AKSA said:Thank you Bob, Edmond, Terry, for interesting results.
In light of the fact that EC is very complex and difficult to precisely set up, and that comparing apples and oranges is a real risk, it's emerging that the only definitive way to test the viability of EC is to build the circuit, measure it carefully vis a vis the PSpice/LTSpice results, then start the listening tests. Argument here proves nothing, and serves only to bruise egos.
Hi Hugh,
I don't think so. Measurements as well as simulations have shown in an overwhelming way that the THD reduction is mainly due to correction of the OPS and in a far lesser degree due to unloading the VAS from a nonlinear impedance. Therefore I'm missing the rationale to build an amp, just to confirm what we already (should) know.
When experts roundly condemn the Dartzeel NHB108 for its adverse distortion performance (presumably having never heard it) and yet this is an amp which won a Stereophile award, then it is clear to me that the gulf between PSpice, measured results and listening tests is wider than ever and all the technical progress of decades past remains inconclusive. Marketing, perhaps, has done this, but I suspect with the complicity of experts who still cannot correlate the objective with the subjective.
[snip]
Cheers, Hugh
That's precisely the problem: you cannot correlate the objective with the subjective, because the latter is subjective.
Cheers, Edmond.
That's precisely the problem: you cannot correlate the objective with the subjective, because the latter is subjective.
If by "objective," you strictly mean electrical measurements and by "subjective" you mean "listening without any sort of controls," then you're right. However, to slightly paraphrase Inigo Montoya, "You keep using those words. I do not think they mean what you think they mean."
Re: Re: Re: Re: Re: Re: Re: Re: nonlinear VAS loading
Well-stated, Jan.
I'll let it be. I'm disappointed that Edmond took my intentions defensively and in the wrong way. Even if there was an "ah but" to this particular contribution of his, it would in no way mean that that contribution was not valuable. Heaven knows, there have been many "ah buts" to my contributions, from which I have learned a lot.
Cheers,
Bob
janneman said:
Edmond,
You have contributed some interesting ideas to this thread. That makes it extra regrettable that you decided on this course of action.
I'm sure you are well aware that the greatest advances always come from interactions and mutual checking of results, conditions and often subtle differences. It is often the subtle differences that provide the insight. Indeed, the best scientists and researchers go through great lenghts to painstakingly describe how and what they do, including full disclosure of methods, circuits etc. This type of peer review greatly increases confidence in results.
Without this intense interaction and review, any reported results remain at the level of anecdotes. Your work deserves better.
Jan Didden
Well-stated, Jan.
I'll let it be. I'm disappointed that Edmond took my intentions defensively and in the wrong way. Even if there was an "ah but" to this particular contribution of his, it would in no way mean that that contribution was not valuable. Heaven knows, there have been many "ah buts" to my contributions, from which I have learned a lot.
Cheers,
Bob
abzug said:Blind listening tests should be the ultimate arbitration, not measurement and not non-blind listening (measurements can only be a substitute once a metric has been found that fully correlates with perception, but despite the work of Geddes and others this seems not very close at this point).
This is a very good point, but the controversy and difficulty of blind listening tests may never allow it to happen.
Even non-blind properly executed A-B tests are difficult to do and require special equipment if it is to be made easy to switch from one device under test to the other. And, of course, the quality of the A-B test device must be very high.
I built such a device, and it was not easy. Among other things, a complete study on relay distortion and hand selection of relays was necessary. We used the device at HE2007 to enable listeners to switch between a tube amplifier and a solid state amplifier so that they could hear for themselves the difference in sound. They did not know which position of the switch corresponded to which amplifier, and could switch at their leisure while listening to different pieces of music.
Cheers,
Bob
Also, additional picked-up RF and LF noise, caused by addidtional cabling, has to be taken into account. It is really not easy to make it well.
Is this only an issue for the power amplifier output switching? I'd assume that for low power line-level switching any quality sealed latching relay with silver plated contacts would contribute no measurable distortion.Bob Cordell said:Among other things, a complete study on relay distortion and hand selection of relays was necessary.
A switching system would be easy to computerize and connect to blind testing software such as abchr, which makes the process much simpler, with control and navigation of listening material quite easy.
No doubt, but surely it's worth it? For line-level connections at least, shielded and, when supported by the equipment, balanced interconnects can be used, with shielded switch/relay-boxes.Originally posted by PMA
Also, additional picked-up RF and LF noise, caused by addidtional cabling, has to be taken into account. It is really not easy to make it well.
Anyway, sorry to be off-topic...
abzug said:
No doubt, but surely it's worth it? For line-level connections at least, shielded and, when supported by the equipment, balanced interconnects can be used, with shielded switch/relay-boxes.
Anyway, sorry to be off-topic...
