Has someone used mjl4281? I'm going to use pair FQA40N25/MJL4281 as suggested some time ago.
See this page: http://www.diyaudio.com/forums/solid-state/255427-simple-quasi-complimentary-mosfet-amplifier-170.html#post4916087
Idefixes? Hugh?
Rgds.
Well, I've just finished popullating the second channel with a MJL4281 and a IRFP26N60L. First tests were not good, as I could not read any potential between R20's legs. What could the possible causes be?Has someone used mjl4281? I'm going to use pair FQA40N25/MJL4281 as suggested some time ago.
See this page: http://www.diyaudio.com/forums/solid-state/255427-simple-quasi-complimentary-mosfet-amplifier-170.html#post4916087
Idefixes? Hugh?
Rgds.
(dismounting and mounting everything again onto the heatsink is pretty much boring!!)
I will do a pin-to-pin voltage comparison with the other - working - board (that uses 2SD1047 and IRFP250).
Bruce, can you post some detailed images of your build?
Whose layout design/PCB did you use or was it your own?
This would help in tracking the problem hopefully.
regards
Prasi
Edit: I should have said "could you post some....."
I am learning English , same way as I am learning PCB design.
Whose layout design/PCB did you use or was it your own?
This would help in tracking the problem hopefully.
regards
Prasi
Edit: I should have said "could you post some....."
I am learning English , same way as I am learning PCB design.
Last edited:
Prasi,
Please never diss your own English; just like pretty much all ESL people in this forum your meaning and correctness in your second/third/fourth language is very good....
Actually the feature of English is that even used inexpertly, the meaning is normally clear.
Oh and your pcbs are superb!
HD
Please never diss your own English; just like pretty much all ESL people in this forum your meaning and correctness in your second/third/fourth language is very good....
Actually the feature of English is that even used inexpertly, the meaning is normally clear.
Oh and your pcbs are superb!
HD
Thank you Hugh,
You are a gentleman!
yes, it is the fourth language I started leaning during childhood.Still enjoying your gifted amp immensely
and its still without case. Need to get a case urgently.As a future project, I plan to do the same stereo +PSU layout design technique with this Quasi amplifier.
regards
Prasi
Hi Bruce,
If you would like to take a shot at AKSA55 (the mentioned stereo amp with PSU, do get in touch, I can supply you the board and/or components.), if you feel upto the challenge!
You will enjoy THE ORIGINAL AKSA design!
The stereo quasi with PSU, you mention, will have to wait, as too many projects in hand , presently.
but let us not diverge from the subject of the thread, and keep it private if you are interested...
regards,
Prasi
If you would like to take a shot at AKSA55 (the mentioned stereo amp with PSU, do get in touch, I can supply you the board and/or components.), if you feel upto the challenge!
You will enjoy THE ORIGINAL AKSA design!
The stereo quasi with PSU, you mention, will have to wait, as too many projects in hand , presently.
but let us not diverge from the subject of the thread, and keep it private if you are interested...
regards,
Prasi
Hi Prasi,
I'm up to the challenge!
Ciao!
Do
Hi Do,
There have been some new developments and I am trying to assimilate the same.
I might have to revise the design (I am in consultation with Mr. Hugh Dean). I do not know (at the present moment) if I shall be able to undertake the task. I really want to do justice to the design and the painstaking efforts taken by Mr. Hugh Dean in developing an amp thats still highly regarded by many, if not all.
Sorry for my post above, I request moderators to delete the relevant post no. 1894.
I truly apologise
. regards,
Prasi
Asymmetrical clipping.
Sorry for this late post but i just came across this thread now. I have built the citation mosfet amp in quasi mode and I have built another similar variation of the citation as well. Much earlier in this thread you all addressed about the asymmetrical clipping which occurs where the negative half cycle clips before the positive by about 2 to 4 volts. Well I have seen this myself and i would like to add some more info to this. In the citation topology the differential input is PNP driving the NPN VAS on the negative rail and bootstrap end on the positive rail. No upper drive transistor and lower drive transistor for quasi operation. In this setup the negative clips before the positive. Now if you install an upper drive transistor just to keep it the same, there is no difference at all. It still clips as above.
Now with another amplifier build where the differential input is NPN driving the PNP VAS on the positive rail end, bootstrap on the negative end, no drive transistor on the VAS end (positive side) and drive transistor on the negative for quasi operation, the amplifier clips exactly opposite. The positive side clips before the negative. If you again add the top drive transistor, no change still clips positive before negative. My question is if BJTs are used in these setups the clipping is equal. I just cannot see why the mosfet setup does this. Same drive circuit in both scenarios. For a simple setup I think these topologies need more complex drive cicuitry to accomodate mosfets than BJTs. Just some more information that i have experimented on a bench not on a simulator program.
Regards.
Billy D....
