Tuesday, May 31, 2011

The KPAC buzz: 10.5kHz in L-R

I have recorded and analyzed the buzz from KPAC FM during a 1.7 second silent interval.  What I hear as the buzz is a 10.5kHz tone mainly in the L-R subcarrier.  I will be sending this information to the station engineer shortly.

I connected the fixed outputs of my Pioneer F-26 tuner to the analog inputs of an Alesis Masterlink set to 88kHz 20 bit sampling (I had intended 24 bit but missed that).  I recorded several snippets when it seemed like there would shortly be a brief silence.  In one of them, I got really lucky, I got 1.7 seconds of pure silence.  Or it would have been silence without the noise and buzz.  I made CD24 discs of the 88.2/20 bit recordings and loaded the files onto my Mac.  I ran the Wave Editor program and created 44.1kHz Wave and AIFF files for later use.  I cropped the best example down to the virtually silent portion, with extent of exactly 1.7 seconds.

Even though I could clearly hear the buzz when playing back the file, it was hard to see anything specific in the Wave Editor Spectrum Analyzer.  But it became more clear once I created a mono signal using L-R subtraction function.  Then, after trying several other digital processing functions in Wave Editor,  I got the idea of using the DSP parametric equalizer.  I noticed a little bump around 11kHz, and aimed for that.  I found I could get nearly perfect cancellation of the buzz by setting the frequency to 10579 Hz (the nearest alternatives in Wave Editor are 9998 and 11194 Hz, which provide considerably less cancellation), the Q to 10, and the amplitude to -20dB.  The peak I am cancelling looks to be a bit lower than the 10579 Hz notch.  When I set the Q to 20, it doesn't cancel quite as well.  So I'm guessing the actual frequency is 10.5kHz, or quite close to that anyway (+/- 300Hz).

Sunday, May 29, 2011

F-26 design

Took a quick look at F-26 service manual and schematic.

#1 The 75 ohm input feeds the front end transformer directly.  The 300 ohm input goes through extra balun first.

#2 The instructions do not indicate any high blend action resulting from use of the muting and stereo threshold controls.  (I do not believe there is any auto-blend either, since I hear very clearly the KPAC buzz which is quickly erased by using the high blend control on the Kenwood KT-6040.)

#3 The front end has two mosfet amps.  First there is double tuned filter, then mosfet amp, then a quad-tuned filter (!?) then another mosfet amp, then another filter before going to mixer.  I think the quad-tuned filter may be partly controlled by the locking circuit.  The filter-amp-filter-amp-filter-mixer topology has been declared "optimal" by my friend T.  There are two tuner gangs used in the oscillator, (the second being oscillator filter?) another optimal approach.

#4 The IM and Spurious rejection ratios are 120dB.  You won't find numbers like that for any varactor tuned tuner.  I am coming more to believe these are the important specs...they indicate how well a tuner is cleanly pulling the needle out of the RF haystack.  Noise figures from bench testing are unrealistic as there is no RF haystack in such tests.  Especially now, when the radio frequencies are saturated with stuff.

Thursday, May 26, 2011

Pioneer 9500II question

I just posted this to FMtuners Yahoo Group:


As part of my newest adventure into Pioneer tuners, I got a TX-9500 II. To
connect it to my usual #1 antenna (a folded dipole, just like #2, but in
different location) I needed to connect it to a 75 ohm cable. So I brought out
the same piece of cut RG-6 I used for testing a Marantz 20B last year.

The sensitivity was horrible, I couldn't get college stations at all, and my
usual KPAC was way noisy, noisier even than the 20B which I tested (or
mis-tested) last year. Very poor quieting.

But then I got the idea of connecting to the 300 ohm terminals, simply by
extending the twinlead from Antenna #2 a few feet by soldering in a section of
twinlead. This brought a massive improvement, though I still find the 9500II is
not as quiet in stereo as my quietest digitally tuned tuners (which feature
advanced MPX chips and technologies, I never bothered with any of the lousy
ones), and I know this is true because I've now done an A/B test with a Kenwood
KT-6040, which was much quieter. From memory, the 9500 II seemed in the same
ballpark, maybe even better than KT-8300 (though I haven't tested the 8300 in a
few years).

