1956 PI COIL SWITCHING

[Eric Foster]

Hi All and JC in particular,

Here is a coil switching circuit devised in 1956. What we want is a way of doing S2 in solid state, without the back emf wiping it out.

Eric.

 

[Wirechief]

Hi Eric, cool that was the year I was born ! Thanks for posting it, I wonder what publication it was in ? CU later Eric.

John T., CET

 

[woody]

My own experiments show that some rather large improvements in sensitivity can come from reducing the series resistor and replacing the back to back protection diodes with an active clamp. The only problem is that you still need around 250 ohms resistance as not to clamp the back emf pulse on the coil. Associated capacitance takes the sample out to 15us but I am working on getting this down by putting the clamp devices in series thus halving the capacitance.

The other trick would be to put a switch in series that could stand off the 500 volt emf pulse, it could be implemented with P and N channel fets using a zener clamp on the gates being driven from the back emf by itself. This too may ass significant capacitance that slows things down too much, one must carry out the experiments to see what one can get away with.

 

[Steve and Bia]

I'm not sure but I thought they did... like serious noise! I don't know but wouldn't 500 volts cancelled at a clamp produce enough RF signals to "glitch" every other detector within a few hundred feet?

It's been a while since I've seen a schematic so excuse my ignorance.

 

[Mr.Bill]

n/t

 

[woody]

It is true that hitting diodes with a quick spike will produce noise and harmonics. I use diodes as frequency multipliers for RF work so it would be interesting different diode types such as ultra fast with hard and or soft recovery characteristics.

 

[robby_h]

Woody,
"...being driven from the back emf by itself."
Been there.
I squared up the clamped spike for another reason and this means squaring it up to a point above zero so as to not include any late time lag. This wave form now obviously responds in width (trailing edge) to the magnetic effect as the coil is lowered to the ground but
the process has taken "time" and so your trigger will arrive late and you will end up not disconnecting the leading edge of the spike. On a clamped spike this leaves you with a non-varying early square section of the spike being presented to the clamp diodes etc. I tried this as a trigger and got better results after correcting the damping etc than if disconnecting the whole spike but am at a bit of a loss as to why.