Sorry, this in in audio path - probably 220pF 630V 5% C0G would be better. Murata is a good quality low distortion brand I have had good luck with.
GRM31A5C2J221JW01D Murata Electronics | Mouser
GRM31A5C2J221JW01D Murata Electronics | Mouser
I lost 5 hrs of sleep because I couldn’t stop listening to this amp last night (early morning). Playing all sorts of music from my collection and rediscovering new things I did not hear before. I actually have a lot of classical CDs but never found them captivating or engaging. That changed with the Alpha 20 on my 10F/RS225 FAST. It was a CD of an orchestra playing Rodrigo with classical guitar centerpiece. The spatial imaging was incredible - I could make out the different background strings and woodwind instruments stage left and back and some were front and right. The stage was wider than the speakers physical placement. It really was a detailed 3d soundstage. Couldn’t stop listening and tried other pieces. The transient perfect nature of the speakers ensured that the guitar sounded very accurate and crisp. Very nice. Really think you will all like this amp a lot.
Last edited:
The ALPHA 20 has become the centerpiece main amp now - pretty happy with the system now. The 10F/RS225 speakers just got some new metal stands, and the speaker lab is undergoing some spring cleaning. System is sounding very nice. I have an Aksa Lender preamp for volume control and an APA-150 Class AB connected to a front loaded horn subwoofer, when needed.
Hi JPS,
Sorry, slow reply, sleeping!
220p/47R in necessary on ALL mosfets; the snubber is used on each mosfet, and sure pass from gate to drain, direct, inside the gate stopper.
XRK is correct, they are in the signal change and his COG Murata is a good suggestion.
X,
WOW!! So this ALPHA is now your gold standard? That's quite an accolade.......
Nice system, my congratulations to your hard work. No CapMX? You are just using a 33mF/0.47R/33mF and an 18Vac beefy toroidal?
HD
Sorry, slow reply, sleeping!
220p/47R in necessary on ALL mosfets; the snubber is used on each mosfet, and sure pass from gate to drain, direct, inside the gate stopper.
XRK is correct, they are in the signal change and his COG Murata is a good suggestion.
X,
WOW!! So this ALPHA is now your gold standard? That's quite an accolade.......
Nice system, my congratulations to your hard work. No CapMX? You are just using a 33mF/0.47R/33mF and an 18Vac beefy toroidal?
HD
Hi Hugh,
Yes, the Alpha 20 is now my main amp. 20 Watts would have been seen as wanting for more on some amps, but for some reason, it seems to be perfectly good here. Well I will have to do a revaluation once I get the 52w Big Boy going, big for now, I am able to put the other amps on the shelf and leave this one as primary. I think the combination of superior spatial imaging, clarity, resolution, and bass authority makes this amp a keeper. Seems to work quite well with conventional 400VA 18v toroidal trafo and CRC 33mF 0.235R 33mF shared with two channels then another 0.235R 33mF RC for each channel independently (F6 stereo PSU). My SMPS seems to not be able to keep up without being stressed at the currents we are talking about here. I was running the DC step up at 29v to make 24v after the cap Mx and so that was 29/24x1.3ampsx2/0.9=3.4amps. Seems within 5amp rating but the smps supply seems to have more noise when pushed this hard. I might try again later with a good common mode choke filter. For now, just use the trafo and linear supply.
Yes, the Alpha 20 is now my main amp. 20 Watts would have been seen as wanting for more on some amps, but for some reason, it seems to be perfectly good here. Well I will have to do a revaluation once I get the 52w Big Boy going, big for now, I am able to put the other amps on the shelf and leave this one as primary. I think the combination of superior spatial imaging, clarity, resolution, and bass authority makes this amp a keeper. Seems to work quite well with conventional 400VA 18v toroidal trafo and CRC 33mF 0.235R 33mF shared with two channels then another 0.235R 33mF RC for each channel independently (F6 stereo PSU). My SMPS seems to not be able to keep up without being stressed at the currents we are talking about here. I was running the DC step up at 29v to make 24v after the cap Mx and so that was 29/24x1.3ampsx2/0.9=3.4amps. Seems within 5amp rating but the smps supply seems to have more noise when pushed this hard. I might try again later with a good common mode choke filter. For now, just use the trafo and linear supply.
Last edited:
I have noticed that Class A amps seem much larger than life. My Glass Harmony was only 28W, but found like 50W, and a 56W water cooled version I build around 1998 seem like a colossus, almost moving the ground. This might be part of the spatial characteristics, because the sound field is very large, giving the ear an impression of great power. I think this might also related to the CapMX too; the general impression is that they confer a narrower sound field, which militate against this huge impression as well.
