I'm also impressed by classD efficiency. Like you said, its just a matter of time until a good sounding classD existed.
Labgruppen has a great idea. Can make class A audio power amp with +/-150VDC rails, with minor heat dissipation. I posted here http://www.diyaudio.com/forums/showthread.php?postid=509817#post509817 and here http://www.diyaudio.com/forums/showthread.php?s=&threadid=45892
But no one replies. Is it too hard to do it? 😀
Labgruppen has a great idea. Can make class A audio power amp with +/-150VDC rails, with minor heat dissipation. I posted here http://www.diyaudio.com/forums/showthread.php?postid=509817#post509817 and here http://www.diyaudio.com/forums/showthread.php?s=&threadid=45892
But no one replies. Is it too hard to do it? 😀
Thanks All ,
Hence it is decided that it is BENIFICAL to use the source resistor , as
majority of experianced amp designers have voted for it .
Lars
I hope now you will go ahead with the PA version of the ZETA .
I have compleated a Zeta with two pairs of Irfp250 as the output devices
without source resistors , I will use the source resistors in the PA version
.
Eirlear during inital testing I had not put the offset null preset , later I
added it but I could vary the offset from 65mv to 95mv only , I reduced
the value of the preset from 10k to 1k but very slight differance 65mv to
80mv , then I removed the preset from the emitters of T1 & T4 and put a
470 ohm preset in the same way on the emitters of T2 & T3 and I could
adjust the offset to 2mv , I did not try 10k preset here maybe that too
would have worked , this means that the offset in my amp was from T2 &
T3 . I have adjusted the Quessent current to 200ma ( 100ma per fet ) , it
is equal on both rails , The transformer is 40-0-40v , filter caps
30,000Mfd per rail .
I did a listening test and feel that it has more gain at higher frequencies
as it is soundimg too sharp , however the Bass is also good !! BUT at
more than 50% output there is audable distortion in the sound which is
more apparent in the highs , it is as if the amp is overdriven even before
Po max .
I connected another amp with the same source and same speakers ,
the sound is absolutely clear .
The temp of the mosfets in the +ve rail is higher than the mosfets in the
-ve rail but within limits .
Please could you let me know what could be the problem ??
Dj Quan
You could try putting a 1k preset on the emitters of T2 & T3 and see if the
problem of offset is sorted out in your amp .
Hence it is decided that it is BENIFICAL to use the source resistor , as
majority of experianced amp designers have voted for it .
Lars
I hope now you will go ahead with the PA version of the ZETA .
I have compleated a Zeta with two pairs of Irfp250 as the output devices
without source resistors , I will use the source resistors in the PA version
.
Eirlear during inital testing I had not put the offset null preset , later I
added it but I could vary the offset from 65mv to 95mv only , I reduced
the value of the preset from 10k to 1k but very slight differance 65mv to
80mv , then I removed the preset from the emitters of T1 & T4 and put a
470 ohm preset in the same way on the emitters of T2 & T3 and I could
adjust the offset to 2mv , I did not try 10k preset here maybe that too
would have worked , this means that the offset in my amp was from T2 &
T3 . I have adjusted the Quessent current to 200ma ( 100ma per fet ) , it
is equal on both rails , The transformer is 40-0-40v , filter caps
30,000Mfd per rail .
I did a listening test and feel that it has more gain at higher frequencies
as it is soundimg too sharp , however the Bass is also good !! BUT at
more than 50% output there is audable distortion in the sound which is
more apparent in the highs , it is as if the amp is overdriven even before
Po max .
I connected another amp with the same source and same speakers ,
the sound is absolutely clear .
The temp of the mosfets in the +ve rail is higher than the mosfets in the
-ve rail but within limits .
Please could you let me know what could be the problem ??
Dj Quan
You could try putting a 1k preset on the emitters of T2 & T3 and see if the
problem of offset is sorted out in your amp .
As some may know, we also do Class D amplifiers with power of 1000 W from just a single pair of TO220 MOSFET's. However the ZETA is an attempt to make an alternative to Class D, for all who still feel an 'old fashioned' Class A/B amplifier is better.
