813se DIY OPT problems

[hemgjord]

Any experienced OPT homebrewers out there who could have some hints as to a strange distortion phenomena?
I've wound a 10k:4 OPT for a 813 triode wired SE amp capable of approx 35W.
The core is approx 23cm2 and there is no signs of saturation.
So OK it is a tough task to wind a big SE OPT with high impedance ratio and I used cheap easy to get materials - 0.6mm OH film for pri to sec insulation and 0.05mm PP foil between primary layers. Instead of vacuum varnish I secured windings with 2-component polyester as I went along.
The primary is just short of 4000 turns and is split in 5 sections.

Given these conditions I did not expect top class performance but I really only expected the frequency range to be a little limited. Measurements confirmed that I was approx 1dB down at 20kHz a small bump up to approx 0dB at 27kHz then it falls steeply and is 3dB down at 35kHz.

Now to my real problem - it does not sound good. Very harsh highs.
Distorsion measurements at 1kHz shows good performance 30w below 3% THD and approx 0.1% at 1W.
BUT when I measure distortion at different frequencies it becomes strange. At 2W output I measure 0.2% at200Hz - 1.8kHz, at 1.9kHz I measure 2,6% and at 2.1kHz it is down to 0.4% then there are some more fluctuations up to 3 kHz then it falls down to 0.2% again and stays there.
These distortion peaks are quite rich on higher harmonics with 4'th and 5'th order being dominant.

Anyone that can offer some explanations as to why this happens. I really would like to know what I should improve before I try again.

[Richard Ellis]

I am REALLY interested in those who can answer your question as my 813 SE project is stalled on the OPTs' and it seems I will have to wind my own.
_________________________________________Rick..... ..

[rdf]

My recommendation, tap off your 813 grid with a high impedance resistive divider and see what the driver is doing.

[hemgjord]

Thanks for your thoughts so far.
I will check the driver at the 813 grid but I would be surprised to find anything strange there. It is a pretty straight forward design with a CF driver. The measurements were made at 2W which is pure A1 operation so it is a quite easy load for the driver.

Richard: the core is 23 square cm. The "C" in your scetch is 45mm and "B" is 52mm.

/Olof

[nhuwar]

Originally posted by Richard Ellis
I am REALLY interested in those who can answer your question as my 813 SE project is stalled on the OPTs' and it seems I will have to wind my own.
_________________________________________Rick..... ..

You don't have to wind your own, eletra-print came make them for you.

Nick

[Bandersnatch]

Originally posted by nhuwar



You don't have to wind your own, eletra-print came make them for you.

Nick

So can Bud Purvine...doesn't mean that is going to be the path, yes?

On the OPT, what is its secondary-shorted self resonant point?

Put a 'scope on the plate and see what it looks like( perhaps cap-couple with a good 1 uF cap ). Dial up your 2W and start sweeping upwards... or perhaps put in a current shunt and see what that looks like.
cheers,
Douglas

[rdf]

Originally posted by hemgjord
It is a pretty straight forward design with a CF driver.

I wondered. Depending on op point, tube and cathode load a CF can easily show rising distortion with frequency.

[hemgjord]

Depending on op point, tube and cathode load a CF can easily show rising distortion with frequency.

The problem is that distortion is not rising with frequency. There is a discontinuity aruond 1.9kHz where THD rises from 0.2 -2.6% within 100Hz then it falls back significantly a few hundred Hz up. Above 3kHz it is back to 0.2% and pretty much stays there as frequency goes up.

My guess is that there is some resonace in the OPT. But how to figure out where and why

If time allows I will do some more measurements this weekend, we'll see......

/Olof

[Bandersnatch]

hey-Hey!!!,
I would still wish to rule out the preceeding stages before applying too much skull sweat to the OPT path. What you're describing is remarkably high Q( as evidenced by its narrow bandwidth).
cheers,
Douglas

[hemgjord]

Hrrmm OK it seems like my sound card was the culprit in this affair.
I measured the signal on the 813 grid and found the same strange beahaviour there.
OK....hooked it up on the input, and guess what...the same!

It seems that it didn't like that I was running both channels into one, can't really see why, but the effect disappered when I disabled one channel.

But the fact remains, it sounds quite harsh in the high frequencies. So even if it measures quite good now there is work to do.

Anyone that knows what OPT parameters can cause this? And can someone point out some good hands on methods of measuring OPT parameters.

Thanks

/Olof