I've already started on new, revolutionary (for me) projects: Using a triad of miniDSP processors with AES digital in and out, I'm going to re-implement my crossovers as linear phase, and possibly correct other phase aspects of my system, by using the non-causal filters which can be implemented with FIR. In principle, the miniDSP's could replace my Behringer DEQ's, but my present plan is to keep both in series, so I display and input more ad hoc "causal" filters (the ordinary kind which can also be implemented with analog electronics) using the Behringer. Since it's all 2496 AES digital, the only loss from having more gizmos inline is jitter, and as I have previously explained, having more quality components inline hardly affects jitter at all, and it's almost completely unimportant anyway at the levels my system would reach with many more such gizmos.
And that gets to another theme I really want to take up, the fake importance of "simplicity." What actually matters is quality, simple crap is still crap. Usually, in order to get higher quality, higher and better aimed effort is required. This effort can go into sorting the simple crap out, finding the very best pieces. Or it can go into trying to use each piece in such a way that it's limitations are fixed by other pieces. Either way is a valid approach to some degree. Though, in my estimation, at the level audio products are now, some complication, such as multiway speakers, is required to get decent quality. Others may disagree. And so what? But I think, more often than not, simplicity surfaces as "you need simplicity," which is either a nag, or a sales pitch, and both of those are just annoying. The sales pitch angle of this is obvious. Sell all your previous crap, and buy this one new (vastly more expensive, of course) piece to replace it all. It's an old line. Whence the expression, "don't spend all your money in one place." But, that is what they are always, and still, trying to get you to do.
And I've started to build a First Watt F5, a Class A amp for my supertweeters (which still seem to add something important...actually by my most recent tuning they are, barely, audible, and this gets me to flat response to 20kHz on my iPhone RTA. Since I'm using damped cloth dome tweeters which happen to have 40kHz response, there's no added "metallic" quality when I do this.
And since I got it back from repair and upgrades, the Krell has been operating perfectly, better than it ever had before. It had always had an overheating left channel--it simply went to 180F at nearly any playback level. That problem is gone, both channels now track within a few degrees, the remaining variation probably caused by airflow. And the clinks have all but disappeared, especially since I readjusted the air conditioning registers on the opposite side of the room to dump all the air before the center of the room (though avoiding the couch) .
But, a few days ago, I decided to play the Reference Recordings HRx sampler, which has music recorded at 176kHz sampling rate in high rez files which can be read by a player like my new Oppo BDP-205. I resample the analog output of the Oppo routed through the Emotiva XSP-1 at -2.5dB into the Lavry AD10 set to peak level 3.88V, the Tact is set to 93.3, and the Emotiva Stealth DAC is set to -7dB for the Krell and connected balanced. (The correct matching level for the Hafler in -3.75, I discovered later, having previously set it to -4.75).
I'm calling the last two levels my current "standard high level." It's about as loud as you would want to play anything, but not clipping the amp (at least, it's not clipping the Krell).
And, just like back in February, the Krell shut down on track 5, the Satie, which curiously was far lower in level than the previous track.
The next night I played the disk again, starting from the beginning, and starting from a cold amplifier again (in past experiments, that has been critical to reproducing the same failures), and it shut down at the same point in the same song. This time, however, I had shut down the air conditioning completely so no air was blowing. I had been thinking I could possibly cure this shutdown problem by moving the amp completely out of the possible air stream, behind the left speaker, where the Hafler and Aragon are now, and bring those amps to the front, where the Krell is. But, since the exact same shutdown occurred with the AC fully turned off, it doesn't look like that would help.
Since then, I've quit bothering shutting the AC down, since it doesn't seem to make a difference, and instead tried disconnecting the right speaker, then disconnecting the right speaker.
With only the left speaker playing, the shutdown still occurred, but in track 7.
With only the right speaker playing, the shutdown did not occur in the entire album.
Though I haven't totally eliminated other variables, I'd say either the left Acoustat, or the left channel of the Krell is at fault, and the next tests should isolate those possibilities. In some sense actually I mean that the Krell is at fault, in the sense that other amplifiers work just fine possibly by ignoring to weird load of the Acoustat. But that's the given that I just have to work around, I don't know why, I like the sound the Krell, the looks, etc.
