John Curl's Blowtorch preamplifier part II

Status
Not open for further replies.
Have yet to meet these folks. MIT has a higher percentage of odd-balls than most places.

The better is education, the deeper is understanding, the wider are choices of solutions, the more of odd-balls are considered as possible alternative solutions. And of course, better understanding which odd-balls can be preferred, and which odd-balls have no foundation at all.
 
AX tech editor
Joined 2002
Paid Member
> so there will be DVDs for all major languages (I'm working on Chinese ;-)

Just let the XEN team in HK know if you need help. :)

Like it or not, Chinese will become the world language 20 years from now.
Just look at how many university students in continental Europe are learning Chinese.
And really fluent speakers.


Patrick

Yes, the language will go where the money goes. Give it a decade or two, max.

jan
 
Let me explain about my position with Marty:
I was introduced to him in the late 90's through my then business partner Bob Crump, who was close friend with Marty. Bob made cables at that time, and we were starting on the CTC Blowtorch. Marty liked Bob's cables a lot. He later liked the Blowtorch as well.
However, something worried me. Getting too close to reviewers may effect their judgement. Oh, I know they will deny it, but our reviews were just to good for me to be comfortable.
In any case, Bob drove up from Texas to Chicago, picking up Marty on the way. They attended an audio 'get-together' put together by another colleague.
Apparently, returning home in the car, Marty 'badmouthed' the set-up in Chicago, and impugned our colleague doing that. Bob, apparently in reaction, said that he didn't think that Marty's set-up sounded that good. Now, I have not heard either, but I DO know, from experience, that just going to another's place and listening to THEIR set-up, might not sound that good, even if it sounds extra fine to the owner.
However, after that, I had all kinds of trouble with Marty, who apparently linked me with Bob. So when I tried to converse with Marty, about the 'break-in' time of the Parasound JC-1, for example, I got all kinds of 'static'. It is rather sad, because I think his publication is 'a breath of fresh air' for the most part. I haven't spoken to Marty in about a decade, why don't you say hello to him for me?
There is a 'lesson' in this, I hope that you note it.
 
Ed will this do? I doubt it, but here goes anyway. A typical dominant pole compensated amplifier (90 deg. phase in the ideal case) this one does not accumulate 144 degrees till over 2MHz. A true delay of 20us would be a straight line (on a linear frequency scale) going through 144 deg. at 20k. My experience in this thread is that those weded to these or other mis-conceptions for years will never let go of them.

BTW - LM3875 just to keep the audio theme.

Here is another gain vs phase shift, a universal R-C coupled amplifier response curve. The top is gain and the bottom is phase shift.

Hope this helps.
 

Attachments

  • gain_phase.jpg
    gain_phase.jpg
    402.8 KB · Views: 172
Here is another gain vs phase shift, a universal R-C coupled amplifier response curve. The top is gain and the bottom is phase shift.

Hope this helps.

Pay attention, Scott! You have the chance to get educated! Don't miss the chance! R-C coupled curve must go backward! :D

PS: I knew that Positron will come back, skip the discussion with explanations, then start over and over. It is not phase shift. It is indeed time delay in 2 pages of explanations that leads to positive feedback and stable oscillations. :scratch:

Let's check if damping of Q helps: I am shutting up.

Edit: pardon, it is not damping of Q. It is cutting of one feedback path.
 
Last edited:
=DF96;3124933]OK, I have thought about that. On that point you are right. I therefore withdraw my comment. A perfect cap would have no DA.

No sweat. :)

However, you may be confusing the concepts of 'infinite', 'ideal' and 'perfect'. Your circuit merely required a zero impedance at the top of the anode load resistor; you can get this from an infinite value cap. DA would merely add a temporary slow change in supply voltage. Your argument appears to be that zero impedance requires a perfect cap, a perfect cap would have no DA, therefore a cap with DA will damage the sound.

But in order to have zero impedance, by definition, the ESR has to be zero. Otherwise the infinite capacitor does not have zero impedance. So ESR, DA, and DF have to be zero.

Another problem is that slow means the electrons are still "flaking off" the insulator at some future point of time, future point of the music. That is voltage and is it only subsonic? See next.

As DA is a slow linear phenomenon with time constants well down in the subsonic region it does not damage sound.

Just curious where did you obtain that idea? And a time constant is what, 63%? Back to the real world on this question. Substitute an electrolytic capacitor for a good poly and hear what happens to the bass, yet the frequency response remains the same. One can over bass, say a preamplifier, with the size and type of capacitors used.

An infinite but non-linear cap would provide your zero impedance. By your logic therefore capacitor non-linearity is not a problem.

Don't think so since non-linearity indicates a property that is not ideal or perfect, such as DA, ESR values. The infinite capacitor might approach zero impedance, but not attain it. I would suggest checking out the problems/graphs with conventional capacitors, "Picking capacitors".

You always need to be careful when using arguments employing infinity, because infinity sometimes behaves like any other number and sometimes behaves like a number with very different properties.