I'm worried that my hacked RG-6 is either bad or I am not terminating the
connection properly. It always seems tricky to me to have to cut a piece of
RG-6, clamp the shield down, and then run the signal wire to a screw. The
pictures always show a piece of RG-6 that has braided shield, whereas all the
ones I've ever cut have a foil-and-wire type shield, which doesn't lend itself
to being folded back as pictured, so I simply twist the foil and wire into a
wire and bend it back over the section of RG-6 that is being clamped. Perhaps
that is the wrong way to do this.

Another possibility, mentioned here before, is that there might be an internal
75 ohm balun that has gotten fried. Does anyone know if TX-9500 II has a balun
to convert 75 ohm inputs to 300 ohms which is actually sent to the front end,
whereas the 300 ohm balanced passes right into the front end? Or do both inputs
go through some sort of balun or impedance matching devices to yet another
impedance like 110 ohms or something.

I've always suspected that tuners lacking an F connector are more optimized for
the 300 ohm connection and the 75 ohm connection needs to go through an internal
impedance matching device, whereas tuners having an F connector are optimized
for 75 ohms, and the 300 ohm connection has to go through internal impedance
matching device. Though there is no logical reason why this would have to be
so, or there might have been some shift that occurred over time (like in the
late 1970's) rather than related specifically to the kind of 75 ohm connection

Perhaps I should go back and retest the Marantz 20B also. It might have been
more sensitive than I thought (I can't remember if I tried direct 300 ohm
connection with it, or I might have handicapped it slightly by using external
balun to convert 75 ohm cable, thus requiring signal to pass through two
baluns). My antennas #1 and #2 are very close on KPAC 88.3, with the #1
possibly having about 1dB more strength according to meter on Kenwood 600T (64
vs 65 dBf).

The 9500II midrange is quite nice. The "buzz" sound in Stereo which I now
believe to be station error of some kind, is not audible at all, instead there
is more noise. So while the 9500II has no buzz, it has a grainy background
hiss, and the 6040 has hardly any hiss but very clear buzz. Unless I engage the
High Blend on the 6040 to kill the buzz, the 9500II is more pleasant to listen
to. The high filter switch on 9500II seems to cut highs in a way I don't like.

Wednesday, May 25, 2011

Pool House dimensions hit the sweet spot

I am now calling the fancy brick shed that is almost under construction (down payment made, waiting for engineering review of concrete slab) the "pool house".  Actually, there isn't any pool yet, but my long term backyard plan is to have a hot tub right next to it at the end of an extended patio.  Pool House sounds a lot fancier than Shed, and in fact the building currently being designed will be much more upscale than the usual shed.

One of the functions of the pool house will be to store the factory boxes I have for many of my most important hifi components...

This Pool House is not primarily intended as a "man cave" (my whole house is currently a man cave, and I intend to keep it mostly that way, perhaps conceding design authority over one bedroom, one bathroom, and some of the area in the kitchen to a friend who hopes to move over in about a year).

But just because I'm an audiophile, and I think it's the right way to do things, I've kept acoustics in mind from the very beginning, and tried to design around a set of well known dimensional ratios for room acoustics.

I originally intended to fit the 1 : 1.28 : 1.54 ratio.  Assuming the (originally planned to be unfinished) ceiling would be almost 10 feet high, it seemed to follow that the (exterior) dimensions should be about 12 and a half feet, and 15 feet.