After playing the Right channel for hours on the right Acoustat, both channels of the Krell had heated up to around 150 (the unused Left channel hotter, actually, driving all the speaker cable capacitance in Class A, and it usually runs slightly hotter anyway), I played both channels, now that they were warmed up, and the shutdown did not occur.
Monday
I've been having an argument with a friend about which tests to do and what they mean. Without getting into the details, this shows that even totally objective testing (you can hardly get more "objective" than an amplifier shutting down, or not) is extremely complex and difficult. The kinds of ad hoc subjective tests that audiophiles generally rely on, are worse.
On Sunday, whether this was a optimal test to do or not, I reversed all connections at the amplifier so that the right amplifier channel was getting the left inputs and playing the left speaker, and vice versa. For all of the subsequent tests, I played the entire disk (FWIW).
Playing the right speaker alone (on the left channel of the amp, but with the right input signal) did not cause shutdown, nor did playing the left speaker alone (on the right channel of the amp), nor did playing both left and right channels on the opposite channels of the amplifier. In each case I waited 6 hours or more for the amp to cool down to normal "stand-by" temperatures, however this might only take an hour or so, standby temperature seems to be about 10F above ambient because it's running 60W, though it doesn't seem like 60W should run that warm.
But right after doing playing the right speaker (on the left amp channel, with right input) I noticed something. I had somehow put the Tact preamplifier in Correction 9 by mistake. I have no idea what Correction 9 does, I mainly use the "corrections" to do measurements since you must select a target correction curve even to do a measurement with the Tact. I could have been measuring the tweeters, I could have been measuring a wire for calibration purposes. Anyway, having the Tact in a unknown correction could by itself cause an amplifier shutdown problem. So, at this point it could be this whole failure episode is nothing but a false alarm. I'll now have to go back and retry the original test with all channels normal. I also do not know when Correction 9 had been selected, it could have been selected on Saturday night after I finished amplifier testing (since the amp had warmed up anyway) and started playing other things, and curiously enough found another test that seemed to show the audibility of polarity (despite my belief it is barely audible, but in this case the MP3 conversion was terrible and possibly non-linear, making polarity changes more audible). I got very involved with that polarity testing, and it's possible that is when I accidentally selected Correction 9, which is very easy to do by accident using the Tact--the correction selector is more prominent than the volume control and if you are not looking closely, you might be selecting a correction rather than setting the volume.
I proceeded to do the "opposite" test, left speaker on the "right" amp channel, with left input, six hours later (to let the amp completely cool down in standby, however 90 minutes might have been sufficient, I was fooled because it seems the standby temperature of the amp is 10F above ambient).
No shutdown.
The next morning I put both speakers on at once in this configuration, and there was still no shutdown.
Finally, that evening, I put everything back to the starting condition, and turned the air conditioning off because that seems to be a highly variable influence, and one that's not needed for the shutdown to occur. But, still, now there was no shutdown.
Already I was beginning to think of something I brushed over near the beginning.
After the first three shutdowns (the first, and two tests) was the first time I removed the left speaker banana to disconnect that speaker (because in the previous test, I had removed the other banana). Right then, I noticed that the knurled knobs holding the thick (2 x 11 gauge) wires into the RCA Gold bananas had come loose. Somewhat loose...they still basically held onto the wire. I had long since resolved to change them. I was planning to replace them the upcoming weekend, but the audio society party, and further amplifier/speaker testing had priority. But right then, I just re-tightened them, that usually held for at least a few weeks. Somehow, those big knurled knobs often come loose, while smaller set screws tightened with a tool almost never do so. I should have known that when I started using these Gold bananas, but caved because they were so easy to use and took very big wires.
Well, if you've been following every line, you will see that there was no failure in any subsequent test, going back to the very first experience. Whereas, it seemed the original problem was quite reproduceable, it happened 3 times in a row on the same recording at the same level.
So it appears this connector was at fault, and a short intermittant in connecting to the Acoustats seems to make the paranoid amplifier think there is a short. (I'm thinking of calling it Marvin, the paranoid amplifier.)
On at least the first two shutdowns, it occurred some seconds after playing the Walton track 4 on the Reference Recordings HRx sampler. Well that Walton is full of very impressive big BOOMs. Those could possibly shake the speaker (not just from the speaker itself, but the subwoofer right nearby). That invisible shaking was possibly just enough, a few seconds later, right as the speaker was settling back into it's normal condition, hit a non-conductive spot on the bare copper wire, causing an instantaneous "open."
