Hi there
This is an idea I have been toying with for a long time, and I finally decided myself to materialize it, in sim at least.
Why?
Ordinary VAS's require at least one voltage translation, sometimes two. I thought it would be advantageous to start directly from ground.
Of course, one cannot actually start from ground, because transistors have a finite Vce, and a correct crossover couldn't happen without some bias.
Basically, it consists of two complementary amplifiers designed to have identical output voltages, except one spends almost 50% of the time in saturation whilst the other is active, and vice versa.
The outputs are then combined to form the final output.
The combining gate must use transistors, because diodes aren't "intelligent" enough in this case.
This is just a concept of course, and the example I gave is rudimentary and simplified: many issues would need to be addressed, like antisaturation circuits for the amplifiers, etc, and this implementation is only one of the many possible, but its purpose is mainly to start the discussion.
I do not necessarily see this concept as primarily applicable to ordinary amplifiers, but maybe to ESL amplifiers, or hybrid amplifiers, for driving a circlotron for example.
Thoughts, opinions, comments?
This is an idea I have been toying with for a long time, and I finally decided myself to materialize it, in sim at least.
Why?
Ordinary VAS's require at least one voltage translation, sometimes two. I thought it would be advantageous to start directly from ground.
Of course, one cannot actually start from ground, because transistors have a finite Vce, and a correct crossover couldn't happen without some bias.
Basically, it consists of two complementary amplifiers designed to have identical output voltages, except one spends almost 50% of the time in saturation whilst the other is active, and vice versa.
The outputs are then combined to form the final output.
The combining gate must use transistors, because diodes aren't "intelligent" enough in this case.
This is just a concept of course, and the example I gave is rudimentary and simplified: many issues would need to be addressed, like antisaturation circuits for the amplifiers, etc, and this implementation is only one of the many possible, but its purpose is mainly to start the discussion.
I do not necessarily see this concept as primarily applicable to ordinary amplifiers, but maybe to ESL amplifiers, or hybrid amplifiers, for driving a circlotron for example.
Thoughts, opinions, comments?
Attachments
And now for something completely different 
But will it be of any use ?
BJT are known for slow recovery from saturation and combined with the high base resistance of the "intelligent" diodes perhaps poor high frequency response.
It looks like the two sides saturation is overlapping leaving the output floating for a while.Without drive it goes where ever the load pulls.
The two input resistors 5k7 and 5k9, any purpose?
Mona
But will it be of any use ?
BJT are known for slow recovery from saturation and combined with the high base resistance of the "intelligent" diodes perhaps poor high frequency response.
It looks like the two sides saturation is overlapping leaving the output floating for a while.Without drive it goes where ever the load pulls.
The two input resistors 5k7 and 5k9, any purpose?
Mona
Aren't Q3, Q4, the wrong way around? Also, the output stage power supply is 15V, only the 47k is returned to 200V. How can you get 200V out in a load that way?
And what is the significance of the Vs you show; as I see it it is defined as a multiple of Vin, so has nothing to do with the circuit. Or am I missing something?
Jan
Last edited:
Who knows?And now for something completely different
But will it be of any use ?
Sure, but I warned that the purpose of this circuit was just to start the discussion: it leaves a lot to be desired, so please, do not nit-pick the details, they can be mended if required, let us concentrate on the concept.
No, that's precisely the purpose of having negative and positive bias/offsets on the two amplifiers, and the combining gate also has a bias (the 4.7Meg resistors) preventing it from going high impedance.
In the pic below, you can see that the base currents of the gate transistors overlap in the zero Xing region, which means the circuit is always in control of the output.
They shift the amplifiers towards the negative and positive, to allow for an overlapping linear region.
They aren't exactly identical because the transistors aren't perfect complementary either, and their asymmetries compensate for the slight gain mismatch between the pos and neg sides
No, it is a bit of a mind twister, but the schematic is correct: Q3 handles the positive part of the waveform generated by Q1, and Q4 the negative part from Q2
No, the 15V supply is the auxiliary bias, and the 200V is the output supply (200+15V in fact).
The following pic shows the voltages on the collectors of Q1 and Q2: Q1 is red, Q2 is green
That is precisely the beauty of this circuit (as imperfect as it may be): with 200V supplies, you can get up to 400Vpp output, with no transistors seeing more than 200V peak, and no active circuitry tied to the HV rails: everything is handled from the ground/aux-supplies side.
Vs is the simulated output of an ideal amplifier having just the gain of the circuit (the gain is 13.25x) . It allows you to see any deviation from the perfect waveform.
