replace them, leaving the older pair as backup or experimentation.

I'm not at all sure that is going to be the case yet.  In fact, one of the speakers has a notable rattle played full range.  But the same is true of one of my pre-existing ones.  The possibility exists that I will end up using one of my old ones and one of my new ones.  Then it will look pretty wierd, with a white 2+2 on one side and a black 2+2 on the other.

Anyway, my first month is chalked out to be a full examination of both sets of Acoustat 2+2's.  That is possibly the main reason I bought them, for testing and experimentation.  Having only 1 set, I am unable to conceive of minor repair and/or modification.  My system would be done for the duration.  Now I have a second set I can experiment with, including taking the grill cloth off.  (The grill cloths on the newer set is after-market, looser, and possibly easier to remove than on the originals.)

The newer set also has the "compact" interface unit which is far easier to remove...in fact you can normally leave it unscrewed and simply pop it off whenever.  Having the compact interface also means that I could use my already Solen-cap equipped original 1+1 set, to which I added a large external capacitor (not an easy modification!) to replace the original 47uF electrolytic.

My original set is the from the original production run, with the impressive looking "large" interface unit strapped to the speaker.  It was refurbished and modified to "C" mod by a famous modifier.  The new ones are both Blue Medallion.  They might also have been modified to C mod but I don't know yet.

Things I want to know are:

Which units are more plagued by low frequency rattles, resonance, and similar issues?  How hard do they have to be driven to exhibit these flaws?  (I intend to do only very safe non-destructive testing in these regards, but I already know several songs that are pretty likely to cause problems)

Are the new sets "C" mod?

How have I tuned the HF control on my original set and what difference does changing it make?  What position is actually best?  Is it better to achieve the HF balance I like with EQ or with treble adjustment?  (I am currently using EQ, I have the control set in the middle which I think means flat though it is not marked.)

How do the frequency responses of the original compare with the new?  Can they be adjusted to be the same, or whatever is best?  Does the C mod vs B mod make a difference in frequency response or sound (if one unit is B mod, which I don't know).

And that's just the start.  My thinking was I belabor many other mostly useless stuff and don't focus on my main midrange speaker, which is old and possibly needs repair or rethinking.

Well, with just one set it's almost untouchable.  I don't even dare adjust the treble control for fear of losing my current assumed best judgement.  I need to measure the control first.  And with large interface and C mod that's quite difficult to do since scads of bolts need
to be removed including those on the brace.

The downside is...I've got so many other things to do these days (holidays!) AND my system was not only sounding totally wonderful and I was just getting used to the vastly improved imaging and center stability after moving the supertweeter towers out.  Now I've got two sets of speakers in the room I can only put in front of each other (hardly any room for anything else) until I do hoped-for experiments as I've described above.  But it's surprising how acoustically transparent Acoustats are.  Having a second pair powered (!) but no signal right in front of the powered ones has surprisingly very little effect on the out-of-room sound...what I hear the most.  I think the supertweeter towers s had more deleterious effect on the out-of-room sound regardless of whether the supertweeters were unpowered.  (They may have contributed more to the out-of-room sound in the first place too, when powered vs not.)

One of the problems is not just that they take a lot of space to store...they are fiendishly hard to move.  Since they're taller than all doorways they need to be tilted just so.  It's pretty much a two-person job.

That means any moving around is going to be, by necessity, a two person operation.  But I think I can assemble them mostly as one person, working on the sides...  (So far, I've always assembled and disassembled 2+2's with help.)

Here's a fairly positive 2+2 review by J Gordon Holt in 1984.  (Harry Pearson was far more negative, coining the term "credit card coloration.")

I dispute Holt's suggestion that the Acoustats have a less punchy sound because of an alleged high frequency suckout.

In fact, I find that with the controls set to neutral they require slightly more cut in the 2-6 kHz (and 12 kHz) regions to sound good.

This yields a fairly evenly but slowly declining response curve from 2kHz upwards.  That's what seems to sound the best, possibly because of the large degree of high frequency reflected sound created with a dipole speaker.

The lack of punchiness, I believe, comes from other issues.

1.  Dynamic speakers sound artificially punchy because of cone resonances, and people are used to that.

2.  The Acoustats have deliberately designed broad positive resonance below 100 Hz by design to compensate for bass loss due to dipole design.

This resonance is not described in any Acoustat literature I've seen.  It is rarely discussed at all.  Most probably don't know about it.

Yes, the Acoustats have to have this resonance to even sound remotely like it has bass.  That is the magic, right there.  A tower of RTR
electrostatic tweeters would have very little bass.

While electrostatics apply force over area (fairly) evenly, that does not necessarily mean the motion directly follows that force.  The motion also depends on the mass of the diaphram, it's elasticity, it's tension, and similar factors.  So yes, there can be resonances.

Linkwitz added electronic bass boost to his dipole speakers.  He published how he calculated how much was necessary, and everything else about building the (analog IC based) circuits required.

Acoustats have that bass boost in the panels themselves, a function of the relatively high mass of the membrane compared with other electrostats.  (The Acoustat membrane is 2-3x thicker than most electrostats.)

So, I think the problem is that the needed resonance in the Acoustat bass adds time dispersion to the the bass, reducing the "punch."

