Hi everyone, first time poster long time reader!
For a long time I've wound many output transformers for guitar amps and due to Low-Fi nature of guitar amps OPTs turnrd out to be good enough, but this time I'm in process of winding my first Hi-Fi tube amp output transformer, and after gathering information from several sources like RDH4, Wolpert's books and mostly important Patrick Turner's website, I realized that I need to find a good GOSS material for this purpose;
After alot of searching for available core materials I could only find some Non-grain oriented silicon steel (NOSS) material from china that nearly there is no data on this material anywhere! The seller only knows that it is M470 NOSS material and he's not even sure about it!!
With this material I know my OPTs will weight like a battleship but I don't have any other option right now.
So I'm wondering if there is any method that I can use to measure permeability vs flux density of an unknown core lamination using maybe an oscilloscope and a test winding with known turns??
Any help would be greatly appreciated!
Sajad
For a long time I've wound many output transformers for guitar amps and due to Low-Fi nature of guitar amps OPTs turnrd out to be good enough, but this time I'm in process of winding my first Hi-Fi tube amp output transformer, and after gathering information from several sources like RDH4, Wolpert's books and mostly important Patrick Turner's website, I realized that I need to find a good GOSS material for this purpose;
After alot of searching for available core materials I could only find some Non-grain oriented silicon steel (NOSS) material from china that nearly there is no data on this material anywhere! The seller only knows that it is M470 NOSS material and he's not even sure about it!!
With this material I know my OPTs will weight like a battleship but I don't have any other option right now.
So I'm wondering if there is any method that I can use to measure permeability vs flux density of an unknown core lamination using maybe an oscilloscope and a test winding with known turns??
Any help would be greatly appreciated!
Sajad
I think you can use the known formula that gives L inductance as a function of relative permeability and number of turns. The issue is that the relative permeability is a nonlinear function of the flux density, in essence it depends on the current.
The best you can do is to wind a test winding of 100 turns, flow through some known current at 50 Hz, and calculate the inductive reactance.
The best you can do is to wind a test winding of 100 turns, flow through some known current at 50 Hz, and calculate the inductive reactance.
I ran some tests on steel cores some while back, using a Variac to apply V to a winding, with a Tek current probe going to a scope. Increasing the AC VOLTAGE, the NON grain oriented cores went from a sine wave current to a triangular wave current, to a spiked triangular wave (the spike arising from the triangular peak).
The grain oriented cores went from a sine wave current, to a triangular wave, to a square wave current, to a spiked square wave. The spike on the end of the square wave.
Another clue is the lamination thickness. Cheap non oriented laminations will be thicker (16 mil typical). While better grain oriented laminations will be thinner. (14 mil typical) All generalizations of course. There may be exceptions.
Hysteresis curves would be another clue. Sine wave voltage applied to a winding and the current I monitored. The attenuated (and isolated) voltage gets integrated with an Op Amp integrator circuit. Both I and integrated V applied to an X-Y scope (with proper isolation from the power line and proper attenuation to protect the scope.)
See online diagrams for various schemes. Grain oriented gives thin square windows.
The grain oriented cores went from a sine wave current, to a triangular wave, to a square wave current, to a spiked square wave. The spike on the end of the square wave.
Another clue is the lamination thickness. Cheap non oriented laminations will be thicker (16 mil typical). While better grain oriented laminations will be thinner. (14 mil typical) All generalizations of course. There may be exceptions.
Hysteresis curves would be another clue. Sine wave voltage applied to a winding and the current I monitored. The attenuated (and isolated) voltage gets integrated with an Op Amp integrator circuit. Both I and integrated V applied to an X-Y scope (with proper isolation from the power line and proper attenuation to protect the scope.)
See online diagrams for various schemes. Grain oriented gives thin square windows.
Last edited:
"Hysteresis curves would be another clue"
You need to measure the B-H curve in any case to check what the saturation flux is.
Iron GOSS can be 1.7T but non-GOSS can be 1.4T or less I imaine for lower grade stuff. If it's an EI core then maybe 1.2-1.4T is the practical limit even for GOSS material.
The technique is available on various web sites and remember to include the integrator time constant in interpreting the results.
Unless you know Bsat you can't design the windings. You may, for an output transformer want to run lower anyway.
