Re: [CAD_CAM_EDM_DRO] Making AC to DC??

jeffdavis516 wrote:

> I am wondering if there is any information out on the web or maybe
> from someone here about the hows and rules in making DC power
> supplies?
>
> I guess my question is if I know what DC volts I want and I know what
> Amps I want, how would I build a power supply to do this?
>
> Is there some formula that tells you the size of transformer, Bridge,
> cap, and what ever else you need?

For DC voltage, you want a transformer output that is .707 of the desired DC
output. So, for a 100 VDC output, you want a transformer with 70.7 V RMS.
That is for use with a bridge rectifier. If you use a center-tapped winding
and 2 diodes, you'd need 70.7 V form center to each end, so a 141.4 V
center-tapped winding would be the right one.

For current rating, you need the full DC current plus a margin for the
non-sinusoidal current draw of a capacitor-input filter of about 50%,
with a bridge rectifer. So, for 10 A output, a transformer rated at 15 A
output should work fine. If the transformer was designed for capacitor-
input filters, then the margin has already been included. If the rating is
RMS or rectifier-duty, then you need to add the margin.

For a center-tapped supply, you only have current in the winding for
half a cycle, so you can live with less margin. But, that would be a
141 V ~10-12 A winding, instead of a 71 V 15 A winding. The 141 V
transformer would probably be made bigger, as it has a larger VA
rating for RMS.

Capacitor sizing is VERY complicated, you have to deal with ripple
voltage, ripple current, dropout ride-through capacity (short power
outages) transformer resistance, etc. But, in most cases, due to
short power interruptions, you want several times the minimum
calculated value.

With full-wave rectification, the capacitor is charged every 8.3 mS
(10 mS in 50 Hz countires). At one amp, a one farad capacitor droops
one volt per second. So, droop = I (Amps) * T (Seconds) / C (Farads)
But, don't use the 8.3 mS figure, as that is best case. I'd want at least
1/10 second ride-through. So, figure for a 10% - 25% droop in .1 sec
for a good safety margin.

Jon