Edmond Stuart said:
Forgive me for adding to John's answer but in this instance I find myself in perfect agreement as this is categorically NOT a GE issue but a physics of semiconductors issue.
Class A has no - that is zero - cross over distortion because it does not ask the transistors to operate in the non-linear region of their Ic curves. So by perfect it means an idealised B or AB stage that still needs to work through the non-linear ("switching") region of the transistors curves. Class A is inherently "perfect" in that sense.
The laws relate to the semiconductor nature of silicon and the requirement for a specific level of energy accross the junction before conduction occurs. Simplistically this relates to the Vbe needed to cause conduction Ic.
Alan, John & Pavel,
Let's stop this discussion. It's all semantics and I was just teasing John a bit. He said PERFECT crossover (without even specifying what kind of output devices). For me, perfect means perfect, that is, without creating distortion. So, his statement that such non-existing OPS is not as good as class-heat is rather pointless. That's all I want to make clear.
Cheers,
Edmond.
Let's stop this discussion. It's all semantics and I was just teasing John a bit. He said PERFECT crossover (without even specifying what kind of output devices). For me, perfect means perfect, that is, without creating distortion. So, his statement that such non-existing OPS is not as good as class-heat is rather pointless. That's all I want to make clear.
Cheers,
Edmond.
janneman said:
Good point Jan. I saw the earlier plot and was wondering how it was generated. John knew because it apparently reproduced one of the setups he has used many times, but others may not be privy to that info.
The reason I mentioned NP0 earlier in this thread was as a proposed replacement for mica in applications such as frequency compensation and input LPFs for power amps. For values in that range (say 10pF or smaller to 1000pF or so), they are cheaper than mica, have lower DA and equivalent high frequency performance. What's not to like?
But for coupling applications they don't seem to make much sense.
.
Finally, one voice of reason.
Thank you, Andy_c for a reasonable input. Yes, I sometimes am more careful with my inputs than many seem to catch. For example, I didn't mention mica, and that is because I have never directly evaluated mica with this test, as I could not get a very large value. I also never previously evaluated NPO ceramic with this test, either, because 25 years ago, when I did the testing, I could not find large enough values to make the test accurate enough to publish.
Without a direct comparison, yet, I would equate mica and ceramic basically. However, I still prefer to use mica for small values, because I can easily recognize it, and nobody can get a non NPO ceramic circuit past me, because it looked similar to an NPO.
I can just look at the RESIDUAL that Demian got in HIS test and know that he has DA. Nonlinear distortion can be evaluated separately with harmonic or IM measurements.
IF you critics were up to speed, you would have at least noted both my paper in 1968 and Walt Jung's and my paper on DA, as it is found on Walt Jung's website for free. Then, you would know about 10% as much as I know about the test, as I have vastly more experience with it, as I ran 100's of tests with just about every cap I could find.
We can address class A and B separately. Please take care in reacting to what I stated about it. Don't presume, or guess. Think it through. That is the point of me saying anything at all about it. I personally don't care if you design A, B, or D.
Thank you, Andy_c for a reasonable input. Yes, I sometimes am more careful with my inputs than many seem to catch. For example, I didn't mention mica, and that is because I have never directly evaluated mica with this test, as I could not get a very large value. I also never previously evaluated NPO ceramic with this test, either, because 25 years ago, when I did the testing, I could not find large enough values to make the test accurate enough to publish.
Without a direct comparison, yet, I would equate mica and ceramic basically. However, I still prefer to use mica for small values, because I can easily recognize it, and nobody can get a non NPO ceramic circuit past me, because it looked similar to an NPO.
I can just look at the RESIDUAL that Demian got in HIS test and know that he has DA. Nonlinear distortion can be evaluated separately with harmonic or IM measurements.
IF you critics were up to speed, you would have at least noted both my paper in 1968 and Walt Jung's and my paper on DA, as it is found on Walt Jung's website for free. Then, you would know about 10% as much as I know about the test, as I have vastly more experience with it, as I ran 100's of tests with just about every cap I could find.
We can address class A and B separately. Please take care in reacting to what I stated about it. Don't presume, or guess. Think it through. That is the point of me saying anything at all about it. I personally don't care if you design A, B, or D.
john curl said:
Hi John,
What measurement are you referring to here? Is it DA?
Mylars by nature have only 2-3 times the DA of polypropylene, but I don't recall how much better than polypropylene polystyrene is. In any case, I don't think there is any where near a 100:1 spread in DA from ploystyrene to Mylar.
Cheers,
Bob
janneman said:
Okay, I found the article that John was talking about on Walt's site here (PDF download). It describes the capacitance test bridge technique.
PMA said:
PMA,
I never said or suggested it was. I was just dropping in a caveat about class A. Like all things in life, class A is not a universal answer. I bet you would be surprized how many well-intentioned amplifier designers didn't get the best out of class A.
This is no different than the naive folk who think they can make a purse out of a sow's ear by substituting $200 worth of boutique capacitors into an otherwise crappy design.
Cheers,
Bob
- Status
- Not open for further replies.