Hi,
Not the regulators open loop bandwidth, but arguably the the Supply should cover it properly.
Hardly. The classic Zenner/Capacitor follower with a Fet or BJT is a prime example for using a very big cap, as normally the output impedance is quite high... Moreover, as this circuit is essentially unconditionally stable (nearly) at least WRT capacitive load on the output there is penalty for doing so. My only concern would be a current limiter for the pass transistor if the cap is really big...
In fact, for fun...
An old preamp (Tubes) of mine used a Fet Zenner follower followed by 10 Ohm per channel and then a 4,700uF/350V "computer grade Cap" and before the zenner follower there was CLC filter with 20H chokes and 470uF/450V Capacitors.
Of course, that was for a completely zero feedback Preamp that was Phono & Line and MC input Phono at that, three 6DJ8 tubes per channel, 72dB Gain MC to line out...
Does something similar make sense with Op-Amp's? Sure, why not?
Ciao T
Not the regulators open loop bandwidth, but arguably the the Supply should cover it properly.
Hardly. The classic Zenner/Capacitor follower with a Fet or BJT is a prime example for using a very big cap, as normally the output impedance is quite high... Moreover, as this circuit is essentially unconditionally stable (nearly) at least WRT capacitive load on the output there is penalty for doing so. My only concern would be a current limiter for the pass transistor if the cap is really big...
In fact, for fun...
An old preamp (Tubes) of mine used a Fet Zenner follower followed by 10 Ohm per channel and then a 4,700uF/350V "computer grade Cap" and before the zenner follower there was CLC filter with 20H chokes and 470uF/450V Capacitors.
Of course, that was for a completely zero feedback Preamp that was Phono & Line and MC input Phono at that, three 6DJ8 tubes per channel, 72dB Gain MC to line out...
Does something similar make sense with Op-Amp's? Sure, why not?
Ciao T
To try and restore the thread and abort the hijacking, I have returned to the OP and the original schematic, if we can dignify it as such!
Putting it into LTspice, there is basically nothing to do 🙁 I could post the plots, but they are just a bunch of colo(u)red straight lines.
It is boring and predictable - it doesn't DO anything! Except lower the op-amp Vcc by some 0,7V. As the op-amp (supply) current is loaded up to 100mA, the voltage falls progressively to 13V, depending on transistor gain and other assumptions.
Given these advanced op-amps excellent PSRR performance, it is hard to accept that this will have any audible or even DRAMATIC effect on the sound.
All this has already been stated by in previous posts by several people.
To summarise:
1) These are NOT emitter followers. That is not an opinion, but simple fact.
2) In all existing literature, I don't think anyone has seriously claimed that adding impedance in the OA supply lines can be of any benefit. (except in the case of local RC decoupling, where the C is obligatory.)
I am sure the OP believes he heard a significant improvement, but I doubt he saw the guerilla, either.
Putting it into LTspice, there is basically nothing to do 🙁 I could post the plots, but they are just a bunch of colo(u)red straight lines.
It is boring and predictable - it doesn't DO anything! Except lower the op-amp Vcc by some 0,7V. As the op-amp (supply) current is loaded up to 100mA, the voltage falls progressively to 13V, depending on transistor gain and other assumptions.
Given these advanced op-amps excellent PSRR performance, it is hard to accept that this will have any audible or even DRAMATIC effect on the sound.
All this has already been stated by in previous posts by several people.
To summarise:
1) These are NOT emitter followers. That is not an opinion, but simple fact.
2) In all existing literature, I don't think anyone has seriously claimed that adding impedance in the OA supply lines can be of any benefit. (except in the case of local RC decoupling, where the C is obligatory.)
I am sure the OP believes he heard a significant improvement, but I doubt he saw the guerilla, either.
Funny, MB's sim. found a bit of low-level spikiness - 0.013% THD at 1Mhz anyway. 😕
That might have to do with the opamp model used - I also picked 1MHz to expose irregular behaviour more clearly. I'd have to agree with cliff that in the audio band, there's virtually zero difference or to use his term, the band is rather boring.