In my experience, it has extraordinary importance. Wiring, shielding, cable types, grounding points, PE-to-case connection yes/no etc. I have always found connection between sound difference and RF/HF content, in the same equipment, and I do have measurements archived.
Yes, it is OT, maybe separate thread, but I do not feel like having time to keep long discussion on this. Just experience and results, like Charles Hansen used to say.
Hi,
The usual value of current of a VAS is about 5 to 10 mA.
Would a higher bias, 30 to 100 mA, because of a lower output impedance (with the slight drawback of higher dissipation) would have the same effect as of a VAS buffer, around 1/2 distortion ?
The usual value of current of a VAS is about 5 to 10 mA.
Would a higher bias, 30 to 100 mA, because of a lower output impedance (with the slight drawback of higher dissipation) would have the same effect as of a VAS buffer, around 1/2 distortion ?
forr said:Hi,
The usual value of current of a VAS is about 5 to 10 mA.
Would a higher bias, 30 to 100 mA, because of a lower output impedance (with the slight drawback of higher dissipation) would have the same effect as of a VAS buffer, around 1/2 distortion ?
Hi forr,
Perhaps. But in the first place, the output impedance is governed by the amount of feedback around the VAS, ie Cdom, the beta of the VAS trannies and the emitter resistor of the VAS.
Cheers, Edmond
Re: Re: Re: Re: Re: Re: Re: Re: nonlinear VAS loading
Hi Jan,
In general, I agree with you, completely. But this time we are dealing with a different situation. First, we have to go back to the actual question of Hugh. For your convenience I'll repeat it here:
Insofar as the EC output stage presents a far more uniform impedance to the VAS, I believe this is where you will obtain the majority of your benefits rather than the prevention of waveform compression at high signal levels.
.......Gospel, or garbage?
So any other question or comment that doesn't address Hugh's point, is highly irritating.
Second, I didn't postulated something new (if it was new, then you are right of course), rather I just confirmed the findings (read: real life measurements) of Douglas Self, and the knowledge we do already have, that unambiguously answers Hugh's question.
So any doubt about the validity of my simulation (read: conformation) is unnecessary and highly annoying.
Third, as a matter of fact, I made it very easy for Bob by running also a simulation of his own amp. So checking the results should be a snap, but he is "not going to waste a bunch of time" (his words).
Lat but not least, you are right, my work deserves better. A far better response for example, instead of careless reading (by some people here) of my posts and a torrent of undisciplined rattling on a keyboard.
I hope you will now understand why I "decided on this course of action."
Cheers, Edmond.
janneman said:Edmond,
You have contributed some interesting ideas to this thread. That makes it extra regrettable that you decided on this course of action.
I'm sure you are well aware that the greatest advances always come from interactions and mutual checking of results, conditions and often subtle differences. It is often the subtle differences that provide the insight. Indeed, the best scientists and researchers go through great lenghts to painstakingly describe how and what they do, including full disclosure of methods, circuits etc. This type of peer review greatly increases confidence in results.
Without this intense interaction and review, any reported results remain at the level of anecdotes. Your work deserves better.
Jan Didden
Hi Jan,
In general, I agree with you, completely. But this time we are dealing with a different situation. First, we have to go back to the actual question of Hugh. For your convenience I'll repeat it here:
Insofar as the EC output stage presents a far more uniform impedance to the VAS, I believe this is where you will obtain the majority of your benefits rather than the prevention of waveform compression at high signal levels.
.......Gospel, or garbage?
So any other question or comment that doesn't address Hugh's point, is highly irritating.
Second, I didn't postulated something new (if it was new, then you are right of course), rather I just confirmed the findings (read: real life measurements) of Douglas Self, and the knowledge we do already have, that unambiguously answers Hugh's question.
So any doubt about the validity of my simulation (read: conformation) is unnecessary and highly annoying.
Third, as a matter of fact, I made it very easy for Bob by running also a simulation of his own amp. So checking the results should be a snap, but he is "not going to waste a bunch of time" (his words).
Lat but not least, you are right, my work deserves better. A far better response for example, instead of careless reading (by some people here) of my posts and a torrent of undisciplined rattling on a keyboard.
I hope you will now understand why I "decided on this course of action."
Cheers, Edmond.
Inigo Montoya
Hi Stuart,
So, what do you think 'they' mean?
Please, give me a hint.
Cheers, Edmond.
SY said:If by "objective," you strictly mean electrical measurements and by "subjective" you mean "listening without any sort of controls," then you're right. However, to slightly paraphrase Inigo Montoya, "You keep using those words. I do not think they mean what you think they mean."
Hi Stuart,
So, what do you think 'they' mean?
Please, give me a hint.
Cheers, Edmond.
soap
defensively? Nobody is interested in red herrings. That's why I refuse my cooperation on this soap.
