An externally hosted image should be here but it was not working when we last tested it.
My Quasi amp : http://www.audio-electronics.go.ro
Marus, I faced the same problem with my amp working off +-63volts. Just try R4, R8 down to 300E or even 270E. I use 330E. R6 can be between 10K and 12K. At the same time increase R12 and R15 to either 56E or 68E, just to be safe. I have struggled with a problem similar to yours when both channels of my amp played flawlessly from +-25 volts to +-50volts but distorted at +-63volts until I made the changes as mentioned above. Since then I have modified the amp extensively and it sounds almost as good as my Stochino amp. Hopefully, this Friday I may be able to compare this with a Krell 400i in the same setup. In different setups, my amp far outperforms the Krell but a proper A/B test with the same equipment and in the same room will settle the score.
So keep trying Marus. Quasi's amp is very stable so there should be no worries while you try different values to your heart's content. I have run this amp with a bias of nearly 1500mA and have shorted one channel while driving it with a line level signal. The rails had no fuses in this test set up. The amp did overheat but continued to perform just as well thereafter.
Quasi, keep up the very good work.
So keep trying Marus. Quasi's amp is very stable so there should be no worries while you try different values to your heart's content. I have run this amp with a bias of nearly 1500mA and have shorted one channel while driving it with a line level signal. The rails had no fuses in this test set up. The amp did overheat but continued to perform just as well thereafter.
Quasi, keep up the very good work.
Re Post #2295
Marus
I have a question about your scope samples you posted in Post #2295. What is the difference between the scope images? I can see the difference in the waveform. I am at a loss what variable(s) are different between the scope images of Post #2295 that would acount for the differences in the waveforms.
Regards,
John L. Males
Willowdale, Ontario
Canada
10 September 2007 22:07
10 September 2007 22:24 Typo correction. jlm
Official Quasi Thread Researcher
Marus
I have a question about your scope samples you posted in Post #2295. What is the difference between the scope images? I can see the difference in the waveform. I am at a loss what variable(s) are different between the scope images of Post #2295 that would acount for the differences in the waveforms.
Regards,
John L. Males
Willowdale, Ontario
Canada
10 September 2007 22:07
10 September 2007 22:24 Typo correction. jlm
Official Quasi Thread Researcher
Hi Samuel,
You may have done this before, but are you able to post your final schematic including the component values? You have obviously made some improvement although I'm still perplexed by the issues that Marus and you have experienced.
Marus, on one occasion a builder put the wrong value resistors in one part of the layout. This happened because the colour bands were quite poor with red and brown being indistinguishable (pretty long word), so are you able to check some of the resistors by measuring them. Also please confirm you used 2SC1845 for the input stage or let me know what you used. Also are you able to post the voltages you've got around the cct. I'm just trying to understand what the issues might be.
Cheers
Q
You may have done this before, but are you able to post your final schematic including the component values? You have obviously made some improvement although I'm still perplexed by the issues that Marus and you have experienced.
Marus, on one occasion a builder put the wrong value resistors in one part of the layout. This happened because the colour bands were quite poor with red and brown being indistinguishable (pretty long word), so are you able to check some of the resistors by measuring them. Also please confirm you used 2SC1845 for the input stage or let me know what you used. Also are you able to post the voltages you've got around the cct. I'm just trying to understand what the issues might be.
Cheers
Q
Keypunch, in my scope samples the difference is the input signal.
First, with a little bit of signal (i don't know the levels) seems ok.
In the second I rise a little bit more the imput signal... and start distorting. And a little more signal and so on ...
Samples are taked at amp output.
Quasi, all (and i mean ALL) my components are NEW , bought specially for this amp. I confirm that i used 2SC1845 for the imput stage and MJE 340 & 350 for the second and IRFP450 exactly like you said ! And all others it's like you said
... that's why i go crazy. This can be true what happend to me.
Now I'm trying to buid a Full Wave Precision Rectifier and a Sine Wave Generator to measure to know what the levels are ...
Quasi, this distorsion it's posible to be from the transistors HFE ? I think this it's the only diference between our amps
Other theory :
This amp, on +/- 80V DC, must output 300W RMS / 8 ohms.
For this it needs (single channel) a 450W PS with 24.000 uF / rail.
But, it can output 50W if i want, not only 300W. And I wanted just 50W. Because i don't have 450W PS and i have just a 100W one i said : If i put this amp on a 100W PS ( 5.000 uF/rail) i can output 50W ... and if this work on 50W it will work on 300W as well . And that was tested with +/-40V DC (from the same 100W PS) and it works, i got 50 pure watts. But why i can't get this with +/-80V. What is wrong ?