Then it became clear that brick wall construction eats a lot of interior space because the walls are about 6 inches thicker (shortening each non-vertical dimension by one whole foot).  Well that changed the calculations entirely, so I started fudging, and with sloppy and probably incorrect math changed the 15 foot exterior dimension to 14 feet 10 inches.  I wanted to have it less than 15 feet anyway to avoid the need for an expansion joint in the foundation (given 6 inch slab, and normally you can have 2.5 linear feet for each inch of slab thickness, or so I have read on the web).  14 feet 10 inches seemed to be changing the non-vertical ratios as needed for a one foot loss to wall thickness, AND to be playing it extra safe wrt the slab design.

But then, after further pondering, I decided it would be even better for ensuring the longevity of my hifi boxes if I moderated temperature (and therefore humidity) swings somewhat by having fully insulated and finished walls and ceiling.  Now that's adding another 4 1/8 inches to each wall, assuming 5/8 interior sheetrock.  Once again blowing up my calculations, though this time the ceiling height was being reduced as well.

But then the contractor asked if I would rather have 8 foot or 9 foot walls.  This was important to the engineering.  Well I had been worried about enough loft space, and after a day of thinking about it decided to go for the 9 foot walls.  I was still uncertain about height of the finished ceiling (I figured about 8-8.5 feet) given 9 foot walls.  But then yesterday I found that with 9 foot walls, I can have 9 foot minimum ceiling height, and it can easily rise to 9.75 or more by attaching ceiling to the underside of the rafter braces, though the lower the braces the better.  The roof is intended to be either 5/12 or 6/12 slope.

So now I finally had all the details for evaluating the acoustics.  And bringing all the numbers together, I hit another sweet spot exactly with a 9 foot 5.5 inch ceiling height.

Here are the numbers.  The exact thickness of each side wall from exterior to inner finished wall is:

4 inches brick
1 inch air gap
tar paper (not expected to affect wall thickness significantly)
3/4 OSB sheathing
3.5 inch framing
5/8 interior sheet rock (hope to get Quiet Rock 528 if affordable)

That adds up to 9.875 (9 7/8) inch thickness.  Two walls reduce each dimension by 19.75 inches from the exterior.

This reduces the 14 foot 10 inch exterior length to 13 feet 2.25 inches on interior.

This reduces the 12 foot 6 inch exterior length to 10 feet 10.25 inches.

Those are fixed by specifications already given to the engineer.  I wouldn't want to change them much anyway.  There is some still some play in setting the ceiling height because the braces can either be raised or lowered a bit.  It turns out, the perfect spot is hit right at 9 feet 5.5 inches ceiling height.  This gives one of the most commonly used dimensional ratios:

1 : 1.14 : 1.39

This is the most cubelike of all the well known and accepted acoustically optimal dimensional ratio sets.  It is a sweet spot.

And I hit this mostly by luck...I had actually been aiming for a different ratio, but because of errors in calculation followed by a new optimization, I got there.

Here is a online great dimensional ratio calculator (the best I have seen, complete with loads of results and documentation):


This is much better than the RoomModes calculator provided by RealTraps because it runs on all computer systems (as a server-side script), includes oblique and tangential modes in the calculations, and provides vast documentation.  BTW the 1:1.14:1.39 ratio set was extensively analyzed by L.W. Sepmeyer in 1965.  This is discussed here:


One slight flaw in my reasoning is the slight rise in the celing from the 9 foot side walls to 9 feet 5.5 inches in the central region.  This is a very slight vaulting (correct word?), fortunately it in the desireable direction wrt the long "listening axis" of the room.  This may have effects mostly on the oblique and tangential modes assumed in the acoustical optimization.  I'm not sure what to do about that.  Would it be better acoustically if I just had a completely flat ceiling at 9 feet? The 5.5 inch rise only takes about 1 foot inwards from the side walls, then the ceiling will be flat (under the brace) from there on to the center.  I would like the extra height to increase loft storage height and volume (and wouldn't mind more actually, 5.5 inches increases loft volume by 10 cuft, and probably soon enough the extra 5.5 inches in height will prove to be essential for something).

