Re: [CAD_CAM_EDM_DRO] Sizing a power supply

This posting is triggered by the mention of keeping televisions from picking up RF leakage from driver supply lines in the quoted
post.

For RF bypass applications a ceramic 0.1 to 1 uF cap is typically appropriate. Electrolytic caps become inductive at higher
frequencies. If you measure reactance vs frequency for different cap types you quickly find that electrolytic and unbiased tantalum
caps behave far from the ideal "Infinite reactance to DC, dead short to infinite frequency" ideal behavior predicted by reactance
formulaes. You will see the reactance decrease with freqency to some point, and then begin to increase again. This is due to the
capacitor construction creating a lumped inductance. At higher freqencies, smaller caps with different physical constructions and
dielectric materials are required to do the job, as they more closely resemble the theoretical ideal capacitor at higher
frequencies.

I usually bypass electrolytics with a smaller cap to avoid this problem. Back when I was into hardcore audio, chasing transient
intermod and slew induced distortions down to the point I couldn't measure them, reading of tech papers and confirming the claims
made via measurment demonstrated that at higher frequencies, a large cap is in effect bypassed by the smaller cap. Large
improvements to quality could be had by bypassing big electrolytics with polypropelene, polycarbonate, or polystyrene caps. Usually
a combo of a 4.7 uF and a 0.1 uF handled the unpleasant aspects of electrolytic behavior at higher frequencies. This doesn't apply
to electrolytic caps in the signal path messing with the slew rate as a symptom of dielectric absorbtion, a whole other can of
worms, only to bypass apps in power supply rails and such.

You may not think of power supply lines having an RF component. Conventional design approaches preach that the power supply is a
short circuit for AC analysis. This ignores supply wiring resistance and inductance, power supply source resistance, and a bunch of
other things. It is correct, but in construction, you have to make it as close to true as you can.

Consider a one mS transient with a 2uS rise time. Welcome to a 500KHz component on your supply line. Ever tried to bypass HF
through an inductor with a cap that is inductive at the frequency in question at the other end of the lumped inductor? You get a
really neat view of this with a good O'scope if you look. This is why you see a basic design rule for TTL that recommends a ceramic
bypass cap per chip, close to the chip, in the 0.1 to 1.0 uF range. It provides instantaneous current for very brief periods, and
bypasses HF components to ground.

Folks in ham radio worked out lots of ways to bypass HF leakage long ago. If you are current starving during transients, add bigger
caps and heavier wiring. If you are supressing transients, use smaller caps good at HF.

Stan.

Jon Elson wrote:

> Jon Anderson wrote:
>
> > Jon Elson wrote:
> >
> > > Anyway, the input to a chopper drive generally HAS to be filtered,
> > or
> > > the inductance of the wires from the power supply will cause enough
> > > transient spikes to blow transistors very quickly.
> >
> >
> > What constitutes adequate filtering? The PS I'm using now is home made
> >
> > per specs for my older CyberPak drivers. There's a decent sized
> > capacitor in there, would that have the effect of filtering?
>
> Clearly, the power supply itself needs a large energy storing capacitor
> to
> tide it over between half cycles of the power line. But, because
> chopping-type
> drives switch the current on and off in the thousands to hundred
> thousands
> of times a second, they need some local filtering to prevent the
> inductance of
> the long wires to the power supply from developing voltage spikes. For
> the
> protection of the transistors, only, a few microfarads would be
> sufficient.
> To prevent every TV in the neighborhood from going on the fritz, a
> larger
> capacitance may be needed. A good value might be 1000 uF right at the
> stepper driver, or on the board.
>
> > And, should I locate a surplus PS in the 70 volt range, how to I tell
> > if
> > it's properly filtered for stepper use, outside of finding a dedicated
> >
> > stepper PS?
>
> 70 volts doesn't help. It is the current draw that makes the
> difference.
> 1 Farad of capacitance drops 1 volt per second for every amp drawn
> from it. It needs to hold a reasonably steady voltage between line half
>
> cycles, which are 8.33 mS at 60 Hz, and 10 mS at 50 Hz.
> Assuming we want no more than 1 V droop at 60 Hz, with full-wave
> rectification, and assuming the charging time is zero (a
> simplification),
> we get C = 8.33 mS * I (amps) / Vdroop = 8.33 mF for a 1 Volt
> droop, for every amp of current drawn. Assuming a current draw
> of 10 Amps, you need an absolute minimum of 83.33 mF, or 83,000 uF.
> It might be a good idea to double that, due to series resistance of
> the caps. So, 160,000 uF would be a good capacitance for a 10 A
> power supply.
>
> Jon
>
> ------------------------------------------------------------------------
> Get a FREE "YourName.Com" Web Address,
> E-Mail, and Home Page...TODAY! CLICK HERE!
> http://click.egroups.com/1/5460/3/_/423600/_/960877406/
> ------------------------------------------------------------------------
>
> Welcome to CAD_CAM_EDM_DRO@...,an unmoderated list for the discussion of shop built systems, for CAD, CAM, EDM, and DRO.
>
> Addresses:
> Post message: CAD_CAM_EDM_DRO@egroups.com
> Subscribe: CAD_CAM_EDM_DRO-subscribe@egroups.com
> Unsubscribe: CAD_CAM_EDM_DRO-unsubscribe@egroups.com
> List owner: CAD_CAM_EDM_DRO-owner@egroups.com, wanliker@...
> Moderator: jmelson@... [Moderator]
> URL to this page: http://www.egroups.com/group/CAD_CAM_EDM_DRO
> FAQ: http://www.ktmarketing.com/faq.html
> bill,
> List Manager