Yes. The key word being "mutual."
When the coupling is perfect, there is no inductance. Only the reflected impedance of the load.
se
You'd be well advised to keep your inductance in your pants and well away from any gauss lines.
se
I don't think so
in the linear model the electrical resistive paths in the core material are always coupled by the mutual inductances - the model is continuous/distributed so a lumped circuit equivalent approximation requires many branches in a "ladder"/mesh? of partial mutual inductances and resistances
the resistances are "real" and have thermal noise which should show up as frequency dependent noise all of the time
the nonlinear core loss mechanisms may quickly dominate as signal level goes up though
Last edited:
You really should have asked about the Ls-Rs model, and how an HP 4284A classes eddy losses as a series resistance within the wire, in addition to the reduction of inductance caused by the eddies..
This implies that any current within the wire not only generates noise within the wire, but the slew rate of the current will cause the eddy currents to generate noise as well. A first derivative based amplitude modulation of a noise source..
I wonder how that would show up on a spectrum analyzer..
Cheers, John
As I mentioned earlier I was disagreeing with the use of the word "excess" noise because it is confusing. Active mixers for instance have modulated noise due to the fact that the operating currents in the Gilbert cell are shuffling back and forth. This signal dependent noise has been studied since at least 1962. The 1/f behavior only will show up (simulation with R, L, C) in a distributed model for the system. This is different of course from when one puts a carbon comp resistor as one leg of a bridge. There you extract the excess resistor noise from the excitation by subtraction. Walt Jung and I built a "perfect" pink noise generator based on this and now I can't find a resistor horrible enough to repeat the experiment.
Sounds like time for an experiment with an ultra-pure sine wave and a transformer. BTW I found some good papers from SLAC on magnetic shielding falling apart at nano-Tesla levels, I would love to see some science applied to this issue it is obviously not nonexistant.
EDIT - I might also disagree a little on this, yes eddy currents are a loss but it is not a given that they generate excess noise (coupled back into the system) especially in a fairly pure conductor.
Attachments
Last edited:
I have to agree with jcx in this case "no signal no noise" does not really apply the noise in the system is a signal. At thermal equilibrium the eddy current losses will be accounted for, i.e. if I take a transformer and resitively terminate it at both ends the losses will show up in the resulting input/output terminal measurements.
Which is another way of saying I have not observed a transformer losing the ability to pass a low level signal even to the point of the thermal noise of its terminations.
This also brings back fond memories of the guy down under that wound 300,000 turns of wire on a bundle of rebar and measured the Schumann resonance (pico-Teslas)
Which is another way of saying I have not observed a transformer losing the ability to pass a low level signal even to the point of the thermal noise of its terminations.
This also brings back fond memories of the guy down under that wound 300,000 turns of wire on a bundle of rebar and measured the Schumann resonance (pico-Teslas)
Last edited:
What then of proximity effect? In practice its just like skin effect but in a multiple layer winding such as found in transformers its a significant loss mechanism. Doesn't it contribute noise at higher frequencies by virtue of increasing the resistive losses in the wires? And if it does then its a signal dependent loss.
Not the same thing this is all covered by a multi-physics model at equilibrium. The skin effect can produce fractional f effects. They are still continuous with level, the skin effect does not "stop" below a certain level.
If I put say 100nV (or whatever) on a transformer some of the results presented say mu collapses and I get nothing out. Please present evidence of this.
I think here we're arguing at crossed purposes. I'm not of the view there's anything discontinuous going on. Just offering support for Mr. Neutron's thesis which you seemed to be disagreeing with.
I'd be as skeptical as you of such 'results'
No, there will be a voltage (see Faraday). But what's the point of a transformer if you've no load on the secondary?
se
I think here we're arguing at crossed purposes. I'm not of the view there's anything discontinuous going on. Just offering support for Mr. Neutron's thesis which you seemed to be disagreeing with.
I'd be as skeptical as you of such 'results'
My disagreement is a very fine point. You induce eddy currents and they dissipate heat so you have a loss that's OK. The claim that these eddy currents experience excess 1/f noise as in a carbon comp resistor and this loss is also coupled into your system is not a given.
Yes, thanks for the clarification. The devil is most certainly in the detail.
In the way I read Mr. Neutron, I don't see him making such a claim (that being about excess or 1/f noise). Just pointing out a signal dependent loss - one which you acknowledge already in accepting fractional f.
If anyone cares to read the AES papers made by Jay Knight during the 60's (when I worked with him) and the 70's, you will find that eddy current loss makes high frequency noise and this is dependent on the lamination thickness, typically 6 mil, back in the early 70's and earlier, 2 mil, later. 1 mil has long been possible, but it is very difficult to handle by the assemblers, I am told. It would be interesting what thickness the laminations are in the input transformers normally made today? Cheap ones may use 10 mil lams, extra good ones might use 2 mil lams, I would bet 4-6 mils for most of the professional models.
Do you mean Jay McKnight?
If so, which paper or papers specifically?
And was he writing about audio transformers or tape heads?
How significant is the noise and what lamination materials did he write about?
se
Last edited:
Oooh se, its great when you get all philosophical
Yes it is still a transformer ontologically, but no its not one existentially. Any help?
I'm using 600:600 quadfilar line output transformers as center-tapped chokes in this headphone amp/preamp circuit.
An externally hosted image should be here but it was not working when we last tested it.
se
- Status
- Not open for further replies.