CAPACITORS IN SERIES
Subject: CAPACITORS IN SERIES
From: richard.quick-at-slug-dot-org (Richard Quick)
Date: Tue, 31 Jan 1995 03:55:00 GMT
>Received: from uustar.starnet-dot-net by csn-dot-net with SMTP id AA11835 (5.65c/IDA-1.4.4 for <tesla-at-grendel.objinc-dot-com>); Tue, 31 Jan 1995 00:28:26 -0700
* Carbons Sent to: nvv-at-mlb.semi.harris-dot-com
> From: nvv-at-mlb.semi.harris-dot-com (Nick van Vonno)
> To: tesla-at-grendel.objinc-dot-com
> Subject: Capacitors in Series
> Cc: nvv-at-mlb.semi.harris-dot-com
NV> As soon as someone admits to being foggy on series-connected
NV> capacitors an automatic caution flag needs to come up. Anytime
NV> caps go in series, and especially if you're about to throw a
NV> 'megavolt' across it
You are partly correct, but you are stretching reality a bit. The
megavolts are on the discharge terminal, not in the tank or supply
circuits. In reality we are talking tank circuit voltages that
rarely exceed 20,000 volts.
NV> a set of parallel resistors needs to be provided. This equalises
NV> the voltage across each capacitor, and consists simply of a
NV> resistor across each cap of the series combination...looks
NV> like a ladder network. Unless this is done the cap string will
NV> not act as a voltage divider, and one or more caps may end up
NV> with a voltage across them that exceeds their working voltage
NV> rating. At least this is the practice with capacitor banks that
NV> are expected to have dc voltages across them. If run off ac only
NV> the caps will voltage divide, but the resistors still sound like
NV> a good precaution to me.
Sure, why not? First I advised that these caps (Sprague .004 -at- 30kVDC
barium titanate dielectric) not be used in a pulsing tank circuit,
singly or in series. They are inefficient. Quoting myself below:
RQ > These are really very poor tank circuit capacitors. They get hot,
RQ > they are low Q, and they fail under pulse excitation. They are the
RQ > best bypassing capacitors I have ever used when placed in series.
If you were to place several low voltage rated capacitors in series and
expect them to pulse, then you are correct: a megaohm resistor should
be wired across the capacitors to equalize voltage. This will prevent
flashovers and dielectric failures. However, due to the high losses in
capacitor Qs with this setup, I would advocate that the proper AC volt-
age rated capacitor be used before I would advise wiring up a "ladder"
rig such as you advocate. The key here is what's efficient.
Second; when used in series as bypass capacitors I have only seen two
failures. Both were in high powered Tesla Magnifiers using the "sen-
sitive" balanced tank circuit. The kickback fired into the safety gap
with the force of a professional pyrotechnic salute (5 kVA input), and
when the smoke cleared the bypass capacitors were split in two. No
resistor was going to save them, but they did die protecting the main
system caps and power supply, which survived.
Last; the tank circuit frequencies we are discussing are much different
from a static DC model. Most of the time we are dealing with freqs
between 60 - 600 kc. At these frequencies the capacitor voltage pretty
much divides without a problem unless you are really pushing them.
... If all else fails... Throw another megavolt across it!
___ Blue Wave/QWK v2.12