[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: DRSSTC Noise source found



Original poster: Terry Fritz <teslalist@xxxxxxxxxxxxxxxxxxxxxxx>

Hi,

I am overnighting two 10nF 500V silver ruby/mica caps from DigiKey (338-1078-ND). Two of these expensive puppies will shunt that 25MHznoise with a solid 1/3 ohm right up into the GHz region >:-))) I am also getting the same in plain ol' cheap ceramic too. If ceramic would work, that saves $30 ;-) But they should squish that noise like an ant!!!

In other news:

I am staying with the 5.1 ohms gate drive resistance. Adding resistance would only slow the turn off which would not help anything. Too much resistance (17 ohms) might expose the IGBTs to dynamic shoot through... The 5.1 ohms might prevent the IGBT from going bonkers without hurting speed so it seems just right.

After the noise is knocked out, I will squeeze the dead time down. I got a bunch of small caps (10-100pF) to play with that. I see no reason to have excessive dead time. Might as well cut it to the minimum.

I added three 100pF caps to the protection card to combat noise. Right now the noise is causing the over current to cut in way too soon. The caps don't "fix it", but they seem like a "good idea". If the noise fix above works like it should, that will fix it at the source.

I stuck LF ferrite beads on the temp sensor and CT leads (DK# - 240-2136-ND). It seems like a "good idea" but might not really be needed if the source of the noise is removed. Also considered a grounded shield between the output leads and the CTs to prevent the 25MHz noise jumping over (capacitivily). But all that is on hold till I see what the mica caps do to the noise. If that fixes the noise well, then all this other stuff is not needed...

Cheers,

        Terry


At 07:40 PM 2/16/2005, Terry Fritz wrote:
Hi All,

I found one giant source of noise on my DRSSTC. The two big poly capacitors (10uF, 600V) right across the H-bridge have too high of impedance and high frequencies. Polypropylene is super good at say 400kHz but at say 25MHz they have very high resistance and loose their ability to stop high frequency noise.

Here is the H-bridge buss voltage during a switch:

http://drsstc.com/~terrell/pictures/BussNoise-01.gif

The big caps are great at low frequency but at high frequency...:

http://drsstc.com/~terrell/pictures/BussNoise-03.gif

I have a 25MHz glitch that goes from 240V to 460V with 50MHz harmonics!!!

If I just travel back to the main filter card, that noise goes down to 300V to 390V (if I ground the differential probe leads to the same buss, I read zero, so these voltages appear to be very real!):

http://drsstc.com/~terrell/pictures/BussNoise-02.gif

So I need to parallel some say 600V caps in there that can work well in the 50MHz range. If I can clamp this noise out, it will have give a giant reduction in noise!! My gate drive cards are driven off this noisy buss but they appear to be able to filter it. If they didn't, I would "know about it" by now ;-))

Off to search for big HF HV caps...

Cheers,

Terry