First, with a little bit of signal (i don't know the levels) seems ok.
In the second I rise a little bit more the imput signal... and start distorting. And a little more signal and so on ...
Samples are taked at amp output.
Quasi, all (and i mean ALL) my components are NEW , bought specially for this amp. I confirm that i used 2SC1845 for the imput stage and MJE 340 & 350 for the second and IRFP450 exactly like you said ! And all others it's like you said
... that's why i go crazy. This can be true what happend to me.Now I'm trying to buid a Full Wave Precision Rectifier and a Sine Wave Generator to measure to know what the levels are ...
Quasi, this distorsion it's posible to be from the transistors HFE ? I think this it's the only diference between our amps
Other theory :
This amp, on +/- 80V DC, must output 300W RMS / 8 ohms.
For this it needs (single channel) a 450W PS with 24.000 uF / rail.
But, it can output 50W if i want, not only 300W. And I wanted just 50W. Because i don't have 450W PS and i have just a 100W one i said : If i put this amp on a 100W PS ( 5.000 uF/rail) i can output 50W ... and if this work on 50W it will work on 300W as well
. And that was tested with +/-40V DC (from the same 100W PS) and it works, i got 50 pure watts. But why i can't get this with +/-80V. What is wrong ?
marus said:
Marus,
Thanks for the clarification. I had been looking at the scope signal values of min, max, mean and RMS at the top of each of the scope images. I was unaware that your issue also showed a worsening distortion as the signal level increased, let alone how low a output level the problem started to show.
Have you tried all of Samuel's suggestions:
1) R4/R8 of 330R, or down to 300R or even 270R
2) R6 of 10K, or between 10K and 12K
3) R12/R15 of 68R, or 56R
I have not caught on if you have tried the R12/R15 suggestion of Samuel.
What type of capacitors are you using for C1 and C12 - MKC, MKT or MKP?
What type of capacitors are you using for C2, C3, C4, and C8 - ceramic, silver mica, MKC, MKT or MKP?
What type of capacitors are you using for C5, C6, C9 and C10 - ceramic, MKC, MKT or MKP?
Are you using 0R33 or 0R47 for your MOSFET output device source resistors?
Regards,
John L. Males
Willowdale, Ontario
Canada
12 September 2007 (07:50 -) 08:42
Official Quasi Thread Researcher
Quasi, yes I have used 2SC1845 (tightly matched) for the input pair. All other transistors are as per your schematic except that the MJE types are replaced with KSE340/350. Output is 3 x IRFP450 from IR quite tightly matched for Vgs. PCB layout is my own. All resistors are hand matched and output stage is 1E (1 watt) x 3 in parallel for each Mosfet.
Apart from the changed values of resistors that I aforementioned, I do have a buffer before the Vas stage.
Power rails from output stage are fed to the input stage through a diode + 100E (2 watt) resistors and local decoupling caps.
Input cap is a high quality MKT of 2.2uF and small value caps are silver mica types.
To be double sure, during the troubleshooting phase, I removed all other embellishments, stripped the amp to the basic published schematic and tried the changes I have suggested to Marus. Only after altering the value of these resistors did the amp function correctly.
I don't know how to post schematics because I use an old version of Electronics Work Bench for sims. My PCB design would be useless as artwork because I hand paint the PCBs before etching and drilling (all labour intensive and manual processes) but my PCBs look almost as good as the computer generated art work ones and perform great. I have been making amplifiers for years and I have no idea of the count. I am saying this just in case someone should say that the problem could be with the PCB etc.,
Can I be of any more help? Please feel free to post. Thanks.
Apart from the changed values of resistors that I aforementioned, I do have a buffer before the Vas stage.
Power rails from output stage are fed to the input stage through a diode + 100E (2 watt) resistors and local decoupling caps.
Input cap is a high quality MKT of 2.2uF and small value caps are silver mica types.
To be double sure, during the troubleshooting phase, I removed all other embellishments, stripped the amp to the basic published schematic and tried the changes I have suggested to Marus. Only after altering the value of these resistors did the amp function correctly.
I don't know how to post schematics because I use an old version of Electronics Work Bench for sims. My PCB design would be useless as artwork because I hand paint the PCBs before etching and drilling (all labour intensive and manual processes) but my PCBs look almost as good as the computer generated art work ones and perform great. I have been making amplifiers for years and I have no idea of the count. I am saying this just in case someone should say that the problem could be with the PCB etc.,
Can I be of any more help? Please feel free to post. Thanks.
