Re: [CAD_CAM_EDM_DRO] Re: Re: Re: Power supply question.....

JanRwl@... wrote:

> In a message dated 03-May-02 12:19:45 Central Daylight Time,
> elson@... writes:
>
> > There is a couple of volts of ripple on the capacitor,
> > depending on how much droop occurs due to the load draining the
> > capacitor between charging pulses from the transformer. If you put 34 V
> > peak on a sizable capacitor, it will explode within seconds. At no load,
> > there will be almost no AC on the capacitor. At full load, you would not
> > want to see much more than 1 V droop (at 120 Hz) under normal conditions.
> > If you use too small a capacitor, the droop will be larger, and the AC
> > ripple
> > current may cause the capacitor to run hot and fail early. Another reason
> > to use a large capacitor is to ride through short power dips.
>
> Jon: Maybe it's just Texas, but down here, we think DIFFERENTLY about these
> details than you seem to. What school of electrical engineering says that
> capacitors "run hot"? I was always taught that a capacitance neither "uses"
> nor "supplies" energy! Ah, well . . . Perhaps things have changed since
> I was a student.

Yes, this is the problem with EE education. They focus on ideal components,
which you can draw on a blackboard, but you cannot buy! All capacitors have
measurable series resistance. At the values and currents of large electrolytics,
a few tenths of an Ohm series resistance will drop a lot of real power in the
capacitor. When you have several amps of AC ripple current, you get I^2 * R
tunring into significant heating. For an example, lets say 5 A ripple (squared) x
.2 Ohm = 5 W. In a small cap, that will cause a good deal of heating.
Some new high-ripple electrolytics even have a hole in them so a heat sink
bar can be shoved into them to draw out the heat. (This is for use in switching
supplies, where the ripple frequencies are in the10s or 100s of KHz.)

So, this is NOT new. In many small-signal circuits, the non-ideal nature
of basic components can be ignored, but in power circuits, they become large
enough to be a constant concern. They can be managed, but you always
have to evaluate the series resistance of capacitors and inductors, for
instance.

Jon