I have find some problems for the old amp. The amplifier had at least two problems.
First was a bad contact in the bass potentiometer interrupting the low freqs and then acting like a high pass filter reducing the level for the 440hz tone and making the distortion more obvious.
Second problem was at the speaker protection relay contacts. Here we have three working states: no sound, normal sound and sound distorted.
All these failures are modified by the level of the signal and the vibrations around the amplifier, like the sound from the speaker.
Now that the problems are fixed, I clean all the contacts, I was not able to pass the ABX anymore.
Problem solved.
But the reality, in my experience, is that these failures are very common for old amps and others can experience similar auditions resulting in false ABXs, at least for simply tones.
First was a bad contact in the bass potentiometer interrupting the low freqs and then acting like a high pass filter reducing the level for the 440hz tone and making the distortion more obvious.
Second problem was at the speaker protection relay contacts. Here we have three working states: no sound, normal sound and sound distorted.
All these failures are modified by the level of the signal and the vibrations around the amplifier, like the sound from the speaker.
Now that the problems are fixed, I clean all the contacts, I was not able to pass the ABX anymore.
Problem solved.
But the reality, in my experience, is that these failures are very common for old amps and others can experience similar auditions resulting in false ABXs, at least for simply tones.
Attachments
These are very good comments, both, about pots and relay contacts. Such failures are quite usual. Especially the relay contacts degradation is treacherous. Several years ago I made a decision not to use the power relays for output speaker protection anymore, just for the reason of contact degradation with time and possible arc interruption. The contacts had measurable distortion after some time. I started to use power MOSFETs in DC protection circuit, instead of relays.
This amp under test here has no output relay or MOSFET protection.
The degraded relay contacts may exhibit something like severe crossover distortion, because they start to be conductive no sooner than certain level of voltage and passing through current is reached.
First Texan file is running. This is at normal bias and with an output attenuator chosen to give identical levels between Vin and Vout. The peak level at the amp output was around 8 volts in the short sample I used to test (the end of the Bach).
On speaker relays... I have had the problem and at the time couldn't quite believe it. I was using an SPL meter and pure sine test signal, and one channel was highly distorted, so bad I thought the speaker was faulty and the midrange cone rubbing. It was really bad, horrendous. I went back to music and was not aware of anything wrong. Back to the sine and there it was again. I turned the level up more and more and suddenly it cleared and I knew instantly what it was. That was the point I abandoned all mechanical switching for speaker outputs (the same applies for low level signal switching as well, I would never go back to mechanical contacts).
On speaker relays... I have had the problem and at the time couldn't quite believe it. I was using an SPL meter and pure sine test signal, and one channel was highly distorted, so bad I thought the speaker was faulty and the midrange cone rubbing. It was really bad, horrendous. I went back to music and was not aware of anything wrong. Back to the sine and there it was again. I turned the level up more and more and suddenly it cleared and I knew instantly what it was. That was the point I abandoned all mechanical switching for speaker outputs (the same applies for low level signal switching as well, I would never go back to mechanical contacts).
Attachments
Here we are then. One file is one of the original files (tocl or tocr, you don't know which) converted to 16/44.1 and the other file is via the Texan running normal bias. The final second on both is cropped because I looked at the properties of the original and it was listed as 1'54" and set the simulation to run for 114 seconds but it seems its actually nearer 1'55".
(although LT seemed to accept a 48k input file, it does not seem to output to a 48k file. The sim appears to run but stalls part way through leaving the output file blank. I recall reading in the LT manual by Gilles Brocard that it has to be 16/44.1)
TEXAN BACH
(although LT seemed to accept a 48k input file, it does not seem to output to a 48k file. The sim appears to run but stalls part way through leaving the output file blank. I recall reading in the LT manual by Gilles Brocard that it has to be 16/44.1)
TEXAN BACH
Low bias Texan. As before the original file could either tocl or tocr 
Haven't listened to these as I have to go out but they look OK and the alignment looks good.
Low Bias Texan Bach
Haven't listened to these as I have to go out but they look OK and the alignment looks good.
Low Bias Texan Bach
At least something now. 6/8, 7/8, no more trials, I will try later, now I am exhausted
The file difference is still extremely low in overall level, but now it is a pure and ugly sounding crossover distortion. Good job, Karl.
Code:
foo_abx 2.0.2 report
foobar2000 v1.3.7
2017-11-09 17:01:53
File A: Gamma.wav
SHA1: 8f6a78d92a6752ea8130de2bd3017537bace4b1f
File B: Kappa.wav
SHA1: df779c38529b3fbad3ae70fdd6233e301ad747c7
Used DSPs:
Resampler (PPHS)
Output:
WASAPI (event) : Speaker (USB Sound Blaster HD), 24-bit
Crossfading: NO
17:01:53 : Test started.