I've always thought that the Acoustats were fundamentally flawed from the vision of being a "full range" speaker.  They are not "full range" to the fully demanding.  They need a subwoofer.

And therein is a big rub, because the subs available in the 1980's were not as good or cheap or plentiful as today.

Today we can have incredibly good subs powered by their own internal digital amps, with DSP processing to optimize them.

Now we can easily have subs that mate with electrostatic panels.

And all the better to use FIR based linear phase high order filters for the crossover.

That's how I can get a sonically wonderful 8th order phase corrected Linkwitz-Riley crossover for the subs and panels at 125 Hz, far above the resonance and wall reflection issues.

So the resonance, even rattles, doesn't matter much to me anymore.  But that also means I could get by with a large 2+2 with thinner membrane and no bass resonance at all.  That would be my preference, but nobody makes such a beast.

Speaker makers want to make full range 'speaker systems' not parts for audio enthusiasts to use in constructing their dream systems.

And I can simply not imagine building a electrostat speaker to my own requirments.

So I am left to working with 2+2's, and now I have two sets so I can be more fearless about experimentation.
 
**** Update December 8

I determined yesterday that all four new Acoustat interfaces are "B" mod Medallion.  My first 2+2's had been modified to "C" mod.  So I have the perfect opportunity to compare the two approaches.  I can compare my old vs new 2+2's, and also swap in my earlier 1+1 interfaces (with jumper change) which are C mod plus Solen 47uF polypropylene cap.

However, the "tunings" of the interfaces may be different too.

I powered the interfaces with AC from nearly the moment they arrived, because I think that discourages cats.  I have NEVER had a problem with cats clawing Acoustats (as J Gordon Holt reported).  The front panel fabric is too "floppy" like a cat-proof screen door, cats recoil against things with that feeling.  They could potentially claw the sides, which are hard.  But I have never seen it.  I think it also helps if the cats hear the new Acoustats playing music as soon as possible.  Cats respect music.  Things that make music are "alive" to cats, and they don't like to mess with living things bigger than they are that are.  I wonder if J Gordon Holt had his cats in the room while he was playing music.  Also, finally, I think the bias transformer system creates a very subtle noise we can't hear but cats can, also lending the sense that the Acoustats are alive.  And I suspect might be able to "feel" the electrostatic field in front and behind the panels.  Their whiskers are very very sensitive to the slightest force.  I think they'd be more inclined to go after planar magnetics.

Oh, wait, maybe once or twice I saw a cat clawing the side  It was just like one scratch to get my attention.  I haven't noticed any mark.  Curiously it was when I hadn't been playing music for more than a day, so it seemed like that cat was telling me to turn the music (I normally have background music of some kind playing) back on.

Speaking of background music, I personally think it's a good idea and vastly preferable to thinks like "news" and "talk radio," which are bound to get you enraged, depressed, or something other than just bubbling along.  News is more swiftly digested and understood in print.

I've always been very inconsistent about it. sometimes having background music on, other times not, it has been not most of the days of my life perhaps.  That's why I automated my system in 2021 and programmed an automatic playlist generator, so I can keep the music playing.

I also worked on the sound of the FM radio so I could keep that playing the classical music radio station all the time.  Previously the sound would bother me in one way or another after an hour or two and I'd have to shut it off.  So now I have:

1) take the signal from the fixed output of my Pioneer F-26 (one of the greats)
2) Run it through a Musical Fidelity X-10 V3, which buffer the impedances and removes ultrasonic crap (I wonder if the V3 isn't similar to the "Noise Filter Buffer" made by audio genius Mitch Cotter.  I would not be surprised if the NFB also used nuvistors as they are wonderful for this task.)
3) Sample with a dedicated ADC, I found the Black Lion Audio Sparrow works extra nice at 24/48.
4) monitor and EQ the signal a bit with Behringer DEQ 2496.  I found a small bass notch below 32 Hz helps, otherwise there can be a very objectionable 20 Hz rumble.
The digital feeds into my system on coax.  When I record FM, I do so with a dedicated Marantz digital recorder, but only with the preceding
digital chain, as I found the analog to digital conversion in the Marantz to be somewhat "fuzzy," and it works best with either 48kHz or 96khz digital inputs which it resamples in any case (ASRC based inputs) to 48kHz.

With those changes, FM became tolerable to listen to all day long, though I still turn it off when I'm thinking or whatever.  If I selected my system wide "mute" control, it comes back on in one hour and 1 minute.  If I changed the home control selector to DVD and nothing is playing on or streaming from my Oppo BDP-205, it can be sllent all day or until I remember to put something on (as I'm about to do right now).


I hooked up the speakers today.  At first there was a problem with the right channel, hardly any sound, and the interface was even labeled "bad."  But as I was removing it to swap with another interface, I noticed one of the wires had come loose.  It doesn't hold to the screw well, since apparently the repairer/modified replaced the original wing nuts with integrated locking washers, and if you get the loop connector in between the wrong things (above the lock washer I presume) it doesn't hold.

Now they're working fine.  I first noticed they were roughly the same as my originals.  But then I noticed a slight increase in harshness from the FM radio signal I'm listening to.  I suspect their brightness controls are turned up more than my originals, and/or it's a C mod (my originals) vs B mod difference, with the B mod sounding brighter.  Since the circuits are different, they can't be compared simply by measuring the resistance setting of the HF adjustment.  I'll just have to experiment with different HF adjustments to see if the new Acoustats can be tuned to match my originals more closely.

