Antek Toroidal power transformer for Step-up, Measurements (part 1/2)

[geraldfryjr]

Yes,Doc,that is normaly the case.
It was what I was scared of,but I had already tested the panels at 9 Kv to 10Kv prior to adding the transformers with no issues or noises.
But since the power supply had quit I couldn't recheck my panels again.
I am working on it now so I will be back with a confirmation shortly.

jer

[arend-jan]

Thanks bolserst for the detailed analysis. Your ways and reports are always a delight.

Regarding the different series/parallel arrangements, perhaps a non-idealized circuit diagram (including all parasitic capacitance, inductances and wire resistance etc) will give more insight?

[geraldfryjr]

I got my supply fixed and I just did some tests.

As it turns out the one panels is not quite up to par yet and had made a few more leakage points along the edge as expected.

The other panel I had spent much more time on it and it will sustain a 10Kv bias with no noises even with the two transformers hooked up.

Except on the initial charging of the diaphram and after everything as been charged and the voltages have equaled out it sits quietly all of the way up to 10kv except for any sudden large Bias voltage changes in which noise would be expected.

The panel that is working correctly has the original diagphram as when it was built in 2003 with the original formula Licron coating.
The diagphram on this panel has a very low tension and is why it takes it more time to stabilize as the bias voltage is raised.

The other panel is the one that I rebuilt tested and had blown up in my testing two years ago and had documented the rebuild in these threads.

I just need to reseal the edges with some Silicone as this seems to be the best stuff to use of all after all of the clear nail polished that I had used.
The nail polish works good but it take a lot of it and it thins more as it cures and the total cure time is very long as well.
One new strip on the older panel with some flowable silicone was all that it took so I will have to do the same to this one as well.

The transformers are the same as I had been using and I am think of getting some of those Antek's to give them a good go around to help confirm this study.

I have a primary turns count of 10 at the moment and this gives me a 1:254 transformation ratio.
And when I blew up the one panel I was running about double that with like 5 to 7 turns on the primary and a 50V to 60V signal into the transformer with my Crown DC300A amplifier.
Bringing my Stator voltage up around 25KV p-p out of the transformer before the edge of the panel gave out (the newer rebuild at the time).

As far as the bias goes I can now continue safe testing at 6.78Kv half of what my supply produces.
I have taken it to 10kv for a few short tests to see if any thing breaks down and it hasn't so far ,but this is just some preliminary test for now.

I will push it farther once I get some baseline data together and switch to an amplifier that is more capable as I am using my little cheapy Aiwa for right now.
As well as considering on a new panel build design for the much higher voltages as I am now pushing them 2 to 3 times what they were initially designed to do.

I can tell you this though is that I am getting aproximately 3db to 4db more sensitivity by increasing the bias voltage from about 6.5 (or so,nominal 6.78Kv) to 10kv as expected.

Sadly I don't have any measurement data from 2 years ago to see if any of the changes are relevant,But I assume they are.

My first successfull SPL measurement this morning was about 85.2db at 1 meter with a pink noise signal feeding a 10 turn primary (1:254) and a 1 ohm resistor in series with the primary at a 10kv bias.
With out the resistor this figure would be a bit higher.
Also the lows at 80hz and highs at 12khz were cut but 12db because the amp can't handle it,so this was the most this amp can do before it shuts down.
But this is about 102db 5 inches from the diagphram.

I didn't want to go of topic But I figured I would give the heads up on my findings and test setup and if I do get some Anteks or run into any other transformer issues I will mention them here.

Thanks !!

jer

[bolserst]

Originally Posted by arend-jan

Regarding the different series/parallel arrangements, perhaps a non-idealized circuit diagram (including all parasitic capacitance, inductances and wire resistance etc) will give more insight?

You are right of course...I really should put together a SPICE model of the 4 transformer arrangement with all parasitics modeled to see if one arrangement has any advantages as far as HF bandwidth. I had been thinking that since all 8 of the primaries are pretty much at the same voltage potential when compared to the HV on the secondary windings, the connection arrangement of the primaries would have little if any affect on the HF bandwidth.

But, we all know what happens when you get to thinking too much

[geraldfryjr]

I was working on my transformer configuration yesterday.
I used my cheapy amp because it gives me a sense of reference for an overload condition that is consistent.

My cores I believe have bifilar 120v windings as well.
With 10 primary turns on each of the two cores and and the 4 120v windings in series as normal.
This gives me a 1:256 transformation ratio so it is quite high,But again I am running smaller panels and the panels capacitance is nil compared to that of the transformers capacitance.
So this test makes no difference if the panels are connected or not.

What I have found is that I could not go past 2Khz with a 10V peak signal without the amp shutting down.

When I tried using 4 cores using only one 120V winding on each core in the same configuration of a 1:256 transformation ratio, I was able to sail on past the 2Khz barrier to 20khz wit a 10V peak input signal with no overload or shutdown,and even to 20Vpeak at 20Khz with my cheapy amp.

This is a very significant improvement.
I have not been able to document any higher voltages because my 10X scope probe is broke and the scope graticule only goes to 20v p-p.

But I was actually able to now use my cheapy amp at a reasonably loud volume without it being so sensitive to overload and shutting down on high peaks (as much).
It was quite loud for pleasure listening at this point with no interuptions.
I Measured 102db at 1 meter with test tones before the amp shutdown at maximum output.

This is around a 16db improvement over my very first measurement in post #33.

I haven't yet measured the capacitance of the new configuration to compare the difference, yet.
But, I do think that it is safe to say that one should seek using transformers with a single 230V winding than using one with two 115V or 120V windings or use 4 cores of the lesser for better performance than using two cores.

Possibly one such as the Farnell transformer that CharlieM had mentioned.


Gotta Love R&D !!!!



jer