Distortion Measurements - Anyone?
Hi,
Can someone who's done a small set of distortion measurements please post them? Just a few values at perhaps 1KHz, 5KHz, 10KHz, 20KHz. (or even at 1KHz and 15KHz) First pass at about 8VRMS output (8 ohm load) and a second pass at something close to 13.5VRMS output.
It would be greatly appreciated.
Also:
For anyone curious about the offset and bias voltage gradient, they are both negative with respect to temperature. After allowing heat sink get to about 45C, and then final adjust on bias and offset, I let it cool down completely. Upon starting it up, offst was at 40mV, bias at 450mV. After 10 minutes offset was 0, bias was at 430mV (where it was at shutdown).
Gary
Hi,
Can someone who's done a small set of distortion measurements please post them? Just a few values at perhaps 1KHz, 5KHz, 10KHz, 20KHz. (or even at 1KHz and 15KHz) First pass at about 8VRMS output (8 ohm load) and a second pass at something close to 13.5VRMS output.
It would be greatly appreciated.
Also:
For anyone curious about the offset and bias voltage gradient, they are both negative with respect to temperature. After allowing heat sink get to about 45C, and then final adjust on bias and offset, I let it cool down completely. Upon starting it up, offst was at 40mV, bias at 450mV. After 10 minutes offset was 0, bias was at 430mV (where it was at shutdown).
Gary
High Distortion
Hoping one of the resident AlephJ experts can help me on this one.
Attached is a picture of my test setup. Not seen is my HP8903B analyzer. It self tests at 0.004% or lower from 20Hz to 100KHz, so it's working fine.
Voltage on rails = +/-27.2V loaded
Load = 8ohms (250W Arcol non-inductive resistor)
Voffset at 0V
Vbias at 400mV (initial)
Voltage across R6(560 ohms) is 4.5V
All IRFP240's matched from same lot. Vgs across all devices operating 4.37V
The "test" heatsink's temperature is about 45C.
Applying a 1KHz signal of 1V RMS results in an output just under 10V (9.62VRMS) about 12W, seems about right -19.6dB gain.
However, distortion is 0.15% at this low level. At 10KHz, it goes to 0.2%. At 20KHz, it's at 0.42%
When I bring the input to 1.4V output just under 14V, about 23W. Distortion at this point is 0.22%. I didn't bother checking the higher frequencies.
Looking at the FFT of the output, the 3rd harmonic is responsible for most of the distortion at higher power, and especially at anything over 3KHz.
I remove and short the input, and readjust bias for 500mV across source resistor. Reapplying signals as above, the distortion at 1V input is still too high, at 0.08%.
The power supply is clean. At no signal input, the ripple is just 40mV. The input 2SJ74BL pair is matched at Idss = 8.2mA.
As I'm using single ended input, I've shorted the input "-" to ground. Using a shielded cable.
Any suggestions...ideas...would be helpful and appreciated.
Hoping one of the resident AlephJ experts can help me on this one.
Attached is a picture of my test setup. Not seen is my HP8903B analyzer. It self tests at 0.004% or lower from 20Hz to 100KHz, so it's working fine.
Voltage on rails = +/-27.2V loaded
Load = 8ohms (250W Arcol non-inductive resistor)
Voffset at 0V
Vbias at 400mV (initial)
Voltage across R6(560 ohms) is 4.5V
All IRFP240's matched from same lot. Vgs across all devices operating 4.37V
The "test" heatsink's temperature is about 45C.
Applying a 1KHz signal of 1V RMS results in an output just under 10V (9.62VRMS) about 12W, seems about right -19.6dB gain.
However, distortion is 0.15% at this low level. At 10KHz, it goes to 0.2%. At 20KHz, it's at 0.42%
When I bring the input to 1.4V output just under 14V, about 23W. Distortion at this point is 0.22%. I didn't bother checking the higher frequencies.
Looking at the FFT of the output, the 3rd harmonic is responsible for most of the distortion at higher power, and especially at anything over 3KHz.
I remove and short the input, and readjust bias for 500mV across source resistor. Reapplying signals as above, the distortion at 1V input is still too high, at 0.08%.
The power supply is clean. At no signal input, the ripple is just 40mV. The input 2SJ74BL pair is matched at Idss = 8.2mA.
As I'm using single ended input, I've shorted the input "-" to ground. Using a shielded cable.
Any suggestions...ideas...would be helpful and appreciated.
Attachments
Somewhere between my journey of F5 then F4 and now A'J I picked up a lot of hum in both channels. even though layout was exactly the same and the only boards swapped were amp boards. I then implemented a star ground after the CL60, without touching the chassis and that fixed it all. So I am wondering if I should re-bias ?
With the diyaudio 4U chassis, can it take more bias and would there be any benefit going above 400mV? Jim mentions going 500mV in his 5U and if in 4U a fan is needed. I dont want to go near that but curious if any gain to upping bias.
Roger57 mentioned shorting input and I noticed that in some circuits this is needed when biasing......is it as simple as connecting "-" to ground?