*****

When I tested the 2+2's on December 7 before buying, I played William Orbit's "You Know Too Much About Flying Saucers."  Even at a fairly moderate level, this invoked rattles on one 2+2 unit but not the other.

Rattles were among the reasons I devised my 125 Hz 8th order phase corrected crossover for my original 2+2's, though not using this song, using another one by Grouse.  4th order didn't get rid of them, so I had to go to 8th order, which helped in other ways too.

I didn't really know the Orbit tune would induce rattles, I just knew it was bass heavy and I liked it.

Now I've tested the new 2+2's playing Orbit through my crossover.  No rattles, even at uncomfortably high levels.  (The bass from my subs sounds very clean.)  I reversed the Acoustat channels and it was still free of rattles.

So perhaps my motto ought to be, "Bring me your old, tired, rattling Acoustats, and they will sound perfect in my system."

The new ones in fact sound great as I used them.  They seem to have more "slam," maybe it is the boosted highs after all, or it could be the medallion bass transformers my older system doesn't have.

Still seems to support my belief that these are really best used above 125 Hz, as you might expect with electrostatics.  But possibly they both Acoustats handled the bass better when they were new, and their mechanical structures more solid, and perhaps membranes tighter as well.

So maybe the correct understanding is that you want these crossed over at 125 Hz for the long run.  Initially you may enjoy the full (even if somewhat fake...boosted with membrane resonance) electrostatic bass, but eventually looseness and rattles will set it.  It's above 125 Hz that it's really indestructible, above that pesky resonance.

It's so ironic that an old friend was an Acoustat Monitor (with tube amps) lover who used the Acoustat monitors on the bass.  His system in 1981 had Hill Plasmatronics on the highs, sometimes Magnepans in the mids, and Acoustat 4's for the bass.  "There's nothing like Acoustat bass," he said.

Well, dipolar electrostatic bass is fine in principle, but you'd need still larger panels to get real bass out of a dipole without any boost or resonance in the membrane structure itself.  And in fact electronic boost might not be bad in such a system, best all digital with soft limiting and phase correction.

Or perhaps dipolar bass is not so fine in principle.  It doesn't activate room modes nearly as much, but perhaps it also has less "impact."  Another friend was always complaining about electrostatic bass, he regards it as fundamentally wrong, but some people get "imprinted" (his word) on it.

In a way, I think he's right.  You don't get adequate bass from Quads or Acoustats or many other "full range" electrostatics.  I haven't heard the highest end Sound Labs.  I only think he's wrong in dismissing electrostatics (not to mention crossovers) altogether.  Electrostatics are great used in the middle, between subs and supertweeters.  In that range there is no single driver that works as well as a point or line source electrostatic.  They just don't do the very extremes very well, most audibly in the bass.  And I think it is wrong to expect them too.  You just need an adequate digital crossover system like mine to get them to mesh well with subs.

For background music standing up and penetrating the house, nothing works as well as a tall line source electrostatic.  Every other room seems adjacent to the concert hall, which is all the more pleasant in background as "not in your face."

***** Update December 9

There is little doubt in my mind now (ie, there is still some doubt) that these new-to-me 2+2's have more dynamic punch.  As much as 300% more punch.  But, along with this, there is also little doubt (though perhaps a bit more doubt) that over time, they sound more fatiguing.  Perhaps I just got tired of examining the 'punchy' sound with relevant recordings at high level.  Or perhaps the two observations are part of the same coin.  As I previously reported, Stereophile complained of lack of punch and specifically blamed a softness in the highs around 2-4kHz (which is exactly what is required for good sound with most speakers--the Linkwitz/Gundry dip).  I wasn't buying their analysis, but perhaps they were right.

There's little question at all that the highs are set to a higher level, though it might be the factory selected level in both cases.  The new B mod 2+2's have the big power resistor with a strap, which has a clear marking for the "0dB" position.  The older 2+2's have a knob which has no marking, but turned right in the middle (straight up) would seem to be the suggested position (though, since it was modified from A mod to C mod, who knows).

Possibly Acoustat decided to crank up the default level for the tweeters after the Stereophile review (they reviewed an A mod version) so the B mod has a higher default tweeter setting.

Or possibly part of this is due to the B mod vs C mod.  The C mod might be less fatiguing as it reduces primary current in the HF transformer.  OTOH, the C mod might be less punchy, as it also introduces a second highpass cutoff in the HF circuit response.  With B mod the HF transformer primary is direct coupled, with C mod it is not.

If I am lucky, the punchiness will not go away when and if I turn down the highs enough to cure the fatiguing problem.

So here are some possibilities:

1.  Punchiness and fatiguing are opposite sides of the same coin.  Once the highs are adjusted the same, both speakers will be identical in both parameters.

2.  Punchiness is caused by one factor, and fatiguing is caused by another.  Some guesses are:

a) punchiness is caused by B mod vs C mod, fatiguing is caused by level of highs
b) punchiness is caused by medallion transformers (B mod unit) , fatiguing by level of highs
c) punchiness is caused by medallions with C mod, etc

I need to assess the level of highs in new vs old Acoustats.

