quickies
Since each are in the same league so far as power goes (100W), which sounds the best, the LeachAmp, ESD 3A, or AKSA? Any others designs which should fit in this list?
I've spent a fair amount of time studying all three, and they seem to be pretty comparable. The 3A is very simple, as is the AKSA, but is it worth it to spend the extra time with the LeachAmp? Also, is the 3A as stable at 100W (42V rails)?
An unrelated question (kinda): I plan on using these amps with ESLs. I've heard of amps directly driving the ESLs without the need for a step-up transformer, but it's hard to find information on them. Not to mention, they don't have the DIY approved sticker the LeachAmp, 3A, and AKSA have. Could an amp effectively be modified to output at 5kV (or something in that range) without changing the entire design?
Since each are in the same league so far as power goes (100W), which sounds the best, the LeachAmp, ESD 3A, or AKSA? Any others designs which should fit in this list?
I've spent a fair amount of time studying all three, and they seem to be pretty comparable. The 3A is very simple, as is the AKSA, but is it worth it to spend the extra time with the LeachAmp? Also, is the 3A as stable at 100W (42V rails)?
An unrelated question (kinda): I plan on using these amps with ESLs. I've heard of amps directly driving the ESLs without the need for a step-up transformer, but it's hard to find information on them. Not to mention, they don't have the DIY approved sticker the LeachAmp, 3A, and AKSA have. Could an amp effectively be modified to output at 5kV (or something in that range) without changing the entire design?
Altaic,
Generally, to drive an ESL, you need very high voltage swings, typically 2000Vpp. The panels are seen by the amp as reasonably high Q capacitors, with almost infinite DC resistance. This is a very tough load for a transistor to drive because the V/I lag creates very high back emf which can easily exceed the SOAR of the transistor and destroy it in microseconds.
Very high voltage bipolars are available, but the higher you go, the slower they are, and their linearity starts to suffer. Both these parameters are highly audible with dynamic speakers, so it's reasonable to assume an ESL would pick them up too. By poor linearity, I mean that their beta is not constant with collector current. Arguably the best devices to use are mosfets, since they are readily available in high voltage ratings (particularly hexfets from IRC), handle appreciable current, and do not suffer hot spots due to thermal runaway. And their thermal characteristics and toughness make them great choices in HV applications.
The problem is they are difficult to drive because of their gate capacitance, but there are tricks to that too.
The foregoing shows us that driving an ESL is not easy because solid state, with few exceptions, is not very compatible with these sorts of voltages and/or drive is an issue. In my view, you are better off using tubes.
You can easily find tubes which operate at very high voltages. A couple of examples are the 211 and the 845. They are also much more linear than either bipolars or mosfets, and generally easier to drive. You'd probably need 50mA of current available at 20KHz in a largish ESL panel. The ratio of voltage and current with tubes is far more suited to this application, and tubes are electrically extremely robust, handling reactive loads and high pulse currents with complete reliability.
A push pull amp with a plate voltage of perhaps 2,500V would be ideal for driving an ESL. But it would require a lot of voltage amplification to get it to 2000Vpp; 60dB in fact, from a 2Vpp signal from a CD player. This sort of amplification would require 66dB of open loop gain with maybe 6dB of negative feedback (necessary to keep distortion to acceptable limits), and this would mean three cascaded gain stages each with 22dB of gain; 6SN7s would be good for the first two, and a large triode for the third. It's all achievable, but remember, a phase splitter is required, and at these voltage swings a transformer phase splitter is probably best - and expensive.
Building a PP tube amp of this dimension is a formidable project. The voltages are horrendous, and everything would need to be enclosed in steel cages for safety. The output triode I recommend sports an anode cap at its top which is dangerous and must be fully enclosed to stop busy fingers. The transformers for the power supply would be large and expensive, and since you are going for broke, you might as well make it Class A. This means 2500V on the plate, a quiescent around 80mA, a 200W dissipation in EACH tube, and a fairly expensive tube at that - I'd recommend the 833A, which will easily go to 350W Pd with forced air cooling. But this tube is not easy to drive, not particularly linear, and quite a lot of negative feedback would be needed to smooth the transfer function of the complete amp. The result of all this steam engine inefficiency would be an ethereal, wonderful sound, yes, but an audio output of around 50W for an amp dissipation of around 500W PLUS FILAMENTS per channel!!
