Micro Switch Testing

[apatel225]

I currently need to test the life of 20 micro switches.   These need to be cycled mechanically, and electrically loaded with 28VDC and 3A on each micro switch.  I want to monitor failures using labview.  Any suggestions on the best way to do this?  Circuitry that would be involved (per switch) and labview hardware/software?  I have a 9205 module which I planned to use.  I want to try and keep this cost effective if I can.  I will be renting a 33VDC, 85A power supply for this.  Any/all suggestions would be welcome! 

Thanks!

[NIquist]

Your 9205 really isn't the best choice for this type of testing.  You want a digital IO module with lots of ins and outs like the 6509. 

I would use digital outs to trigger solenoids to physically press the switches and then use external circuitry to convert the 28 volts to a TTL signal connected to digital in ports to verify switch connections.

BTW, most of the switches out there are good for millions of cycles so you either better be patient for your results or look into HALT/HASS testing:  http://www.google.com/search?q=halt+hass&sourceid=ie7&rls=com.microsoft:en-us:IE-SearchBox&ie=&oe=

EDIT:  One other thought.  You can always make a solenoid driven bar that presses your whole bank of switches at once.  Drop the bar and close them all, scan them for a high signal.  Then lift the bar opening the switches, scan them for no voltage.  Rinse & repeat... and repeat... and repeat...

[apatel225]

NIquist,

Thanks for the response.  The reason we're doing this testing is because the switches will be going in products which are rated for high temperatures.  The switches we are testing are rated for a million cycles, but only at room temperature, and high temperature life cannot be provided.  We expect the life to be reduced, but we need to see how much.  We are going to be actuating the switches using a cam setup because we need to control the rise/fall rate of the switch buttons, and also the rate of actuation.  My concern right now is that the external circuitry needed might get a little involved.  However, I am not exactly sure how to go about the circuitry in order to get that TTL signal which the DAQ system could read.  Any insight into this?

Thanks!

[apok]

We use pneumatics to mechanically cycle a push button switch, that being said...you need a signal output (typically a digital output to some sort of transistor driver). Now when you say, "monitor failures..." I don't understand exactly what parameters you specifically measure/monitor?

is the switch...SPST, SPDT,DPDT....?

[apatel225]

the switches are SPDT, there will be 2 brands, 10 switches each.  5 will be wired normally closed, and the other 5 normally open.  The DAQ system will be used to monitor the electrical leg of the switch.  Basicaly as a counter.  When the switch fails, the labview system will show the cycle count that the switch failed at.

[johnsold]

How are you defining a failure? Switch contact resistance increase by 10%? You may not get completely open or completely shorted contacts after large numbers of operations at high temperatures and maximum rated load.  Do you have a bank of load resistors and a big fan?

While the result may be binary (pass or fail), the testing may need to be analog.

What about switch bounce? What is the range of durations of the actuated and non-actuated phases for the switches? Do you have an optical encoder or some similar means of monitoring the position of the cams?

Lynn

[apatel225]

we are looking for complete electromechanical failure. meaning it fails one way or another.  As long as the switching element no longer "switches"  wether it be in the normally closed, or normally open state.  Some bouncing is to be expected.  We are not looking for a precise count.  however, an excessive (thousands) of bounces would indicate a problem anyways and we'd have to evaluate that on a case by case basis. 

[pjr1121]

To ensure a 3A current at each switch, you can build a "Current Source" or "Current Sink" circuit (schematics are easy to find online).  Note that in reality, resistor & transistor values vary from part to part so each circuit you build will not be identical (something you may need to consider).  As far as measurement, you could build a voltage divider circuit to measure the voltage drop across a reference resistor.  If you are truly only interested in determining the end of switching, choose your resistors so the voltage drop is in the 4-5V range and wire this to a 5V TTL input (this range will be considered HIGH).  If you want analog measurements, you will need to be careful to calibrate each circuit (since they'll each vary a little).

Remember that resistances change with temperature so make sure the circuit is what you want at the measurement conditions (use power resistors to reduce the heat generation from the 3A of current).

[apatel225]

I was thinking of doing something like this.  V1 would be responsible for taking most of teh 3A, while the other two resistors would be the voltage divider.  V3 would be the one that provides 5V to the DAQ.

What would the advantage be of using the current sink, rather than what i've shown here?  Is there a better way of doing this?

The circuitry would not be exposed to temperature (kept outside of the chamber)

All options welcome at this point.

[apok]

switch bounce can be an issue with your counts....I would use a pair of SR nand latches attached to the throws with pullups (common tied to gnd). latch#1 set tied to latch#2 reset and latch#1 reset tied to latch#2 set.