audio-talk

Nick · Posted: Sat May 30, 2009 11:16 am Post subject: 813 Electrostatic amp

Thought I would start a thread of its own about this.

Got most of the bits now, waiting on 4 valve sockets (and caps) and a transformer for 4 813 heater supplies. Colin has lent me the output transformers from his Kit6550 to use as a center tapped choke, and the SP wound interstages to use as a phase splitter.

My thoughts so far (and please if anyone has other ideas let me know) is to use the same driver circuit I used for the 300b. DC coupled two stage 5687. It can provide enough to handle the 160v p/p that I think the 813 will need. That will have its own little 310v supply.

Output stage will be the two 813's triode strapped running at about 80ma each. So thats 320ma in total. I have a nice old oil filled transformer thats good for 735-0-735 @ 365ma so that should do. The power supply will be simple ish, using a string of 6 UF4007 as rectifiers for simplicity. PSD predicts a peak current pulse of 2A with a width of 2.5ms so well inside their specs. I have a couple of 1.5H 500ma hammond chokes I can use. They are not spec'd for 1kv, but if I don't earth their chassis they won't know, or I could put them in the ground line. So a CLCLC with 50uf caps between them. The first cap will havbe a ripple of about 700ma so I will use some stacked electrolytics for that, and I have a couple of 50uf 1500v solens I can follow them with. Should give me about 950v with about 50mv of ripple. So should be ok for class A PP.

Looking at the triode curves for the 813 800va @80ma is about right for the power supply I can provide. So a 1k cathode resistor would be about right. But I have worries about how closely matched the 813's will be, so given the driver TX has two independent secondaries, I am going to use the Blumlein garter arrangement so it will reduce the current imbalance through Colins TX's. That burns another 80v of B+, but its about right, 160v on the cathode, -80v on the grid, 800v across the valve, that given me the 860v I started with. Hopefully it will work out like that in practice.

The output will be taken from the anodes via a pair of two 10nf 2kv caps (I think a single cap would do, but I am being a bit careful) to a string of 10 1M resistors each side to ground, and then to the terminals to go to the speakers. I haven't got a 813 spice model, but using a pair of GM70 which look about the same sort of curves it looks ok to produce about 2kv P/P which should do something to the statics Smile

Toppsy · Posted: Sat May 30, 2009 11:45 am Post subject:

pre65 · Posted: Sat May 30, 2009 11:53 am Post subject:

I've just purchased 50 of BY255 diodes for a pittance so have a few ( Wink

) to spare if they are of any use for this project.

http://www.semikron.com/internet/gecont/pdf/1020.pdf

Nick · Posted: Sat May 30, 2009 11:58 am Post subject:

Nick · Posted: Sat May 30, 2009 11:59 am Post subject:

Why do I do this, moment it gets hot, I start building a amp with 200w of heater power. And in the winter I build 300b amps. Something wrong there.

Andrew · Joined: 24 May 2007 Posts: 1938

Hi Nick, there's an 813 model in SB's tetrode library.

-- Andrew

Nick · Posted: Sat May 30, 2009 12:08 pm Post subject:

Thanks, why didn't I think to look there.

Nick · Posted: Sat May 30, 2009 12:11 pm Post subject:

Can't find it in mine, could you send the model if you have it Andrew?

pre65 · Posted: Sat May 30, 2009 12:29 pm Post subject:

I also have some EM516 if they are better.

http://www.semikron.com/internet/ds.jsp?file=1056.html

Andrew · Joined: 24 May 2007 Posts: 1938

Done....

Mike H · Posted: Sat May 30, 2009 4:11 pm Post subject:

Ok so I see from sirquit that R3 & 4 is the electrostatic wotsit. And there seems to be a centre-tap to earth as well. How does that work?

Nick · Posted: Sat May 30, 2009 4:20 pm Post subject:

Nick · Posted: Sun May 31, 2009 2:59 pm Post subject:

Playing a bit more in spice, the 5687 gives far too much gain, given the amp had about 70dB of voltage gain I don't want unexpected 1kv spikes out of the thing when switching inputs Smile

So I have changed the front end to a d3a, maybe a c3g if I can find the ones I have and the sockets for them. better gain, 1v in 1kv out, a nice 60dB Smile

I also used the numbers Colin got from the designer of the statics about the capacitance them present, and have added the equivalent circuit of the speakers on the end. the peak voltage is now down to about 1k6 P/P with the load, but I thought it was interesting to see that the normal problems with statics and HF impedance is far less of a issue here. So it will be interesting to see what the top end sounds like.

I think the real response is going to be controlled by the interstage.

Nick · Posted: Sun Jun 07, 2009 6:38 pm Post subject:

Well thanks to Colin for letting me plunder his 807 breadboard for bits (I will put them back honest). I have most of the 813 amp built, just about to pluck up courage to test the output stage. The driver stage and power supply seems to work so far.

andrew Ivimey · Joined: 11 Jun 2007 Posts: 2468 Location: Bedford

Music to my eyes - that looks the business, I love it - phroaaaaaaaaaar!