Thinking about it now, this might even explain some shutdowns last year. But not the very first, because I was still using the Canare 4S11, which had terminated ends (unless...I had already removed those ends for some reason...those Canare's have been used in several different ways since I bought them).
I was certainly using the Radio Shack bananas from the time I set up the ABX system in January. However, by that time, I had fixed the "shutdown" problem by making repairs to the speakers. It was about that time, however, that the overheating problem materialized, and that couldn't happen from the connectors "opening" up...unless very rapidly.
Anyway, lessons learned:
1) Of course, don't use those Radio Shack Gold jacks! Or re-tighten every week. But don't count on remembering that, so just don't use them. This is ESPECIALLY true for the speaker connection, because speakers are known to have significant vibration and movement. Actually, I don't think I have ever had a problem with these connectors on amplifiers, and I think I may still be using them on the tweeter amp and bedroom amps, without issue, for years.
2) On the speakers, use the best connectors! I installed Furutech locking bananas on Tuesday. I had recently removed them from some other connection; I hadn't had enough of them to cover all connections when I set up the ABX in January.
3) When doing tests, don't overlook tiny changes. If you change ANYTHING not related to the test protocol, because it needed changing, consider that change worthy of an immediate test.
4) Think about other things, about everything.
5) Temporary "opens" in the Krell to Acoustat wiring can cause the Krell to sense short and shut down. So, if a shutdown occurs, that should be one of the first things to check...the tighness of the speaker connections. The Manual actually says to check the speaker connections if a shutdown occurs, but I always thought they meant to see if those wires were shorting, not losing contact.
6) It may be good, in some way, that the Krell FPB is so paranoid or "finnicky," because it forces you to fix things and do them the right way. Fixing things like intermittent connecters improves the sound, also, even for the amps that don't shut down.
But it's hard to feel that way, when Wednesday night, after just playing some Steely Dan, the amp was cooling down, and before I could play something else, had shut down again.
And, 3 times in a row, shutdown, actually as I was attempting to play Buffalo Springfield about 9dB below my standard level.
So, the moral is, Don't Draw Conclusions! It may not be over yet.
Maybe the Radio Shack Gold jacks are not guilty as charged, but I would not use again on the Acoustats.
I may be mainly using my Hafler. In fact, when I'm playing Krell, and it shuts down, I should not pout, just switch over to the Hafler as I can do with a press of the "B" button on my ABX.
[The saga continues in a later post. I've discovered now that doing an Italian Tuneup fixes the Krell, at least temporarily.]
And that gets to another theme I really want to take up, the fake importance of "simplicity." What actually matters is quality, simple crap is still crap. Usually, in order to get higher quality, higher and better aimed effort is required. This effort can go into sorting the simple crap out, finding the very best pieces. Or it can go into trying to use each piece in such a way that it's limitations are fixed by other pieces. Either way is a valid approach to some degree. Though, in my estimation, at the level audio products are now, some complication, such as multiway speakers, is required to get decent quality. Others may disagree. And so what? But I think, more often than not, simplicity surfaces as "you need simplicity," which is either a nag, or a sales pitch, and both of those are just annoying. The sales pitch angle of this is obvious. Sell all your previous crap, and buy this one new (vastly more expensive, of course) piece to replace it all. It's an old line. Whence the expression, "don't spend all your money in one place." But, that is what they are always, and still, trying to get you to do.
And I've started to build a First Watt F5, a Class A amp for my supertweeters (which still seem to add something important...actually by my most recent tuning they are, barely, audible, and this gets me to flat response to 20kHz on my iPhone RTA. Since I'm using damped cloth dome tweeters which happen to have 40kHz response, there's no added "metallic" quality when I do this.
And since I got it back from repair and upgrades, the Krell has been operating perfectly, better than it ever had before. It had always had an overheating left channel--it simply went to 180F at nearly any playback level. That problem is gone, both channels now track within a few degrees, the remaining variation probably caused by airflow. And the clinks have all but disappeared, especially since I readjusted the air conditioning registers on the opposite side of the room to dump all the air before the center of the room (though avoiding the couch) .
But, a few days ago, I decided to play the Reference Recordings HRx sampler, which has music recorded at 176kHz sampling rate in high rez files which can be read by a player like my new Oppo BDP-205. I resample the analog output of the Oppo routed through the Emotiva XSP-1 at -2.5dB into the Lavry AD10 set to peak level 3.88V, the Tact is set to 93.3, and the Emotiva Stealth DAC is set to -7dB for the Krell and connected balanced. (The correct matching level for the Hafler in -3.75, I discovered later, having previously set it to -4.75).
