It's 'just' engineering. Throw enough money at it and pretty much everything is possible ;-)
But I don't see that the problems with an analog solid state amp magically disappear by making it class D.
BTW I just found out the hard way that you should not leave high voltage connectors open ended. The mating plug construction is an integral part of the isolation.
Jan
But I don't see that the problems with an analog solid state amp magically disappear by making it class D.
BTW I just found out the hard way that you should not leave high voltage connectors open ended. The mating plug construction is an integral part of the isolation.
Jan
Attachments
Actually, it might not be a bad idea, if you can get those high voltage FETs or IGBTs to switch fast enough.
They are optimized for switching.
There would probably still be significant losses.
Jan
A feedback loop around the transformer (the good ol analogue way) has been on my mind some times...
A voltage divider at the secondary, or maybe better a series resistor and apply current feedback.
Doable... maybe the phase shift is to large at high frequency?
Below 100Hz i dont play since i go active.
A voltage divider at the secondary, or maybe better a series resistor and apply current feedback.
Doable... maybe the phase shift is to large at high frequency?
Below 100Hz i dont play since i go active.
If we knew the transfer function of the transformer and assume that the transfer function is somewhat stable over the listening range.
Then we could model that and compensate with 180deg phase shifted harmonics in a DSP in the preamplifier stage.
I know that they use this trick in subwoofers, and they can reduce the distortion with 20dB. More complicated but doable would be to model it over a wide range of amplitudes in the DSP...
Then we could model that and compensate with 180deg phase shifted harmonics in a DSP in the preamplifier stage.
I know that they use this trick in subwoofers, and they can reduce the distortion with 20dB. More complicated but doable would be to model it over a wide range of amplitudes in the DSP...
Back to Jan's direct drive. It seems this is a single ended class A design. The least efficient amp possible. Since the the load is by definition AC coupled (no DC though this cap) you could cut the idle power consumption by a lot by driving the positive and negative circuits, instead of using the positive side as a current source. AC coupling the drive signals should not be an issue. Turn on/off transients will of course be an issue.
Current feedback with the transformer would be great. There are two transformers so the ground referenced sides are accessable. Another alternative is the AP transformer drive which is known to work quite well : https://patents.google.com/patent/US4614914A/en
Current feedback with the transformer would be great. There are two transformers so the ground referenced sides are accessable. Another alternative is the AP transformer drive which is known to work quite well : https://patents.google.com/patent/US4614914A/en
Yes Demian, it's class A but note that the current source load is dynamically driven (the Nelson Pass method) so the efficiency is almost double what a pure class A would be. But class AB would of course be better.
My initial idea was to place a class AB output stage after the current design (throttled down) so I could drive up to 100mA or more into non-segmented, pure capacitance speakers like the Martin Logans.
Just haven't got around to it, and my current priority is to double the output voltage so the ESL 63's can be driven to full output.
Jan
My initial idea was to place a class AB output stage after the current design (throttled down) so I could drive up to 100mA or more into non-segmented, pure capacitance speakers like the Martin Logans.
Just haven't got around to it, and my current priority is to double the output voltage so the ESL 63's can be driven to full output.
Jan