You need to measure the B-H curve in any case to check what the saturation flux is.
Iron GOSS can be 1.7T but non-GOSS can be 1.4T or less I imaine for lower grade stuff. If it's an EI core then maybe 1.2-1.4T is the practical limit even for GOSS material.
The technique is available on various web sites and remember to include the integrator time constant in interpreting the results.
Unless you know Bsat you can't design the windings. You may, for an output transformer want to run lower anyway.
I've done the variable 60 Hz V on a winding of some (economical) commercial OTs, with a current probe and scope, to see how high the current spike gets at the max rated ( at 60 Hz ) power or V level. Usually the current spike is unacceptably high, (like a power xfmr) so I rate the core usage for some V (freq. corrected) below that level for my OT usage. Some (all?) of the commercial OTs tend to push the B limits to get good power ratings. I figure I don't want to listen to big saturation spikes, but the reduced "clean" V per turn gets disappointingly low then. (more turns needed) I generally would use a commercial OT rated for approx. 50% more power than the Amp will be putting out, IF you want the max blast level to be clean. Rarely do most Amps get to that level in usage anyway though. Some distortion may be un-noticed at painful ear piecing levels.
Last edited:
L = AL . N^2, where AL is everything other than N^2 in that equation. However, as already mentioned above, this doesn't give you any idea regarding the saturation limits.
If you want to measure permeability on the scope, B should be proportional to Vrms and H should be proportional to the current (ampere-turns). The slope of the X-Y pattern would tell you permeability, onset of saturation etc.. Since the saturating portion of the curve is not going to be used anyway, it makes little sense to go beyond Bsat / Hsat.
If you want to test saturation without applying high voltages, just reduce the frequency, as the flux is proportional to V/f. This would also make the core losses / heat due to eddy currents etc. negligible.
It may be worthwhile checking with any valve audio people in your locality to see if any 'failed' hi-fi OPT's are hanging around as door stops, or sitting on a shelf. I probably have 5-6 bad hi-fi OPT's, given that decades of use and failed valves can cause primary winding faults. Of course most such OPT's get tossed, but you never now. Also those OPT's likely have a make-model that can be checked to see if they used GO cores.
Another 'cheap' option is to check solid-state PA amps of circa 35-120W rating as they typically include a very sizable OPT, albeit with inappropriate impedance windings (unless you want to use them 'in reverse' to couple an existing PA amp with highish impedance output to a speaker impedance).
Another 'cheap' option is to check solid-state PA amps of circa 35-120W rating as they typically include a very sizable OPT, albeit with inappropriate impedance windings (unless you want to use them 'in reverse' to couple an existing PA amp with highish impedance output to a speaker impedance).
Thank you smoking-amp,
I will try to plot the B-H curve and see the results. I guess u is the tangent of curve at any point, right?
Regarding max flux density in past I used 1T as max for NOSS and output transformers that I designed based on this turned out to work as intended, but anything higher than that would shift the saturation frequency higher than design; so I will use 1T for my designs again.
Thanks again.
Sajad
Is this correct? I thought it was reverse 😅
BTW can you send a link to B-H plot test configuration?
Thank you
Sajad
Thank you so much,
After a little search I found this: Link
In the link it mentions that for a digital oscilloscope the integrator is not necessary, is it correct?
The link above that you show has it all. Substitute E(t) = E sin (2*pi*f*t) into the equation for B(t) and you'll also see how flux is proportional to V/f.
A B field results (according to the material's permeability) for the given H field is applied. After saturation, changes in H field do not reflect in proportional changes to the B field, resulting in the flat portion of the BH curve. The curve need not trace back on itself because of hysteresis.
Since the final goal is to set the lower cutoff frequency, determined by the source resistance Rpp and the primary inductance L, perhaps it makes sense to design for a given inductance. Then calculate the optimal geometry for the peak current. The optimal geometry parameters are the core cross-section A and the average magnetic path length l (lower case L). From these parameters you can select a suitable laminate material. Then use a 100 turn test winding, measure the inductance with medium current (voltage) and you will get L. From these you can calculate n number of (primary) turns.
Of course, if you know the relative permeability, you can reverse the design order.
Of course, if you know the relative permeability, you can reverse the design order.