5k:8 ohm vs. 10k:16 ohm output transformer. What is the difference?

[popilin]

"That leakage inductance is too high for any of the two transformers.
I would suspect manufacturing defects such as wrong interleaved winding technique."

Well no, 84.6 mH gives 9568 Ohms reactance at 18 KHz, which is just below the 10 K Ohm primary Z, which it acts in series with at the high freq. end to give about a 3 dB drop as expected. Its just not good for a 5K primary.

Is 18 KHz a new standard?
In my SET designs always try, at least, 20 Hz to 50 KHz
Leakage inductance depends on magnetic coupling between primary and secondary.
So interleaved winding is the natural way to reduce it.

And these OTs don't have any interleave winding, so it would be hard to mess that up any further.

No interleaved winding is worse than wrong interleaved winding technique.
This explains it pretty well.

 

[smoking-amp]

LL testing is difficult on a well-equipped manufacturer’s bench. I am curious how this is done in the DIY world. How was the reported LL testing done, at a single test frequency and the frequency responses extrapolated by the Chrowhurst rule of thumb; 2*Pi*F*L?

The DIY leakage test is to put an L meter across the primary and short the secondary (use highest freq. available on meter, like 1 KHz or 10 KHz). Unless you have a network analyzer to resolve the other parasitics, thats how it usually gets done. One could measure the high end 3dB drop off with a signal generator & load R and then calc it.

The L meter and short technique is very useful for resolving the coupling of portions of a winding too. For example, you can determine the coupling of the 60% and 40% UL primary sections individually for a P-P OT to determine balance (you have to take into account the turns squared for dissimilar section comparisons). The CXPP OTs for example have very poor balance with one 60% winding flapping in the breeze (thats what you get with insufficient interleaves and a non split bobbin).

---------------------------

In my SET designs always try, at least, 20 Hz to 50 KHz

Thats what it SHOULD be, but you can forget Edcor in any form for that performance level. Even the P-P OTs only make 20 KHz, forget about global feedback with them. Local feedback is the only option there these days.

 

[Pano]

Please use the quote function of the forum, it is the accepted format.
Either chose the quoted text with your mouse and click the quote balloon icon

___ or do it by hand with BB code. Quote tags look like this: [quote]My hovercraft is full of eels[/quote] which shows as:

My hovercraft if full of eels

(BB code tags are case insensitive)

Thank you.

 

[smoking-amp]

I just went and measured all three cases of the GXSE 15-16-10K using a signal generator (50 Ohm) and load resistor (8 Ohm) and scope. Surprisingly I get a 30 KHz top bandpass for all three (spacer gapped, butt gapped and interleave stacked). Then I took the load resistor off and found there is a 30 KHz resonance peak there, so this is rather misleading data. Not sure how to measure this thing now, or whether the resonance is actually helpful for the bandpass. Could make for some shrill notes at 30 KHz, but assuming nothing is up there, maybe this is an amazing way to compensate HF roll-off .
There is also another shunt resonance (abrupt signal drop) at around 50 KHz. A big phase shift occurs at the 30 KHz point, so is probably not useful bandwidth as far as any global feedback goes.

 

[DualTriode]

My guess is that the resonance peak is due to the tuning of inductance and the capacitance between the coils and between secondary coil and iron.

Lots of variables happening there BTW.

You need a SET to hook up to that transformer test rig.

DT

 

[smoking-amp]

Correction:

I just realized my LC meter powers up in 120 Hz mode, not the 1 KHz mode I thought. After re-measuring the leakage L (in 1 KHz mode now) for the GXPP15-16-10K in the 3 core versions, I am seeing a narrow 42 to 44 mH range of leakage L. This puts the upper freq. end at 37.7 KHz for native 10K Ohm mode and 18.9 KHz for 5K Ohm mis-applied mode (8 Ohm load). Not bad at all. I get about 40 KHz with the signal gen. in 10K mode (16 Ohm load). And 30 KHz in 5K mode (8 Ohm load). I still can't explain that one. The resonance looks to be closer to 40 KHz on closer inspection (at end of sig gen dial, calib was off).

