Welcome! Need support, you got it. Or share your ideas and experiences.
Return to MagneQuest/Peerless Forum
24.22.235.48
In Reply to: re: a transformer's REAL operating impedance posted by Tre' on February 16, 2015 at 07:55:46
We can only analyze models, which are simplifications of reality - so the question is, how complicated a model do you want to understand? :^)
The Level I model assumes a perfect transformer loaded by a perfect resistor.
The Level II model, expanding from the tube, would include the transformer DC resistances, self-inductance, eddy-current losses, and hysteresis losses in the bass; in the treble it would include the DC resistance, leakage inductance and parasitic capacitance. In both cases the load would remain a pure resistor.
Then you get a choice of what to refine for Level III - do you consider a specific speaker impedance characteristic, or do you build in the tube's cathode resistor/capacitor, plate choke/capacitor if parafeed, and power supply impedances? Or both?
Even at Level II, I think Spice is the most practical tool.
Follow Ups:
In practice, it is impossible to separate the loading effects of a speaker from the performance of the amplifier itself. However, it has always been my goal to optimize the amplifier first, and the system second. To accomplish this, more and more of my "acoustic" testing over the last few years has employed resistively-padded headphones rather than speakers. I don't have the equipment or environment to measure speakers, but when headphones sound good and speakers don't (with a particular amplifier), I know the speakers are a bad match.
--------------------------
Buy Chinese. Bury freedom.
Agreed, that once you include real world speaker loads and their behaviors the complexity of modeling those interactions becomes CRAY supercomputing level analysis.
I suppose you could limit the universe and use "bogey" numbers from a library of driver specs--- but even here, then when we recognize the existance of a cabinet and it's effects--- way too complex--- at least for the level of modeling that I'm interested in exploring.
MSL
Builder of MagneQuest & Peerless transformers since 1989
An understanding (or even screen shots at different frequencies) of Level II (something like what V\S did for only the inductance) would be good.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
I too think level II or components therefrom could form the basis for a useful "tool".
Would need to include only the variables that the end user would have some shot of actually obtaining the values of--- so that they could plug in the numbers to play.
Of the variables in Level II that Paul Joppa listed... I think the following could be included;
DCR (winding resistances)
self-inductance
anode's r sub p
the following variables are not likely to be available to the end user...
eddy current losses
hysteresis losses
Leakage L
lumped capacitance
and I agree that the "load on the secondary" should be represented as a pure resistance.
If it is possible I think a neat tool or model would be to (as VS did) show the "real" loadlines of the TUT (transformer under test) and superimpose these loadlines on the "flat line" loadline gleaned from a set of static plate curves.
From there I think it should be possible to calculate the distortion generated by the anode from the "real" loadlines and contrast that data to the plate curve results.
Should also be able to calculate when voltage and current clipping would occur.
these are just some thoughts...
please feel free to comment or add variables that you think would make the model richer--- keeping in mind that (at least on first pass) we should want to keep the variables which the end user has some realistic chance of obtaining.
MSL
Builder of MagneQuest & Peerless transformers since 1989
Post a Followup:
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: