Re: [CAD_CAM_EDM_DRO] Re: What size capacitor should I use?

caudlet wrote:

>There are two reasons to parallel caps (well actually three). Cost.
>In our world of scrounged parts it's ususally easy to find caps at
>the voltage required at smaller values. When you start looking for a
>35,000 to 80,000 mfd cap with a 100 volt rating they get more scarce
>($$$$). Next is the physical size. Smaller value caps are typically
>shorter and have less cross section per MFD so three 15,000 100VDC
>caps will probably end up being somewhat smaller in total. The last
>thing is ESR (Effective Series Resistance). All caps have
>capacitance (duh), resistance and inductance. The larger the cap the
>more ESR it has. Take 3 smaller caps in parallel each have smaller
>ESR each and in parallel they have 1/3 the ESR of one. ESR in linear
>DC power supplies is of little concern since it would limit the
>ability to filter higher frequency waveforms but are used extensively
>in switchers as filters. Besides, I wanted to give everyone
>something else to stew about (:-)
>
>

Hmmm. Smaller caps shouldn't end up being smaller overall than a single
large one. Capacitance is proportional to the area of dielectric,
divided by the distance between conducting plates. The breakdown
voltage is inversely proportional to the separation distance (thinner
dielectric will break down at lower voltages) Given a particular
dielectric material, you need some thickness of dielectric to get the
necessary voltage rating, and some area to get the necessary
capacitance, so you ned the same amount of dielectric no matter how you
slice it. (there are edge effects and packaging questions that might
change this a couple of percent one way or the other)

As for filtering, it's pretty common to see two or more different types
of capacitor in parallel, to filter different frequencies. A small
value tantalum cap alongside a large value electrolytic will filter out
high frequency noise, while the electrolytic provides low frequency
filtering.

>Oversizing the power supply for the load offers no tangable benefits
>other than making the builder feel better. I think even the 66%
>number is conservative. I find in the actual world the total worst
>case current draw (all motors under normal load) is closer to 50% of
>the nameplate values on a chopper system.
>
>

Oversizing the supply is one thing, using higher rated caps is another.
There are several specs that aren't usually discussed: temperature,
lifespan, ripple current, and tolerance. Most electrolytic caps are
rated for a capacitance, maximum voltage, maximum temperature, and
lifespan at max volts/current. "Long-Life" caps are rated for 1000 to
5000 hours, some are available up to 8000 or more hours. The lifespan
is inversely related to the voltage/current, so the lower the operating
voltage, the longer the cap will last.

So, the necessary operating current and voltage need to be overspecified
for longer lifespan, and the tolerance on capacity suggests that you
overspecify just to get the capacitance you want (go over by whatever
the tolerance is, ie 20% over for a +- 20% cap).

>Watts is watts. The average current is a factor of the instantaneous
>current times the duty cycle.
>

Now I have a question:
When discussing the sizing of a supply, people seem to use "rated
current" and "max current" interchangeably. So, which is it? I have
three motors (and may add a fourth and possibly a fifth) that have a
continuous current rating of 9.2A, and max instantaneous current of 37A
(which will be limited by the Gecko to 20A). So, assuming the 2/3 rule,
is that 2/3 of the maximum rating (40A), or 2/3 of the continuous rating
(18.4 A)? And of course I assume that we take 2/3 of the total for all
of the motors expected to run off the supply...

Additionally, how does one go about choosing fuse ratings for the
Geckos. Some of the schematics show a 5A fuse in line with each Gecko,
but that's obviously too small for my motors (or is it?). Slow blow,
fast blow, etc. etc.

If I can get answers to questions like this, I can write a more explicit
power supply sizing document for the files.

- Steve

(funny how the EE background and the real world don't actually relate to
each other sometimes :)