I respect your opinion and, in part, I think the same of you.
But I also think that dx trust is an exception to that. it is unconditionally stable and COLD.
My problem was due to the cap of the zobel net. 22nf is too small in that position and is gonna make the amp unstable.
Consider also that the amp is uncoditionally stable also without a zobel net., it is that too small value that made it unstable.
Ok, thanks for the reply. I think a better way to do this would be to fit 2.2 or 2.7k for both R9 and R10. That way you keep the resistance ratio you found to be good but keep load impedance up. The impedance should be higher than 1k, I think. Alternatively, try a standard type BD139 in place of BD139/16. If this is not as effective, you can always return to your 1k values, though the amplifier will probably seem weaker at high volume settings, if you understand what I mean.
Regarding DX's ideas on amplifiers; we should keep his often said preferences for sound quality in mind when appraising his designs. I'll have to paraphrase his words, so this is not a quote but you get the idea when he related some advice given to him: Remember the good ones, they are the ones on the edge of oscillation. Scary, eh?
I built a DX ES version and it did sing eventually, when I realised that not all BD139/140 are created the same. I had about 8 branded types to hand, even some Philips examples from the the 1970s. There was a considerable range of Hfe and output capacitance between them and the 'scope results were surprisingly different too. There were other difficulties with getting the sound just right for me but this was never about making laboratory quality amplifiers with flat response or ultra low distortion - DX is clearly focused on sound quality as it appeals to him but he felt it necessary to drop that design for the more stable DX ST version which so many built and gave very favourable feedback over a long period.
Many here admire his efforts as an amateur designer, advocate, very experienced builder, researcher and experimenter. The forum owes a lot of its popularity to his input, personal news, views, rants and Youtubes about anything at all affecting his audio journey and participation here. Nice work, Carlos.
So, you mean class A, single supply, perfectly stable.....and sound great?
I think the "sound great" part there implies some audible low order distortion and a little instability, actually. You can't simply identify distortion or hear instability as what they are, but if you happen to like the resulting sound effects, you are not alone, and that is often the case for many audiophile folk.
So, if it's the sound effects you like to hear and in well known and liked class A amps, you should study, compare and try prototypes of Sugden's commercial products like the A21 and perhaps A28 amplifiers and early variations. These are part class A, part AB and you can find the schematics all over the web, Including here: J. E. Sugden Class A amplifiers - The 1960s and 1970s
I think the "sound great" part there implies some audible low order distortion and a little instability, actually. You can't simply identify distortion or hear instability as what they are, but if you happen to like the resulting sound effects, you are not alone, and that is often the case for many audiophile folk.
So, if it's the sound effects you like to hear and in well known and liked class A amps, you should study, compare and try prototypes of Sugden's commercial products like the A21 and perhaps A28 amplifiers and early variations. These are part class A, part AB and you can find the schematics all over the web, Including here: J. E. Sugden Class A amplifiers - The 1960s and 1970s
I described those Sugden amps as part class A, part class AB and at 0.9A, that is not very low bias. A JLH class A amplifier for example, requires only 1.25A for full class A operation @ 8Ω with a similar supply voltage. The relevant point is that most of what you hear will be well below peak clipping levels in class A and anything beyond low level sound is irrelevant to the class of operation anyway.
Given a larger power supply and heatsinks and a suitable control circuit, there is nothing to stop you raising the bias up to the point of burning the output stage, if you must have full bias up to clipping levels.
Rather that stand rigidly on a design principle though, listen to the sound quality before presuming what it will sound like. You may hear differences, as with any different models, but it won't be bad or necessarily because one does not have full bias here.
Given a larger power supply and heatsinks and a suitable control circuit, there is nothing to stop you raising the bias up to the point of burning the output stage, if you must have full bias up to clipping levels.
Rather that stand rigidly on a design principle though, listen to the sound quality before presuming what it will sound like. You may hear differences, as with any different models, but it won't be bad or necessarily because one does not have full bias here.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.