**** Saturday Afternoon

I now have the measurements.  As expected, the new 2+2's have significantly extended highs, which only begin to turn downwards from a 4kHz shelf around 14kHz, whereas the old ones begin turning down around 8kHz, mind you this is with the same eq notches I applied for better sound for the first pair, notches at 3kHz and 12kHz.  If I continue to use new pair, I might just readjust the notches rather than the speaker HF level.  But I do plan to see how HF level works, for the first time, that at least might be what makes this whole exercise worthwhile.

Also, the new 2+2's have significantly larger level, at least 2dB higher measured right in front of the speaker in the same relative place.

It looks to me like the new ones have either a rebuilt or a less deteriorated HV level.  Possibly it was modified/repaired with slightly higher voltage level than stock.  But it could all be a matter of relative deterioration.

My older pair was worked on by a famous Acoustat modifier around 1998 or so, I think.  It seems he would have repalced the HV diodes or other parts if they were deteriorated then.  So this makes the difference look more like the new ones having higher-voltage-than-stock.

But even that 2dB level can hardly describe the difference in impact the speaker have, listening to pink noise.  The new 2+2's sound 10dB louder.  At the same level where the old 2+2 sound as smooth as riding way above Cloud 9, the new ones sound like you're riding just below the peaks of Cloud 9, riding through one mountain of cloud after another.  The dynamic contrasts are frightening.  Just listening to pink noise.  Yet it measures about equally flat, just needs a small bit of EQ readjustment perhaps, and/or lower HF level.

So I think it's the effect of a higher bias voltage, getting more of that "real electrostatic" sound.  Maybe it was more like they were new too, but I think also it's a kind of nitro upgrade done by the same guy (also an Acoustat modifier, if less renowned) who put on new socks and may have done other upgrades, including Cardas jacks and new possibly more acoustically transparent socks.  I think the upgrade in
bias voltage is minor compared with new, but large compared to elderly bias supply.  Possibly the diodes were 5-10% higher voltage than new, is my guess, but 30% higher than old.

I have, never used, a suitable HV 15kV probe.  I burned out one of my meters years ago using a 5kV probe in my first attempt to measure Acoustat bias voltage.

Along with it, there's a slight tick about every 20 seconds in one of the new speakers.  It takes ear nearly to speaker to hear it.  It's quite tolerable like that, though since I bet it results from a tiny bit of arcing it generates ozone and it might get worse, or even self-heal.  I smell no ozone and I doubt it's a serious generator.

While the highs do need re-EQ if not HF adjustment, I'm liking the more dynamic sound.  My old 2+2's were just too laid back.  That was their problem.

Perhaps that difference will survive HF level, EQ, and speaker level adjustments.  Or perhaps not.

*****

Just turning down the panel level by 2dB relative to the bass (unchanged) seems to have eliminated the "listening fatigue" problem.  Somehow, the new panels play louder.   I've also noticed that the upper panels on my older 2+2's seem to have diminished level compared to the bottom panels.  A more systematic examination should be done.

I've also noticed that the manual for the A version of the 2+2's says the "flat" position is 3 o'clock, not the 12 o'clock I had been presuming.  So it looks like I've had the HF level too low.  Which seemed unintuitive to me because I had to add some EQ notches in the treble anyway.  But it seems the new speaker has more extended response, and perhaps that's what the HF control being higher does.  The combination of my existing HF notches and the more extended response sounds about right, more transparent and punchy than before.

This may just be about the adjustment, and not the many other differences between my two pairs of 2+2's.











Monday, November 20, 2023

Supertweeter Changes

My next generation supertweeter, to be unveiled on October 31, 2024, will have multiple lines of protection:

1) Two series fuses, fast and slow blow.  I'm going to start with 1a fast and 0.5a slow.

2) I'm going to stick with the 0.47uF series cap, but upgrade to a polystyrene or polypropylene instead of polycarbonate, and 400v rating or better.

3) I may put in a shorting lamp prior to the fuses so it doesn't cause nuisance blowing.

4) I will stick with current 150W amplifier, which is robust.  As long as I have series capacitor like 0.47, I need about 100W or more.

5) I will program the miniDSP used by tweeter with only one option, the 17kHz high pass, so it can't accidentally be turned to flat.

6) The amplifier will be Insteon switched, and only turned on for "Feature Music."  It will be on a power strip that also powers a purple or ultraviolet indicator of some kind, so you can see if tweeter amp is turned on or not.

7) There will be blinking level lights set to some high level like 1 watt, or perhaps multiple LED's.

8) No Behringer will be required.  The miniDSP will be moved over to where the DAC is, to keep center of soundstage clear.

9) The tweeter in back will be soft mounted to the back of the Acoustat interface, firing rear, then timed to match front arrival or not (NE19VTS tweeters.)  This is all I expect to do at first.

10) If tweeter in front, maybe later, Dynaudio in protection cages soft mounted to front of base.  Dynaudio might also be mounted in back facing forward, depending on how acoustically transparent the Acoustats seem to be at 20-30kHz.