Or maybe I should raise the ceiling height just a bit more to compensate for the vaulting.  It's well within structural considerations to raise ultimate ceiling height to about 10 feet in the central region, but then that starts getting too cubelike (2 almost identical dimensions) for good acoustics.  So 9 feet 5.5 inches is a good compromise given what I know now.  A flat ceiling at 9 feet also looks pretty good acoustically, it comes close to the 1 : 1.2 : 1.5 ratio set (actually 1 : 1.2 : 1.46) which might even be slightly better than 1:1.14:1.39.  The whole range from 9-9.5 feet is excellent, when you get down below 9 feet certain other acoustic problems occur.

Here is a blog in which Ethan Winer claims that vaulting is great (well, so long as you buy his Real Traps to run along the peak).


The real world is never as simple as some of these acoustical optimizations assume.  For example, the walls are not perfectly rigid, rather somewhat flexible and damped.  And then there is the framing and sheathing, etc., behind the walls.  All these are active to some degree.  And then there are the windows.  And then there is the stuff you put in the room.  Most of this doesn't have much effect, but all together it has a substantial effect that is way beyond the power of free simulators to estimate.

Often when people try to build to the correct ratios, or modify to them, they find it doesn't make as much difference as expected.  The correct ratios are not magic numbers that make all the room modes go away, they are simply magic numbers that prevent (to some degree) the room modes from adding up in especially undesireable ways, as often happens in the real world.  Perfect ratios or not, a fantastic amount of properly designed acoustic bass traps would be required to get to truly accurate sound reproduction; even to meet international standards for accurate musical evaluation (based on RT60 and the like) my 1:1.14:1.39 room would have to have about 360 effective sabins of broadband damping, according to the calculator above, and the damping itself can cause problems if not properly engineered and installed for the room at hand. Good luck.  A well designed $500 bass trap can give you about 16 sabins at 50 Hz. while an ineffective trap for $250 might give only 0.5 sabins at 50Hz (I have actual products in mind).   As the bass trap sellers say, you'll run out of room before you run out of the need for more bass traps.  (And money also, but they don't say that.)

Update on KPAC noises

I have gotten the Pioneer 9500 II tuner to work properly by using a direct 300 ohm connection straight from 300 ohm antenna.  For some reason, perhaps opened balun in tuner itself, or lousy cut RG-6 cable, I couldn't get decent performance through the 75 ohm connection (there is no F connector, you have to cut open a cable in a tricky way).

For the last couple of days I've been comparing the 9500 II and my Kenwood KT-6040.  The Pioneer is noticeably noisier overall in stereo, but still acceptable for background or maybe serious listening.  The Kenwood is quieter, but makes the buzz sound much more noticeable.  In A/B comparison, the noise from the Pioneer is about comparable to the buzz from the Kenwood.  Actually I find the buzz intolerable, but on the Kenwood the buzz can be almost eliminated with high blend which doesn't reduce highs (only high frequency separation).  The Pioneer only has a "noise filter" switch which reduces highs enough to bother me.

It seems like the stereo noise from the Pioneer may be simply a less transparent rendition of the buzz.  The Kenwood reproduces it perfectly, whereas the Pioneer blurs it slightly, making the buzz into a peaky sounding hiss which is actually less annoying than the buzz itself.

Thursday, May 19, 2011

More thoughts about KPAC noises

I posted my experiment on KPAC to the Yahoo Group FMTuners, and there has been a lively discussion.


Here's some bits of followup posts I made:

I had trouble finding the Kenwood manual too and was going from lousy memory. If reading at friends house was about 64dB, that means 69.2dB according to the table, not 70 as I reported previously FWIW. I think 85dBf is a bit low for the meter markings to top out, but IIRC you can also extrapolate about 20db higher than the last marking and it's still mostly OK. I think I was pushing 110dBf into it and it was still moving higher though beginning to compress somewhat. This was many years ago and I can't remember exactly. I know at some level I could produce with my ST-1000A the meter started to compress (or maybe the ST-1000A).