Click here to see
Rajeev: I have no explanation for your findings, i did not experience the same with my setup. Did you check for oscillations at high powers? I did not experience any oscillation myself, but the stability of course depends on the layout. If you want me to investigate further, please send me your PCB, so i can take a listen and some measurements. Maybe the distortion is introduced by the parallelling of output devices, which i did not try initially.
The positive side MOSFET's get slightly hotter because the P channel FET is helping the negative side MOSFET's so they have slightly less power to dissipate. If you add the dissipation from the P channel to the negative N channel FET's, total dissipation is of course the same for negative side and positive side.
Click here to see
Rajeev: I have no explanation for your findings, i did not experience the same with my setup. Did you check for oscillations at high powers? I did not experience any oscillation myself, but the stability of course depends on the layout. If you want me to investigate further, please send me your PCB, so i can take a listen and some measurements. Maybe the distortion is introduced by the parallelling of output devices, which i did not try initially.
The positive side MOSFET's get slightly hotter because the P channel FET is helping the negative side MOSFET's so they have slightly less power to dissipate. If you add the dissipation from the P channel to the negative N channel FET's, total dissipation is of course the same for negative side and positive side.
Hey Lumanauw;
Yeah that would be cool and tricky.
It would be easier to use a multi-tap transformer and use solid state switching to select higher voltage windings when more power is required (NAD and others do this already).
Then you could have a class A amp with variable rails. This would improve the idling dissipation a great deal. I.e. you could riun 20 volt rails normally then switch in 40 / 60 volts progressively as required by the input signal.
I run a bench power supply 2 - 30 volts / 30 amp, with relay switching to a multitap transformer. This was necessary because at 12 volts / 30 amps the output devices were trying to get rid of 600 watts. By using a lower transformer tap this dissipation is around 100 watts.
Cheers
I can't understand why people want a class A amp anyway. A class AB amp well set up will be provide excellent performance. But hey, I'm not everyone else, so each to their own.
Yeah that would be cool and tricky.
It would be easier to use a multi-tap transformer and use solid state switching to select higher voltage windings when more power is required (NAD and others do this already).
Then you could have a class A amp with variable rails. This would improve the idling dissipation a great deal. I.e. you could riun 20 volt rails normally then switch in 40 / 60 volts progressively as required by the input signal.
I run a bench power supply 2 - 30 volts / 30 amp, with relay switching to a multitap transformer. This was necessary because at 12 volts / 30 amps the output devices were trying to get rid of 600 watts. By using a lower transformer tap this dissipation is around 100 watts.
Cheers
I can't understand why people want a class A amp anyway. A class AB amp well set up will be provide excellent performance. But hey, I'm not everyone else, so each to their own.
Is it full range or subwoofer amp?lumalauw,
the offset is present if the input is grounded or not. the offset is positive after the VAS and negative in the output. which means the input stage is already trying to counter the DC offset.
rajeev, I'll try the 1k trimmer later. but I think it would have no effect as even 9if I set the trimmer right on the ends, there's still the offset.
it is for full range. I would have gone buying one of those instead of building the zetas but I don't have the money.
the offset is present if the input is grounded or not. the offset is positive after the VAS and negative in the output. which means the input stage is already trying to counter the DC offset.
from what I recall, N channel devices can handle more power than P channel devices.
rajeev, I'll try the 1k trimmer later. but I think it would have no effect as even 9if I set the trimmer right on the ends, there's still the offset.
it is for full range. I would have gone buying one of those instead of building the zetas but I don't have the money.
Ahh, we found an important clue. You are right. The front stage is trying to correct the final stage. So, something is wrong with the final stage.
First make sure none component is broken or malfunctioning. After you are sure, maybe it is a problem of bias. Try to recalculate bias on every transistor on the ZETA schematic and measure in your cct, are they the same number?
Lack of bias could be a problem. Try to raise quiscent current in the suspicious points.
If thats ok, maybe you have a problem with oscilation. This could be caused by pcb routing. Since we are working with mosfets, it becomes more critical.