Like the physical properties of a capacitor would be infinite size, zero resistance of the foil so infinite thickness etc.

Theory was what we were taught in College, 60s. Has that changed?

Cheers.
 
If this is true an short pulse applied to the input of a circuit, while attenuated at the output should be visible as an "echo" repeating at whatever that loop period time is. The suggestion is that high feedback levels will do this to a greater degree. I have never been able to see such a phenomenon. Perhaps at microwave frequencies its possible. I have never seen an audio amplifier with usable gain at 1 GHz+.

Phase shift occurs everytime "A", "A1", "A2" passes go, mixing with the new input. We can be talking -20db down, -30db, -50db. Thus inner detail would be affected.

Cheers.
 
Last edited:
I have been in the company of quite a few who have high level degrees, published papers, can do high level math in their sleep...... But I have only been in the company of a very few who can come up with a great product AND design, build and bring it to market. And since the Blowtorch is not just a "great" product but IMO a pinnacle product, that sets the designer in a truly top level league.
 
Last edited:
Let me explain about my position with Marty:
I was introduced to him in the late 90's through my then business partner Bob Crump, who was close friend with Marty. Bob made cables at that time, and we were starting on the CTC Blowtorch. Marty liked Bob's cables a lot. He later liked the Blowtorch as well.
However, something worried me. Getting too close to reviewers may effect their judgement. Oh, I know they will deny it, but our reviews were just to good for me to be comfortable.
In any case, Bob drove up from Texas to Chicago, picking up Marty on the way. They attended an audio 'get-together' put together by another colleague.
Apparently, returning home in the car, Marty 'badmouthed' the set-up in Chicago, and impugned our colleague doing that. Bob, apparently in reaction, said that he didn't think that Marty's set-up sounded that good. Now, I have not heard either, but I DO know, from experience, that just going to another's place and listening to THEIR set-up, might not sound that good, even if it sounds extra fine to the owner.
However, after that, I had all kinds of trouble with Marty, who apparently linked me with Bob. So when I tried to converse with Marty, about the 'break-in' time of the Parasound JC-1, for example, I got all kinds of 'static'. It is rather sad, because I think his publication is 'a breath of fresh air' for the most part. I haven't spoken to Marty in about a decade, why don't you say hello to him for me?
There is a 'lesson' in this, I hope that you note it.

Sorry to hear that. I will say hello to Marty for you John. I think Marty thinks well of the blowtorch, as he has mentioned it in articles.

Cheers.
 
Pay attention, Scott! You have the chance to get educated! Don't miss the chance! R-C coupled curve must go backward! :D

PS: I knew that Positron will come back, skip the discussion with explanations, then start over and over. It is not phase shift. It is indeed time delay in 2 pages of explanations that leads to positive feedback and stable oscillations.

Ah, first year college textbook, "Semiconductor and Tube Electronics, an Introduction". :p

Could you explain "stable oscillations" again Wave? :violin:

Cheers.
 
Last edited:
Ah, first year college textbook, "Semiconductor and Tube Electronics, an Introduction". :p

The graph on the page that you use to teach Scott shows integrating R-C network, while your remark is about R-C coupling that is differentiating network. What it has to do with "propagation time" nobody knows.

Could you explain "stable oscillations" again Wave? :violin:


I will try. When you read the answers and try to digest information presented in them it can be viewed as a phase shift, i.e. you start learning immediately, but the process takes time, so your understanding is getting better and better. The same happens with integrating R-C networks: when you apply input signal voltage on the capacitor starts following the signal immediately. It is the case of phase shift between input and output signals. You can call it "time delay", but it is incorrect: how do you go to measure the time? On which level of voltage on the capacitor you decide that it is enough to wait?
However, if you add some comparator after the R-C network you can call it "time delay" machine, because voltage on output will appear only after some certain time that depends on the level of input signal, time constant of the R-C network, and threshold level of your output device. And if you apply negative feedback to such device you will inevitably get stable oscillations.
Do you see the difference? One R-C device is not enough to cause oscillations because it shifts phase 90 degrees only on infinite frequency. The lower is the frequency, the less is phase shift. But if to add some threshold device after the R-C network you can no longer talk about phase shift, instead you are getting real time delay. And if amplification factor through negative feedback loop is slightly greater than 1 you will absolutely necessary get stable oscillations. Exactly what Demian said you from the beginning. And it does not oscillate because of presence of RC network and feedback! It oscillates because you skipped the information: the comparator reacts after voltage on the capacitor raises up to the certain level. From zero to this level the information was _skipped_.


The same, when you skip the information presented to you on the forum
turning the system with phase shift into the system with time delay. That also leads to stable oscillations. I hope you will understand it now, that will damp oscillations, but if you skip it it will go and go around in endless cycles...
 
Last edited:
Positron said:
Substitute an electrolytic capacitor for a good poly and hear what happens to the bass, yet the frequency response remains the same.
Almost certainly caused by non-linearity, not DA.