Yeah I should probably call the station pretty soon now. Prior to doing this test (which took ten years of procrastination to get around to and still might not have happened) I didn't think I had a very good case. I figured I was much farther from the transmitter (>30 miles not 21) and it was quite likely I had a reception problem.

And furthermore I never noticed the buzz before getting my Kenwood KT-6040 in March 2011, and I think the buzz is not caused by a legal subcarrier, more likely equipment fault like ground loop. The 6040 is said to have adequate subcarrier rejection for modern subcarriers, whereas many of my older tuners do not.

The swishing sound could be a subcarrier thing, I think, perhaps even the result of a fully legal subcarrier (though KPAC does not have HD, and there are no nearby stations with HD either). The KT-6040 (and also Pioneer F-26) eliminate it almost, but not quite, entirely at my home (65dBf), you have to put your ear right to the speaker to hear it with those tuners. I'm just now testing a Pioneer TX-9500II that may have serious problems, and you can't miss the swishing sound, and that is how it has been with many older tuners in the past. Only when you get to the modern era tuners* like Yamaha TX-1000 (or F-26 !) do you get performance not unlike the KT-6040 for which the swishing sound is only a problem for perfectionists like me and takes serious concentration to hear. Perhaps they have some subcarrier which is just slightly beyond specs, or maybe this is "spec" performance.

But with the quiet background of the Pioneer F-26, the buzz which I just started noticing recently is a huge annoyance, though probably average person with average tuner wouldn't notice. Plus, lacking high blend I can't eliminate it as easily as I could with KT-6040. That's what got me moving on this issue. Except for the buzz, I get the most beautiful sound.

Tube tuners like KM-60 also don't project much of the swishing sound. I think that is because they have lousy HF separation, or at least mine does. It's mainly a problem for early transistor tuners where they started pushing the HF separation up, but without having sophisticated MPX circuits or antibirdie filtering.


I don't think KPAC has RDS, though maybe I don't understand RDS. It does not
show any content or category information on my car stereo (2006 Prius with JBL
stereo upgrade) like other stations do. Just 88.3, last I checked.

I've always been suspicious, however, that they stick some non-standard
subcarrier to run or help manage their repeater station KTXI. And classical
radio stations in my experience have often done plain old SCA to make a few
extra bucks.

The funny swishing noises have been there as long as I've been in San Antonio,
1992, but most of these subcarriers existed then anyway if under different

Aren't the rules now for subcarriers pretty wide open? I though they can now
put out whatever they want within the 200kHz station bandwidth, as long as it
doesn't overlap the baseband and stereo modulations at 23-53kHz.

Monday, May 16, 2011

Noises now believed to be station problem, not reception problem

Synopsis: I brought two high performance FM tuners over to friend's house.  He lives near (or at least much nearer to) the KPAC transmitter, and in a different direction.  I heard the same two problems there using his antenna that I hear at home on my antenna: a buzz and a swishing sound.  These noises are fairly low in level (I'm guesstimating -45dB for the buzz and -55dB for the swishing sound) but strong enough to be very annoying on this otherwise wonderful sounding station which plays beautiful music with minimal compression.  I do not hear these sounds on other stations, even lower strength (at my home) college radio stations.  The noises are found in stereo mode only; in mono they go away entirely, and a good high blend feature can reduce them almost entirely.  The signal strengths at both locations (around 65dBf at my house, 69dBf at his) should be quiet enough for very quiet reception in stereo barring interference problems.  The fact that the noises are virtually identical at both locations using top notch equipment is strong evidence that this is a station problem and not an interference problem.  If it were an interference or signal condition problem, I would have expected his location to be rid of these noises.