Bob Cordell said:I'm disappointed that Edmond took my intentions defensively and in the wrong way. Even if there was an "ah but" to this particular contribution of his, it would in no way mean that that contribution was not valuable. Heaven knows, there have been many "ah buts" to my contributions, from which I have learned a lot.
Cheers,
Bob
defensively? Nobody is interested in red herrings. That's why I refuse my cooperation on this soap.
Edmond,
Thank you for your answer, and your attempts to explain to me the worthiness of EC, much appreciated.
This conflict with Robert is absurd. You are both very decent technocrats. Please, put it behind you and move on!
May I ask another question? Ordinarily the Vbe distortion of the double EF output stage is largely produced by the output device itself, whose Vbe runs from 600mV at no signal to around 1.8V at 5A. These voltages are easy to divine from the device curves, of course. But there is also a contribution from the drivers, which might run to 400mA collector current with a four pair output stage. This might only be from 600mV to 800mV, but it does contribute to Vbe compression nonetheless.
What about using a CFP driver stage, wherein the pre-driver controls the Vbe which stays essentially unchanged due to the very small current changes in this first device? (Some attempt should be made to keep this CFP stage on at all times.) This would reduce driver stage distortion to virtually zero...... and might deliver very high, EC style input impedance to the stage as a whole.
My concern is the stability of EC. It appears to me that it usually requires revision of the lag comp to the VAS, and this is a PITA.
Cheers,
Hugh
Thank you for your answer, and your attempts to explain to me the worthiness of EC, much appreciated.
This conflict with Robert is absurd. You are both very decent technocrats. Please, put it behind you and move on!
May I ask another question? Ordinarily the Vbe distortion of the double EF output stage is largely produced by the output device itself, whose Vbe runs from 600mV at no signal to around 1.8V at 5A. These voltages are easy to divine from the device curves, of course. But there is also a contribution from the drivers, which might run to 400mA collector current with a four pair output stage. This might only be from 600mV to 800mV, but it does contribute to Vbe compression nonetheless.
What about using a CFP driver stage, wherein the pre-driver controls the Vbe which stays essentially unchanged due to the very small current changes in this first device? (Some attempt should be made to keep this CFP stage on at all times.) This would reduce driver stage distortion to virtually zero...... and might deliver very high, EC style input impedance to the stage as a whole.
My concern is the stability of EC. It appears to me that it usually requires revision of the lag comp to the VAS, and this is a PITA.
Cheers,
Hugh
AKSA said:
What about using a CFP driver stage, wherein the pre-driver controls the Vbe which stays essentially unchanged due to the very small current changes in this first device? (Some attempt should be made to keep this CFP stage on at all times.) This would reduce driver stage distortion to virtually zero...... and might deliver very high, EC style input impedance to the stage as a whole.
YES 😉 !
More than that, there is no need for EC over an well designed BJT output stage (from soundfingerprint POV).
Cheers,
Mihai
Re: soap
WRT buffered VAS, I have been looking into this a little further.
I did say I agreed "to a point" and it appears, as always there are
exceptions.
I have done quite a few different OP topology sims now (dozens) for
my own design purposes. My findings at least, indicate that with
unbuffered (dual EF) there is a definite limit on distortion floor that
I can't get past.
If I want to get to much less than 0.001 and across whole audio
band, the additional buffer is needed.
It varies slightly from design to design and obviously how much OLG
there is, but overall this appears to be the case.
I was looking for around -120dB / 4 ohms / 20kHz into 200W or so
and found that it was not possible sans buffer.
I was also looking for the ability to drive down to 2ohm loads with
minimal THD impact, same deal, without the buffer too much of the
load is transferred back.
Also I need the option to use the OP stage bridged, in which case
the addition buffer is pretty much mandatory.
Further to this one thing that has not been discussed here is the
real world situation of non ideal semis compared to spice models.
Modulating capacitances, drooping beta with higher currents etc etc,
We can safely assume that real world measurements will come up
fairly short of these sims and this can only point in defence of more
OP load isolation.
Add to this the extremely reactive load nature of speakers, and
I think we start to see a different picture than the perfection
of sims.
So, I think that there is a real case for the buffer, but it depends
what your design criteria is. I am designing for real world loads of
med/large studio monitors and even PA driven at relatively high
powers.
But - as the saying goes - YMMV 🙂
cheers
Terry
Edmond Stuart said:
defensively? Nobody is interested in red herrings. That's why I refuse my cooperation on this soap.
WRT buffered VAS, I have been looking into this a little further.
I did say I agreed "to a point" and it appears, as always there are
exceptions.
I have done quite a few different OP topology sims now (dozens) for
my own design purposes. My findings at least, indicate that with
unbuffered (dual EF) there is a definite limit on distortion floor that
I can't get past.
If I want to get to much less than 0.001 and across whole audio
band, the additional buffer is needed.
It varies slightly from design to design and obviously how much OLG
there is, but overall this appears to be the case.