I'm calling the last two levels my current "standard high level." It's about as loud as you would want to play anything, but not clipping the amp (at least, it's not clipping the Krell).
And, just like back in February, the Krell shut down on track 5, the Satie, which curiously was far lower in level than the previous track.
The next night I played the disk again, starting from the beginning, and starting from a cold amplifier again (in past experiments, that has been critical to reproducing the same failures), and it shut down at the same point in the same song. This time, however, I had shut down the air conditioning completely so no air was blowing. I had been thinking I could possibly cure this shutdown problem by moving the amp completely out of the possible air stream, behind the left speaker, where the Hafler and Aragon are now, and bring those amps to the front, where the Krell is. But, since the exact same shutdown occurred with the AC fully turned off, it doesn't look like that would help.
Since then, I've quit bothering shutting the AC down, since it doesn't seem to make a difference, and instead tried disconnecting the right speaker, then disconnecting the right speaker.
With only the left speaker playing, the shutdown still occurred, but in track 7.
With only the right speaker playing, the shutdown did not occur in the entire album.
Though I haven't totally eliminated other variables, I'd say either the left Acoustat, or the left channel of the Krell is at fault, and the next tests should isolate those possibilities. In some sense actually I mean that the Krell is at fault, in the sense that other amplifiers work just fine possibly by ignoring to weird load of the Acoustat. But that's the given that I just have to work around, I don't know why, I like the sound the Krell, the looks, etc.
After playing the Right channel for hours on the right Acoustat, both channels of the Krell had heated up to around 150 (the unused Left channel hotter, actually, driving all the speaker cable capacitance in Class A, and it usually runs slightly hotter anyway), I played both channels, now that they were warmed up, and the shutdown did not occur.
Monday
I've been having an argument with a friend about which tests to do and what they mean. Without getting into the details, this shows that even totally objective testing (you can hardly get more "objective" than an amplifier shutting down, or not) is extremely complex and difficult. The kinds of ad hoc subjective tests that audiophiles generally rely on, are worse.
On Sunday, whether this was a optimal test to do or not, I reversed all connections at the amplifier so that the right amplifier channel was getting the left inputs and playing the left speaker, and vice versa. For all of the subsequent tests, I played the entire disk (FWIW).
Playing the right speaker alone (on the left channel of the amp, but with the right input signal) did not cause shutdown, nor did playing the left speaker alone (on the right channel of the amp), nor did playing both left and right channels on the opposite channels of the amplifier. In each case I waited 6 hours or more for the amp to cool down to normal "stand-by" temperatures, however this might only take an hour or so, standby temperature seems to be about 10F above ambient because it's running 60W, though it doesn't seem like 60W should run that warm.
But right after doing playing the right speaker (on the left amp channel, with right input) I noticed something. I had somehow put the Tact preamplifier in Correction 9 by mistake. I have no idea what Correction 9 does, I mainly use the "corrections" to do measurements since you must select a target correction curve even to do a measurement with the Tact. I could have been measuring the tweeters, I could have been measuring a wire for calibration purposes. Anyway, having the Tact in a unknown correction could by itself cause an amplifier shutdown problem. So, at this point it could be this whole failure episode is nothing but a false alarm. I'll now have to go back and retry the original test with all channels normal. I also do not know when Correction 9 had been selected, it could have been selected on Saturday night after I finished amplifier testing (since the amp had warmed up anyway) and started playing other things, and curiously enough found another test that seemed to show the audibility of polarity (despite my belief it is barely audible, but in this case the MP3 conversion was terrible and possibly non-linear, making polarity changes more audible). I got very involved with that polarity testing, and it's possible that is when I accidentally selected Correction 9, which is very easy to do by accident using the Tact--the correction selector is more prominent than the volume control and if you are not looking closely, you might be selecting a correction rather than setting the volume.
I proceeded to do the "opposite" test, left speaker on the "right" amp channel, with left input, six hours later (to let the amp completely cool down in standby, however 90 minutes might have been sufficient, I was fooled because it seems the standby temperature of the amp is 10F above ambient).
No shutdown.
