On Mon, Mar 01, 2010 at 02:22:06PM -0500, Gene Heskett wrote:
> >
> But probably right at the motor supply terminals of the board, not 3 feet 
> away in the psu.

I'm with you on the "stop interference at the source" approach, but
having the transorb on the board will catch fast spikes induced in that
cable.

> It's a pi section filter, with 75,000 uf on each side of a choke that 
> probably has 3 to 4 ounces of decent silicon steel in its E core but I have 
> not measured its true inductance.  The hf noise isn't much greater at the 
> xylotex terminals than it is at the psu, another 50mv or so.  Barring failure 
> of the caps, I think I am in pretty decent shape.  Those 75,000 uf caps are 
> also beyond my ability to measure, but were top quality stuff 45 years ago, 
> in humungous screw terminal cans.  Even I, as acutely aware of the ESR 
> requirements for such duty as I am, was pleasantly surprised at the seemingly 
> zero ESR those old caps have.

Ah, I didn't realise you had a pi filter. With any decent inductance
between two of those caps, and good ESR, your setup is a very serious
spike killer.

However, H-bridges can generate their own spikes, given the partially
inductive load, and a transorb on at least the supply rail can be
life-extending.

> >You'd want something with tighter tolerances than the old 5Z27 devices
> >floating at the bottom of my junk box, for lower headroom.
> 
> No doubt.  How are these transorbs thingies for long term stability?  Only 
> true zeners are stable over time, and those stop at 4.7 volts, anything above 
> that is actually an avalanche diode, and they will drift low over time, rate 
> dependent on how much its average power dissipation is.

Haven't seen drift in the datasheet, even for a slightly more recent
device. A couple of those old 5Z27s checked out within their barn-door
specs, and they've been lying about for about 30 years now. Even
in-circuit, they're only conducting when there's a spike. If they are
very high energy spikes, then some cooking might go on.

Erik

-- 
manual, n.:
A unit of documentation. There are always three or more on a given item.
One is on the shelf; someone has the others. The information you need is
in the others.                      -- Ray Simard


------------------------------------------------------------------------------
Download Intel® Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
_______________________________________________
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Reply via email to