I was looking for around -120dB / 4 ohms / 20kHz into 200W or so
and found that it was not possible sans buffer.
I was also looking for the ability to drive down to 2ohm loads with
minimal THD impact, same deal, without the buffer too much of the
load is transferred back.
Also I need the option to use the OP stage bridged, in which case
the addition buffer is pretty much mandatory.
Further to this one thing that has not been discussed here is the
real world situation of non ideal semis compared to spice models.
Modulating capacitances, drooping beta with higher currents etc etc,
We can safely assume that real world measurements will come up
fairly short of these sims and this can only point in defence of more
OP load isolation.
Add to this the extremely reactive load nature of speakers, and
I think we start to see a different picture than the perfection
of sims.
So, I think that there is a real case for the buffer, but it depends
what your design criteria is. I am designing for real world loads of
med/large studio monitors and even PA driven at relatively high
powers.
But - as the saying goes - YMMV 🙂
cheers
Terry
abzug said:
Is this only an issue for the power amplifier output switching? I'd assume that for low power line-level switching any quality sealed latching relay with silver plated contacts would contribute no measurable distortion.
A switching system would be easy to computerize and connect to blind testing software such as abchr, which makes the process much simpler, with control and navigation of listening material quite easy.
No doubt, but surely it's worth it? For line-level connections at least, shielded and, when supported by the equipment, balanced interconnects can be used, with shielded switch/relay-boxes.
Anyway, sorry to be off-topic...
Switching the power amp outputs is the most difficult, to be sure, but it all has to be done carefully with high-quality components. Among other things, I took special pains to switch both the hot and return for both line level and speaker level so as to not inadvertantly introduce ground loops and the like.
Cheers,
Bob
Re: Re: soap
Thanks, Terry. These are good points. I'm not sure that Edmond and I actually disagreed on the merit of Triples or EC. I think we both strongly endorse them. I think we just disagreed on how to further explore certain findings.
I guess that my POV is that for an amplifier without EC, I would likely never design it without a Triple; the extra transistor in the Triple is a pretty small guy that need not dissipate a lot. I think it is a bargain, and certainly mitigates the effects of beta droop in the power transistors. It may not do much for Vbe-related compression.
For an amplifier with EC, it is a bit hard for me to imagine an EC output stage architecture that does not in fact end up providing largely as much current buffering from the VAS as a Triple.
Cheers,
Bob
Terry Demol said:
WRT buffered VAS, I have been looking into this a little further.
I did say I agreed "to a point" and it appears, as always there are
exceptions.
I have done quite a few different OP topology sims now (dozens) for
my own design purposes. My findings at least, indicate that with
unbuffered (dual EF) there is a definite limit on distortion floor that
I can't get past.
If I want to get to much less than 0.001 and across whole audio
band, the additional buffer is needed.
It varies slightly from design to design and obviously how much OLG
there is, but overall this appears to be the case.
I was looking for around -120dB / 4 ohms / 20kHz into 200W or so
and found that it was not possible sans buffer.
I was also looking for the ability to drive down to 2ohm loads with
minimal THD impact, same deal, without the buffer too much of the
load is transferred back.
Also I need the option to use the OP stage bridged, in which case
the addition buffer is pretty much mandatory.
Further to this one thing that has not been discussed here is the
real world situation of non ideal semis compared to spice models.
Modulating capacitances, drooping beta with higher currents etc etc,
We can safely assume that real world measurements will come up
fairly short of these sims and this can only point in defence of more
OP load isolation.
Add to this the extremely reactive load nature of speakers, and
I think we start to see a different picture than the perfection
of sims.
So, I think that there is a real case for the buffer, but it depends
what your design criteria is. I am designing for real world loads of
med/large studio monitors and even PA driven at relatively high
powers.
But - as the saying goes - YMMV 🙂
cheers
Terry
Thanks, Terry. These are good points. I'm not sure that Edmond and I actually disagreed on the merit of Triples or EC. I think we both strongly endorse them. I think we just disagreed on how to further explore certain findings.
I guess that my POV is that for an amplifier without EC, I would likely never design it without a Triple; the extra transistor in the Triple is a pretty small guy that need not dissipate a lot. I think it is a bargain, and certainly mitigates the effects of beta droop in the power transistors. It may not do much for Vbe-related compression.
For an amplifier with EC, it is a bit hard for me to imagine an EC output stage architecture that does not in fact end up providing largely as much current buffering from the VAS as a Triple.
Cheers,
Bob
My findings at least, indicate that with unbuffered (dual EF) there is a definite limit on distortion floor that I can't get past.
That is absolutely correct, with dual EF distortion is above -70…80 dB with conventional direct path. With triple EF it would be on -100dB level. I’m talking about OFB design.
- Home
- Amplifiers
- Solid State
- Bob Cordell Interview: Error Correction