Thx
Neville
With the diyaudio 4U chassis, can it take more bias and would there be any benefit going above 400mV? Jim mentions going 500mV in his 5U and if in 4U a fan is needed. I dont want to go near that but curious if any gain to upping bias.
Roger57 mentioned shorting input and I noticed that in some circuits this is needed when biasing......is it as simple as connecting "-" to ground?
Thx
Neville
Hi Roger57,
There is a distortion vs power graph (@1KHz) on the last pasge of the AlephJ manual:
http://firstwatt.com/pdf/prod_aj_man.pdf
The graph shows a distortion of about 0.2% @10W.
There's also a graph for distortion vs frequency at 1W.
Cheers,
Dennis
There is a distortion vs power graph (@1KHz) on the last pasge of the AlephJ manual:
http://firstwatt.com/pdf/prod_aj_man.pdf
The graph shows a distortion of about 0.2% @10W.
There's also a graph for distortion vs frequency at 1W.
Cheers,
Dennis
Assuming there's no stray voltage being picked up at the input stage, you can just disconnect the input. On any amp I test, I always short the input, and have an RCA made for that purpose.
Still not sure what you are asking, your distortions measurements seem really quite good...
Dennis,
After I posted last night, and realizing I must have missed something, I looked up the Aleph J manual on firstwatt. After looking at the graphs, I did not know the Aleph J had those levels of distortion. Didn't RTFM. Although my 8903B provides an accurate distortion measurement, it's THD. So far, I've looked only at my scope's FFT output which doesn't have real good resolution. Over the next day, I'll use my ARTA/M-Audio system to provide a more accurate FFT of the output at various power levels, and post them here.
Thanks for your reply.
2 questions:
To which procedure are you referring when you say AC bias? The ACS? As in this page?
http://www.diyaudio.com/forums/pass-labs/51808-more-aleph-2-questions.html
If these distortion measurements are about right, what's the point of adjusting AC bias?
Thanks for asking.
As I mentioned to Dennis in a previous post, my screw-up, I didn't read the manual on the Aleph J, so it's come as a bit of a "cold water in the face" reality check.
In your build guide, I missed the little snippet regarding the power level you used in your 0.045% result at 1KHz (1 watt)
I was hoping to get something at least 20W or more with this build at reasonable distortion. "Reasonable" distortion is ideally all second order (as my tube amps have) But right now, I'm concerned about the third order harmonic overpowering the second at any power level over about 15W output.
Will this be a concern? Maybe not as I was planning on building a pair to bi-amp my 803D's. They are only 90dB efficient. With bi-amping, and at reasonable power levels, perhaps I'll never hear that 3rd harmonic wiping out the 2nd...
Papa commented on the transition from second to third order harmonics in the BA-3
article (near the end of the "BA-3" subsection):
http://www.diyaudio.com/forums/diyaudio-com-articles/194809-burning-amplifier-ba-3-a.html
Cheers,
Dennis
article (near the end of the "BA-3" subsection):
http://www.diyaudio.com/forums/diyaudio-com-articles/194809-burning-amplifier-ba-3-a.html
Cheers,
Dennis
I could have answered your other question but missed it. Regarding bias, and as Jim pointed out in his original build guide, there is an advantage to higher bias, to a point. The distortion does improve, and substantially, but at the cost of power dissipation. If you're running 24V rails, just 75mV more across source resistors (i.e., 475mV vs 400mV) results in about 15 additional watts of dissipation. Might not seem like much...but almost 20%.
NP Quote: "I recall Jean Hiraga wrote that the best sound was a modest mix of second and third harmonic with the second being dominant. Individual preferences vary, but I think there is considerable truth to that. "
When I built my Class A tube amplifiers (6L6 push-pull parallel) I noted that at lower power they had fairly even 2nd & 3rd harmonics. However, when opening them up to their 40W/chn capacity, the transition to dominant 3rd was too early (push-pulls seem to have dominant 3rd)
So I tweaked the drive stage in LTSpice and changed the anode resistors to have a slightly higher 2nd - about 2.5dB. Testing showed about the same. The result was quite noticeable, and at the output levels I listen to, made the difference I was looking for. Overall, the amplifiers (have built 2 of them) produce 0.3% distortion at 40W. Not so bad for tube. However, the distortion of my amps is quite flat, rising to just 0.42% at 40W.
The point here is - I agree with NP's quote. I'm curious about the single-ended sound vs push-pull, so I will build...and listen.
Roger57,
Please post your FFT results when you have them. That would be really
interesting to see. Thank you for your comments and measurements,
both for you tube amp and the Aleph-J.
I haven't read anything about the H2 to H3 transition of the Aleph-J before.
I wonder what mechanism comes into play here?
Cheers,
Dennis
Please post your FFT results when you have them. That would be really
interesting to see. Thank you for your comments and measurements,
both for you tube amp and the Aleph-J.
I haven't read anything about the H2 to H3 transition of the Aleph-J before.
I wonder what mechanism comes into play here?
Cheers,
Dennis
ooooops.....the above should have said:
However, the distortion vs frequency of my amps is quite flat, rising to just 0.42% at 20KHz
For the record...