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

Re: Rectifier's reverse current (was Re: NST->DC Re: CGA Monitor Flyback ???)




-----
>
>
>Original Poster: "Jim Lux" <jimlux-at-jpl.nasa.gov>
>
>>
>>No problem.. I do it all the time for capacitor charging, etc.  Just watch
>>out if you make a bridge rectifier charging a capacitor that you are aware
>>of the path for reverse current which will destroy the rectifiers.
>
>Jim,
>
>1. what "reverse current" did you mean?
>2. Is it the one at bang time?

That's the time.  When the cap discharges, you get a voltage reversal, and
the bridge rectifier is forward biased for the reverse voltage (on the
output of the bridge).  I destroyed several microwave oven diodes doing this
before I realized that it wasn't the transformer dying, nor the cap failing.
Only a few of the junctions in the stacks fail at a time, so the effect is
that the max charge voltage starts decreasing (because the PIV is steadily
getting less).

>3. Have you got "special" solutions for protecting the diode bridge?

A resistor in series with the charging circuit to limit the current works
nicely.  A few hundred ohms is what I use. I figure the rectifiers can take
an impulse of 100Amps as long as its not repetitive.
I use the same resistor as a safety dump, by shorting the output of the
bridge. Even if the NST is still on, the short keeps the DC voltage to zero.


>
>At bang time, when the capacitor voltage reverses, a considerable amount of
>current will flow ALSO through the tank cables and the diode bridge: at
that
>time, the TC primary winding is literally feeding the tank diode bridge
(!). I
>thought the only solution is to add a series inductance to the cables
>(which is,
>by the way, the usual way of doing).

The series inductor do also... In my application, I didn't care about the
loss from the few hundred ohms (charging at 100 mA, the voltage drop is only
a few tens of volts)..

I'm always leery of inductive pulse current limiting, because circuit
parasitics can make resonant circuits where you don't expect them.
Resistors, although lossy, can't resonate.

I suppose, without thinking about it too much, that another approach is to
put a capacitor or an RC snubber in to suck up the transient.