Hi D,
I see that your amp has an open loop bandwidth of only +-200Hz.
I'm currently focussing a lot on this in my design. I'm breaking my head trying to reach an open loop bandwidth op +-10Khz without using resistive loading of the VAS.
Can someone tell me if this is really that necesarry, some people are telling high open loop BW is important. But, most Solid State amps have rather small open loop BW.
i saw the diagram. It will not work! It only works when the Collector voltage is higher then the basis voltage. Because of the ripple at the collector, the collector is not always higher then the basis voltage. Besides that the whole current has to pass this transistor that produce lot of heat. After all the power supply has to have a low impedance for AC which is acchieved by the usual Condensator but in this case the impedance is the emitter resistance of the emiter which is too high..
Open loop high bandwidth, low gain, already low distortion with little feedback gives the best result.
This is with the VAS local feedback, no resistive loading of the VAS needed, if you like that more. THD20k is the same as before, but higher at lower frequences.
dado
Attachments
Here we go again.
Show us some low distortion with little feedback.
Just two days ago while working on a design this being a current feedback one I saw how damaging this low feedback myth really is. With the amp in question I found that lowering the loopgain to get wider bandwith and lower feedback causes the amp to have considerable more higher order harmonics from 4th up. TIM is also much higher.
Jlh has proved and shown this before back in the late 70s, if youre going to use feedback, better it be plenty or rather than just dont use it at all. A certain amount of feedback is necessary before the higher order harmonics are lowered so in fact low feedback is the worst you could do as it will cause them to be higher instead of lower.
Interesting, the local feedback to IPS is done with the 10Meg resistor if see it correct?
I'm gonna play arround with this a bit.
Greetz
Here we go again.
Show us some low distortion with little feedback.
Just two days ago while working on a design this being a current feedback one I saw how damaging this low feedback myth really is. With the amp in question I found that lowering the loopgain to get wider bandwith and lower feedback causes the amp to have considerable more higher order harmonics from 4th up. TIM is also much higher.
Jlh has proved and shown this before back in the late 70s, if youre going to use feedback, better it be plenty or rather than just dont use it at all. A certain amount of feedback is necessary before the higher order harmonics are lowered so in fact low feedback is the worst you could do as it will cause them to be higher instead of lower.
OK... Now define how much is low, how much is enuff vs a lot ...
Yeah ...
Yes this is R20 10Meg resistor making local FB from the Q15 collector(cascode transistor) to the Q12 base(VAS transistor).
dado
You wouldnt be able to no matter if you were a prof or a street sweeper.
Fact is fact and myths only really have a place in childrens books and witchcraft.
OK... Now define how much is low, how much is enuff vs a lot ...
Yeah ...
That depends on the design, to quote values is not productive but if youre thinking lets just degenerate some more or kill loopgain with loading and increase bandwith youre just fooling yourself.
Here is the schematic of the low power TT amp with cascoded VAS.
It will be very nice if some of PCB experts(alex mm?) try to make the PCB for this amp. I think it is very simple amp with excellent data, and I am not so good in the PCB design. More powerful version could be a challenge to make the PCB.
Important is to have the drivers and Vbe multiplier on the same small separate heat sink(cca 5W dissipation for low power and 8W for high power).
dado
It will be very nice if some of PCB experts(alex mm?) try to make the PCB for this amp. I think it is very simple amp with excellent data, and I am not so good in the PCB design. More powerful version could be a challenge to make the PCB.
Important is to have the drivers and Vbe multiplier on the same small separate heat sink(cca 5W dissipation for low power and 8W for high power).
dado
Attachments
Dado care to post the spice file ?? Id like to have a closer look with my models. The figures you obtain are very good and Im a little surprised that this design and using TMC can have such figures. The concept is not new, Pioneer used it in the late 80 and Sony has been using the concept for the past 15 years and still do in their latest models but just more refined, neither used TMC though. Figures obtained by Sony with bipolars usually around .003 20hz to 20 Khz.
To summarize what are fortes of those two amps. Nothing new just proven circuit put together to get stable good sounding amp. Quite easy obtainable components. Very high Negative Feedback is used.