Some Serious Comments
Hi Rajeev
SO all of u guys have allotted enough accomodation to source resistors on ur PCB's, therefore Source resistors would feel Quite happy with u Guys.Have u allotted them the space on rent or a permanent settlement is made upon.
The High Frequency Distortion may arise due to the asymmetrical drivedue to P-Channel drive mosfet The Slew Rate in this Zeta Amp is not Symmetrical in any aspect. Secondly Then there may arise a situation when u push the amp hard for first time.
Give ur amp a BURN in Final WEED out Test i.e connect a heater element of about 8 Ohms to the out put and drive the input with low frequency sine wave say 50Hz along with 20KHZ intermodulated type with full drive at threshold of clipping for about 3-4 Hours .
Caution ,Use Force cooling with Extremely good Heatsink .
Heat the amp like a Burning fire with heatsink temp about 90deg celcius.After Performing this test listen again and feel the Difference.
Hi lars Clausen
Good to hear about ur PA version of ZETA .
Looking Forward to see it.
regards
Kanwar Abhishek Singh
rajeev luthra said:
Hi Rajeev
SO all of u guys have allotted enough accomodation to source resistors on ur PCB's, therefore Source resistors would feel Quite happy with u Guys.Have u allotted them the space on rent or a permanent settlement is made upon.
The High Frequency Distortion may arise due to the asymmetrical drivedue to P-Channel drive mosfet The Slew Rate in this Zeta Amp is not Symmetrical in any aspect. Secondly Then there may arise a situation when u push the amp hard for first time.
Give ur amp a BURN in Final WEED out Test i.e connect a heater element of about 8 Ohms to the out put and drive the input with low frequency sine wave say 50Hz along with 20KHZ intermodulated type with full drive at threshold of clipping for about 3-4 Hours .
Caution ,Use Force cooling with Extremely good Heatsink .
Heat the amp like a Burning fire with heatsink temp about 90deg celcius.After Performing this test listen again and feel the Difference.
Hi lars Clausen
Good to hear about ur PA version of ZETA .
Looking Forward to see it.
regards
Kanwar Abhishek Singh
lumanauw said:
I originally I thought it's the input stage that's causing the problem. the inverting input seems to have a low input impedance as it is getting some current at the feedback resistor (100k in the schem, the 1k shows no current flowing) if the input stage is really trying to correct it, the offset should be zero whatever the assymmetry of the output stage. the only thing I might have wrong is the 499R resistors in the input stage. it is indicated as 499R, but I can't find any so I used 510R. I'll try 470R when I buy some around next week. I'll also get eight new IRFP250's rather than the old IRFP460's I already have. (as I also blew four output stages the last time I experimented with it.)
I don't have problems with oscillation as I have a 250MHz scope and I don't see anything wrong down to 50mV.
I used 470 ohms I think 510 ohms should also be fine,
Now I am worried about the problem in my amp , up 50% vol all is fine later too much distortion , I will check it later
I hope Your amp will be ready and I will get your feedback
cheers
Now I am worried about the problem in my amp , up 50% vol all is fine later too much distortion , I will check it later
I hope Your amp will be ready and I will get your feedback
cheers
rajeev luthra said:
you'll never know what could happen.
😉
I guess we're the guinea pigs of lars' zeta. 😕 😀
if I really can't fix the DC offset, I'll go the DC servo route. (dad doesn't approve the "send the boards to lars" route.
)Heater element??????
Is this proper way way to test an amp if so better put a couple of eggs on top as well. If it runs well at a high temprature then better add some chicken too.
No smoke as yet what more to add???
Regards
Rahul
Lars
Thanks for pointing it out that the problem could be due to oscillations I
will try to see where it is
Dj Quan
You did not read my post 282 properly , please also read what I had
written to Lars ,
I had asked you to put the 1k preset on the emitters of the T2 & T3 ie the
pnp diff pair, most proberbly this will sort out your problem , as I had the
same problem .
Ampman
Thanks
Does this "BURN in Final WEED out Test" really work , Is this test for
mosfet amps only or for bipolar amps too ? will it sort out oscillations
also ??
