I believe the analysis was a scan of handwritten equations, don't remember if it was Edmond's bbf or who else put them up
I duplicated the wye-delta equations in MathCad, started on a FDNR in Ltspice
but neither were finished before I was distracted elsewhere - haven't felt motivated to revisit TMC
I duplicated the wye-delta equations in MathCad, started on a FDNR in Ltspice
but neither were finished before I was distracted elsewhere - haven't felt motivated to revisit TMC
"Megajocke" did a hand written analysis that he scanned in this forum and perhaps what you recollect. It was perhaps a bit round-about and not explicitly a wye-delta so I didn't find it until I was finished, but I was reassured that my transform had produced more or less the same results.
If you ever do find the motivation to revisit improved feedback ideas then please keep me informed. I could use help in my quest for "Bode optimal feedback"
Best wishes
David.
If you ever do find the motivation to revisit improved feedback ideas then please keep me informed. I could use help in my quest for "Bode optimal feedback"
Best wishes
David.
Hi Tony,
I agree with your affirmations now. After my last post, i continue to search information about ceramic capacitors and i found in another thread right here in this forum a test where NP0/C0G are compared with X7R and even with X5R. Sorry, i didn't remember of bookmark it.
There, issues like non linearities and captation of sound and vibration (like a microphone) are well showed. The X5R type it's the worst of them. The majority 100uF ceramic caps that i found, are of this type.
I was really impressed with the small size of an of 100uF capacitor, but, without a minimum quality for use in audio, they are discarded.
Thanks for your feedback , Tony.
Rubens.
I agree with your affirmations now. After my last post, i continue to search information about ceramic capacitors and i found in another thread right here in this forum a test where NP0/C0G are compared with X7R and even with X5R. Sorry, i didn't remember of bookmark it.
There, issues like non linearities and captation of sound and vibration (like a microphone) are well showed. The X5R type it's the worst of them. The majority 100uF ceramic caps that i found, are of this type.
I was really impressed with the small size of an of 100uF capacitor, but, without a minimum quality for use in audio, they are discarded.
Thanks for your feedback , Tony.
Rubens.
No problem rubens 🙂 It still might be interesting to try if the ceramics are cheap enough and you have the equipment to do distortion measurements for comparison.
Tony.
Tony.
Yes, same thing.
It's probably to late for me to think straight, but your statement "simulation of a real amplifier design is closer to reality than theory" doesn't make any sense. Last time I've checked, simulation is theory, so it can't go against theory.
Hi Waly,
I disagee with your above statement. First off, let's agree that simulation is certainly not perfect - for starters, it does not include parasitic capacitances and inductances, and does rely on transistor models that are not perfect.
So simulation is a simplification of the real world.
However, therory is a far graeter simplification, just in order to make the analysis tractable. This does not mean that theory has no place - it certainly does. One example is that it renders a lot more insight into the fundamental causes and effects of those aspects being covered by the theoretical model.
In getting a picture of reality, there are steps to be taken in getting closer to the real thing, and all of those steps should be taken and understood. The first step is theory, the second step is simulation (which gets us closer), and the third step is actually building the design and measuring it. That last step is reflective of the reality that you can achieve when you build it with a given set of parts and a given assembly approach and a given measurement approach.
Even the third step has shortcomings from "reality" in that there are implementation imperfections that can hide or blur the relationships in design that you wish to understand for the more general case.
A simple example is hum and noise pickup. Another example might be distortion due to the imperfections of a passive component that might be gone in another design where a better choice of component is made (for example, thermal-dependent nonlinearity in a resistor). Yet another example might be crossover distortion in the output stage of the real amplifier, where one is somewhat at the mercy of time-dependent bias conditions. This might be different in a different real design where different heat management is used. This doesn't mean that build and test is of great importance; it just means that even with real build and test we must be cautious about applying those findings to the general case.
Cheers,
Bob
The more interesting capacitor tests and evaluations are among the following:
COG
Mica
Polyester (Mylar)
Polypropylene
polystyrene
Capacitor tests among electrolytics and ceramics like X5R and X7R are relatively uninteresting because these are known crappy capacitors.
