PMA said:
Hi.
Good question.
And the answer is very simple (but not technical).
With my amplifier I can resolve a full stereo image in front of me and hear all the sounds with clarity and they are stable.
I can accuratly place the position of a sound - with a good recording I can hear the movement of the players.
I can also listen to one instrument or sound, and follow it even when it is all but hidden in amounst other louder sounds (the sound of brass cartridges hitting a marble floor within the roar of continious automatic fire for example - The Matrix, Lobby Sceen).
On Jean Michelle Jare's Concert's in China about the third track there is an auditary firework fizzing and moving from left to right and back again. My amp keeps the sound defined and I can track it constantly.
Any other amplifier I have listed to leaves a hole and everything is muddy and ill defined.
The big problem of course is that whilst we can all talk about this audio stuff we can't actually hear the result unless we build them or know someone who has one.
I built the amps in the form you see in the picture 10 years ago. Since then most people have derided my design, particually the experts - but none of those have actually listend to them.
I fully appreciate that there are a lot of very good amplifiers out there on the market and for people to build. And some will probably be considered better than mine.
However at the end it's my preferences and own hearing which is what I am comfortable with and like, and I fully accept that other people have other values and tastes.
Different strokes, etc.
Best wishes,
Susan.
===
I will reply seperatly to the other questions in the morning.
Meanwhile I thank everyone for their positive comments.
Hi Susan,
Thank you for your seminal post; this is indeed full of content!
I agree with all the sentiments here expressed; particularly Jan's.
The big ticket on this amp appears to be the quality of the transformers. Interestingly, even though even order distortions are cancelled by the diff drive at the output stage, from your comments they do not seem to skew the distortion spectrum towards bad sound, with odd order predominating. I suspect this could be because the transformers add a little low order distortion, restoring the distribution. Does it sound warm, like a tube amp, or somewhat sharper?
What sort of quiescent do you run through center tap for 35W into 8R, Susan?
The low impedance (600R) drive is something of an issue; you would need an active transimpedance device at the input. If source impedance were nominally 600R, and input trafo step up of say 20 times, then Zout would be around 240K, rather high. I assume that this is sufficient to drive the gates effectively? Of course, you might actually be using a source impedance much lower than this, say 50R, in which case Zout would be typically 20K, more than adequate. But, like the Zen amps, this design does require pretty stiff drive.
I commend all to Susan's website at http://www.susan-parker.co.uk/zeus-out-tx.htm. Very clear descriptions!
Again, Susan, thank you for sharing this!
Cheers,
Hugh
Thank you for your seminal post; this is indeed full of content!
I agree with all the sentiments here expressed; particularly Jan's.
The big ticket on this amp appears to be the quality of the transformers. Interestingly, even though even order distortions are cancelled by the diff drive at the output stage, from your comments they do not seem to skew the distortion spectrum towards bad sound, with odd order predominating. I suspect this could be because the transformers add a little low order distortion, restoring the distribution. Does it sound warm, like a tube amp, or somewhat sharper?
What sort of quiescent do you run through center tap for 35W into 8R, Susan?
The low impedance (600R) drive is something of an issue; you would need an active transimpedance device at the input. If source impedance were nominally 600R, and input trafo step up of say 20 times, then Zout would be around 240K, rather high. I assume that this is sufficient to drive the gates effectively? Of course, you might actually be using a source impedance much lower than this, say 50R, in which case Zout would be typically 20K, more than adequate. But, like the Zen amps, this design does require pretty stiff drive.
I commend all to Susan's website at http://www.susan-parker.co.uk/zeus-out-tx.htm. Very clear descriptions!
Again, Susan, thank you for sharing this!
Cheers,
Hugh
Bocka, this design would not be optimum for RF, BUT I once built a similar RF amp with the features you want. Please consult an Amateur Radio Handbook for instructions on how to make a good RF amp. It is completly different from audio, and even my fastest designs would make lousy RF amps, even though I could probably run AM radio through them easily enough, with a little adjustment.
You can still use dual transformers, BUT the advantage of high input impedance is lost above a few MHz, because of the high input capacitance of the fets. Common source pp design will give you more POWER gain, which is what you need.