The next issue is to reduce the ripple a little more. I would suggest using a 33mF/15mH/33mF, replacing the 0.235R resistor with a small inductor. I'm pretty sure you can buy them reasonably cheaply from Hammond or Edcor; I have used them on the Glass Harmony and they offer far more filtering of the 100Hz ripple (31.4R impedance at 100Hz versus 0.235R!!). I had mine built here in Melbourne, expensive, you should try to find one with DCR less than 0.33R. The Hammond 159ZG is 15mH, passes up to 4A with 0.25DCR. It would be ideal for this job.
I am over the moon about this amp. I think we can owe the sound to the AKSA/Lender and the efficiency and grand scale of the NP dynamic CCS, which is one of the cleverest 'audio' things I've seen in a long time.
We are moving towards a SS version of a tube supply of yore; when you listen to this approach you will realise that iron is inevitable, a bit like impedance is futile.......
Cheers,
HD
The next issue is to reduce the ripple a little more. I would suggest using a 33mF/15mH/33mF, replacing the 0.235R resistor with a small inductor. I'm pretty sure you can buy them reasonably cheaply from Hammond or Edcor; I have used them on the Glass Harmony and they offer far more filtering of the 100Hz ripple (31.4R impedance at 100Hz versus 0.235R!!). I had mine built here in Melbourne, expensive, you should try to find one with DCR less than 0.33R. The Hammond 159ZG is 15mH, passes up to 4A with 0.25DCR. It would be ideal for this job.
I am over the moon about this amp. I think we can owe the sound to the AKSA/Lender and the efficiency and grand scale of the NP dynamic CCS, which is one of the cleverest 'audio' things I've seen in a long time.
We are moving towards a SS version of a tube supply of yore; when you listen to this approach you will realise that iron is inevitable, a bit like impedance is futile.......
Cheers,
HD
Last edited:
Hi Hugh,
Normally I find all sorts of mains (60Hz/120Hz/180Hz/240Hz) in my linear powered amps. But Alpha 20 actually has decent PSRR. So even with a basic linear 400VA 18V trafo
and CRC, the amp is running clean and quiet. Here is the FFT on one channel with both channels drawing 1.3amps each. Not much mains peaks to worry about so not sure if a CLC is needed.
Normally I find all sorts of mains (60Hz/120Hz/180Hz/240Hz) in my linear powered amps. But Alpha 20 actually has decent PSRR. So even with a basic linear 400VA 18V trafo
and CRC, the amp is running clean and quiet. Here is the FFT on one channel with both channels drawing 1.3amps each. Not much mains peaks to worry about so not sure if a CLC is needed.
Hmm, that's impressive....... 120Hz ripple is at -104dB below fundamental, this is well below audibility.
You're right; the psrr is better than I'd expected, and the hum only gets into the VAS collector from the negative rail, NOT from the LTP because of the RCR stage supply.
It seems you don't need an inductor!! Man, this is cheap as chips, impressive.
HD
You're right; the psrr is better than I'd expected, and the hum only gets into the VAS collector from the negative rail, NOT from the LTP because of the RCR stage supply.
It seems you don't need an inductor!! Man, this is cheap as chips, impressive.
HD
JPS64,
Re your question in post #761.
Connect the snubber network to each mosfet as shown.
Keep the two resistors close to the mosfet gate (to minimize the track inductance), and ensure the capacitor has a high enough voltage rating (100v or better would be good). A NP0 ceramic is ideal in this use.
Regards,
Paul Bysouth
Re your question in post #761.
Connect the snubber network to each mosfet as shown.
Keep the two resistors close to the mosfet gate (to minimize the track inductance), and ensure the capacitor has a high enough voltage rating (100v or better would be good). A NP0 ceramic is ideal in this use.
Regards,
Paul Bysouth
Attachments
Yes, Alibear, exactly that. The two devices of the amplifier, a pmos and a nmos, will both be dissipating 48W, a total of 96W. This is a LOT of heat to dispense with.
However, X and I have come to the conclusion that the correct quiescent for this amp at 24V rails is 1.6A. This mandates a source resistor of 0.39R, rated at 2W or preferably 3W. It should not get too hot...... and reduces EACH mosfet dissipation to 38.4W, 76.8W total. This is manageable, even with a large 350mm x 100mm heatsink.
Cheers,
Hugh
However, X and I have come to the conclusion that the correct quiescent for this amp at 24V rails is 1.6A. This mandates a source resistor of 0.39R, rated at 2W or preferably 3W. It should not get too hot...... and reduces EACH mosfet dissipation to 38.4W, 76.8W total. This is manageable, even with a large 350mm x 100mm heatsink.
Cheers,
Hugh
Wow, that looks fantastic JP! Quite the masterpiece and a unique looking amp. I worry about having a PWM fan drive on board. Is there sufficient de-coupling to ensure that the kHz PWM pulses don’t get into the audio path? I am wondering if they should be off-board?
Thanks so much for all your efforts on a totally professional looking board. Hope the flu is all but over now for you.
Cheers,
X