Cheers,
Bob
I guess we disagree by the notion of "theory". To me "theory" is the work required to conceptualize, understand and, perhaps, eventually generalize a certain circuit topology. In this respect, simulation barely helps, not more than (poor analogy, but you'll get the drift) a hammer helps understanding how nail joints are working.
Hand analysis is just another tool in the design process. Obviously simulation is quicker and much more precise, because both methods are ultimately relying on the same set of mathematical tools (linearizing and solving a set of linear equations) and computers shine at such jobs. Though, how would ever simulation reveal (to use the current example) the equivalence of TMC to TPC plus a lead-lag compensation? Simulation helps verifying such hypothesis, but will never discover them by itself.
You may argue that such "discoveries" are irrelevant, as long as the results are ok, anyway. If so, this boils down to a fundamental disagreement in the design methodology. I personally hate taking a certain topology and then start patching feedback loops and compensation caps here and there, running simulations, and hoping the loop gain/phase will behave. Ending up with 3291 embedded positive/negative feedback loops and claiming 0.000001% distortions, before even considering that global stability is more than each loop's stability, and that there's Sj(C(n,j)) ways a circuit can get unstable or marginally stable. This is essentially what Edmand is doing (no pun intended, his work is respectable), and admittently I'm not a fan of this design methodology.
Last edited:
I was always wondering why mica got such a good rap in audio. Speaking about the (in)famous and evil dielectric absorbtion (DA), mica is among the worst, certainly worse than most modern plastic film dielectrics.
And why placing electrolytics in the same group with the high K ceramics? Good quality aluminum electrolytics, correctly biased, have virtually unmeasurable distortions.
Am I detecting some prejudices here? Or are we talking the "sound of capacitors" (which I can't argue about)?
Last edited:
Waly,
We are in violent disagreement in this respect, and that is clear to anyone who reads my book.
Theory and understanding are VERY important. That is the synthesis aspect of design.
I obviously don't argue that "discoveries" are irrelevant, and I urge you not to put words in my mouth.
I seem to recall that, theory vs reality was the main context of the discussion.
I was largely focusing on what I thought you meant more as the analysis aspect of theory. For example, doing analysis on paper with handwaiving and some equations and some math. This aspect is useful in getting a start in analysis, and in reviewing various hypotheses, but is not sufficiently accurate to credibly assert certain things. Here is where simulation is important to more reliably analyze and check and verify hypotheses. TMC vs TPC is a good example. Handwaiving theory and the like is a decent start, but it is crucial to simulate to prove your point.
"Theory" can also be misused by making certain assumptions or applying certain constraints that another designer might not do. For example, in comparing TMC vs TPC the component values assumed for each are very important. If someone assumes the same relative values for the two capacitors for each approach, that over-constrains the design and can lead to some wrong conclusions. If you assert that C1 and C2 must be the same for both approaches (I'm not saying you do), that being one's philosophical view of an apples to apples comparison, then that is an example of where there might be a disagreement.
A classic example of misuse of "theory" was Otala's assertion that wide open-loop bandwidth was necessary for low TIM. The building and measurement of real amplifiers quickly proved that wrong. Actually, there was not much simulation going on back then, but proper simulation also would have proved the assertion wrong as well. Moreover, the proper application of theory also largely showed it to be wrong, but many to this day cannot agree on what the proper application of theory is there. That argument largely had to be proven by the proof in the pudding, where amplifiers were built and actually measured using Otala's own measurement technique.
Too many people also confuse theory with handwaiving.
I cannot count the number of learned PhDs who have written theses and papers who came to the wrong conclusions and assertions using their version of theory (and who failed to verify their assertions by simulating and/or testing an example circuit).
Cheers,
Bob
Manufacturing techniques have improved and a couple of maufacturers produce Mica that have lower DA figures which are comparable to plastic film.
For those preferring cheapest Ebay Mica and claiming better, thats the sound of capacitors view.
Yes we are, and BTW, quoting your otherwise excellent book in this context is exactly... handwaiving. Fact is, simulation doesn't create knowledge and understanding, but it's more like an (usually distorted) reflection of reality. Active vs. passive views, if you prefer.