I brought the Pioneer F-26 tuner because it is the best sounding and most transparent sounding tuner I have.  It is also low noise, perhaps the lowest noise from any tuner I have ever heard that isn't doing any kind of auto blending.  My F-26 was shipped direct to me from the top rated (by Tuner Information Center) tuner restorer, Absolute Sound Labs just a few weeks ago.  It sounds and works perfectly.  It also reveals the buzz sound more clearly than any other tuner...I believe that is because it is so transparent.  Other tuners reveal the buzz but often hard to hear amid hiss and random noises.  The F-26 was said to be one of the 3 greatest tuners of all time (the others being Sansui TU-X1 and Kenwood LO-2t) by tuner expert "anonymous Dave" at Tuner Information Center.  But in detailed notes, he believes the Pioneer has the most transparent midrange of the 3, so you could consider it *the* best.  (Since making that statement a few years ago, he's now onto the latest 2011 Accuphase shown at CES, he now thinks that may be the best.)  With any older tuners like mine, or even brand new ones, the devil is often in the actual condition and alignment.  With a recently restored and aligned unit, still under 6 month warranty, I have that devil handcuffed for awhile.

I brought the Kenwood 600T because it is also a great tuner and because it has the best meters of any tuner in my collection.  It has a signal strength meter calibrated in 10dB units (not exactly dBf, but Kenwood provided a chart for estimating dBf from the Kenwood dB numbers shown).  In my tests with a Sound Technology ST-1000A FM alignment generator, the dB meter is very linear and seemed about as accurate as I could determine.  You can read it to the closest 1 dB by visual interpolation.  It also has multipath and modulation meters.  It reveals the swishing sound a bit better than the Pioneer because it has a exaggerated high frequencies (or maybe a less good 15kHz lowpass filter or less good antibirdie filtering).

Friend lives in 2nd (upper) story in an apartment building, and uses a BIC Beam Box antenna and Pioneer F-9, and uses Carver preamp and poweramp, and some extra Carver gadgets (which were mostly turned off during these tests).  I also brought my own headphones and headphone amplifier (actually a Emu 0404 USB in standalone mode).

We first connected my Pioneer F-26 to his antenna and his stereo system.  It was a bit harder to hear on his AR-9 speakers (less transparent than my Acoustat 1+1's) but after a minute or so I was certain I was hearing the same buzz sound I hear on this station only.  I did not hear the swishing sound (it's always hard to hear on Pioneer F-26 for some reason, possibly because it has excellent SCA rejection.)  Then I connected the headphone amp and headphones to the tuner.  After about 15 minutes of concentrated listing (just before 12 midnight) I suspected I heard the swishing sound, but wasn't totally certain.  Unfortunately my dynamic headphones are not as transparent as my electrostatic speakers.  I also had a hard time hearing the swishing sound on the same headphones at my house (in fact, I didn't hear them at home on headphones for the first 15 minutes or so also).

We then hooked up the Kenwood to his antenna and electronics.  We found the signal level on the Kenwood meters was 60dB (which actually corresponds to about 65dBf).  Trying different settings on the Beam Box, we found that another position gave us 64dB (which would correspond to about 69dBf).  It seemed to make no signifcant difference on the multipath meter, which occasionally bounced a bit from the very bottom for all settings) so we went with that new position for the remaining tests.

The Kenwood also revealed the buzz though it was harder to hear because the Kenwood has considerably more hiss.  After about 10 minutes of non-equalized listening, I was pretty sure now I was hearing the swishing sound as well.  It's very irregular, you have to wait for situations where the music has been slowly fading away to a very low level.  When that happens, it seems the noise level (from the station or recording itself) rises (probably due to some compression at the station...they don't use a lot of compression, but apparently they use some).  When the noise level has risen like this, and if then there is a brief total silence, THEN you can hear the swishing, and even then it's often hard to hear unless your ears are right at the tweeters.

Then my friend engaged his normal EQ, which greatly boosts highs.  At that point I became quite certain I was hearing the swishing sound.  As this setup (AR-9 with boosted highs) seemed more revealing of high frequency problems than my headphones, I didn't bother with using my headphones with the Kenwood.

