switch/relay for fast, clean interruption of current 2

To carry out testing of fuel cell/battery components I need to develop a way to interrupt the current being applied to the cell in a quick, clean way. The DC current to be interrupted will be in the range of 0.1-10A (the voltage will be 0.5-1.2 V). The resulting voltage response will then be monitored on an oscilloscope. In order to obtain the desired information it is essential for the switching speed for interrupting the current to be <1 microsecond. Also, the break must be clean with no bouncing, noise, capacitance effects, etc. so that the voltage response seen on the oscilloscope is due solely to the processes occurring in the fuel cell/battery and not due to the switch/relay circuitry. It looks like a mercury wetted reed relay is a possible option but I’m not sure it can meet all of the requirements. Any help in this matter concerning electronic components and/or circuitry would be appreciated. Thanks.

 

[melone]

Why do you have switch? Do you have multiple circuits / setups that you need to test?

 

[nbucska]

Some MOSFETs parallel, switched with external 5-15 W
pulse. <nbucska@pcperipherals.com>

 

[BrianG]

Do you really need a switching time of less than 1 micro second? This will not be achieved by any sort of relay, mercury-wetted or otherwise. If this is a primary requirement then you must look at solid state switches such as FETs.

 

[nbucska]

Sorry, I meant 5 to 15 VOLTS not W. <nbucska@pcperipherals.com>

 

[eduardom]

You need to interrupt the current without bouncing. A mercury switch can do this in less than 10 nonoseconds.
The problem to solve is to cancel the efects of parasitics.
I made a nanopulse generator using a common level sensitive mercury switch. The resulting pulse was really clean, showing no bouncing, but the mercury switch was mounted in a Zo= 50 ohm coaxial transmision line. This technique solves the problem of stray capacitance and inductance. To pick the transient you need a digital high speed scope, because pulse repetition rate is very low (3 pulse per second). If you think this solution can help you, please let me know. Perhaps I can design a test circuit for you, with some additional data on the problem.

















0

 

[ranjeed]

Hi eduardom,
I'm interested in the layout and circuitry of your nanopulse generator. Could you tell me more about pulse width and risetime you achieved?
I'm going to build a pulser for risetime calibration of my sampling scope.
Any diagrams and hints are very welcome.

Best regards

Ranjeed

 

[nbucska]

Eduardom:

The switch still can bounce, 10A will evaporate the Hg
in no time. I don't believe in the NONOsec. mechanical
switch.


<nbucska@pcperipherals.com>

 

[hacksaw]

forget relays, just use power FET's. you can get any wave shape and duration you need and you don't have to worry about turn-off problems that scr's experience in dc circuits.

the mercury wetted contacts are fine for short duration app's, but they have current limitions unless you go with the heavy duty plunger types.

don't think you need the pulse gen. config. describe previously but rather a programmable power supply. in fact there are p/s that allow for remote adjustment on a programmed basis. some even allow for direct computer control.