17:02:39 : 01/01
17:02:52 : 02/02
17:03:06 : 03/03
17:03:23 : 04/04
17:03:36 : 05/05
17:03:46 : 06/06
17:03:55 : 06/07
17:04:18 : 07/08
17:04:18 : Test finished.
----------
Total: 7/8
Probability that you were guessing: 3.5%
-- signature --
4d4c80c0ac4ca70c3989c2a8f0b0f612cc5e3b41The file difference is still extremely low in overall level, but now it is a pure and ugly sounding crossover distortion. Good job, Karl.
This is the crossover+intermod distortion extracted from Mooly's latest gamma and kappa test files, 36dB amplified (63x) to be clearly audible.
http://pmacura.cz/xovergammakappa.zip
http://pmacura.cz/xovergammakappa.zip
Great result Pavel That distortion file is intriguing, when I play that on my laptop speakers and move my head to the right, the image goes fully to the left and vice versa. Bizarre !
Foobar is not smiling on me today
You have to try very hard indeed to get a result like this
That distortion file is intriguing, when I play that on my laptop speakers and move my head to the right, the image goes fully to the left and vice versa. Bizarre !Foobar is not smiling on me today
You have to try very hard indeed to get a result like this
Attachments
Thanks for the xovergammakappa track Pavel, this is more closely like I thought it was the crossover distortion. But this time we have the total distortion as you say.
Low bias Texan. As before the original file could either tocl or tocr
Haven't listened to these as I have to go out but they look OK and the alignment looks good.
Low Bias Texan Bach
Good job Karl!
What about THD and IMD?
Can you post distortion numbers for the normal Texan and the low bias tracks please?
I can not discern between Alpha and Beta, not from speakers or headphones, arbitrary results.
With speakers and at 24.5 to 26 seconds I can hear little whisper (tickling in my ear) coming from the tweeter for Gamma track.
With headphones, as other times, I'm sure that hear it but I not pass ABX.
This is for Gamma vs Kappa ABX test:
Code:
foo_abx 2.0.4 report
foobar2000 v1.3.16
2017-11-09 18:49:28
File A: Gamma.wav
SHA1: 8f6a78d92a6752ea8130de2bd3017537bace4b1f
File B: Kappa.wav
SHA1: df779c38529b3fbad3ae70fdd6233e301ad747c7
Output:
DS : Headphone (VIA HD Audio)
Crossfading: NO
18:49:28 : Test started.
18:59:37 : 01/01
18:59:47 : 02/02
19:00:03 : 03/03
19:00:18 : 04/04
19:00:43 : 05/05
19:00:51 : 06/06
19:01:10 : 07/07
19:01:21 : 08/08
19:01:59 : 09/09
19:02:46 : 10/10
19:02:56 : 11/11
19:05:34 : 12/12
19:05:34 : Test finished.
----------
Total: 12/12
Probability that you were guessing: 0.0%
-- signature --
2d4705da6009dae9db70ccd9d05d09aaf6ea74f0Another great result 
Here is the distortion for the Texan at low and normal bias.
I do wonder how realistic the Texan simulation is with regard to using the LT1001 opamp model rather than using a full discrete model of the 741.
However
I've still got my Rotel files to post (that's a real amp, not a sim). I'll get cracking and write that up.
Here is the distortion for the Texan at low and normal bias.
I do wonder how realistic the Texan simulation is with regard to using the LT1001 opamp model rather than using a full discrete model of the 741.
However
I've still got my Rotel files to post (that's a real amp, not a sim). I'll get cracking and write that up.
Attachments
These are very good questions, Dan. The original tests are in the 1st post of the thread, together with a question what you should do. The original test was recorded with OPA549 chip amplifier and the goal was to discern its crossover distortion in the foobar ABX test. The original files were in 44.1/16 (Steely Dan). For the reason that there were no statistically valid positive test result with these first test files, we started to explore audibility of the distortion of the individual 1kHz and 440Hz tone. The 440Hz tone test is at
http://pmacura.cz/440_549.zip
As we did not get a positive result even with the sine files, Mooly suggested to create and created some test files through LT Spice simulated Texan amplifier, set for normal and low bias, both creating crossover distortion, worse with low bias. Distorted 1KHz tone was easily discernible, but Steely Dan music not.
So I prepared 48/24 files with violin sound, linked again in the 1st post of the thread. We were not able to discern the OPA549 distortion again, and I asked Mooly to prepare this violin test through the simulated Texan. At normal bias, we did not hear the distortion, at low bias, some of us heard the distortion. The latest Mooly's files are in posts
normal bias
Test your ears in my new ABX test
low bias
Test your ears in my new ABX test
You or other members are asked to listen to the files and post a Foobar ABX protocol, in case they have a positive test result. The protocol should be the complete Foobar generated txt file with digital signature, for validity verification purposes.
I will add Mooly's files to the 1st post later.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.