Samuel Jayaraj said:
Quasi,
I have seen a few of Samuel's hand drawn PCBs. They are simply amazing. Samuel has a real talent to draw his own PCBs. I have to assume for many years all PCBs were hand drawn as there was no, or not cost feasable, technology to autoroute/draw PCBs. One would have to wonder if many companies contracted Samuel's PCB talents before technology was more cost effective for creating PCB artwork.
Samuel,
Is Electronics Work Bench a Windows based program? If Electronics Work Bench is a Windows program I can give you a few suggestions how to post the schematic with your buffer addition to the quasi schematic.
Regards,
John L. Males
Willowdale, Ontario
Canada
12 September 2007 (20:03 -) 20:24
Official Quasi Thread Researcher
Samuel Jayaraj said:
Hi Samuel,
That is ok. Does EWB have a print function? I assume you have the Windows 98 CD or Diskettes? If so then it is possible to create output you can then post.
Regards,
John L. Males
Willowdale, Ontario
Canada
13 September 2007 01:05
Official Quasi Thread Researcher
I have win98 and I use print screen to capture the screen contents.
Then open IrfanView (free download) and paste.
I am still trying to understand the cropping and moving of the required image. Save as .png
Then use DIYaudio "browse" to attach the pic.
remember to attach a second time if you use "preview".
Then open IrfanView (free download) and paste.
I am still trying to understand the cropping and moving of the required image. Save as .png
Then use DIYaudio "browse" to attach the pic.
remember to attach a second time if you use "preview".
more tests ... the same result
New power supply : (2x100W)
No diference ...
I think i'll quit ...
New power supply : (2x100W)
An externally hosted image should be here but it was not working when we last tested it.
No diference ...
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
I think i'll quit ...
THD1 reminds me of the behaviour of two amps in cascade with overall feedback. If the final stage clips the first stage output zooms up to its max level to try and compensate the loss of the feedback signal.
Something is not properly biased ? Check currents in all parts of the circuit and put it down on the circuit diagram. Maybe you'll find it easier to troubleshoot.
Cheers.
Something is not properly biased ? Check currents in all parts of the circuit and put it down on the circuit diagram. Maybe you'll find it easier to troubleshoot.
Cheers.
Re: more tests ... the same result
Can it be a possibility that the FETs are fakes.....
Just a viewpoint.
marus said:New power supply : (2x100W)
An externally hosted image should be here but it was not working when we last tested it.
No diference ...
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
I think i'll quit ...
Can it be a possibility that the FETs are fakes.....
Just a viewpoint.
Marus,
One question occurred to me to ask you that is so obvious, but i never thought of until now. Did you use a different transformer you for the +-40VDC and +-80VDC rail tests but the exact same PSU (bridge, capacitor, etc) circuit?
You might want to scope the loaded at idle and unloaded PSU output and/or just the transformer output with no PSU circuit to see if perhaps there is some unwanted noise/ripple. I know it may be a off chance that the PSU or a PSU element is the cause of the problem, but your PSU circuit may need different values at different supply voltages if the PSU circuit is the source of the problem.
If the PSU/transformer tests are not reveling if you can try R6 at 6K8 for +- 40VDC and 12K7-13K7 at +-80VDC and what the results are. You might need to increase R4/R8 to 420R0 - 470R0 with R6 at 12K7 - 13K6. R12/R5 may also need to be changed to 86R6 for +-80VDC rails.
Regards,
John L. Males
Willowdale, Ontario
Canada
14 September 2007 (07:40 -) 08:19
Official Quasi Thread Researcher
One question occurred to me to ask you that is so obvious, but i never thought of until now. Did you use a different transformer you for the +-40VDC and +-80VDC rail tests but the exact same PSU (bridge, capacitor, etc) circuit?
You might want to scope the loaded at idle and unloaded PSU output and/or just the transformer output with no PSU circuit to see if perhaps there is some unwanted noise/ripple. I know it may be a off chance that the PSU or a PSU element is the cause of the problem, but your PSU circuit may need different values at different supply voltages if the PSU circuit is the source of the problem.
If the PSU/transformer tests are not reveling if you can try R6 at 6K8 for +- 40VDC and 12K7-13K7 at +-80VDC and what the results are. You might need to increase R4/R8 to 420R0 - 470R0 with R6 at 12K7 - 13K6. R12/R5 may also need to be changed to 86R6 for +-80VDC rails.
Regards,
John L. Males
Willowdale, Ontario
Canada
14 September 2007 (07:40 -) 08:19
Official Quasi Thread Researcher