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

[kevin gilmore]

I doubt those resistors are rated for voltage, so don't mount them to the chassis, or try and find a voltage rating

measure the high voltage transformer for leakage between the windings for the 900V

[Tinkerer]

Datasheet says limiting element voltage is about 1900 VDC.

Edited February 12, 2017 by Tinkerer

[kevin gilmore]

they are wirewound, not sure how much inductance that is going to add, or whether its going to cause any trouble.

I would probably use these

http://www.vishay.com/docs/50051/lto100.pdf

[sorenb]

15 minutes ago, kevin gilmore said:

they are wirewound, not sure how much inductance that is going to add, or whether its going to cause any trouble.

I would probably use these

http://www.vishay.com/docs/50051/lto100.pdf

 two of those in series to get to 1kV for the 120ohms ...well only 100ohms are available ...so four it is

 

Edited February 12, 2017 by sorenb

[Tinkerer]

Can you put a 100 and 2 10's in series for the exact value? Also, mouser only gives it a 375V rating instead of the 500V in the data sheet so I'm not sure what's up with that.

[kevin gilmore]

I was thinking that you would not need heatsinks. Kind of expensive to use 3 in series

might be easier to use 2 of the smaller ones in parallel depending on available values

i had to use 2 in series for the esx

[Tinkerer]

What about a pair of these in series? They're pretty cheap and close enough to be within tolerance

http://www.digikey.com/products/en?keywords=AP10162RJ

 

Edited February 13, 2017 by Tinkerer

[kevin gilmore]

I like those

[JoaMat]

Just switching to 100 watts resistors will that protect us from exploding output board and associated PSUs?

[sorenb]

On 2/13/2017 at 2:24 AM, JoaMat said:

Just switching to 100 watts resistors will that protect us from exploding output board and associated PSUs?

I would assume that having some more robust 120ohm's will easen things quite a bit, as it will take a considerably amount of time before a pair of those break down, and at that time the current limiter from the HV900 should have kicked in and the situation being under control until the driver output has settled

 

 

Edited February 14, 2017 by sorenb

[Tinkerer]

You mentioned the 12V diodes on the output should be probably be uprated as well, right? And 22AWG 2KV wire should be fine to connect the new 100W resistors from the sink to the board?

[sorenb]

as far as I remember I use 12V+5V back2back 1W's ...probably not critical 
22awg is probably plenty

For testing you can:
Power up: bring up the drivers - wait until output settles - bring up the output stage
Power down: bring down the output stage - wait until the HV900 < 200V - power off the drivers
That works for me.

Keeping the HV900 < 200V you can investigate the troubling behavior, having meters on the current limiter resistors of output stage and HV900s

As far as I understand from Kevin he hasn't had problems in the lab, but he used some film on ceramic resistors at 10W's for current limiter.
The DigiKey you found  is 100W/700V rated and will most likely not break down.

[Tinkerer]

I've been working on checking the blown parts in the HV900's slowly. Some parts good I didn't expect and some bad I didn't, like all the voltage reference chips. Did a bunch of alterations on the amp itself for quick connect transformer leads for testing and two stage startup as well from Soren's recommendations.

 

But my question was about the new power resistors for the output boards. They don't fit on the existing bracket and will need a good few inches of wire to reach the pretapped M3 holes on the main sink. That adds up to about .25 microhenry of inductance for each pair. How much does that matter?

[sorenb]

Finally received Front/Back plate for the 5U Dissipante - almost took forever.
Decided to have a groove dividing the front plate into an upper 2U and a lower 3U, to have it look less bulky; it's still one 5U piece.
The "pi" sort of symbols the topology

To be continued ...

Edited April 13, 2017 by sorenb

[JoaMat]

More sort guld!  Indeed an elegant front plate.

[Tinkerer]

Nice. I don't see any holes in the back plate for transformer mounts. Were you able to fit everything inside without resorting to that?

 

 

[JimL]

Cool, but...a normal bias output?  Isn't that a little like killing a fly with a small nuclear device?    Not that there's anything wrong with that.

[sorenb]

5 hours ago, Tinkerer said:

Nice. I don't see any holes in the back plate for transformer mounts. Were you able to fit everything inside without resorting to that?

everything but output boards are mounted on bottom plate

[johnwmclean]

Very sleek, I like how you’ve divided the front panel.

[Kerry]

Very nice.  I like your design

[wink]

Shoulda used Euler's equation instead of boring Pi....   

 

Edited March 15, 2017 by wink

[mwl168]

Very elegant

I don't see a power-on indicator on the front plate?

[wink]

... should get the pi sign to light up....   

[chinsettawong]

The front and back plates are beautiful.  Congrats!

[Tinkerer]

Just a heads up that Mouser just got some of these in a couple days ago. http://www.mouser.com/search/ProductDetail.aspx?r=588-AP10162RJ Bit cheaper than DigiKey at the moment.

 

FInally got all my stuff sorted and a big parts order to get the PSU's and output boards up and running. As far as I could tell, the things that blew were consistently.

 

12V diodes on the output boards all the time (some fused open is why I didn't realize this earlier)

The diode side of the 4N25 half the time

Big Cree devices on the output boards. Survived on the PSU's.

Both 120 ohm current resistors on the Output boards always blew, the right less obviously as the damage was usually underneath.

On 900V PSU's, 10M90S behind the DN254 every time. Other 10M90S and DN254 survived every time.

SMD stuff, STN0214 always died, STN9360 survived rarely

Voltage references always died without exception. Not fun on the wallet.

BC557's died every time without exception

 

Might save someone a bit of time if something goes wrong. Like you don't have to go hunting in the rectifier bridges or anything.