The next morning I put both speakers on at once in this configuration, and there was still no shutdown.
Finally, that evening, I put everything back to the starting condition, and turned the air conditioning off because that seems to be a highly variable influence, and one that's not needed for the shutdown to occur. But, still, now there was no shutdown.
Already I was beginning to think of something I brushed over near the beginning.
After the first three shutdowns (the first, and two tests) was the first time I removed the left speaker banana to disconnect that speaker (because in the previous test, I had removed the other banana). Right then, I noticed that the knurled knobs holding the thick (2 x 11 gauge) wires into the RCA Gold bananas had come loose. Somewhat loose...they still basically held onto the wire. I had long since resolved to change them. I was planning to replace them the upcoming weekend, but the audio society party, and further amplifier/speaker testing had priority. But right then, I just re-tightened them, that usually held for at least a few weeks. Somehow, those big knurled knobs often come loose, while smaller set screws tightened with a tool almost never do so. I should have known that when I started using these Gold bananas, but caved because they were so easy to use and took very big wires.
Well, if you've been following every line, you will see that there was no failure in any subsequent test, going back to the very first experience. Whereas, it seemed the original problem was quite reproduceable, it happened 3 times in a row on the same recording at the same level.
So it appears this connector was at fault, and a short intermittant in connecting to the Acoustats seems to make the paranoid amplifier think there is a short. (I'm thinking of calling it Marvin, the paranoid amplifier.)
On at least the first two shutdowns, it occurred some seconds after playing the Walton track 4 on the Reference Recordings HRx sampler. Well that Walton is full of very impressive big BOOMs. Those could possibly shake the speaker (not just from the speaker itself, but the subwoofer right nearby). That invisible shaking was possibly just enough, a few seconds later, right as the speaker was settling back into it's normal condition, hit a non-conductive spot on the bare copper wire, causing an instantaneous "open."
Thinking about it now, this might even explain some shutdowns last year. But not the very first, because I was still using the Canare 4S11, which had terminated ends (unless...I had already removed those ends for some reason...those Canare's have been used in several different ways since I bought them).
I was certainly using the Radio Shack bananas from the time I set up the ABX system in January. However, by that time, I had fixed the "shutdown" problem by making repairs to the speakers. It was about that time, however, that the overheating problem materialized, and that couldn't happen from the connectors "opening" up...unless very rapidly.
Anyway, lessons learned:
1) Of course, don't use those Radio Shack Gold jacks! Or re-tighten every week. But don't count on remembering that, so just don't use them. This is ESPECIALLY true for the speaker connection, because speakers are known to have significant vibration and movement. Actually, I don't think I have ever had a problem with these connectors on amplifiers, and I think I may still be using them on the tweeter amp and bedroom amps, without issue, for years.
2) On the speakers, use the best connectors! I installed Furutech locking bananas on Tuesday. I had recently removed them from some other connection; I hadn't had enough of them to cover all connections when I set up the ABX in January.
3) When doing tests, don't overlook tiny changes. If you change ANYTHING not related to the test protocol, because it needed changing, consider that change worthy of an immediate test.
4) Think about other things, about everything.
5) Temporary "opens" in the Krell to Acoustat wiring can cause the Krell to sense short and shut down. So, if a shutdown occurs, that should be one of the first things to check...the tighness of the speaker connections. The Manual actually says to check the speaker connections if a shutdown occurs, but I always thought they meant to see if those wires were shorting, not losing contact.
6) It may be good, in some way, that the Krell FPB is so paranoid or "finnicky," because it forces you to fix things and do them the right way. Fixing things like intermittent connecters improves the sound, also, even for the amps that don't shut down.
But it's hard to feel that way, when Wednesday night, after just playing some Steely Dan, the amp was cooling down, and before I could play something else, had shut down again.
And, 3 times in a row, shutdown, actually as I was attempting to play Buffalo Springfield about 9dB below my standard level.
So, the moral is, Don't Draw Conclusions! It may not be over yet.
Maybe the Radio Shack Gold jacks are not guilty as charged, but I would not use again on the Acoustats.
I may be mainly using my Hafler. In fact, when I'm playing Krell, and it shuts down, I should not pout, just switch over to the Hafler as I can do with a press of the "B" button on my ABX.
[The saga continues in a later post. I've discovered now that doing an Italian Tuneup fixes the Krell, at least temporarily.]
No comments:
Post a Comment