You may dislike the math as much as you want, fact is, it would be very hard to publish a relevant paper in a reputable electronics journal, without revealing the underlying math (that is, providing an analytical understanding of the foundation). Over 60 years of IEEE and IEE publications are witnessing this fact.
The reasons why many people enjoy simulation, in particular in the DIY field? It doesn't cost a penny, and it's easy to build a "designer" reputation that nobody could deny, unless substantial amounts of time and money are invested. One just needs to blur the edge between simulation and the real world by using the appropriate verbiage.
DA is a material property, barely related to the manufacturing technology.
But then, I don't think that NP0 vs. mica vs. styroflex vs. mylar makes any difference. I can afford to believe this, as I am no audio professional designer that need to make a difference in the marketplace.
actually circuit simulation is accepted to "validate" circuit theory equations in some IEEE journals - hardware realizations, 'scope shots required in others
I think most would agree in discrete audio amps the Spice models and the physical realization's unmodeled/parasitic elements are major limitations
but lots of tools have ranges of application despite limitations, Spice can be very valuable to make specific tests of equations Or hand waving - and can't resolve all of either type of controversy
but physical audio amps can't either - parts variability - various of those physical implementation introduced elements. measurement tool misapplication, outright error can invalidate some "definitive statement" being "irrefutably proved" by a specific piece of hardware, some one person's posts about its "audio quality"
posted Spice sims, with source, for freely available simulators allow a high degree of community review, tire kicking, quick changes, application of differing and sometimes physically impractical "measurements"
where the tool is applicable circuit simulation is the fastest method today for circuit designers to examine, illustrate, teach, share circuit ideas, illuminate theory
I think most would agree in discrete audio amps the Spice models and the physical realization's unmodeled/parasitic elements are major limitations
but lots of tools have ranges of application despite limitations, Spice can be very valuable to make specific tests of equations Or hand waving - and can't resolve all of either type of controversy
but physical audio amps can't either - parts variability - various of those physical implementation introduced elements. measurement tool misapplication, outright error can invalidate some "definitive statement" being "irrefutably proved" by a specific piece of hardware, some one person's posts about its "audio quality"
posted Spice sims, with source, for freely available simulators allow a high degree of community review, tire kicking, quick changes, application of differing and sometimes physically impractical "measurements"
where the tool is applicable circuit simulation is the fastest method today for circuit designers to examine, illustrate, teach, share circuit ideas, illuminate theory
Hi Waly,
Sorry I set you off. I just re-read the first line of my post. I meant to say we are in violent agreement.
You keep putting words in my mouth, though. I certainly do not dislike the math. It has its place, and it is an important one. It just has limitations as well. As I've tried to point out over and over again, math, simulation and measurement all are very important; each has its role, its advantages and its disadvantages.
Your last paragraph is truly an insult to the whole of the DIY community who do simulation. It is completely wrong and you should be ashamed of making such a statement. Maybe you should just become a politician with slander like that. Ridiculous statements like that add nothing to the dicourse.
Bob
I think your definition of knowledge is narrow and academic.
A pilot does not need to know how to design an airplane in order to fly one (or use the training simulator for that matter), and I don't need a PhD in order to design a decent amp. The fact that a simulator is being used (or an airplane) to achieve something that those individuals could not do without their respective tools does not lessen the validity of either effort, or the final result - a point you seem to completely missed.
Interestingly, I lead a business a few years ago, and one of the PhD's said 'lets just tweak the mask and put a few wafers through the fab to check our idea out ...". Luckily, from my position as GM I was able stop him (it was the 3rd or 4th attempt at solving this problem) and send him and his team back to the simulator to do the job properly. After that, he made some progress.
Last edited:
Its not quite as simple as that, there are a couple of forms of mica having different properties. Some are not so well suited for capacitors and is most probably the problem with the DA properties. There is a premium form called Muscovite mica which also comes into two forms, the best being ruby, the other called green. With Ruby very good mica capacitors are made. Green is also used for capacitors but is inferior, (used in common Mica caps). Dont expect to pay the same for Ruby though, they are expensive compaired to other plastics and lesser quality micas.
Im not going to go into the technology during manufacturing part, that would fill another half page but I think you get the idea that it is not so straight forward as you thought.
- Status
- Not open for further replies.