I re-checked the primary L for the gapped case in 1KHz mode and I now see 27.45 H, and 80 H for the interleave stacked case. That puts the primary L at the right figure for the 40 Hz low end spec, as long as the DC current doesn't reduce the inductance. Hannah magnetization curves for DC show the M6 dropping by over 1/2 in permeability at half flux (max DC), but that would be moderated by the air gap.

Seeing as the gapped case has about 1/3 the primary L as the interleave stacked case, that would be like 2 Ohms gap + 1 Ohm core (magnetic analogy) versus 2 Ohms gap + 2 Ohms core with max DC, so the DC primary L should be around 3/4 with max DC or 20.6 H giving a low pass small signal f of 53 Hz. (I think my earlier 80 Hz figure, from my now year old notes, was measured at the full 15 Watt power)

Using the 10K as a 5K OT would then lower that small signal f to 53/2 or 26.5 Hz and the full 15 W power f to 80/2 or 40 Hz. The max DC current however in 5K mode would be limited to 0.7X of a real 5K OT. Looks like the high freq. end is OK now.

 

[smoking-amp]

Late corrections to the above post:

("GXPP" above should have been "GXSE")

I forgot to include the increase in core permeability with larger signal level (over the small LC meter test level) in the above low pass freq, small signal, calc (overcomes the DC loss of permeability). Looks like that would be just enough to allow the GXSE to make the 40 Hz small signal LF spec (instead of the 53 Hz figure above). Doesn't help the max power saturation spec though, 8 Watts looks to be the abs. max at 40 Hz.

 

[smoking-amp]

I just measured the LF small signal response on the gapped GSXE15-16-10K using a 10K source impedance from the sig gen and a 16 Ohm load. Depending on signal level (which affects the material Mu) I get 18 to 26 Hz for the 3 dB low pass (that's without DC, .3 VAC to 7 VAC sig). Taking into account the DC effect would raise this freq. to about 26 to 37 Hz. So the small signal LF spec is looking OK.

I re-checked the magnetizing current spiking and it sets in at about 35 VAC (60 Hz) which would correspond to about 8.5 Watts at 40 Hz 16 Ohm load with DC max current consuming 1/2 of the flux swing. Can get up to around 48 VAC where it looks like a typical power xfmr running to max sat. and that would correspond to 14 Watts at 40 Hz 16 Ohm load (at max DC current = 1/2 max flux). So I guess one could say it just squeaks by on its specs. That seems to be the sat level used by guitar OTs. Maybe I was overly harsh in down rating the unit earlier, but for Hi-Fi use I would still call it an 8.5 Watt OT at 40 Hz.

To get any more accurate will require building a CCS to supply the max DC current during all the tests. Since it seems to be just making its specs, I will call it quits here.

 

[smoking-amp]

For the mis-application initially raised: (use as 5K instead of 10K and 8 Ohm load instead of 16 Ohm load), the main issues are the doubled wire resistances (primary 159 Ohm increased to 375 Ohm, and secondary about .5 to 1 Ohm), and the 0.7X reduced DC max current versus a real 5K OT. The high freq. limit should have reduced by 1/2 too, but it only came down to 30 KHz from 40 KHz. ?? The low freq small signal limit will drop by 1/2 to 20 Hz approx. Power rating at the low freq. end should drop by 1/4 due to the reduced DC current avail (class A op, 0.7X, hence lower AC current too) and lower Zpri (I*I*R = .7 x .7 x .5 = .25), and with more losses from wire resistance. So like 2 Watts at 20 Hz. A DC bucking winding and CCS can bring the LF power spec back up some though (1 x 1 x .5).

 

[DualTriode]

smoking-amp,

Provocative, Interesting results!

Thank you for re-thinking and going back through to test things. Not everyone would do that.

DT