Discussed in a earlier post as the inherant unavoidable 🙄 THD of a AD826 at 0.01258 exceedingly close to 0.013
Cheers / Chris
Posts removed. Keep this to technical discussion or posting privileges will be suspended or revoked- this is directed at several people, if your post was pulled, you're one of them.You know what ?
I saw Georges question, ( first post ) and checked out what hew was talking about. People who bought the item seemed to like it, thought it was a decent upgrade on what they had and were pretty well pleased with the investment.
There were a few replies and Chris stepped in as the seller and gave an explanation.
In the opinions of the knowledgeable members it was pretty much blown out of the water immediately.
The more he tried to explain the more he got into deeper and deeper water.
Then there were accusations of mere ' money making ' and then several other insults thrown at him about his general knowledge of electronics.
Fair enough - but he never once said he was an Engineer and.......he was not rude to any of you if you look back at the posts. Astonishing really.
We are all at different levels here and we NEED you clever guys sometimes to lift our understanding of basic and sometimes more technical aspects of this hobby
I know I do.
It's when it starts to get rude that I object - and I've seen loads of this on this thread..... and I find it shocking.
There's a clever Scot on this forum called AndrewT whose not afraid to point out your errors, whose patience can run short sometimes too but he always ...ALWAYS tries to tell you how to research further to help you understand.
He's no fool but he'll give you a few chances.
That's the way to do it.
This thread has been ' put down after put down ' and amazingly no one has tried to build one - even if it's just to say ' look I tried it and it's awful - and here's the reason why '
It's been very personal in my opinion.
Hey, he's just a guy with an idea - nothing wrong with that.....is there ?
I guess I've made it personal, I've been sarcastic and a bit rude toward the end but I'm a bit fed up of watching this fella get kicked.
Apologies to all offended - I just think there is a better way of doing this
I saw Georges question, ( first post ) and checked out what hew was talking about. People who bought the item seemed to like it, thought it was a decent upgrade on what they had and were pretty well pleased with the investment.
There were a few replies and Chris stepped in as the seller and gave an explanation.
In the opinions of the knowledgeable members it was pretty much blown out of the water immediately.
The more he tried to explain the more he got into deeper and deeper water.
Then there were accusations of mere ' money making ' and then several other insults thrown at him about his general knowledge of electronics.
Fair enough - but he never once said he was an Engineer and.......he was not rude to any of you if you look back at the posts. Astonishing really.
We are all at different levels here and we NEED you clever guys sometimes to lift our understanding of basic and sometimes more technical aspects of this hobby
I know I do.
It's when it starts to get rude that I object - and I've seen loads of this on this thread..... and I find it shocking.
There's a clever Scot on this forum called AndrewT whose not afraid to point out your errors, whose patience can run short sometimes too but he always ...ALWAYS tries to tell you how to research further to help you understand.
He's no fool but he'll give you a few chances.
That's the way to do it.
This thread has been ' put down after put down ' and amazingly no one has tried to build one - even if it's just to say ' look I tried it and it's awful - and here's the reason why '
It's been very personal in my opinion.
Hey, he's just a guy with an idea - nothing wrong with that.....is there ?
I guess I've made it personal, I've been sarcastic and a bit rude toward the end but I'm a bit fed up of watching this fella get kicked.
Apologies to all offended - I just think there is a better way of doing this
"In what way do you think that is "real" or relevant?"
Referring to #56, p6, there is a marked appearance of small distortion peaks coinciding with signal threshold crossings in the opamp and apparent supply variation with the modified supply, compared to the unmodified one.
It's understood that the higher simulated frequency emphasises any distortion effect, owing to much lower feedback levels than audio at say, 1kHz
I can't comment on the subjective likelihood of what the OP or any listener might hear, if anything, with perhaps 0.0013% THD!0? We have slender enough evidence with no regard to likely sharp phase shift, harmonic structure or any of those subliminal qualities sought and quantified by audiophiles here, that could well have a more humanly detectable effect than the distortion level ordinarily suggests.
