Kevin Gilmore Circlotron Output, Large Enclosure [high-slew] Electrostatic Amplifier

[Tinkerer]

EDIT 3: Fixed the balance problem. It's just the driver board has to warm up all the way for it to get near zero on the right channel. Still need to fix the buzz/hum. Will post a video in a second. Sounds like 60hz to me but you guys will know better.

Output stage gets damn toasty though. It's the only thing on the front 12x6x3 heatsink and after less than ten minutes, that is hot to the touch at 38mA. Driver/HV900 sink is warm and GRHV/HV900 rear sinks are stone cold. Lot of mass to those things.

 

EDIT: And current front panel with small variacs mounted. Not the ideal solution but it works and looks pretty spiffy.

Edited November 20, 2017 by Tinkerer

[Tinkerer]

Link to hum. You'll probably need to turn up the volume a bit since it was recorded with a crappy phone mic. https://www.dropbox.com/s/zomvnhpob9d4dx5/hum.mp3?dl=0

 

Any help you guys can give hunting this down would be great. If I can just kill this hum and be sure I don't have to reduce the current even further on the Output boards because boy do those things get hot in a hurry then I'm in the clear.

[sorenb]

if you lower the picth, aiming for A4 @ 432Hz, you'll have a perfect B2 pedal point ...might make a good contrast to the fluttering appearing a little later 

Anyways ...most likely the wiring harness around the output board that picks up the rectified mains or bad connections somewhere .... look at the schematic for the output board; the batteries, resistors (and the current limiter you have added) are supposed to be an integrated part hereof, meaning shortest possible wiring - maybe start with red and blue wires...

[Tinkerer]

Would only be able shorten a few things a few inches, so like 12" instead of 16" on some of the HV900 - leads. The way the boards are laid out necessitate most of those lengths, which are pretty darn long. I might have to try shielding some stuff instead, like the really long transformer leads. The audio input stuff already is and are connected to chassis ground. Any good bulk braid cheap?

 

Heck, I was reading that sort of thing might even be that the diode bridges on the HV900's might be damaged and leaking. They're about the only parts I've never replaced on those things no matter how much they've blown because they still work. Need to check if they're leaking AC I guess.

 

Connections are good. Just rechecked all those from stem to stern.

[JoaMat]

Early stage work on a Circlotron (Aug. 2015).

The prototype exploded during initial take-off roll.

[Tinkerer]

On 12/10/2017 at 8:57 AM, JoaMat said:

The prototype exploded during initial take-off roll.

Exploding is one thing it's really good at.

 

Well, I went and put some 1 and 3/8" tinned copper braid shielding on the the transformer secondaries. 1/4" quick connectors to ground leads to star ground. Made stuff significantly quieter but hum still there. I'll get around to the primaries and such next. I got 25 feet of the stuff so might as well cover everything. But I wanted to ask if using some kind of high voltage coax cable to replace the unshielded 900V lines might be a good idea? Since that seems to be what is picking up most of the noise in the first place.

[JoaMat]

Electrolytics in red….

….done today - probably on its way to trash can.

[JoaMat]

Putting something like cpc1117 cross C2M1000170D gate and source would that be usable as an emergency stop? Alternative cross gate and output (before 5.1k resistor).

Edited December 22, 2017 by JoaMat

[kevin gilmore]

That should work

also since it's normally closed could be used for startup and shutdown

[Tinkerer]

So, been kinda beating my head against a wall with this like usual, shielding stuff and whatnot but I found something by accident that may be helpful. I was bringing up the GRHV main transformer on a variac to make sure things still started up right on the first half of the amp after all the redoing of cables and was getting the exact same noise as when I would bring up the output stage as I raised the voltage. Same drifting from right to left channel buzz to both, same tone. After some digging around, I had forgotten to reconnect the PSU star ground to the chassis. So everything was floating. Once I connected it back, dead silence again. And that got me thinking about the HV900's/Output floating and doing the same thing.

 

So what I thought was, is there any way to do a ground plane with the HV900's/Output with ultra thick boards and/or really wide spacing around HV rails? Because I think that would probably solve the final noise problem I'm having.