Sorry for this late post but i just came across this thread now. I have built the citation mosfet amp in quasi mode and I have built another similar variation of the citation as well. Much earlier in this thread you all addressed about the asymmetrical clipping which occurs where the negative half cycle clips before the positive by about 2 to 4 volts. Well I have seen this myself and i would like to add some more info to this. In the citation topology the differential input is PNP driving the NPN VAS on the negative rail and bootstrap end on the positive rail. No upper drive transistor and lower drive transistor for quasi operation. In this setup the negative clips before the positive. Now if you install an upper drive transistor just to keep it the same, there is no difference at all. It still clips as above.
Now with another amplifier build where the differential input is NPN driving the PNP VAS on the positive rail end, bootstrap on the negative end, no drive transistor on the VAS end (positive side) and drive transistor on the negative for quasi operation, the amplifier clips exactly opposite. The positive side clips before the negative. If you again add the top drive transistor, no change still clips positive before negative. My question is if BJTs are used in these setups the clipping is equal. I just cannot see why the mosfet setup does this. Same drive circuit in both scenarios. For a simple setup I think these topologies need more complex drive cicuitry to accomodate mosfets than BJTs. Just some more information that i have experimented on a bench not on a simulator program.
Regards.
Billy D....
Hi Billbo,
This quasi uses a bootstrap to drive the upper mosfet from the VAS.
This means that the mosfet gate can be driven ABOVE the positive rail; using a CCS, there is a couple of volts restriction preventing the gate driven higher than 2V lower than the rail.
With a bootstrap you can drive the gate easily to the same voltage as the drain, and this means that the source can be raised another couple of volts, increasing headroom for the amplifier. This is the primary reason we used a bootstrap - it offers more headroom (three or four more watts of audio), but it actually sounds very good on voice, and it is a passive system, rather than a CCS, increasing reliability.
Cheers,
Hugh
This quasi uses a bootstrap to drive the upper mosfet from the VAS.
This means that the mosfet gate can be driven ABOVE the positive rail; using a CCS, there is a couple of volts restriction preventing the gate driven higher than 2V lower than the rail.
With a bootstrap you can drive the gate easily to the same voltage as the drain, and this means that the source can be raised another couple of volts, increasing headroom for the amplifier. This is the primary reason we used a bootstrap - it offers more headroom (three or four more watts of audio), but it actually sounds very good on voice, and it is a passive system, rather than a CCS, increasing reliability.
Cheers,
Hugh
Last edited:
Asymmetrical clipping
All said is good. Although the point I am addressing is the clipping changes from negative to positive when the VAS is revered. Even if you do not use bootstrapping the clipping is not equal with mosfets. But in BJTs the clipping is equal in all aspects.
Regards.
Billy D...
All said is good. Although the point I am addressing is the clipping changes from negative to positive when the VAS is revered. Even if you do not use bootstrapping the clipping is not equal with mosfets. But in BJTs the clipping is equal in all aspects.
Regards.
Billy D...
Hi Billy,
Yes, and you put your finger on precisely why a mosfet at top and npn quasi at bottom was the reason we chose this odd topology. Spice indicates that the clipping is pretty identical pos and neg, and this is very elegant for a quasi.
When you invert the signal at the OPS predrive, the concertina action of the inverter consumes headroom and it's a problem you can only reduce with low bias outputs, like a bipolar npn, but not eliminate. The JLH has this same problem for precise reasons, in fact it's less efficient than the Ranchu/AKSA quasi. If you create the inverting drive earlier in the amp, for example in a Siemens amp from the 80s where it's done at the input LTP stage, you can eliminate this headroom issue completely, but then the upper npn OP device clips earlier anyway.
BTW, an asymmetrical clip has no effect on the sound quality as long as you keep within clip, same for any amplifier. I'm not sure if you are being negative about this amp, or positive. It is an interesting amp design to reduce headroom issues.
A quasi is a mass of compromises.
HD
Yes, and you put your finger on precisely why a mosfet at top and npn quasi at bottom was the reason we chose this odd topology. Spice indicates that the clipping is pretty identical pos and neg, and this is very elegant for a quasi.
When you invert the signal at the OPS predrive, the concertina action of the inverter consumes headroom and it's a problem you can only reduce with low bias outputs, like a bipolar npn, but not eliminate. The JLH has this same problem for precise reasons, in fact it's less efficient than the Ranchu/AKSA quasi. If you create the inverting drive earlier in the amp, for example in a Siemens amp from the 80s where it's done at the input LTP stage, you can eliminate this headroom issue completely, but then the upper npn OP device clips earlier anyway.
BTW, an asymmetrical clip has no effect on the sound quality as long as you keep within clip, same for any amplifier. I'm not sure if you are being negative about this amp, or positive. It is an interesting amp design to reduce headroom issues.
A quasi is a mass of compromises.
HD