So it seems to me you have to make a choice!
(Assuming of course that this 28v PS you have is suitable for a Class A amp - not the same thing as a PS for a Class AB amp.)1. Pay Hugh to develop an AN4R circuit that uses +/-28v rails.
2. Build the standard AN8R circuit - but change your 4ohm spkrs for some 8ohm spkrs.
or
3. Change the power transformer on your PS to one with lower secondary voltages ... so as to deliver +/-21v rails.
@rdb64
Indeed that would be the best option, although I wouldn’t get too concerned. If you have the impedance and phase curves for your current speaker that would be most useful.
Of note, my colleague has an 8R version of the AN39 with +/-30V rails. The speakers he used were sensitive at about 90dB with impedance dips to 4 ohms (at 100-300Hz). No clipping distortion and no wooly bass issues at 85-90dB at the listening position of about 8 feet.
Here is another data point to consider. Post 2763 right here. The Legacy Focus speaker we had listened to was certainly not an easy load with a rather overly complex crossover network. (3) 12 inch woofers per side btw.
Best,
Anand.
Indeed that would be the best option, although I wouldn’t get too concerned. If you have the impedance and phase curves for your current speaker that would be most useful.
Of note, my colleague has an 8R version of the AN39 with +/-30V rails. The speakers he used were sensitive at about 90dB with impedance dips to 4 ohms (at 100-300Hz). No clipping distortion and no wooly bass issues at 85-90dB at the listening position of about 8 feet.
Here is another data point to consider. Post 2763 right here. The Legacy Focus speaker we had listened to was certainly not an easy load with a rather overly complex crossover network. (3) 12 inch woofers per side btw.
Best,
Anand.
Another option... or two...
4. Build the 8R knowing that even though it's in the back of your mind that you're not perfectly balanced for max current and max voltage to clipping at 4R... that you'll be perfectly fine in the vast majority of situations.
Do some basic calcs to see if it's worth the bother.
If you've got some particularly unique speakers and/or have a massive room and like it LOUD! You could...
5. Go nuts and build two "8R" to run in parallel. That might be sexy. Check with the master to see if there are any potential detriments, but I can't think of any (other than the obvious cost and space).
6. Read this for a few more tips. Emphasis added is mine.
4. Build the 8R knowing that even though it's in the back of your mind that you're not perfectly balanced for max current and max voltage to clipping at 4R... that you'll be perfectly fine in the vast majority of situations.
Do some basic calcs to see if it's worth the bother.
If you've got some particularly unique speakers and/or have a massive room and like it LOUD! You could...
5. Go nuts and build two "8R" to run in parallel. That might be sexy. Check with the master to see if there are any potential detriments, but I can't think of any (other than the obvious cost and space).
6. Read this for a few more tips. Emphasis added is mine.
AN4-39 REVIEW - AndyR's SYSTEM
Two weeks ago I was invited by AndyR to listen to his 4R Alpha Nirvana using two quartet Satori line array with open baffle 6.5" speakers and two Satori 1" ribbon tweeters. There is a separate pair of 15" sub-woofers which chime it at <110Hz. The active system is DSP with 48dB/octave slopes, and a high spec DAC using a variety of conventional and HD files. I brought along a good friend, Russell, who owns a Maya 200W, Magico and Harbeth speakers, both three ways. At my suggestion Andy replaced each NMOS output with a NPN Darlington from Sanken (2SD2560) for a little more slam and impact. Russell is an avid audiophile and very skilled at assessing sound quality.
Now a couple of weeks on we both look back on an amazing experience. My commercial amps are AB, mosfet and bipolar, and I use a very good SMSL D6S DAC which can go to 32bit 768KHz, SOTA. I have been listening to audio systems sixty years and learned a thing or two, Russell perhaps even more.
Russell was astonished at two qualities; natural sound, and stunning resolution. I was struck by the natural presentation; he and I tend to hear similar things. Despite having highish THD the harmonic profile seems to generate a sense of natural sound; I have a few ideas about this but I won't explain them because it might not be scientific. However, there is no doubt that the scale, the resolution and sense of 'being there' (apologies to Peter Sellers), and the clarity is in abundance and it's hard to imagine anything could best it.
The imaging was palpable, you could assess the height of the singer, the depth beyond of the percussion, and the width of the stage. Much of this is easily done with good systems, but height is unusual and I found myself agreeing with Anand on this one. Andy's listening room is about 2.5m deep, but 7m wide, speakers apart about 3m. (Andy to correct.) The room is well damped - carpet and some absorbent panels at the front, so there was no reverberation or unmusical sound reflections.
The other quality we noticed was silence. Of course since digital sources have developed, noise has considerably reduced, but with this system there is absolutely NOTHING at the speakers, not even with the ear cupped into the speaker cone. This benefit accrued from Andy's assiduous work with earch loops, also on this thread. I believe this is one of major issues in the improved resolution; we could hear mice breaking wind under the performers with this remarkable system.