A difference, is a difference. Perhaps, if the OP would care to comment on why and how the base resistors were selected, we might gain a better insight into the likelihood of "realness" or the relevance of my query, which as we see nothing of the sim, makes it very hard to respond to. 😀
Unfortunately, most other comments here have simply been criticism of the DC supplies as not being good text-book ones. This seemed to me just as silly as anything else in the thread since it is obviously sabotaged for an effect, whether the OP or posts recognise it or not.
There is weakness in both positions, as I see it....Have at it!
Referring to #56, p6, there is a marked appearance of small distortion peaks coinciding with signal threshold crossings in the opamp and apparent supply variation with the modified supply, compared to the unmodified one.
It's understood that the higher simulated frequency emphasises any distortion effect, owing to much lower feedback levels than audio at say, 1kHz
I can't comment on the subjective likelihood of what the OP or any listener might hear, if anything, with perhaps 0.0013% THD!0? We have slender enough evidence with no regard to likely sharp phase shift, harmonic structure or any of those subliminal qualities sought and quantified by audiophiles here, that could well have a more humanly detectable effect than the distortion level ordinarily suggests.
A difference, is a difference. Perhaps, if the OP would care to comment on why and how the base resistors were selected, we might gain a better insight into the likelihood of "realness" or the relevance of my query, which as we see nothing of the sim, makes it very hard to respond to. 😀
Unfortunately, most other comments here have simply been criticism of the DC supplies as not being good text-book ones. This seemed to me just as silly as anything else in the thread since it is obviously sabotaged for an effect, whether the OP or posts recognise it or not.
There is weakness in both positions, as I see it....Have at it!
Last edited:
Hi
I have modified AD825, opa627 and opa1611, all with good result. The same characteristic of better stereo spatial delivery, consistently apparent in each.🙂
I look forward to trying on other op amps, and will report findings.
Cheers / Chris ,
Chris,
Objectively, what you are doing has two effects I can think of:
1) You lower the powersupply voltages - for example the AD825 when it was in fashion tended to sound a bit forward and harsh on 15V rails but much better on 12V rails...
2) You create a powersupply that has a small extra impedance in line with the actual PSU and this impedance is non-linear, the more current you draw the more it falls (as the transistors Gm rises)...
To tease out which is which you could try running the same Op-Amp one with your circuit and one without but lowered supply rails to match the first. The comparison should allow you to determine if the different sonic is a result of the change in supply voltage or is due to the transistor/resistor circuit.
A second comparison would be to replace the circuit with resistor that produce the same voltage drop quiescent as your circuit. This would illustrate if the non-linear nature of the circuit has any effect.
The final test I would do to get behind what is happening is to replace the circuit with diodes to get a similar drop in rail voltage, this one just for fun, select diodes with a lot curvature to their V/I curve in the current region where it will operate.
I think the results of these would show you at least some inkling why you happen to hear what you hear. Of course this only needs doing if you want to understand more about the circuit and what it does and why.
If instead you are just happy to have some results that you like, simply keep doing that.
Ciao T
Objectively, what you are doing has two effects I can think of:
1) You lower the powersupply voltages - for example the AD825 when it was in fashion tended to sound a bit forward and harsh on 15V rails but much better on 12V rails...
2) You create a powersupply that has a small extra impedance in line with the actual PSU and this impedance is non-linear, the more current you draw the more it falls (as the transistors Gm rises)...
To tease out which is which you could try running the same Op-Amp one with your circuit and one without but lowered supply rails to match the first. The comparison should allow you to determine if the different sonic is a result of the change in supply voltage or is due to the transistor/resistor circuit.
A second comparison would be to replace the circuit with resistor that produce the same voltage drop quiescent as your circuit. This would illustrate if the non-linear nature of the circuit has any effect.
The final test I would do to get behind what is happening is to replace the circuit with diodes to get a similar drop in rail voltage, this one just for fun, select diodes with a lot curvature to their V/I curve in the current region where it will operate.