We didn't get around to using the rabbit ear antenna I had also brought.  My friend thinks his Beam Box Model 10 is better than such antennas, I'm not so sure.  But I don't think it would have made any difference in these tests.  The initial signal strength of 60dB (as per 600T meters) was exactly what I get at home, which I think considerably farther from the transmitter.  (Since writing this, I've determined my house is 21.7 miles from the 100,000 watt transmitter; I'd guesstimate friend's house is 10 miles but haven't gotten his coordinates yet; I had been expecting my home to be farther and his closer.)  Even the 64dB level we got after changing the Beam Box setting was less than you would expect for being 10 miles from a 100,000 watt transmitter (though that corresponds to 70dBf signal level which should be essentially "full quieting" in stereo for a good FM tuner, barring interference problems).

Both of these problems are eliminated by going into mono, and even using high blend (on some tuners, the F-26 has neither mono nor high blend controls) so it must some sort of noise that leaks into the L-R subcarrier at the station or wherever the stereo encoder is.

Investigations will continue.  Once I have digitally recorded and analyzed the noise, at least the continuous buzzing noise, I will send a letter to the station.

I've only actually noticed the buzz sound in the last 4 months when I got a Kenwood KT-6040, which was my favorite tuner for awhile (before I got the Pioneer F-26).  Perhaps I failed to notice it earlier because on less transparent tuners it is not as obvious, and also I had just spend a year listening to the Sony XDR-F1HD which does agressive blending and filtering to remove all noises.  Something like the swishing sound I have noticed for 18 years since I moved to San Antonio.  It does seem that with better tuners and better antennas the swishing sound gets less, while the buzz gets more and more clear.

In very quick testing, it didn't seem like the Carver FM stereo postprocessor TX1-11 (takes L and R audio signals from a tuner and then works some magic on them to reduce noise) made any improvement to either the buzz or the swishing.  Funny because I believe high blend would have made an improvement (though I couldn't test that either, because neither tuner I brought has a high blend switch..but in past tests at home high blend has greatly reduced the noises). Carver does claim the TX1-11 does not reduce stereo separation, so it is not the same as high blend.  This is further evidence, though I wouldn't suggest it's very strong, that the noises are *within* the stereo MPX itself, not something caused by interference or multipath, or else the Carver (if it works as claimed) should have reduced it.

Wednesday, May 4, 2011

Why are digital tuners quieter (well, mostly) than older analogs?

[After listening tests on Sunday and Monday, I posted this to the FMTuners group on Yahoo, one of my favorite haunts.]

Recently I've been using a number of great digital tuners, Kenwood KT-6040 (my
latest fave), Sony XDR-F1HD with modified output stage, Sony 730ES, Yamaha
TX-1000. Just from memory, I'd rate noise (with 60dB signal strength, slightly
suboptimal) from quietest to noisiest (with the last 3 being very close):

XDR-F1HD virtually silent, except for some clicking (impulse noises?)
Yamaha TX-1000
Kenwood KT-6040
Sony 730ES

I most recently acquired the KT-6040, it is a very sweet sounding tuner. At
first really blew me away as more musical sound than XDR-F1HD (but that's true
of the 730ES also). I don't think Yamaha sounds quite as sweet, but it was
quieter when tested last. The XDR-F1HD analog output modification (including
corrected EQ) really improves sound, but it's still overprocessed sounding, not
necessarily the best for serious listening.

All of the above, however, exhibit a periodic "clicking" sound. I'm not sure
what causes this, possibly digital communications of some sort. It happens a
few times every minute.

ALL of the above have less background hiss than the analog tuners I've tested.
But just a few days ago I brought back the Kenwood 600T, which is one of my
quietest analogs, and I noticed in contrast to all of the above...the clicking
sound is simply gone. Very nice! I could get used to that!