Susan, I think that your design is straightforward and refreshing. I have worked all my professional career to REMOVE transformers and caps from the audio circuit, but most examples in the marketplace still sound worse than my old Telefunken radio. For example, I have a Class D amp of similar power and when I use it in the same application, it is NASTY! Virtually unlistenable. I was even a consultant, after the fact, on the design. There wasn't much that I could recommend to improve it, however.
Folks, we have to keep our priorities straight. We should try to make good sounding designs, not just good measuring ones, or just theoretically possible ones. You know, infinite feedback means 0 distortion.
You can still use dual transformers, BUT the advantage of high input impedance is lost above a few MHz, because of the high input capacitance of the fets. Common source pp design will give you more POWER gain, which is what you need.
Susan, I think that your design is straightforward and refreshing. I have worked all my professional career to REMOVE transformers and caps from the audio circuit, but most examples in the marketplace still sound worse than my old Telefunken radio. For example, I have a Class D amp of similar power and when I use it in the same application, it is NASTY! Virtually unlistenable. I was even a consultant, after the fact, on the design. There wasn't much that I could recommend to improve it, however.
Folks, we have to keep our priorities straight. We should try to make good sounding designs, not just good measuring ones, or just theoretically possible ones. You know, infinite feedback means 0 distortion.
Only one suggestion
Susan,
I have only one suggestion to make :
I agree totally with you that the only way to assess a design is to build and listen. So to put the other guys to task who says no before trying, may I suggest you make some enquiries with your suppliers and offer us a set of transformers at a reasonable price (whatever that means).
I guess if it is not too unaffordable, you would not have too much difficulty finding enough 'customers' for a batch of 10 or even 20, which helps to keep cost down.
Once there is feedback from different corners, the sceptics would shut up.
Me, I am not qualified to comment. I have not built one. But I also do not rule out that it can sound pleasant. It is an interesting design in any case, and the audio world is definitely a not a poorer place with it.
Patrick
Susan,
I have only one suggestion to make :
I agree totally with you that the only way to assess a design is to build and listen. So to put the other guys to task who says no before trying, may I suggest you make some enquiries with your suppliers and offer us a set of transformers at a reasonable price (whatever that means).
I guess if it is not too unaffordable, you would not have too much difficulty finding enough 'customers' for a batch of 10 or even 20, which helps to keep cost down.
Once there is feedback from different corners, the sceptics would shut up.
Me, I am not qualified to comment. I have not built one. But I also do not rule out that it can sound pleasant. It is an interesting design in any case, and the audio world is definitely a not a poorer place with it.
Patrick
Hi Susan,
What's inside your amp?
In your case that really is a redundant question; but oh how much thinking you must already have done about so little !!!
I'm moving over to your thread here, and I note that there already are some worthy contributions. I'd love to hear your amp, but alas I live north of Belfast so it is not likely I can call round for a demo.
In jcx's other thread I suggested that your input transformer would need to be something rather special, I was thinking of a need for about 1:20+20 of grain orientation + split windings. I see you have described it as 1;5+5, which allows you to use bell wire for the input cable. However, I think figure-8 pairing has a natural impedance closer to 100 rather than 600 ohms, though 600 ohm driving should not have too great an effect at audio frequencies if the cable is only of 'domestic' length. If signal source push-pull/balanced driven at say series 56 ohms per core wrt to ground, the cable should have an unchallengable hf phase response, and could be loaded with a resistive portion across the transformer primary as well as secondary. ( I wrote this yesterday and see that Hugh too has come up with lower impedance and higher ratio, maybe there is room for more investigation here.)
I have no wish to upset or dampen your enthusiasm, but as with PMA I too would wonder about phase accuracy, because an amplifier can sound good enough for listening satisfaction (and easily better a Quad set-up) without it being as accurate as is possible with other designs that are equally capable of resolving swaying or breathing musicians. You mention a 300kHz bandwidth; is this -3dB, and did you ever run an audio frequency phase check ? I suspect transformer core losses, which become both measurable and so audible when you A/B compare mains types pressed into audio usage when compared with properly grain orientated versions.
With regard to driving the loudspeakers directly from source + c.t. transformer connections used as a choke I recon that per half hf voltage due to loss of coupling between halves and per half inductance might be theorectically less natural than via your separate core coupled winding, but it would be worth auditioning in case it fortuitously compensates for transformer imperfections.
