[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [TCML] BPS Testing
Lau, Gary wrote:
Well, something is going on. It might be ferro resonance although I must
admit that I haven't thought about your situation long enough to say it
is the only explanation (but who knows, maybe I'll end up there as
well). I'm curious, have you measured this Jefferson's open/short
circuit testing on both the low and high voltage sides? That would
probably tell the tale. Even some DCR winding measurements would be
interesting. Who knows, maybe you've got a mismarked NST? Wouldn't that
be cool! No chance, I know. The only reason that thought enters my mind
is because it appears to be just about double what I would expect. Any
other NST's lying around?
When I was using a static gap, I ALWAYS ran with 140VAC. If the knob goes up to 11, why stop at 10?
Beyond that, I seem to have been blessed with a 15/60 that's endowed with super-NST powers. I think it may have been bit by a pig that was exposed to ionizing radiation... Although the shunts are unmodified, I have scoped the NST secondary current, and clocked it (with a digital scope) at over 200mA RMS (not peak, not .707* peak, not a typo). Although the short circuit secondary current is still in fact 60mA, when the input voltage is maxed out and when using LTR caps, ferroresonance is the only explanation for this behavior. I *LIKE* it! This must be why I'm able to use larger than typical cap values (.02uF w/ static, .04uF w/ SRSG).
The mains current, even with PFC caps, was close to 20 Amps.
Another case for a 120mA NST.
I've been wondering about that but I'm not (yet) convinced current is
the only way that can happen. Those bangs are evenly spaced (bang and
then another soon after), but there are a couple instances where they
occurred right next to each other. I don't see anyway the transformer
could have recharged again in the time frame allotted.
When ferroresonance occurs, the NST charging current can far exceed the normal, faceplate rating. Even with the BPS testing that you have just performed, you noted that occasionally you see one bang following another in quick succession. The only way for that to occur (assuming that the gap breakdown voltage is reasonably consistent) is if suddenly, the NST is pumping out a lot more current.
Something else I should point out which I found interesting. If I count
the number of quick bangs that occur after the normally timed bang, my
bps drops to 70 bps (as calculated). In that first set of data for the
15/60 test, I measured 82 bps. There were 12 quick bangs counted
(82-12=70). It may be that the gap simply stopped and restarted for
"Lord knows what reason". But it could also be ferro resonance was the
cause (which occurred 12 times). Well, when I test with a variac this
weekend at 140Vac, then I should see far more of these occurrences. I'm
expecting 82 bps by calculation. I expect to measure near 95 bps. But if
it's much higher, then it will make some interesting theory.
I believe anything can be modeled if we know the variables. Example, my
12/60 NST has 1/2 shunts removed (Franceformer). In order to input this
NST correctly into the model, I had to perform voltage and current tests
on the NST to get anything even near reality (however, my voltage
divider used I'm not thrilled with and I may be off a little. I've also
measured coupling on the NST in nearest 10V increments. Interesting how
k drops off after 80V input. This was before shunt removal, so is
probably good for unmodified NST's (the basic curve).
If the cap is charging but fails to achieve the gap breakdown voltage Vgap before the charging reverses, that energy in the cap is not lost, but must go somewhere as the cap voltage "charges" towards zero, and then on to -Vgap. That unspent energy manifests itself as (.5 I**2 L) NST secondary current, and I'm guessing that any additional charging current must be on top of that, the sum saturating the current shunts. That's my theory. And unfortunately, such behavior is not consistent from NST to NST or bang to bang, and is impossible to predict or model.
Thanks for your thoughts Gary.
Tesla mailing list