I think the results of these would show you at least some inkling why you happen to hear what you hear. Of course this only needs doing if you want to understand more about the circuit and what it does and why.
If instead you are just happy to have some results that you like, simply keep doing that.
Ciao T
The above suggestions have already been made and at least the basic simulation of the OP circuit to find the difference has been posted.
The question "Did you mod any opamps this way" was directed at AndrewGM who was arguing on subjective values which was not yet qualified. Hence the fair question. I don't think there is any doubt that you have listened to your devices, Chris.
What I was interested to know was what your criteria for selecting the base resistors was. Can you measure a DC voltage difference when you experiment with values and if so, where? - or is any of this actually not too critical?
The question "Did you mod any opamps this way" was directed at AndrewGM who was arguing on subjective values which was not yet qualified. Hence the fair question. I don't think there is any doubt that you have listened to your devices, Chris.
What I was interested to know was what your criteria for selecting the base resistors was. Can you measure a DC voltage difference when you experiment with values and if so, where? - or is any of this actually not too critical?
Hi,
I doubt even more advanced simulations would show much. Most Op-Amp Models are criminally optimistic about their capabilities (actually, so are the other models).
You would probably have to use something substantially better than an AP2 to be able to measure anything and we would be looking at what exactly to measure.
But such things only affect obsessive 'scope jokeys, hence my suggestions of listening tests to tease out the factor that actually makes the difference...
Ciao T
I doubt even more advanced simulations would show much. Most Op-Amp Models are criminally optimistic about their capabilities (actually, so are the other models).
You would probably have to use something substantially better than an AP2 to be able to measure anything and we would be looking at what exactly to measure.
But such things only affect obsessive 'scope jokeys, hence my suggestions of listening tests to tease out the factor that actually makes the difference...
Ciao T
I do not mind that the circuit produces a difference in sound.
I do not mind that the circuit produces a difference in measurements.
I do not mind the inventor selling his product.
I do mind selling a product based on a misunderstanding of the science and turning that misunderstanding into advertising lies. This is further compounded by the inventor's refusal to modify his unsubstantiated claims of the science. Continuing to SELL the product and LYING about the science behind it crosses the line of criminality.
We on DiyAudio should never condone criminality.
We on DiyAudio should endeavour to correct errors we see/find in claims made to conn customers.
The only way we meet that responsibility is to post on the DiyAudio Forum what we see as the truth.
I do not mind that the circuit produces a difference in measurements.
I do not mind the inventor selling his product.
I do mind selling a product based on a misunderstanding of the science and turning that misunderstanding into advertising lies. This is further compounded by the inventor's refusal to modify his unsubstantiated claims of the science. Continuing to SELL the product and LYING about the science behind it crosses the line of criminality.
We on DiyAudio should never condone criminality.
We on DiyAudio should endeavour to correct errors we see/find in claims made to conn customers.
The only way we meet that responsibility is to post on the DiyAudio Forum what we see as the truth.
Thanks Ian and ThorstenL, AndrewT and AndrewGM
A transistor connected this way is for all intents is or has similarity to a transistor connected as a diode as far as I can tell, but from what I am seeing on my scope at its settings of 2mv ac and 0.1us ( AndrewT I get nothing meaningful to work with with scope set to DC ) there is benefit in loading the base with resistance, vs a short.
My scope shows what I assess as being a residual ( giving me a visual indication of something to work with ) at the above settings that improves its wave form as resistance is introduced from collector to base, the optimum being close to what I use with the op amp circuit. But there is a second result ... as the load is increased the wave form residual improves again.
Tonight I connected up a BC547 in the same format with 8.90v DC+ and with 1402 ohms of resistance, between Collector and base C to E measures 0.93v B to E measures 0.74 and I set 235 ohms of resistance providing load to Gnd from the emitter. I decreased resistance to 159 ohms to assess greater load.
I also tried a 2N3055 with similar result - making load resistance 39 ohms from a 10v supply.
Cheers / Chris
Last edited:
- Status
- Not open for further replies.