The infinite capacitor might approach zero impedance, but not attain it.
An infinite capacitor, even if non-linear, would still attain zero impedance.

I would suggest checking out the problems/graphs with conventional capacitors, "Picking capacitors".
I am well aware, and have said myself, that picking the right dielectric technology etc. is important. This means the right technology for the particular job. 'Audio' is not a job. 'Audio coupling cap setting LF rolloff' is a job. If I were writing the blurb for an amp, or an amp philosophy, I would use terms like distortion or IM - not 'smearing' or similar vagueness.

Positron said:
Phase shift occurs everytime "A", "A1", "A2" passes go, mixing with the new input. We can be talking -20db down, -30db, -50db. Thus inner detail would be affected.
Still sticking to your flawed model of feedback, I see. What is 'inner detail'? How does it compare with 'outer detail'?
 
Last edited:
Almost certainly caused by non-linearity, not DA.

Prove it.

An infinite capacitor, even if non-linear, would still attain zero impedance.

Wrong since a voltage presented by DA would prove the impedance is not zero. Zero impedance would mean zero voltage. This has been mentioned to you before.

I am well aware, and have said myself, that picking the right dielectric technology etc. is important. This means the right technology for the particular job. 'Audio' is not a job. 'Audio coupling cap setting LF rolloff' is a job. If I were writing the blurb for an amp, or an amp philosophy, I would use terms like distortion or IM - not 'smearing' or similar vagueness.

Read Walter Jung "Picking Capacitors". Next, "smearing" is a form and identifies the type of distortion caused by DA, and is not IM or HD.

Cheers.
 
Positron said:
Wrong since a voltage presented by DA would prove the impedance is not zero.
I said non-linear, not DA. I have accepted your point about DA. I was just pointing out that using infinite capacitors to make a point can lead people astray. An infinite non-linear capacitor would do no harm, but this does not mean that non-linearity is not harmful in finite capacitors.

Next, "smearing" is a form and identifies the type of distortion caused by DA, and is not IM or HD.
DA is linear so no distortion is generated. All you get back out of the cap is the long-term average of the voltage across the cap. For music this will be zero.
 
The graph on the page that you use to teach Scott shows integrating R-C network, while your remark is about R-C coupling that is differentiating network. What it has to do with "propagation time" nobody knows.

I will try. When you read the answers and try to digest information presented in them it can be viewed as a phase shift, i.e. you start learning immediately, but the process takes time, so your understanding is getting better and better. The same happens with integrating R-C networks: when you apply input signal voltage on the capacitor starts following the signal immediately. It is the case of phase shift between input and output signals. You can call it "time delay", but it is incorrect: how do you go to measure the time? On which level of voltage on the capacitor you decide that it is enough to wait?
However, if you add some comparator after the R-C network you can call it "time delay" machine, because voltage on output will appear only after some certain time that depends on the level of input signal, time constant of the R-C network, and threshold level of your output device. And if you apply negative feedback to such device you will inevitably get stable oscillations.
Do you see the difference? One R-C device is not enough to cause oscillations because it shifts phase 90 degrees only on infinite frequency. The lower is the frequency, the less is phase shift. But if to add some threshold device after the R-C network you can no longer talk about phase shift, instead you are getting real time delay. And if amplification factor through negative feedback loop is slightly greater than 1 you will absolutely necessary get stable oscillations. Exactly what Demian said you from the beginning. And it does not oscillate because of presence of RC network and feedback! It oscillates because you skipped the information: the comparator reacts after voltage on the capacitor raises up to the certain level. From zero to this level the information was _skipped_.


The same, when you skip the information presented to you on the forum
turning the system with phase shift into the system with time delay. That also leads to stable oscillations. I hope you will understand it now, that will damp oscillations, but if you skip it it will go and go around in endless cycles...

I think you missed a few posts of mine. ;)

Cheers.
 
Wrong since a voltage presented by DA would prove the impedance is not zero. Zero impedance would mean zero voltage. This has been mentioned to you before.

No, he is right. DF means presence in parallel with capacitor of rated capacitance of one more "phantom" capacitor with resistance in series. If capacitance is rated as infinite that means that it's impedance is zero on any frequency except zero Hertz, so no matter what you add in parallel to zero Ohm, it's impedance on any frequency will still be zero.

However, if you have own definitions of DF and infinity you are welcome to present them here and now.
 
No, he is right. DF means presence in parallel with capacitor of rated capacitance of one more "phantom" capacitor with resistance in series. If capacitance is rated as infinite that means that it's impedance is zero on any frequency except zero Hertz, so no matter what you add in parallel to zero Ohm, it's impedance on any frequency will still be zero.

However, if you have own definitions of DF and infinity you are welcome to present them here and now.

A quick link to correct you. You do understand that there are viewers who may actually believe you?

Equivalent series resistance - Wikipedia, the free encyclopedia

Cheers.
 
Last edited:
Status
Not open for further replies.