Peter,
With looking through the books texts I have here, That is the way you
design this filter using a swinging choke. With mine, I needed 400 Vdc @
80mA maximum. So Lcrit = E / Imax which is 400 / 80 = 5 Henrys. Now the
choke is a 5-20 henry swinging choke. So to get the minimum current for
the bleeder resistor to have the 20 Henry portion we take 400 / 20mA =
20 Henrys or 20H x 20mA = 400V. So for 20 mA at 500 Volts, our bleeder
resistor would be 400 / .02 = 20K ohms. Then to size the wattage, 400 x
.02 = 8 watts. A 10 or 15 watt resistor would work here. As long as
Lcrit is met, the swinging chokes inductance will swing and level out
with the DC load without saturating. If Lcrit is below this value, it
will try to run as a capacitive input filter. Actually, I will be adding
a page on filters to the transformer tutorial when I get to feeling like
it as promised. I have some done, but nothing ready yet to publish it.
Last, Tony, I0JX, has a page on this at:
http://www.qsl.net/i0jx/supply.html
Will Matney
peter.chadwick@Zarlink.Com wrote:
Is actual resonance what you want? The argument I've seen runs as follows;
In a choke input supply, there's a critical minimum value of inductance for
regulation. At low currents, this is very high, but if the choke is tuned
slightly HF of the ripple frequency, it looks like a very much bigger
inductance. At resonance, it will look like a resistor, of course. As the
load current increases, the critical value of inductance decreases, and the
choke will start to lose inductance anyway.
From this, the choke needs to be tuned a tad HF at minimum load. This
certainly ahs worked for me. Incidentally, by using negative lead
filtering, you can rectify the ripple across the choke for a low current
negative bias supply.....
I tend to go for negative lead filtering anyway to ease the choke
insulation requirements.
73
Peter G3RZP
__________ NOD32 1.817 (20040719) Information __________
This message was checked by NOD32 antivirus system.
http://www.nod32.com
_______________________________________________
Amps mailing list
Amps@contesting.com
http://lists.contesting.com/mailman/listinfo/amps
|