Well, that might go along with the more bulletproof front end. The downsides
are a non-ignorable level of background hiss and the sound quality. I mainly
keep the Kenwood for it's wonderful signal strength meter, I don't really much
like the sound of it anymore. In contrast to other tuners, and especially other
Kenwoods (including KT-8300) I have, it sounds warm and slightly diffuse. The
usual Kenwood sound seems to be very clear and open (possibly with excess highs
or even stridency or thinness), which I've noticed in past Kenwoods (KT-917 and
KT-7500) as well and from what other people tell me. But the more open and
musical sounding KT-8300 is simply too noisy for me to enjoy on my 60dB strength

That got me to the question above. It seems like the older tuners, if anything,
had better front ends and IF's. Are the digital tuners quieter because of
detector or MPX? The newer MPX chips (like LM3410) do use some sort of
walsh-function-like or harmonic canceling scheme which older MPX's lacked, and
some digitals go even farther with analog multiplier (such as TX-1000).

If the noise difference is entirely in the MPX, one could just add outboard MPX
to get same effect. But that wouldn't explain differences in mono specs anyway
(though I haven't done much mono listening to tuners...perhaps the mono specs
aren't really that significant under my conditions).

Could difference be in detectors? Newer tuners typically use PLL or quadrature
detectors, don't they? Are those detectors fundamentally quieter than the older
ratio and foster seeley detectors?

Could the difference is in the transistors or chips used in the detectors
getting quieter?

If the quieting were only in the front end, one would think the KT-6040 would be
the ultimate with very few competitors, since it has GaAsFET front end. But it
does not particularly seem quieter than other digitals I have, and may be
surpassed by some. I don't think front end is the biggest factor in background
hiss level, anyway, except for extremely weak stations.

Tale of two antennas

My living room has two indoor antennas near ceiling height, one is near the windows and the other is near the door.  I put the door antenna in last, as a backup for 2nd tuner.  Then it turned out, that it actually seemed better (quite unexpected).  These have been there for several years now, unchanged.

To measure the two antennas this week, I moved Kenwood 600T into living room.  The 600T has a marvelous linear signal strength meter calibrated in 10dB increments.  From tests using a Sound Technology ST-1000A I can say the Kenwood meters are quite accurate, or at least quite linear (since my ST isn't calibrated recently either, and I don't have necessary impedance adapter for total accuracy).

So here are the measurements:

Living Room Door Antenna
88.3     60dB
90.1     43dB
91.7     47dB

Living Room Window Antenna
88.3     59dB
90.1     47dB
91.7     53dB

OK, now it looks like except for the tiny difference to 88.3, the window antenna is better after all.  But the problem is, 88.3 KPAC is my favorite station...

Now I am listening to KPAC, and on window antenna, and it probably isn't as good there.  The background hiss has turned slightly sizzly, and so has the omnipresent birdie (on all tuners to some degree except XDR-F1HD).

All the above stations are to the west, but possibly KPAC is more to the northwest than others, explaining why one antenna is better for some but not others (?).  All antennas are on the eastern wall of the room with same orientation, but perhaps the differences result from different sets of interfering objects.

I've also hooked up a 100Mhz oscilloscope to the scope outputs of the 600T.  The scope does not seem to reveal any meaningful differences...except that possibly the door antenna gives a slightly flatter and nicer trace on the screen than the window antenna.  All scope traces look reasonably good.  The cause of the birdie is not evident.  The best luck I've had with that was when I hooked up spectrum analyzer and observed small signal at 88.5.  There is nothing but noise at 88.5 using every tuner in my collection.

Monday, May 2, 2011

Aha! The best listening position

After finding the bass somewhat better at a listening position slightly farther back, I tried going all the way baaack to my old deep field listening position another 2 feet back.  Even better bass, really enjoying the bass sounds from the DVD-Audio of Santana's Supernatural here.

But even though I could be classified as bass freak, I'd still rather have the closer-up imaging, now that I hear what it can sound like.

So I tried something way different.  I moved the listening chair all the way up front, as close as possible without losing the separate ottoman of my listening chair.

And THAT position has it all, great bass, even more incredible imaging.  Funny it always seemed that moving so close destroyed the center image.  But now, not so.  What gives?  Is it just that I have learned to hear a center image with such a wide angle (greater than 70 degrees I think) now, when it never seemed I could do so before?