Looking for a mercury-wetted reed relay that can produce nanosecond pulse

Thread Starter

radetonator

Joined Aug 1, 2011
50
Does anyone know which manufacturer still produces mercury wetted reed relay capable of generating 5ns high voltage pulses? Thx:)
 

AlexR

Joined Jan 16, 2008
732
Two questions;
1. Why does it have to be a relay?
2. Why 5 nSec?
5 nSec is faster than most common logic operates at, you have no chance of finding a relay that operates that fast.
Relays, (even reed relays) generally take milliseconds to operate. Figures for reed relays I have seen show on closure time of about 5mSec, that's 6 orders op magnitude slower that you require!
 

Wendy

Joined Mar 24, 2008
21,906
Many logic families could have trouble producing a pulse that fast. I suspect you haven't dealt with speeds like this before, 1ns is 9 inches of travel at light speed. You would need a circuit with a minimum freq response of 200Mhz, and for a square wave it would need to be closer 2 Ghz.
 

Thread Starter

radetonator

Joined Aug 1, 2011
50
Some scholars used mercury relay to generate nanosecond pulse. It is real. I am also considering to build pulse generator based on avalanche transistor. But i think using mercury relay is the most easiest way to do it.
 
Last edited:

Wendy

Joined Mar 24, 2008
21,906
Avalanche oscillators are fast by definition. Any mechanical system is not.

I have experience with power mercury switches, for high current applications.

You have any sources of literature for fast mechanical relays?
 
Last edited:

crutschow

Joined Mar 14, 2008
23,508
I assume 5ns is the rise-time of the output signal, not the operating speed of the relay (which is obviously impossible). If so, it can be done, but it requires well controlled impedance transmission line connections directly to the relay contacts. This is best done with stripline or microstrip design techniques on a controlled impedance circuit board, not a trivial task.
 

Thread Starter

radetonator

Joined Aug 1, 2011
50
I assume 5ns is the rise-time of the output signal, not the operating speed of the relay (which is obviously impossible). If so, it can be done, but it requires well controlled impedance transmission line connections directly to the relay contacts. This is best done with stripline or microstrip design techniques on a controlled impedance circuit board, not a trivial task.
can you recommend a relay for this application?:D
 
Hi. Are you still interested in this? Mercury wetted relays were used to make
"tail pulse generators". The "Mech-Tronics 1000 mercury relay pulse generator"
and the "Ortec 480 Pulser" are both examples in NIM format. They give a pulse risetime
(or fall) of a nanosecond or so followed by a slow recovery. By building simple RC
circuits you can produce 5ns width pulses easily. I had a Mech-Tronics 1000 fail a few
years ago - but repaired it with the exact relay it was built with which I got from some surplus house. Most manufacturers don't like mercury anymore - but there are plenty from Clare, western electric etc on the surplus market. You can build nanosecond electronics without stripline - but you need to be careful. If you're still interested, I can look out the exact relay type.
 

P-MONKE

Joined Mar 14, 2012
83
Many logic families could have trouble producing a pulse that fast. I suspect you haven't dealt with speeds like this before, 1ns is 9 inches of travel at light speed. You would need a circuit with a minimum freq response of 200Mhz, and for a square wave it would need to be closer 2 Ghz.
As we're here - the speed of light in a vacuum is closer to 1 foot per nano second - it's a physicists rule of thumb.

Not trying to be a smart-arse, just increasing the collective knowledge. ::ducks down behind desk:: ;)
 
Top