 

Acoustat version info

Acoustat 2+2 panels came in several versions:

3-wire original

5 wire with red bias wire

5 wire with yellow wire with spiral red stripe: improved coating, made under Hafler ownership

https://www.diyaudio.com/community/threads/acoustat-answer-man-is-here.183168/post-7410305

 

Wednesday, November 8, 2023

Supertweeters Removed

A week after my return from my vacation in the San Diego area (notably, where much of my audio insanity derives from, where I worked at a high end store and met many high end audiophiles, though I only visited only one such audiophile on this trip, the other notably and disappointingly having given up on the habit) I noticed the super tweeter DEQ unit, which mostly operated as an RTA visual display indicator for the above 17kHz super tweeter signal, as well as an extra digital level control, had a greyed out display.

I measured the output of the supertweeters using the RTA app on my phone, and sure enough there was no output at all.  When the display goes out on a DEQ, the I/O functions could still be working, but probably not.  Probably there's a power supply failure that leads the entire DEQ device to be dysfunctional, not just the display.  That is how it has seemed to work in every failure I've paid careful attention to so far.  BTW this unit was connected with digital input and digital output, merely serving to pass (and slightly modify) digital signals from the super tweeter miniDSP which does the ultimate high pass crossover, and the Emotiva Stealth DC-1 which does the conversion to analog and final level adjustment, which I've tended to set around +7dB since the super tweeter signal is so low and also highly attenuated by the tiny capacitors I connect the physical drivers through.

Damned!  (And ironically, I'd just been talking about my super tweeters  in San Diego, as my still practicing audiophile friend shared his stories with Dynaudio drivers, though he used D28's and not the D21AF's I have been using as front facing supert weeters.)

I did a quick bypass of the DEQ unit.  Then it was still dead in the right channel, but working in the left.

The way these are hooked in leaves me little opportunities for easy testing.  I ultimately had to move the massive sand filled 40 inch high Target stands with LS3/5A boxes on top, which is not getting any easier for my 67 year old body despite the (very limited) strength training I do.   Which I didn't bother to do until exhausting all other opportunities.  And then it was clear on the proven good signal of the right channel which plays fine on the right super tweeters that the left super tweeters were not working at all.

This was true even bypassing the capacitors.  Measurement suggests both super tweeters on the right channel, including the now unobtanium D21AF in front, now have open windings.

Double Damned!

But now perhaps is also the best opportunity to figure out what it sounds like with the super tweeters and their stands removed.

You would probably not be favorably impressed by the fact that I just hadn't done that experiment before.  I'd only compared the situation with the super tweeters powered and not, and powered always seemed somehow magically better.  Bass seemed clearer, and highs seemed more balanced, never harsh or closed in or honky.

Always, the difference was very slight, and possibly only sighted bias.  I doubted I'd be able to pass a blind test.

But I wanted to make the supertweeters as good as they could be before trying the Acoustats 2+2's with them completely removed, which I knew was going to have a lot going for it.

When I had previously (and for over 11 years) using Acoustat 1+1, they were narrow enough that I could place the super tweeters and their stands on the outside of the Acoustats.  In that position, there was very little negative effect...perhaps even a slight positive, from having the extra stuff on the outside.  It seemed to make the image even wider from left to right.  But because of the added width of the 2+2's, and my need for a pathway through the room from the front door to the kitchen, meant that I had to put the supertweeters on the inside, if I was to use that assemblage at all.  I knew from the start it was acoustically problematic.  But super tweeters are so me.  I didn't want to give them up.

But now that I've done so, there's no going back.  It's clear to me from 4 years of listening to the previous arrangement that taking the supertweeter "towers" out of the way opens up the center of the image.  The effect sonically is much the same as it is visually.  Moving the super tweeters out of the way opens up the complete image.  Previously it was parcelled and often cartoonish.  Now it is expansive and every little bit from left to right and front to back can be mapped out.  It has more slam too.

This opening up is even audible in other rooms where the stereo image, as such, isn't.

The problem as I suspected all along was that the heavy 40 inch sand filled target stands, and LS 3/5A boxes bearing super tweeters, was just getting in the way of the natural dipole response of the Acoustats, complicating the natural 3d line source approximation with lots of spurious reflections.

So back to the drawing board on super tweeters, in more than one way.  I managed to snag 3 more of the high technology Tymphany NE19VTS-04 tweeters I was using in the backs, anyways, on ebay, they were unobtainable in stock anywhere else.  I bought them as a Vifa product (at Madisound, IIRC, who doesn't even list them now).  Later they appear to have been also sold as Peerless.  It's Tymphany either way.  The Dynaudio D21AF's in front are probably too heavy to deal with in future, and about as unobtanium as it gets (but miraculously, just while writing this paragraph, I managed to snag a replacement unit on eBay, so hopefully my loss is now corrected, I will still have a pair of working D21AF's for comparison if nothing else.  And, btw, in comparisons the Tymphany's seem about as good as the Dynaudio's, at about 1/20 the weight, a true engineering miracle that it seems nobody else but me has recognized (they have essentially flat on axis response to 40kHz...you want at least that or there's no hope at all.  I always suspected the D21AF's would handle any power available from my 150W amp, but in this situation both D21AF and NE19VTS died in the same catastrophic situation...I'm still not sure if the amp failed or what, perhaps just a full signal digital signal combined with all that gain I was applying, was just too much, I need to be more careful about that, also thinking of using smaller amp, 50W would be sufficient).

I plan a new supertweeter perhaps mounted only in the back of the Acoustats, possibly firing backwards only (and timed to arrive with the front) or both ways, on a strap of metal bent to fit under the Acoustat base board to the middle of the transformers.

I honestly believe the supertweeter effect is greatest in the reverberant field anyway, and not in the audiophile obsession--leading edge transients.  Otherwise there would be almost universal discontent with super tweeters.  Many of the leading edge tweeters (Scanspeak comes to mind) have pronounced peaks above 20kHz, which presumably extends the ultrasonic response much as I was doing.  Now if this were all about the leading edge, it would be almost impossible to get it exactly right.  I tried very hard (using ARTA program measurments) to get it as exact as possible, and most supertweeter users have never done anything like that.  And move your head 1mm and it's different, the supposed leading edge may start out inverted, etc.  Anyway, I think it's much more about the reverberant field reflected sound that leaks into your ears from the side.  It's not that critical, down to the 0.1mm and so on, to get the supertweeters exactly coincident, and impossible in most cases anyway.

In actual soundfields, leading edge transients are rare.  All is buried in the jiggling ambient noise.

What adding a super tweeter does is make this ambient noise sharper.  That actually makes it sound softer because it's simpler.

Anyway, what counts is getting the supertweeter added sharpness within a few ms of the actual signal.  Probably even 7ms (about 7 feet in distance) or so from back reflections is fine, but I could also try compensating it for simultaneous first arrival at the listening position.  I could also try a separate front firing unit, with a different delay...

Anyway, for now I'm enjoying a more wholesome sound, full size players, wholistic, musical, etc., from not having any super tweeter stuff in the critical area in between the speakers.

You might say, I should have tried this before.

Update on Friday.

I haven't yet moved the LS 3/5A's based supertweeter assemblages completely out of the living room, which I presume, following ancient Linn advice (yes, no kidding, I first heard this in the day from the Linn Rep where I worked in 1977, though I honor it roughly in the breach) will make it just that much better still.  Also, I need to get it out of the way for appearances sake.  And it will take some doing because every place where equipment of this value could be put is already filled up.  Especially inside the house,   I need to carve out junk of lesser value and do something with it.  Another thing I observed on my trip is that it's critically important to get on with the transfer or disposal of junk equipment or you'll still be stuck with it when you're too old to move it very easily.  So I've got years of work cut out for me here, it looks like.  And perhaps it's a good excuse, just about as good an excuse as owning cats, to be getting up and moving around now and then.  So I've moved all the new odds and ends to be investigated onto the Coffee Table in the living room, and now I can move the LS 3/5 A's (which are so heavily modified as to be not really that anymore, they combine D21 with B110 and external driver connections and internal magnet shielding...because I was previously using in proximity to a CRT TV in Kitchen.  But they originated as now extremely valuable LS 3/5As, and I believe I still have enough parts to make them all whole again and like original except a capacitor or two which might have been lost from the crossovers.  Which I probably still have, if I could only find it within dozens of potential boxes, often hard to get to.  So that's how things go.  This time around, I thought they made a dandy acoustically solid base for the supertweeters, and they did, except too large and massive and complicated when positioned in between the speakers combined with the similarly large Target 40 inch stands.  I'm now convinced the best supertweeter arrangement is on axis in back of the Acoustats, and possibly back firing alone is sufficient combined with compensating anti-delay (delay for midrange and bass only) for the first bounce (to and from the back wall).  That, combined with a new lower powered amp and/or protection/monitoring system, etc, will be the NEXT GENERATION supertweeter arrangement.  AND, I should probably wait about another year or so before deploying, so as to have the current sound firmly set in my mind.  AND I've needed a pause in supertweeter deployment anyway (and bear in mind my supertweeters are truly super...I hear nothing coming out of them as such...I only believe they have a beneficial effect on how other things sound) so I can optimize the actual audible part of the HF spectrum.  Back in 2021 I deployed a chairside EQ and began experimenting with midrange EQ's (which are now dialed into my midrange unit) that went up to a cut around 12kHz.  The measured effect in pink noise was a 3dB/octave rolloff which became noticeable around 2kHz, but instead of just being a limited Linkwitz/Gundry "dip" it was a slice, basically continuing to 17kHz (at which point the Acoustats swiftly drop off).

This is dialed in as a set of 3 PEQ's actually, each one of which was tweaked by ear.  But such tunings are never certain.  I've long wanted to re-analyze the possibilities.  I may be rolling off too much of the audible highs, long before it even gets to for-me inaudible highs above 16kHz.  So you would think I should optimize what I can hear before bothering with what I can't, but my practice of audio, which presumes the possibilities of apparent subliminal effects (there has been controversial and unreplicated research showing brain wave changes when the bandwidth is increased from 20 to 40 kHz).  But then perhaps too much I waste my time on things which might actually be of little value compared to the things that may well (and objectivists would believe) have important value.

So much as I keep wanting like a junkie to get the supertweeters back up again somehow, with some combination of bubble gum or something, I should resist that temptation, and even set make a rule that I won't again try super-tweetering for about a year, both to get an all new better and safer approach up and running, and to familiarize myself with the sound without it, and to get the audible parts of the Acoustat EQ understood if not dialed in better.