Thanks for pointing it out that the problem could be due to oscillations I
will try to see where it is
Dj Quan
You did not read my post 282 properly , please also read what I had
written to Lars ,
I had asked you to put the 1k preset on the emitters of the T2 & T3 ie the
pnp diff pair, most proberbly this will sort out your problem , as I had the
same problem .
Ampman
Thanks
Does this "BURN in Final WEED out Test" really work , Is this test for
mosfet amps only or for bipolar amps too ? will it sort out oscillations
also ??
djQUAN .. guinea pigs ... 😀 No i have tested the amplifier in real life, and i did not encounter any of the problems you have. I have even posted scope photos previously in the thread. .. I am sure you have seen.
But maybe i should have waited posting the schematic until i could offer a PCB that will simply work without any trouble. When you just start with the schematic for people to copy, there may be too much room for 'interpretation' 😉
I agree the problem with your amplifier is definitely in the mid to output stage, especially i would look around R15 and R16 where you seem to have a large inregularity.
All the best from
Lars
But maybe i should have waited posting the schematic until i could offer a PCB that will simply work without any trouble. When you just start with the schematic for people to copy, there may be too much room for 'interpretation' 😉
I agree the problem with your amplifier is definitely in the mid to output stage, especially i would look around R15 and R16 where you seem to have a large inregularity.
All the best from
Lars
rajeev luthra said:
ah! had another look.
I'll give it a try when I buy some new parts. all work is stopped when I blew the output stage. need a case to put everything in first, much safer.Lars Clausen said:
yeah, I saw the test results and thought it might be a fun amp to build.
I'll try to check around again but no testing with power. although I just remembered something, you did the tests with the low power version right? have you done tests with the high power version?so far, thanks for everybody's help and for sticking around even if my amp still doesn't work.
People make interesting misteaks they cannot seam to find by themselves.
Having someone else look at your work seems to help find the mistakes.
Having someone else look at your work seems to help find the mistakes.
There is possibility of changing R4+R7 and R5+R6 to a single 200ohm VR, with its wiper connected to CCS. But this should be done after you are SURE nothing else is causing the DC offset.
LC's original design of R8 should be enough for properly working amp.
Hi, LC,
The lower array have 10ohm in T14 drain to translate Vgs for lower output mosfets. Will putting another 10ohm from +of C8 to output (to translate Vgs for upper output array) help in anything? Or we should put 1 more N source follower after R28 with 10ohm load to output, then the drop of this 10ohm drives the Vgs of upper output array?
Ampman,
It seems that for you it is not for testing, more to "break-in" to change sound. You could use XLO Burn-in CD.
You really torture your amp 😀. But its not the most painfull for power amp, since the load is only resistive and the transistors working in full (I assume nearly saturated) mode.
The difficult load is Z, you put R, L and C that gives quite phase shifts, and you play it in 2/3 of full output rating. This is the point where the amp's output transistor peak stress.
LC's original design of R8 should be enough for properly working amp.
Hi, LC,
The lower array have 10ohm in T14 drain to translate Vgs for lower output mosfets. Will putting another 10ohm from +of C8 to output (to translate Vgs for upper output array) help in anything? Or we should put 1 more N source follower after R28 with 10ohm load to output, then the drop of this 10ohm drives the Vgs of upper output array?
Ampman,
It seems that for you it is not for testing, more to "break-in" to change sound. You could use XLO Burn-in CD.
You really torture your amp 😀. But its not the most painfull for power amp, since the load is only resistive and the transistors working in full (I assume nearly saturated) mode.
The difficult load is Z, you put R, L and C that gives quite phase shifts, and you play it in 2/3 of full output rating. This is the point where the amp's output transistor peak stress.
djk said:
I admit I have done mistakes before but I was able to find it within a few days or so. this one, I've been working for weeks and I hope that I did everything right and not missed anything.
lumanauw said:
the 10R resistor in my zeta is changed to 39.9R with no effect (except for reduced dissipation in T14)