Attachments
Last edited:
Yes, I see now, putting in the voltages/currents in zero posistion shows it.
Mona
Attachments
Looks like some kind of current mirror input stage with current gain and folded cascode sliding bias OPS scheme??
Beware of cross conduction dissipation at high frequencies.
Beware of cross conduction dissipation at high frequencies.
No, it bears a superficial resemblance with current mirrors, but in fact they are just ordinary CE amplifiers. They are temperature-compensated, because the stability of the operating points is important, and slightly level shifted, to ensure a correct crossover, but that's all.
There is no current gain, cascode, folded or not, or sliding bias either: just what's said on the box: complementary amplifiers followed by a gate.
As for the cross-conduction, the worst it could do is short the two 47K from the HV together, with very unspectacular consequences.
The circuit amazes me, basically since I really can't figure out the purpose of it. I can see that it works but what good does the output stage do? It doesn't lower the output impedance and it just adds distortion.
Regards
Regards
ESL driver, piezo-stack driver, circlotron driver to name a few, and I am certain others could find many more examples
Well, it just amplifies, as the "A" in VAS implies, it does not mow your lawn or lower impedance, because it has never been designed for that, but it could, quite easily (lower the impedance I mean, the lawn would be negotiable too, but more difficult): if you include followers using the same principles, you will end up with a complete unusual class B amplifier, capable of putting out hundreds volts at any arbitrary number of amperes, but I don't see what good it would do....
Regarding the distortion, it is the true and sad reality that any circuit (even the proverbial straight wire in fact) distorts, but this one is not particularly bad (or good) in this respect: you can check it visually on the first pic, where the output trace (green) is overlaid on the ideal waveform (red).
Since you don't see that much red, the circuit cannot be that bad (and I made minimal efforts to adjust the coefficients).
Anyway there is no kind of GNFB at all: if this block is included in a complete project, it will automatically include GNFB (except if the builder has an audiophiliac aversion to any kind of feedback)
OK, Elvee. I still don't get it. Perhaps you are getting tired of my intellectual numbness but what is the electronical purpose? You mention circlotron driver and other stuff but what does that output stage do?
I mean, I could invent a VAS with an always closed relay at the output. It would indeed work but the relay desn't really have a purpose.
I'm very fond of new stimulating ideas and ways to accomplish things and would be glad being enlightened in this matter.
Regards
I mean, I could invent a VAS with an always closed relay at the output. It would indeed work but the relay desn't really have a purpose.
I'm very fond of new stimulating ideas and ways to accomplish things and would be glad being enlightened in this matter.
Regards
First, let me remind you of what a great thinker once said:
"Something having no definite purpose is the most powerful of secret weapons, because it can be used for any purpose"*
There are hundreds if not thousands of amplification schemes/topologies around, and most of them have more than one single application:
In general, it will be possible to use a number of topologies for a specific purpose, and all of them will have peculiarities, advantages or issues, which is why it is useful to have a broad range to chose from.
This is just one more topology, maybe a drop in the ocean, or a fart in the wind, but there are certainly situations where it will be superior.
Take for example an OTL ESL amplifier: if you build it the conventional way, you will need V/I converters on the input side, and current mirrors on the HV rails.
Complementary transistors are limited to ~600V Vce, which means the p to p output voltage will also be limited to 600V, 300V peak.
With tricks, like H drive, you can go higher, but with the GND-centric VAS, you just need plain amplifiers on the low voltage side, and the transistors will see only half of the p to p voltage, which means an output of 1200Vpp or 600V peak. And you are still free to use the other tricks to go higher if required.
There are many fields of electronics where an easily obtained HV drive is useful: PZT drive, deflection amplifiers (CRT have gone, but many scientific instruments like SEM still use vacuum techniques).
You could even try to apply it to a conventional audio amplifier: all you have to do is to add followers.
I am not sure about the results, because the class B switching will be performed upstream of the OP transistors, but at least it would eliminate the hassle of base spreader, temperature compensation, etc.