So I think I'll choose Halloween 2024 as the first date I will permit myself to run the supertweeters again, in some new setup, hopefully a much perfected one, with no extra stuff in between the speakers, and just a little bump above the back transformer, and a working power limiting system to prevent future burnouts, and probably a different amplifier (in my own collection I might try the 60 watt Marantz 15b which has 'just enough' bandwidth and presumably Sid Smith got the internal slewing and limiting right, this amp is legendary for good sound.  I've always wanted an excuse to use them.  But I also keep lusting after the few $200-300 Chifi Class A amps still available on ebay.

Whatever amp will probably be right out front and center where the Aragon 8008 BB is now.  I've done enough A/B'ing to say there's no audible difference between the Aragon, as currently biased (on the high Class AB side: about 28mV instead of Klipsch specified 12mV across the collector resistors) and the Hafler 9300's.  And under no situation is the Hafler underpowered.  The speakers begin to sound dynamically limited at exactly the same point with either amp.  So I don't need the greater (measured) 600W (spec 400W) into 4 ohms that the Aragon provides, the 350W (spec 250W) that the Hafler provides is sufficient.  And the Hafler 9300 is better in every measured way, 0.002% THD instead of 0.02% for example, and even better HF damping factor (it was even better than the Krell FPB 300).  The Trans Nova circuit in the 9300 is about as simple and fast as it gets.  However, Parasound HCA-1500A is almost identical, and slightly superior in near clipping distortion levels.  With the relays and stuff it's just less of a 'perfectionist' amplifier IMO than the 9300.  But winning amplifier of earlier epochs basically need not apply.  Before 1982 or so it was impossible to make amplifiers this good, then better parts made it easy.

The problem is, I'm not sure I want to sell the 8008 BB, and I've got no other place to put it, certainly not without removing the 2" spiked Mapleshade carpet feet.

I'll keep the A/B setup, but only temporarily put other amps to the side during A/B testing, rather than in front, which will be dedicated to the supertweeter amp in future setups starting next Halloween.  (I also have parts to build a Pass Class A...an effort which now seems unimaginable.  The whole purpose of the Pass amp was for the super tweeters in the first place.  I have years worth of repair work to catch up on first.  And at the current rate, they will be many years more before much is done.)

I've got plenty to do until then to give the supertweeters very much priority.  I've now got 3 essential but dead DEQ's, all needed a power supply refurb.  I should get the first one going to get back to having the chairside EQ which was so helpful to EQ experimentation.

I need to get things moving through "the laboratory" again and back working again.  There's an incredible pile of to-be-fixed stuff in there now.

An alternative to the Chifi class A amplifier would be Schiit Aegir.  The power is just right.  I'm not going to consider such a thing until I've sold an equivalent amount of stuff on eBay, starting right now.  But it looks pretty cool.  The design goal is correct, there is really no need for a class A transistor amplifier.  But high bias is generally good (this amp is quite high bias for the power, and features their special circuit which sounds good in the description and the Amp scores well at Audio Science Review...in fact about the best thing I can see for the price, as it is surrounded by $5k amplifiers in the upper end of the ratings).  The upper bandwidth is huge and there are no coupling capacitors or servoes.  The one non-purist thing is a relay, but those are fairly ubiquitous, and this one is CPU controlled.

Saturday, October 7, 2023

Creating Music playlist and playing in Roon

Here is a 30 second video showing me clicking on the "MakeMusic" script icon (built upon my universal playlist generating program mplay) which generates a playlist called Music (it takes about a second, the script window stays open for 5 seconds so you can see it worked) and then playing this playlist in Roon.





Wednesday, August 23, 2023

mplay Generation 2: playlist generating programs

I've been working on automatic playlist generating programs now since January 2021.

I originally had just one program, mplay (make playlist).

Now there are more programs: splay (shuffle multiple playlists together), tplay (tell about what actually played, and truncate playlist or play history), and shufflelinks (shuffle multiple "playlist" folders-of-links together).

These programs can be compiled in scripts, making playlist generation either more intelligent (such as removing the items that didn't actually play from the playlist history before creating a new playlist) and/or combining playlists (or folders of links that serve as playlist for programs that don't handle playlist) together in different ways.  (Playlists can also be simply concatenated together using the system command cp or equivalent so I didn't write a program for that.) 

Also mplay itself has advanced considerably since the 2022 software release on Sourceforge.  I hope to release this new version before the end of 2023, possibly fixing some of the current "gaps" between what it does and what it really should do.

One key new feature is the ability to specify ALL of the files in the specified folders.  As in "just make a playlist with all of the files in these folders."  You don't have to guess or predetermine the total number of files that would be included.  Since the playlist history is unchanged by this operation, it is neither read nor updated (however, it might be better if it reads to playlist history first, to put unplayed items at the head of the playlist, currently it just ignores if itms have already been played or not).

You can also apply an file age criterion in order to make a playlist consisting of only files newer than a specified number of days.  That can be combined with the ALL option to easily create a playlist of ALL the new files.  A script can concatenate this premier playlist at the beginning and/or end of a main playlist.  (Another useful feature not yet implemented would be to make a playlist of ALL the as yet unplayed files.)