I'd be saying 'Is it all worth it?' The hifi virus is alive and well in all of us, but have we yet reached the stage where caution is thrown to the wind, and all sense of balance is lost?
That said, there are some marvellous dynamic loudspeakers out there these days. In the last decade they have come a long way, while the ESL 'sound' has remained pretty much as it was (flamesuit donned!!). Why not just buy a really good dynamic speaker and drive it with a quality SS amplifier?
Food for thought........
Cheers,
Hugh
www.printedelectronics.com
Generally, to drive an ESL, you need very high voltage swings, typically 2000Vpp. The panels are seen by the amp as reasonably high Q capacitors, with almost infinite DC resistance. This is a very tough load for a transistor to drive because the V/I lag creates very high back emf which can easily exceed the SOAR of the transistor and destroy it in microseconds.
Very high voltage bipolars are available, but the higher you go, the slower they are, and their linearity starts to suffer. Both these parameters are highly audible with dynamic speakers, so it's reasonable to assume an ESL would pick them up too. By poor linearity, I mean that their beta is not constant with collector current. Arguably the best devices to use are mosfets, since they are readily available in high voltage ratings (particularly hexfets from IRC), handle appreciable current, and do not suffer hot spots due to thermal runaway. And their thermal characteristics and toughness make them great choices in HV applications.
The problem is they are difficult to drive because of their gate capacitance, but there are tricks to that too.
The foregoing shows us that driving an ESL is not easy because solid state, with few exceptions, is not very compatible with these sorts of voltages and/or drive is an issue. In my view, you are better off using tubes.
You can easily find tubes which operate at very high voltages. A couple of examples are the 211 and the 845. They are also much more linear than either bipolars or mosfets, and generally easier to drive. You'd probably need 50mA of current available at 20KHz in a largish ESL panel. The ratio of voltage and current with tubes is far more suited to this application, and tubes are electrically extremely robust, handling reactive loads and high pulse currents with complete reliability.
A push pull amp with a plate voltage of perhaps 2,500V would be ideal for driving an ESL. But it would require a lot of voltage amplification to get it to 2000Vpp; 60dB in fact, from a 2Vpp signal from a CD player. This sort of amplification would require 66dB of open loop gain with maybe 6dB of negative feedback (necessary to keep distortion to acceptable limits), and this would mean three cascaded gain stages each with 22dB of gain; 6SN7s would be good for the first two, and a large triode for the third. It's all achievable, but remember, a phase splitter is required, and at these voltage swings a transformer phase splitter is probably best - and expensive.
Building a PP tube amp of this dimension is a formidable project. The voltages are horrendous, and everything would need to be enclosed in steel cages for safety. The output triode I recommend sports an anode cap at its top which is dangerous and must be fully enclosed to stop busy fingers. The transformers for the power supply would be large and expensive, and since you are going for broke, you might as well make it Class A. This means 2500V on the plate, a quiescent around 80mA, a 200W dissipation in EACH tube, and a fairly expensive tube at that - I'd recommend the 833A, which will easily go to 350W Pd with forced air cooling. But this tube is not easy to drive, not particularly linear, and quite a lot of negative feedback would be needed to smooth the transfer function of the complete amp. The result of all this steam engine inefficiency would be an ethereal, wonderful sound, yes, but an audio output of around 50W for an amp dissipation of around 500W PLUS FILAMENTS per channel!!
I'd be saying 'Is it all worth it?' The hifi virus is alive and well in all of us, but have we yet reached the stage where caution is thrown to the wind, and all sense of balance is lost?
That said, there are some marvellous dynamic loudspeakers out there these days. In the last decade they have come a long way, while the ESL 'sound' has remained pretty much as it was (flamesuit donned!!). Why not just buy a really good dynamic speaker and drive it with a quality SS amplifier?
Food for thought........
Cheers,
Hugh
www.printedelectronics.com
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.