*The thinker is Pierre Dac, probably helped by his friend Duval, and the invention in question is the Biglotron. The (poor) free translation is mine
"Le Biglotron est le nom d'une machine, fruit dans la série de l'imagination du professeur Slalom Jérémie Ménerlache, que Pierre Dac avait inventée quelques décennies plus tôt et qu'il présentait dans ses conférences scéniques ou radiodiffusées. C'est un autre nom donné au schmilblick, une machine « qui ne sert à rien, et par conséquent peut servir à tout » et de ce fait doit assurer à la France la suprématie pour des décennies dans des domaines « variés, allant de la physique thermonucléaire jusqu'à la pratique de la gynécologie dans l'espace »."
https://www.youtube.com/watch?v=9bs1WVDkaRY
Last edited:
That's exactly what I wanted to hear. A practical purpose, namely to let the output transistors work with double their voltage ratings. Very nice.
But on the other hand, the first stage still has to be capable of full voltage swing.
I'm a practical man and I think it's always nice to get a concrete example of what the circuit may be able of.
Actually, I have performed that trick earlier using another technique. For some reason I substituted an output tube with a BJT that only could handle 120V. I extended the range by adding two IRF610 in a cascode like manner so the whole thing withstood around 500V.
BTW, is there any practical way to build oneself a couple of ESL speakers using alu-foil or other cheap stuff?
But on the other hand, the first stage still has to be capable of full voltage swing.
I'm a practical man and I think it's always nice to get a concrete example of what the circuit may be able of.
Actually, I have performed that trick earlier using another technique. For some reason I substituted an output tube with a BJT that only could handle 120V. I extended the range by adding two IRF610 in a cascode like manner so the whole thing withstood around 500V.
BTW, is there any practical way to build oneself a couple of ESL speakers using alu-foil or other cheap stuff?
Very interesting but I don't agree with <<qui ne sert à rien, et par conséquent peut servir à tout>> 🙁
Cicéron c'est Poincaré, mais si c'est Poincaré n'implique pas que c'est rond ! 😀
Mona
No, it doesn't, look more closely
Standard practice, indeed
You should ask on the dedicated forum
Yes, thinking unboxed is very pleasurable. I see now that your little circuit works, it's quite clever. Simple things are normally hard to invent.
But this illustrates a bit of the "engineer problem". People who master technical stuff usually has problems with describing their babies in common language. You mentioned earlier "voltage translation". That didn't say anything to me. Perhaps you assume that everyone visiting this thread already knows what it's all about, but explaining in simpl words is a virtue.
I would have put it this way:
"This is a nice little cicuit I have fiddled with some time. The mening with it is to let the VAS work with twice the voltage that the transistors are rated for".
A second thing. The switching will probably generate cross over distortion. Human hearing is very sensitive to those kind of artefacts, but maybe they are low enough.
My "basic" solution you mentioned doesn't give any distortion at all.
But it is an inventive little circuit, perhaps you can get a patent for it and it will probably be useful for someone, some day.
Good luck.
But this illustrates a bit of the "engineer problem". People who master technical stuff usually has problems with describing their babies in common language. You mentioned earlier "voltage translation". That didn't say anything to me. Perhaps you assume that everyone visiting this thread already knows what it's all about, but explaining in simpl words is a virtue.
I would have put it this way:
"This is a nice little cicuit I have fiddled with some time. The mening with it is to let the VAS work with twice the voltage that the transistors are rated for".
A second thing. The switching will probably generate cross over distortion. Human hearing is very sensitive to those kind of artefacts, but maybe they are low enough.
My "basic" solution you mentioned doesn't give any distortion at all.
But it is an inventive little circuit, perhaps you can get a patent for it and it will probably be useful for someone, some day.
Good luck.
A picture is worth a thousand words, and the picture was present from the beginning
I am interested in discussing the concept of ground-centric amplification, because I find it interesting and powerful, and I have offered a simple example, as a basis for discussion.
I found advantages and justifications to the modest example I provided, mainly a posteriori, and because I was asked for, but basically, that's not what I am interested in: I am a practical guy too, but being too centered on utilitarianism prevents you from seeing the big picture, and I know at first sight when topologies have a potential, even when no practical advantage or application is directly apparent.
Sure, and if you reread carefully the whole thread, you will see that I have mentioned that kind of problem myself, as I did in fact for most of your objections.
As with all topologies, there are issued to be addressed, but that is a technical problem, and technical problems can always be solved, given enough means.
When the means to benefit ratio becomes too large, the balance is unfavorable, and that is often the case, but that is another question
That is a bold statement, and let us just say that it brings its own kind of problems, but at least they are well documented, since it is the default ad hoc solution for such a problem
I could certainly have gotten a vanity patent on it (unless there is prior art), but I am not interested in vanity patents, and now that it has been disclosed, it is definitely out
I hope and I think it is a real possibility
- Status
- Not open for further replies.