But I've found it useful also to mix the new files into a playlist of older files.  There are two basic approaches to this kind of mixing:

1) Dense Shuffling selects one (or some specified number) of items from each playlist, which are then ordered randomly (or not) into a new segment of the output playlist, with additional segments added until all the items from every playlist are included.  If any playlist is exhausted early, it is reset and reshuffled.  This works very much like mplay itself (though currently only for non-audio files, since audio files are currently treated as if they were all in one big folder, with no other option).

2) Sparse Shuffling mixes a smaller playlist into a larger playlist so that each item in the smaller playlist is included exactly once (or some specified number of times) in the end result, using random placement.

Currently splay does only Dense Shuffling, and shufflelinks does only Sparse Shuffling.  It would be good to have both available in both programs, and/or to combine the programs to work on any combination of playlists and links-folders, to generate either a new playlist or links-folder.

tplay is very useful in the case where I've been playing an audio playlist in Roon, but switch to playing other items.  If I go back to the playlist in Roon, it does not remember the last item played but simply starts all over from the beginning.  Using tplay, I can remove the items that have already been played from the playlist.  Or, I can remove the items that were not played from the playlist history and create a brand new playlist.

Saturday, July 15, 2023

Thinking about HDCD again

Using DVD-5000 to decode HDCD on played on PD-75 as digital transport

 I've been testing a pair of the DVD-5000's I originally bought for use as living room DACs.  They are both for sale on ebay as I write this.  I am reminded of the fact they are very good, and according to objectivist standards should not be audibly different from the very best DACs.  They also have the ultimate R2R chip made by Burr Brown, the PCM 1704, which was discontinued in 2012.  By objectivist standards, it should be no different than a good Sigma Delta chip.  But it is different in objective ways, for what that's worth.  I love the DVD-9000 for HDCD decoding but don't use it for anything else.  I would have been tempted to use DVD-9000 for midrange DAC until I discovered that I needed 3 DACs of exactly the same design for my tri-amplified system.  Even if I adjusted the delay times for one particular sampling rate, when I changed sampling rates those adjustments would no longer be valid, because the latency of nearly every device varies with sampling rate.

Faced with that problem, and the (at the time) skyrocketing price of DVD-9000's, I opted to get 3 DVD-5000's instead.  Those would be basically the same thing, I predicted.

But for probably no good reason, I just never fell in love with the DVD-5000's the way I did for the DVD-9000.  And it was much more convenient to have multiple DACs with level controls, and smaller.  So I ended up with 3 Emotiva Stealth DC-1's, which are quite fine also (technically they measure much better than the DVD-5000's, but once again that difference should not be audible).

I still use DVD-9000 for decoding HDCD's (because my Oppo BDP-205 doesn't do that).  So I tried using the DVD-5000 (which doesn't actually play discs, but connected as a DAC to the digital output of a Pioneer PD-75, optically) and I found it does indeed do very well with HDCD's, just like my DVD-9000.

But it is different.  When playing an HDCD the output level never exceeds the output level from an ordinary CD.  The DVD-9000 actually plays the level expanded portions of an HDCD up to 6dB louder than an ordinary CD.  In contrast, the DVD-5000 does what every other HDCD player I've ever tested (other than the DVD-9000), it lowers the average level, just so so the most expanded peaks of HDCD reach the maximum CD level, but no more than that.

When my brother-in-law George first heard the HDCD on a DVD-9000, he was (unusually I'd say, because he never matches levels by measuring) shocked and appalled that that HDCD boosted levels 6dB above normal CD levels.  He felt that was unfair.  HDCD was all just a cheat, he insisted.

Well, as I've started to explain in many other previous essays, it's much more complicated than George and almost everyone thinks, because of inter sample overs.

(And of course, George was also wrong that HDCD players boost the level beyond that of CD's.  Though for the longest time I wondered if earlier HDCD players did do that boosting, and it changed when players stopped using the PMI chips.  I had it reversed.  The DVD-5000 uses PMI chips and doesn't do boosting beyond CD levels.  The DVD-9000 uses a software implementation--like all later players--and does boos the level.  So it does not necessarily have anything to do with using the chip, it was just a particular approach the designers of the DVD-9000 took, which possibly was available to the users of the chip as well, but I haven't seen one that does.)

The ISO's on regular digital recordings can match those of even peak level expanded HDCDs.  And what's more, HDCD's don't seem to have such as big ISOs.  So it comes out about as a wash, the HDCD is just giving a kind of engineered peak, and regular CD's are giving us extrapolated peaks.  That dynamic range in the analog output was there and needed anyway, HDCD's (when boosted) are just using that dynamic range for real music dynamics, rather than extrapolations from what may be just high frequency ringing.

Now I also know, that SACDs may in fact be the worst offender of all.  They can have the highest peak output above the nominal 2V level.

But I'm wondering if it was the boosted HDCD level (compared to other players) that led me to fall in love with the DVD-9000 in the first place.

(I think HDCD done as intended (and with the filter control*) would have been a great idea...as an open standard, and would have enabled a way to bypass the loudness wars, using the HDCD encoding to provide the uncompressed version which for which the compressed version would be heard without it.  Now we're simply stuck with HDCD as necessary for reproduction of a significant number of fabulous sounding recordings, which are better heard when available in high resolution.)

(* The filter control somehow seems less necessary as HDCD's don't seem to have the transients that provoke excessive ISO's.  Somehow they achieved that effect without a post-filter change.  It will be necessary to do more investigation.)