John Curl's Blowtorch preamplifier

[john curl]

Did everyone come to understand that 'circuit hook' is a property of dielectric absorption? You know, change in capacitance with frequency. They CAN'T compensate for this with a regular cap. This caused problems in their attenuators and probes. Think about it. Then think about potential audio problems as well. You never know!

[PMA]

Originally posted by john curl
Rogers is what? Specifically.

http://www.rogerscorporation.com/acm/litintbl.htm

[john curl]

Which Rogers material, specifically?

[john curl]

For the 'silent majority': I might not have given enough info as to how and why this is interesting and important.
With RF frequencies, there is a problem with the parasitic capacitance of the resistors and the circuit boards they are wired on.
For example the input of a typical osicilloscope is something like 1meg ohm and 20 pf.
Usually, for high frequency work, we put a passive 10:1 reduction probe with 9meg ohm and 2 pf across it. Do you see the correlation? The caps cancel each others effects at high frequencies and the bandwidth can be many hundred mega-Hz.
It is the same with the internal attenuators that change the range of the input of the scope.
Now, when the parasitic capacitance (in this case, 20pf) changes its value with frequency, because of DA effects, then it is virtually impossible to compensate it with a capacitor of higher quality, because it has the same capactance with frequency, and it will NOT track!
Now think of a circuit board made of cheap materials, with 10's to 100's of resistors, each adding a changing capacitance with frequency. Then, even at audible frequencies, who knows if this could be harmful to the sound.
Remember, 'circuit hook' was visually observable on a scope with one or two resistors. We never see audible distortion in amps until it is too late. We typically look at and measure effects 100-10,000 times lower than this. Now think again at the number of connections to a circuit board, as well as its traces.

[PMA]

It is almost impossible to completely compensate the R//C - R//C divider. It is much easier to do it for the R-C//C - R-C//C structure. I have known this for at least 25 years

Regarding that Rogers PCB materials - another thing that helps ...... ....... not pure PTFE, so it is easier to drill and mill .........

[john curl]

PMA, I am trying to demonstrate something, NOT design something. Second, Rogers makes MANY circuit board materials. I know what you are saying, but you have not been clear about it to others, until now.
I suspect that your college instructors once found you hard to please.

[PMA]

John,

I have learnt from you:

1) to give the hints only
2) never to give away my secrets

Thanks for these lectures

[john curl]

PMA, I am not allowed to be rude to you, so please try not to be rude to me.
Personally, I make a living as a CONSULTANT. That means that I am normally paid for conveying my knowledge of audio electronics. IF I were inclined to GIVE everything away that I normally convey to my clients, then my clients would just 'tune in' and not pay me. I try, with a minimum of words, to convey some of the important aspects of audio design that might not be well known to everyone. I try to back up 'controversial' things, like cleaning leads, etc with my own experience.
I might add that I did NOT start this thread. Others did, and I came in only after much misunderstanding of one of my designs was put here by others. Now, I am trying to add some small, but important, aspects in making (in my opinion) successful audio products. Everyone has the perfect right to use the cheapest solder, the cheapest circuit board, the cheapest caps, resistors, and semiconductors and TRY to make a really successful audio product, due to your brilliant design topology and artful SPICE simulation. Go for it! I have, over the decades, and each and every time, I am pulled back to addressing issues would normally be discounted as unnecessary as a design engineer. When I do pay attention to some details, then extra successful audio products come forth, such as the 'Blowtorch' or the JC-1 power amp.
I can assure you that topology sophistication and simulation alone will not separate the great products from the just-OK products.

[MikeBettinger]

Originally posted by john curl
For the 'silent majority': I might not have given enough info as to how and why this is interesting and important.
With RF frequencies, there is a problem with the parasitic capacitance of the resistors and the circuit boards they are wired on.
Now think again at the number of connections to a circuit board, as well as its traces.

John,

I'd like to respectfully point out that all circuit nets are not as affected as others by this and similar effects. This is one of the reasons I keep mentioning that understanding and controlling the layout is a prominent driver of any designs performance.

Regards, Mike.

[john curl]

Mike, I don't disagree, but all else being equal, should not the best DA be better? We often don't have complete control of our parts placement.