Did you ever measure the output resistance/impedance at say 1kHz ? Have you managed to achieve a decent loudspeaker damping figure without NFB ? Have you tried listening whilst adjusting the quiescent bias and found a level at which further increase does not improve reproduction, or did you set for a reasonable dissipation ? Have you ever subjected the amplifier to distortion test set measurement ?
Like johnnyx I could ask myself if I need to change direction, as inded I have already have following Pavel's new circuit release, but as I prefer transformerless power, regard 100W-8 ohms a requirement, and as it is possible to minimise feedback induced problems, then I would need to know phase, distortion and output impedance figures of your design before I could be convinced. I would rather have an accurate feedback amplifier and then pre-amp modify its signal for listening that suits my ears, than have an amplifier that is easy to listen to but which compromises waveform accuracy in a way that can never be subsequently adjusted out.
I must state that I have not heard this kind of amplifier, and so I cannot state whether it is cleaner sounding than a feedback amplifier or not, so please do not take my points as being either an attempt to be definitive or derogatory.
You like the Line Source too ! What is your driver ? Mine is Visaton FRS-8, far from perfect, but cheap and clean over mid range and needs a central top end tweeter, though I'm always on the lookout for other suggestions.
Cheers ............. Graham.
What's inside your amp?
In your case that really is a redundant question; but oh how much thinking you must already have done about so little !!!
I'm moving over to your thread here, and I note that there already are some worthy contributions. I'd love to hear your amp, but alas I live north of Belfast so it is not likely I can call round for a demo.
In jcx's other thread I suggested that your input transformer would need to be something rather special, I was thinking of a need for about 1:20+20 of grain orientation + split windings. I see you have described it as 1;5+5, which allows you to use bell wire for the input cable. However, I think figure-8 pairing has a natural impedance closer to 100 rather than 600 ohms, though 600 ohm driving should not have too great an effect at audio frequencies if the cable is only of 'domestic' length. If signal source push-pull/balanced driven at say series 56 ohms per core wrt to ground, the cable should have an unchallengable hf phase response, and could be loaded with a resistive portion across the transformer primary as well as secondary. ( I wrote this yesterday and see that Hugh too has come up with lower impedance and higher ratio, maybe there is room for more investigation here.)
I have no wish to upset or dampen your enthusiasm, but as with PMA I too would wonder about phase accuracy, because an amplifier can sound good enough for listening satisfaction (and easily better a Quad set-up) without it being as accurate as is possible with other designs that are equally capable of resolving swaying or breathing musicians. You mention a 300kHz bandwidth; is this -3dB, and did you ever run an audio frequency phase check ? I suspect transformer core losses, which become both measurable and so audible when you A/B compare mains types pressed into audio usage when compared with properly grain orientated versions.
With regard to driving the loudspeakers directly from source + c.t. transformer connections used as a choke I recon that per half hf voltage due to loss of coupling between halves and per half inductance might be theorectically less natural than via your separate core coupled winding, but it would be worth auditioning in case it fortuitously compensates for transformer imperfections.
Did you ever measure the output resistance/impedance at say 1kHz ? Have you managed to achieve a decent loudspeaker damping figure without NFB ? Have you tried listening whilst adjusting the quiescent bias and found a level at which further increase does not improve reproduction, or did you set for a reasonable dissipation ? Have you ever subjected the amplifier to distortion test set measurement ?
Like johnnyx I could ask myself if I need to change direction, as inded I have already have following Pavel's new circuit release, but as I prefer transformerless power, regard 100W-8 ohms a requirement, and as it is possible to minimise feedback induced problems, then I would need to know phase, distortion and output impedance figures of your design before I could be convinced. I would rather have an accurate feedback amplifier and then pre-amp modify its signal for listening that suits my ears, than have an amplifier that is easy to listen to but which compromises waveform accuracy in a way that can never be subsequently adjusted out.
I must state that I have not heard this kind of amplifier, and so I cannot state whether it is cleaner sounding than a feedback amplifier or not, so please do not take my points as being either an attempt to be definitive or derogatory.
You like the Line Source too ! What is your driver ? Mine is Visaton FRS-8, far from perfect, but cheap and clean over mid range and needs a central top end tweeter, though I'm always on the lookout for other suggestions.
Cheers ............. Graham.