Theoretical question about interstage transformers

[popilin]

I have seen partial cancellation of both 2nd and 3rd harmonic with a 6P41S and 6N2P driver.

Great! Thank you for sharing!

More experimental evidence for those who want to cover the sun with their hands.

However it was very sensitive to bias condition of both tubes.

Not something I would want to try to duplicate in mass production, or try to maintain over a long period of time.

I can vouch for that, a bunch of simulations with a lot of different valves, every one at different bias points, the result was surprising, the winner was another cinderella, the ECC81/12AT7.

Unfortunately I have just an old scope and a couple of multimeters, OTOH LTSpice makes calculations better than me and distortion cancellation is more than evident.

 

[45]

You did not understand anything! Please, read again post#81 and post#82

I understand that you read books and run simulations. What should I read again?

Sorry Juan but this is enough.

 

[waltube]

to 45: It is a old story.

I have read this thread quickly. I have also ready some proto on lab, if in next days I found some time I will run some test to show what happen.

The theory is always working. I never understood why, in some "esoteric" solution to drive a power tube someone spent time and test to get a great swing (possibly with low distortion) on driver tube connected to a interstage transf. with the descending ratio!! (this because we need to drive the power tube with a low impedence) Then this signal will be amplified again and , at the end, with OT trafo (descending) , finally I drive the speaker. Too much work and money; and it is easy that the Interst. Trafo cost more than output!!! For what? Also for a interstage with a high impedance where the ratio rises; they have always problems!

This is my opinion
Walter

 

[50AE]

It has nothing to do with DC flux but with a nonlinear anisotropic magnetic material, SE transformers are more linear than PP due to the air gap, anyway transformers add odd-order harmonics and it is measurable.

My rule of dumb is to use transformers/inductors as little as possible, i.e. PTs and OPTs, less is more... and cheaper!!!

Because in SE mode, the DC flux moves the bias point into the first quadrant of the magnetization curve. When AC swing occurs, the flux increases with the first half period, then decreases with the other half period. Isn't this a situation for assymetrical distortion?

In PP, ideally there is no DC flux, hence the magnetizing current increases symmetrically for both half periods. Odd order distortion.

Hi grade steel has low distortion in comparison to the rest. THD, basically H3 and very little H5, only raises to very acceptable entity below 30Hz where there is basically nothing at music level in most cases. IMD caused by that is even smaller as shown before. Properly designed transformers will reach that point around clipping where most of the distortion will come from somewhere else....
If then one really wants to make core distortion negligible just uses 49%Nickel. For PP this is a good choice: very low distortion, low losses, high initial relative permeability at 10000, 1.6T saturation. It comes in handy EI shapes but doesn't have the issue of GOSS EI's because the steel balance in it is non-oriented. Game over.

That sounds like a good info, but made me curious. Why does the problem about the orientation with GOSS disappear with nickel?

 

[indra1]

... If then one really wants to make core distortion negligible just uses 49%Nickel. ...

I wonder why so little love for Parafeed SE IST. Too heavy?

 

[TheGimp]

Now you have both Transformer and Capacitor in the signal path.

Make everyone unhappy

 

[indra1]

... Make everyone unhappy

Sometimes I wonder if it is caused by fleeting hype. I see the Pass SIT-3/DEFiSIT use both yet gather positive comments.

 

[45]

That sounds like a good info, but made me curious. Why does the problem about the orientation with GOSS disappear with nickel?

In GOSS EI cores only about 2/3 of the laminations are oriented in easy mag. direction the rest it's 90 deg which is the worst for losses and permeability. When they cut an E the gaps in reality become the I's if I remember well. So there if the I and the 3 legs of the E are in the easy mag. direction then the rest of the E is not, being at 90 degs. In C and toroidal cores this doesn't happen because they are bent into shape along the easy mag. direction. So they are 100% oriented. The consequence of being not fully 100% oriented I have already told earlier in this thread.
The non-oriented steel has lower permeability and a bit worse hysteresis but doesn't have this problem. When you mix it with Nickel in the right balance you can get the best of both worlds or specific performance that cannot be achieved with GOSS.

50AE I actually wrote it in another thread...

"Without going too much into details, C cores are more efficient because grains are 100% oriented. They can work up to 1.6 T induction if good grade (possibly 1.7T in best grades) without relevant distortion at low frequency, at least down to 30Hz. In EI cores only about 2/3 of the grains are oriented and saturation doesn't surge uniformly. Saturation hots spots do have a visible influence on distortion and inductance. To get a performance close to that of C cores one has to stop at about 1T or accept some more distortion below 35-40 Hz for running up to 1.2T. So for the same core area and turns one can apply 30-60% more volts resulting in more headroom or can reduce the size of the core for the same headroom. Toroidals too are like C cores."

 

[popilin]

Because in SE mode, the DC flux moves the bias point into the first quadrant of the magnetization curve. When AC swing occurs, the flux increases with the first half period, then decreases with the other half period. Isn't this a situation for assymetrical distortion?

In PP, ideally there is no DC flux, hence the magnetizing current increases symmetrically for both half periods. Odd order distortion.

i) Magnetization M, and magnetic field B, are related but not the same thing.

B = H + 4π M

Then magnetization curve, M vs H, and magnetic hysteresis curve B vs H, both have the same shape but are not the same thing.

ii) I think that you call flux to magnetic field B, that is not correct because it is a flux density.

iii) From Maxwell's equations

dB/dt = (c/S) Uac

It follows that AC voltage and magnetic field B have a π/2 phase shift, then, when AC swing occurs, magnetic field B increases with the first half period and decreases with the other half period, regardless of SE or PP operation.


I think that you are trying to say is that in SE operation magnetic hysteresis curve is asymmetrical and in PP operation it is symmetrical.

This is true in pathological cases when you are near saturation, for well designed transformers the magnetic hysteresis curve can be amazingly symmetrical.