Re: TDC DC

[P3D Michael Kersenbrock <michaelk@xxxxxxxxxxxxxxxxxx>]

> >Straight full-wave rectification plus a capacitor would yield about 160 VDC.
> 
> Assuming no load. With a resistive load, there will be an AC component 
> ("ripple"), and the RMS voltage will be lower.

True.  I only allowed 10-volts of drop for diode and other losses, but
I was assuming big herky diodes directly tied to the power cord and
large capacitors with reasonably low ESR.

Even if it were only 150V, it'd still fry a 120V lamp.  If the ripple
was such that the RMS voltage was 120V, that'd mean there was essentially
zero filtering (neglecting the diode-drops), so there'd be no de-flickering
(which I personally don't think matters, esp. for this application because
the 120Hz "flicker" is viewed in a darkened room).

I'm not suggesting to do this, I'm mostly saying "forget it". With as much as 
1500W of lamps being used in a TDC, generating low ripple DC isn't a reasonable 
thing to do -- but is what needs to be done to gain the benefit desired.

> I would consider big 120VAC capacitors to be fairly risky to use without
> fairly elaborate precautions.

I agree.  But without a big one (they'd be more like >= 250VDC capacitors) there
seems no benefit.  Recall that it was desired to use "DC" rather "AC", and there's 
no point in going to DC unless one wants to get ripple somewhat low.  Staying with
AC would cut the metal migration side effect as well as be more reliable (no
bridge rectifier or cap).  DC that's still pulsing at 120Hz (unfiltered DC) is 
the same  as 120Hz pulsing AC that the bulb gets normally (so far as lamp 
performance).  There is no reduction in flicker without filtering, but filtering
at the 1500W (or 1KW) level) isn't something one probably wants to do.  At
very least, it could be *VERY* dangerous!

> I thought there were devices on the market to limit the startup surge of
> incandescent light bulbs.

Yes, the NTC thermistors that I mentioned do that.  But they do have a
little bit of loss when running, and some may object to that.

When I was a young pup engineer, I worked on a project that had a (2D)
projector as a controlled peripheral.  As a test, a number of test cyclers
were made.  They turned the projector on and off 24-hours a day.  I don't
recall the duty cycle, but it was something like twenty minutes on then off.
All of the projector bulbs tested lasted about their rated hours, which is
what the old-timer engineer told me, even though I thought the "surges" would
take it out sooner because of the cycling.  He said that although it may
burn out when powered on, that's only the breaking of the camel's back. The
"fault" that burned out with the surge would have soon burned out anyway.

The power surges (voltage going up while running) possibly could be
controlled by using a tyristor + feedback circuit.  Wouldn't help
voltage *drops*, but could provide a limit.  For projectors like the
Viewmaster one, where the bulb power is much lower, either a larger PC
UPS could be used or a cheap surplus constant-voltage transformer could
be used (just threw one of those away recently when I moved -- they're
VERY -explitive deleted- heavy).   :-)

Mike K.

P.S. - I've two sets of spare bulbs for my projector.  That way if I get a
       burn out, I'll still have a spare set while I wait for replacemnents
       to come.  :-)


------------------------------