[kevin gilmore]

I can certainly do a version with a ground plane.

there is no ground on the 900v board, so it would be the entire top side and whatever clearance we decide. say .2 inches

here it is without the keepout for the caps and .15 inch clearance

 

Edited January 14, 2018 by kevin gilmore

[Tinkerer]

If I had a wishlist on a HV900 V2, it would probably be:

Each still split into a separate board for four total (mostly so I don't have to make a new set of brackets or drill new holes in the heatsink and easier to service if one blows)

Transformer inputs on the front like you posted instead of the side like in the first run boards (shorter transformer secondary leads and easier to get to)

Integrated current limiter (would help shorten leads for whole output stage)

Ground plane (I know it's not a ground part of the circuit but wouldn't it still need a terminal to ground?)

Change the 4.42K Xicon to 4.32K Dale resistor so it's closer to 900VDC even instead of a little higher.

But you know better than me. I can just put stuff together and troubleshoot a little bit, read specs on parts or whatever. Design itself is not something I really understand.

[kevin gilmore]

which current limit circuit?

on the singles board, to put the ac input on the front requires a size change

[Tinkerer]

The normally open power mosfet mounted on the +900 out with a resistor picked for gate voltage to hit at a little under 50mA current to the output board. Just a second safety pretty much. IXTH6N100D2 I think.

 

On the mounting front for the singles, longer is perfectly fine for me, it's just wider I can't do. My side sinks are three separate 6" wide aluminum blocks stacked side to side and the current HV900 singles take up pretty much all of one 6" block per mount horizontally.

 

 

[JoaMat]

Isn’t that an earlier PSU version with the long zener string?

[kevin gilmore]

yes I think you are right.

Edited January 15, 2018 by kevin gilmore

[JoaMat]

Thanks.

This is a very long shot but what happens to the hum if you swap one of the AC from transformer to rectifier diodes (HV900). Will there be a change in hum if the AC to the two “rectifier bridges” are in 0 degree or 180 degree phase?

 

[kevin gilmore]

here is the new single, a little longer same width

 

 

 

circlotronhvpowersinglenewver3swt.zip

Edited January 15, 2018 by kevin gilmore

[Kerry]

Looks really nice.  I think the IXTH6N100D2 is a to247 package.

I also see you moved the diode and 10R resistor near the output.  Any reason?

 

 

[kevin gilmore]

will fix the package in a minute.

moved the resistor and diode to limit turn on charging current on the caps

new picture and board file

someone please check for errors

Edited January 15, 2018 by kevin gilmore

[Tinkerer]

Dumb question but what's that fifth thing on the bottom side that's not the STN0214 or STN9360?

[kevin gilmore]

Cp1117 optional part for doing controlled power up. For example tube circlotron. If you don't want it, do not populate the part.

[Tinkerer]

7 hours ago, JoaMat said:

Thanks.

This is a very long shot but what happens to the hum if you swap one of the AC from transformer to rectifier diodes (HV900). Will there be a change in hum if the AC to the two “rectifier bridges” are in 0 degree or 180 degree phase?

 

I remember I tried that at the primary end once at it didn't change noise behavior in either channel, but not sure how I'd do it at the secondary. With 4 secondary pairs in a channel set and no oscilloscope, I'd probably need to go through all 16 combinations and hope one cancels like how a balanced cable works. Is there a smarter way to go about it?

[JoaMat]

After a second thought I’m beginning to believe that it shouldn’t make any difference at all. After the diodes the “output” should be the same after  swapping the leads from transformer, or?

[Tinkerer]

Went ahead and did one last thing as a check. Moved one of the HV900's directly under an output board with another flipped directly next to it for minimum lead length and had all leads HV900 shortened to <5 inches. Got the exact same sound. No louder. No less.

The new HV900 boards look good to me. Parts match what they should. I didn't go through every trace though. If they look clear, anybody else interested in having them made? I'll chip in for two sets to start it along. Somebody else would probably need to run the buy though, since my experience with PCB houses at this spec is nill.