Andy took a long time to build his Alpha Nirvana. It only has six semiconductors in each channel, but because of the high quiescent current (close to 2.5A on his system I believe) the thermal considerations require a lot of time and effort. He built is first AN using forced air PC cooling units; both worked extremely well but over time the sound of the fans grew intrusive so he resolved to rebuild using one 5U case with each heatsink serving the power supply and channel for each channel. My assessment is that the operating points of the outputs are on the limit; my view is that dissipating so much heat we should monoblocks; perhaps 4U would be OK but each heatsink could then serve ONE output device, not two. You can see his excellent work at #3045 post in this thread, in fact.
From the point of my design career since the mid-nineties, I suspect I will never do anything better than the Alpha Nirvana. I continue to try, of course, but as it stands it seems to me to be one of the best amps I have ever designed. I recognise the fast and thorough help with XRK to develop this exceptional amp.
Thank you Andy!! I wonderful visit to listen your wonderful Alpha Nirvana, beautifully constructed.........
Cheers,
Hugh
Two weeks ago I was invited by AndyR to listen to his 4R Alpha Nirvana using two quartet Satori line array with open baffle 6.5" speakers and two Satori 1" ribbon tweeters. There is a separate pair of 15" sub-woofers which chime it at <110Hz. The active system is DSP with 48dB/octave slopes, and a high spec DAC using a variety of conventional and HD files. I brought along a good friend, Russell, who owns a Maya 200W, Magico and Harbeth speakers, both three ways. At my suggestion Andy replaced each NMOS output with a NPN Darlington from Sanken (2SD2560) for a little more slam and impact. Russell is an avid audiophile and very skilled at assessing sound quality.
Now a couple of weeks on we both look back on an amazing experience. My commercial amps are AB, mosfet and bipolar, and I use a very good SMSL D6S DAC which can go to 32bit 768KHz, SOTA. I have been listening to audio systems sixty years and learned a thing or two, Russell perhaps even more.
Russell was astonished at two qualities; natural sound, and stunning resolution. I was struck by the natural presentation; he and I tend to hear similar things. Despite having highish THD the harmonic profile seems to generate a sense of natural sound; I have a few ideas about this but I won't explain them because it might not be scientific. However, there is no doubt that the scale, the resolution and sense of 'being there' (apologies to Peter Sellers), and the clarity is in abundance and it's hard to imagine anything could best it.
The imaging was palpable, you could assess the height of the singer, the depth beyond of the percussion, and the width of the stage. Much of this is easily done with good systems, but height is unusual and I found myself agreeing with Anand on this one. Andy's listening room is about 2.5m deep, but 7m wide, speakers apart about 3m. (Andy to correct.) The room is well damped - carpet and some absorbent panels at the front, so there was no reverberation or unmusical sound reflections.
The other quality we noticed was silence. Of course since digital sources have developed, noise has considerably reduced, but with this system there is absolutely NOTHING at the speakers, not even with the ear cupped into the speaker cone. This benefit accrued from Andy's assiduous work with earch loops, also on this thread. I believe this is one of major issues in the improved resolution; we could hear mice breaking wind under the performers with this remarkable system.
Andy took a long time to build his Alpha Nirvana. It only has six semiconductors in each channel, but because of the high quiescent current (close to 2.5A on his system I believe) the thermal considerations require a lot of time and effort. He built is first AN using forced air PC cooling units; both worked extremely well but over time the sound of the fans grew intrusive so he resolved to rebuild using one 5U case with each heatsink serving the power supply and channel for each channel. My assessment is that the operating points of the outputs are on the limit; my view is that dissipating so much heat we should monoblocks; perhaps 4U would be OK but each heatsink could then serve ONE output device, not two. You can see his excellent work at #3045 post in this thread, in fact.
From the point of my design career since the mid-nineties, I suspect I will never do anything better than the Alpha Nirvana. I continue to try, of course, but as it stands it seems to me to be one of the best amps I have ever designed. I recognise the fast and thorough help with XRK to develop this exceptional amp.
Thank you Andy!! I wonderful visit to listen your wonderful Alpha Nirvana, beautifully constructed.........
Cheers,
Hugh
Last edited:
Thank you, Hugh - it was a pleasure to show off my ANs to you & Russell after reducing the bias to 2.5a.
Only one correction - on room size; the room is:
* 3.6m deep - not 2.5m
* and only 5.8m wide - not 7m.
Andy
Hi Mr Hugh.
What is the maximum safe temperature limit in the body of the mosfet that should work? No in data sheet...
@Andy do you measure the temperature on case of the body mosfets ? also on the heatsinks? 5U box?
Once we have the same current about 2.6A.
What is the maximum safe temperature limit in the body of the mosfet that should work? No in data sheet...
@Andy do you measure the temperature on case of the body mosfets ? also on the heatsinks? 5U box?
Once we have the same current about 2.6A.
Attachments
Last edited: