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

Re: Any Geekgroup MMC failures out there??



Original poster: Bart Anderson <classi6-at-classictesla-dot-com> 

Hi Terry,

Well, I've had one cap bite the bullet. This happened about a year ago. I 
was the 1st to receive caps from The Geek Group many moons ago. These are 
the 942C series 0.15uF caps. The one that that died was the cap next to one 
of the brass bolts I used for a connection. There's a hole in the cap about 
1/8" from the end. I don't believe it an arc from the bolt to the cap 
occurred since the hole is on the opposite side of the cap, however, 
something odd occurred. It was "not" from over volting the cap but I 
believe RMS current. I believe I was running 5 strings of 12 for 0.0625uF 
total at the time. The caps were set on a pcb with 10M resistor soldered 
across the cap on the opposite side of the pcb. Each cap from side to side 
had at least 1/2" clearance. Plenty of room there. I suspect I exceeded the 
rms current capability (25 amps).

I've since rebuilt the entire cap bank onto an 8" pvc tube. Each resistor 
is now soldered across the cap and each cap is simply placed on screws 
protruding from the pvc tube (this way I can maneuver cap arrangements 
anyway I like). It is currently 3 strings of 18 caps for a total C of 
0.025uF. I haven't had a failure since.

I'm not sure why the cap failed. It's still on the pcb board along with 5 
neighboring caps. I just threw it up in a cabinet and left it there ever since.

Take care,
Bart

Tesla list wrote:

>Original poster: Terry Fritz <teslalist-at-twfpowerelectronics-dot-com>
>
>Hi,
>
>I was just wonder if anyone had any failures of Geek group caps for "no 
>good reason"?  I have not heard of any problems, but it seems like a good 
>time to ping the users out there to see if any problems have been 
>noted.  Sort of just checking to see if there are any holes or unknown 
>problems we need to check into.  Of course, if one does not follow the 
>tips or runs them out of what is recommended, that is your problem 
>;-))  But if you think you did everything right and they blew, I would 
>like to hear about it to see if any common trends or problems are out 
>there we don't know about that need fixing
>
>Perhaps additions to the "tips" too.
>
>
>Cheers,
>
>         Terry
>
>
>At 12:15 AM 1/17/2004, you wrote:
>
>>Well, let's put it this way....  We have never heard of a failure of a 
>>"Geek Cap" that wasn't directly related to poor design or 
>>construction....  The design chart on our mmc page is tried and true.  If 
>>you build an MMC using the suggested design on our site and follow our 
>>construction guidelines, your MMC should last forever in hobbyist duty.
>>
>>If you really do feel uncomfortable using an "underated" capacitor, then 
>>go ahead and make it more beefy. :)  But, IMO, if "our" caps can hold off 
>>5000V, one could argue that 10 in series is plenty overdesigned.  IIR, a 
>>second high voltage test with a newer cap held off 5300V at the first 
>>breakdown, then more than 6000V at the second (that's a 2.7J bang inside 
>>the cap!).  Then the current draw got pretty high.  You have that info 
>>handy, Terry, or do I have to go digging?  (one of these days we'll just 
>>add that piece of information to our site)
>>
>>http://www.thegeekgroup-dot-org/mmc/
>>
>>FWIW, my first MMC was made from 1600V Panasonics (I think I managed to 
>>hit Terry's first bulk buy in early 2000), total voltage rating of 
>>19.2kV.  After a few hours of runtime, one dead NST, one dead "Terry 
>>Filter" (shorted the MOVs), a lot of toasted AOL CDs, and a lot of awed 
>>spectators, the MMC caps are still fine.  And IMO I've been gentle on 
>>those compared to some of the torture we've put our Geek Group MMCs 
>>through!  The Geek Group MMCs have seen many coil setups and many power 
>>levels, including running near-resonant value caps with a 14.4kV PDT 
>>overvolted to ~280V input on a 360BPS SRSG coil (half the Gemini project) 
>>and also running a static gap 12/180 setup with no topload (we tuned at 
>>about 2 turns - the gap was much less happy than the MMC).  So all this 
>>is just saying that, yeah, these are pretty tough little buggers.
>>
>>Cheers
>>
>>Mark Broker
>>Chief Engineer, The Geek Group
>
>
>
>
>