olin_piclist@embedinc.com (Olin Lathrop) wrote:
> On May 27, 2006, at 4:48 PM, Spehro Pefhany wrote:
> > You can swap out the diode, in the current position, for a TVS or
> > beefy zener, and the circuit will work as-is.
> 
> Yes, that should protect the transistor, but I still think that's the
> wrong place to deal with the relay load dump. The relay dump current now
> has to go from the relay, thru the bottom lead, thru the TVS, thru the
> power supply, and then back to the relay. The voltages at the transistor
> should be safe, but there will still be large voltage spikes elsewhere.

Where? The power supply is fine -- you're not trying to back-drive it --
the decaying current is still flowing in the "normal" direction from its
point of view.

If you're trying to get the current to decay quickly in order to minimize
the hold time of the relay/solenoid (and the OP eventually said that he
wants "short pulses"), this is exactly the correct approach.

Just pick a TVS/zener that has a rating 2x to 3x the supply voltage, and
pick a transistor that has a Vce rating with a comfortable margin above
that. For example, with a 50 V supply, use a 150 to 200 V TVS and a
transistor rated at 300 V.

When the transistor cuts off, the collector will immediately* rise to the
TVS voltage and the coil current will drop rapidly. When the TVS stops
conducting, you'll see some self-resonant "ringing" of the coil at the
transistor collector, but this won't go above the TVS voltage, and the
negative peaks will be clamped by forward conduction of the TVS.**

Pay attention to the average dissipation in the TVS -- calculate the
amount of energy dumped into it each time you switch off the transistor,
multiply this by the number of times per second you want to do this, and
make sure the total doesn't exceed the power rating of the device.

> There should be something at least to deal with the high frequencies
> right at the relay. A cap accross the relay will limit dV/dt so that
> other components can handle the spikes better and to reduce RF emissions.

Possibly. In most cases, the self-resonant frequency of the coil isn't
all that high to begin with. If you make the capacitor too large, then
the transistor needs to be able to handle the spike current of charging
it at turn-on, and this can be another source of RFI.

> Then a reverse diode to catch the current. If that causes the relay to
> turn off too slowly, put a resistor in series with the diode, but make
> sure the transistor is rated for the supply voltage plus the relay
> current times the transistor.

That last word should be "resistance".

Yes, that works too, but for a given back voltage, the resistor will
still have a slower dropout (exponentially decaying voltage) than the
TVS approach (constant voltage).

-- Dave Tweed

 * Actually, a half-cycle of the self-resonant frequency of the coil.

** This is the only time the power supply gets back-driven, but by
   this point, the energy involved is miniscule.
-- 
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist