Re: The death of a classic - Second look

["Tesla list" <tesla-at-pupman-dot-com>]

Original poster: "Finn Hammer by way of Terry Fritz <twftesla-at-qwest-dot-net>" <f-h-at-c.dk>

Terry, all1

Thankyou, Terry for digging into this.
I am sorry, I don`t have anymore of these slightly special caps. But I
have a few of the 56nF/1600VDC types that are acting as "stand in`s"
right now.
I wonder whether they would be of any use?
The 22nF types that you checked were special orders with a smaller
footprint.

I`ve been thinking a lot about this cap failure, and come to the
conclusion, that I just grew plain careless, thinking that the cap could
take anything.

At the museum, we have decided to install a runtime monitor, and I am
starting to think that a new, slightly more conservative cap should be
made, still MMC.

I have 900 pcs. 68nF/1600VDC caps in order to arrive mid september. They
are former Phillips, now BC-Components, MMKP 376 types, they should be
identical in internal construction and materials.

Arranged in strings with 30 caps, as opposed to 20, and 15 strings would
result in the same 34nF, and would have to be bullet proof.

And still would provide me with 2 caps at 1/3 the price of one new
Maxwell cap.

On second thought: I still think MMC is the way to go, and I am willing
to give it another try, even in the museum world.

Cheers, Finn Hammer

Tesla list wrote:
> 
> Original poster: "Terry Fritz" <twftesla-at-qwest-dot-net>
> 
> Hi All,
> 
> Under the scope, each of the blown areas had a little hole in the center
> were it shot through.  So it does appear to be all over voltage related.  I
> saw no evidence for #3 below.  The metalization in non-arced areas was
> perfectly normal.  There is some corrosion inside the caps, but I think it
> is due to all of that polypropylene being burned when the arcs occurred.
> The burnt poly seems to have an oily residue that attacks metal a bit.
> This is a secondary thing that would not be an issue if the caps had not
> arced so much.
> 
> These caps do show ionization damage to the dielectrics.  I can't rule that
> out as a cause of the problem.  The edges of the plates have damage lines.
> It is possible that long term ionization damaged the poly to where it
> started arcing through.  It is also possible it is a chemical reaction of
> normal ionization and the chemicals from the burnt poly.
> 
> So we have two possibilities:
> 
> 1.  The caps were run over voltage which blew thousands of holes in the
> dielectric and released chemicals inside the cap.  As time went on, these
> chemicals and normal ionization continued to degrade the cap abnormally
> even though the voltage was no longer high.  There is a chance that if it
> were not for the corrosive chemicals, the cap would have survived a normal
> life.
> 
> 2.  Ionization at the edge of the plates cut away at the dielectric
> eventually causing a large number of punctures.  Then, the above chemical /
> ionization process eventually destroyed the caps catastrophically.
> 
> So, either the high voltage damage in the past finally caught up with it,
> or long term (10 hours) ionization did.  Over voltage damage in the past is
> just one of those things...  However, if it is ionization damage, that
> would be a real concern.  There is not much I can do to prove or disprove
> either with Finn's caps.  One would need long term run caps that were not
> subjected to over voltage breakdown to rule that out as a cause.
> 
> I am going to send the caps on to Scott so he can look at them too.  It
> would take a relatively simple experiment of running a few caps at
> 1400VACrms (2000Vpeak) for 10 hours and checking the dielectrics afterwards
> to see how badly they were damaged.  Just an NST and variac could do that
> but you would have to be careful not to resonate the NST/cap combination.
> I have a lot going on these days ;-) but sounds like an experiment that
> needs to be done.  It would be nice to use 'new' caps like Finn's MMC used
> since ionization damage could differ significantly from manufacturer to
> manufacturer.  I don't know if Finn has like 5 unused caps left?  But I
> have Panasonics here too.
> 
> Cheers,
> 
>         Terry
> 
> ================
> Hi Finn and All,
> 
> Today I got the failed strings from Finn's cap.  It sure does smell!  I
> love the smell of burning electronics :o)))
> 
> I tested a number of the resistors and they were all perfectly in spec.
> They showed no signs of stress.  The resistor I was concerned with before
> was in the middle of a pretty bad fire.  The fire damage to the boards is
> very profound.  Seeing the boards first hand and testing the resistors, I
> can find no evidence the resistors caused the failure.  They seemed to be
> perfectly good and adequate.
> 
> I disassembled a number of caps that were not involved in the fire or were
> in the "good" string.  The is no evidence of end plate or current heating
> damage.  They appeared to run cool without any thermal stress.
> 
> However.........  All of the cap's internal plates and dielectrics are
> "machine gunned".
> 
> http://hot-streamer-dot-com/temp/finn-001.jpg
> 
> There are literally thousands of holes blown in the dielectric layers that
> have self healed.  Some areas are very bad.  Considering the large number
> of caps in Finn's MMC, catastrophic failure was almost certain.  One cap
> had a section blown out of the case but did not burn up.  It was a disaster
> that never happened..., but it tells a story of serious problems...
> 
> I will have to get to a high power scope tomorrow, but there are three
> possibilities:
> 
> 1.  Over voltage on the cap caused millions of dielectric breakdowns.  A
> "few" did not self heal and ignited the caps
> 
> 2.  Ion damage to the dielectrics caused weak spots that eventually failed.
> 
> 3.  High currents in the internal metalized plate degraded and failed the
> thin metal center internal layer.
> 
> Finn stated a very possible clue:
> 
> "Gap voltage is 32 kV (most of the time, but I am sure I`ve hit it with up
> to 45kV for quite some time as well)"
> 
> 45kV is 40% over the caps rating.  "I" would only go to 25% on EMMCs.  If a
> few caps in a string breakdown and go to zero volts, it can put even more
> voltage stress on the others in the string causing further breakdowns.  I
> note that some areas have slight corrosion suggesting that the damage is
> older than a few days.
> 
> When I can see the areas under high magnification tomorrow, I can determine
> which it was.  But it sure "looks" like the first possibility.
> 
> BTW - No doubt about MMC's ability to self heal now!!!  Even though it
> "eventually" failed.  Finn's MMC easily has millions of self healing events!
> 
> Unfortunately, the rest of Finn's cap is certainly heavily damaged too.
> Time to take it out of service and put it in a trophy case so It can be
> admired for generations to come :-))  Maybe if you explained to the museum
> that it was one of the first of it's kind, they would accept it as a
> display item.  100 years from now, it will be a big piece of Tesla coiling
> history!
> 
> Cheers,
> 
>         Terry