High power version:
Transconductance stage - Cascoded LTP with floating bias and Current Mirror
The tail current was set to 11.5 mA to get good Slew Rate and very low noise LTP transistors used
Transimpedance stage - Enhanced Cascode with TMC
Voltage amp(OPS) - Triple EF for high input impedance
Low power version:
Transconductance stage - Simple LTP with Current Mirror
The tail current was set to 11.5 mA to get good Slew Rate
Transimpedance stage - Enhanced Cascode with TMC
Voltage amp(OPS) - Triple EF for high input impedance
Power Supply is as important as the amp it self. Separate Capacitance multipliers for the Input Stage and Output Stage should be situated very close to the amp circuitry on the same Printed Circuit Board. In both amp Backer Clamp was used to prevent OPS saturation.
Any comments or suggestion is very welcome.
dado
High power version:
Transconductance stage - Cascoded LTP with floating bias and Current Mirror
The tail current was set to 11.5 mA to get good Slew Rate and very low noise LTP transistors used
Transimpedance stage - Enhanced Cascode with TMC
Voltage amp(OPS) - Triple EF for high input impedance
Low power version:
Transconductance stage - Simple LTP with Current Mirror
The tail current was set to 11.5 mA to get good Slew Rate
Transimpedance stage - Enhanced Cascode with TMC
Voltage amp(OPS) - Triple EF for high input impedance
Power Supply is as important as the amp it self. Separate Capacitance multipliers for the Input Stage and Output Stage should be situated very close to the amp circuitry on the same Printed Circuit Board. In both amp Backer Clamp was used to prevent OPS saturation.
Any comments or suggestion is very welcome.
dado
Some interesting comments earlier on about feedback.
Bruno Putzeys 'The F-word' (Jan didden kindly posted up a link to this article over in the Blowtorch thread) , should be required reading. Unfortunately, the low feedback, wide bandwidth as a route to sonic Nirvana is not based on engineering fact - and the earlier OP was correct in stating that we've known about this through JLH et al for at least 30 years or more, yet it still persists.
That said, my e-Amp does offer the opportunity through the use of jumpers to select TMC or MC and also to 'wide-band' the loop gain to >40 kHz by resistively loading the VAS. Loop gain is reduced by about 6-8 dB as a result. Currently its set for high loop gain and TMC.
I make no claims on the sonics either way, and my interest in using this approach purely so that I can change the loop gain characteristic and, if there is a clear difference, hear it.
I modified my original Ovation 250W power amp about a year ago for TMC and no VAS loading - so the loop gain is moderately high and the -3 dB Fo at a few kHz. The two amplifers sound very different, and yet currently are using the same comp scheme. So, there is a difference in the sound of amplifiers, and even when they use similar feedback schemes.
Bruno Putzeys 'The F-word' (Jan didden kindly posted up a link to this article over in the Blowtorch thread) , should be required reading. Unfortunately, the low feedback, wide bandwidth as a route to sonic Nirvana is not based on engineering fact - and the earlier OP was correct in stating that we've known about this through JLH et al for at least 30 years or more, yet it still persists.
That said, my e-Amp does offer the opportunity through the use of jumpers to select TMC or MC and also to 'wide-band' the loop gain to >40 kHz by resistively loading the VAS. Loop gain is reduced by about 6-8 dB as a result. Currently its set for high loop gain and TMC.
I make no claims on the sonics either way, and my interest in using this approach purely so that I can change the loop gain characteristic and, if there is a clear difference, hear it.
I modified my original Ovation 250W power amp about a year ago for TMC and no VAS loading - so the loop gain is moderately high and the -3 dB Fo at a few kHz. The two amplifers sound very different, and yet currently are using the same comp scheme. So, there is a difference in the sound of amplifiers, and even when they use similar feedback schemes.
Last edited:
Why to use resistive load of the VAS if you can use local VAS feedback, you are throwing useful gain down the sewer. If I remember correctly Bruno Putzeys say that he likes wide Open loop bandwidth if distortion is substantial and in this case to have uniform distortion over whole audio band. With very low distortion that is not important.
How do you describe the difference between loaded VAS and not loaded?
dado
Precisely because it allows me to experiment between lower loop gain, lower VAS stage local gain and the high loop gain options. High loop gain is not a prerequisite for success - I think the sound of both my amps attest to that.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.