[Occam]

Hey Shayne,

Well I am something of a pedantic, prickly wanker. But simply put, this Lab forum is not a democracy. I am concered with those posts in which the poster is unwilling to explain to my satisfaction their conclusions which I consider 'less than thouroughly thought out', i.e. misinformation. In places other that the Lab, one is free to declaim 'Powerconditioners with series inductive elements all suck'. That may well be their opinion, and they're entitled to it. But I'm under no obligation to give it any credence here in the Lab circle. And there are vendors and posters here that are more knowledgeable than me. But there is, in any real lab environment, standards of rationality that should apply. My issue is not that I think them wrong (I certainly make mistakes and misstatements), but my applicable standard of the scientific method. And certainly, that is arbitrary. At some point I simply decide that there is no point in pursuing the exchange, as I lack the energy; or as my Mama told me -

'Never wrassle with a pig..... you just get filthy, and the pig enjoys it.'

I'm awaiting those 'battling engineers' parole from their timeouts, and am eagerly anticipating their input on this thread.
I endeavor to make this circle a somewhat academic environment, and like real academia, the conflicts are vicious and heartfelt, as, in truth, so little is really at stake.

[JohnR]

What a frustrating thread. It would be interesting to see a realistic model of a power supply. A capacitor and diode aren't it -- just accounting for the capacitor's ESR shows that noise does get onto the supply rail. You know, you puts a current through a resistor, you gets a voltage across it...

Anyway, I don't think that was what this thread was about, my apologies

[Occam]

John,

The best (easiest and free) off the shelf linear ps design modeling software I know of is Duncan's PSUD -
http://www.duncanamps.com/psud2/

I assume you're familiar with it, and I do realize its not amenable to noise injection, but it potentially serves as a 'callibration tool' for implementing a more sophisticated spice based model.  Spice (I believe?) models transformers as coupled inductors with control over parametric characteristics. I have used psudII and it was quite accurate in confirming the 'mystery of the exploding rectifiers' , forecasting ripple levels.... and it was quite easy to characterize the actual transformer. If one were to develop a Spice model that tied out with the PSUDII analysis for the fundamental power frequency, expanding the Spice model to a enhanced filter component both pre and post transformer might be a substantially easier when evaluating with various types of noise, common, transverse....

Better yet, cadge a working, accurate generalized Spice model of a linear supply from some kind soul. Are there any SPUGs (spice user groups) out there? But as most everything other than poweramp uses different forms of voltage regulation, the task is substantially more complicated. I realize that some might opine that voltage regulators eliminate the need for powerconditioning, I can only offer that a large majority of the NY AudioRave must be suffering from mass hysteria

[JoshK]

Well I can't speak for the rest of the ravers, but I have done very few drugs in my days and have decent hearing still (I can hear 18Khz just fine) despite a few heavy metal concerts in my youth. Felicia is no joke, it is most certainly audible, although Felicia herself thought that a few of the iterations were near the same in her opinion, the overall outcome was much better than wall.

[JohnR]

Here's something that could be used for a start on a better model:



It's from: http://www.beigebag.com/case_xfrmer_2.htm -- I'm not sure that there shouldn't be some interwinding capacitance (CWW) across the whole circuit, surely some of the distributed capacitance is outside the series LL and RW.

Then I guess we could just add some R and L to the filter caps and supply wiring and see what happens. Using PSUD as a sanity check is a good idea

[Occam]

John,

Thanks for the Beige Bag link.  The magnetics article was quite enlightening, and it was great to see a pure engineering evolution via those infinitely flexible primatives that largely converged with an 'old school' discussion -
http://www.geofex.com/Article_Folders/xformer_des/xformer.htm
But that same flexibility of Spice sure has a steep learning curve, and I certainly understand the desire to simplify the overly complex. Inevitably, even if you have the 'macros', the more you model, the more you're forced to actually examine the minutae of those macros. At what point must you go beyond that n order appoximation, i.e. is that distributed capacitance adequately described as bulk....