isolation problem

[Justanotherguy]

Our area has been given the task of installing monitoring equipmnt on our snow plows. For the most part it was straight forward, but we ran into a hurdle with a couple circuits that need monitoring. The system is using PWM to control pump speed on the de-icer circuit, and the speed of the motor on the sand spreader. The monitoring unit needs an on/off 12v to determing if and how long the truck is using de-icer or sanding the road.

when the mechanics first brought this to my attention they were reading the voltage on the lines and it varied so they assumed the circuits were voltage controlled, i suggested a relay that would turn on with voltages between 3-15, they ordered several of Crydom CX240d5 relays. The next day they came back to me stating that the output of the relay was varying with the input, I tested the relay on my bench and it was fine. Today i hooked up the o-scope on the lines and it was indeed PWM conrtrol.

I thought simple fix, We can just put a cap on the input of the relay to level out the pwm signal. On second thought this will mess with the signal to the motor/pump. Then I thought OK, we can put the cap on the output of the relay, it did indeed give me a constant 12v output and didn't effect the input signal. After some more thinking, I figured maybe the relay would be effected by the constant on/off due to the pwm input. After some more thinking... I thought I could maybe use some kind of opto isolator on the signal line and then use the cap on the input of the relay so when the signal line shows a square wave, the opto isolator output shows a square wave but the cap brings it closer to a constant dc signal enabling the relay which in turn puts out the needed 12v for the monitoring equipment.

I looked for an opto isolator, read that most of the invert the signal, which isn't going to work for us. Now trying to find some kind of isolation device that a high gives a high, but the voltages are at 12v instead of 5 or 3.3v.


Any input would be greatly appreciated.

 

[KrisBlueNZ]

Hi there and welcome to Electronics Point

Thanks for the thorough description of the problem, and the schematics. This REALLY helps.

Could you just use a diode between the PWM signal and the relay coil, with a capacitor across the relay coil? When there's ANYTHING happening on the PWM, the diode would keep the capacitor charged up, and that would keep the relay activated. When PWM activity stops, the cap would discharge and the relay would drop out.

Something like a 1N400x would do, or a 1N540x to play it a bit safer. And you've already worked out the capacitor value. It needs to keep the coil voltage above roughly 2/3 of the rated voltage at minimum PWM width, so you need to know the PWM frequency. Using a capacitor much higher than you need will skew the equipment hours measurement slightly because turn-off of the relay will be delayed more than necessary.

 

[hevans1944]

... Then I thought OK, we can put the cap on the output of the relay, it did indeed give me a constant 12v output and didn't effect the input signal. After some more thinking, I figured maybe the relay would be effected by the constant on/off due to the pwm input. ...

Are you sure about your "relay" part number: Crydom CX240d5? See datasheet here. This is a solid-state "relay". No moving parts, but it operates only to switch an AC load from up to 240 VAC line. It uses back-to-back SCRs as power line switches. Current through an SCR must go to zero before it will commutate and turn off, hence the need for an AC power source to the load. However, the CX240D5 can cycle on and off as much as you please without ever wearing out and it has an optical isolator on the control terminals. If that is indeed the part you are using, and a capacitor on the output somehow makes it work, go with it.

Your schematic shows the "relay" switching a DC power source, presumably a 12 V battery source, to control perhaps an elapsed-time meter. Crydom makes several models of solid-state "relay" for that purpose, but none with mechanical relay contacts that I am aware. See their catalog page here. You can choose either a MOSFET or a bipolar transistor switch in the output. Bipolar might be more robust in terms of sensitivity to voltage transients on the power supply. MOSFET outputs are available that switch more current. For your application, the MPDCD3 would probably work fine. The datasheet is here.

Whether you use the AC version (CX240D5) or the DC version (MPDCD3), all of these solid-state switches feature an optically isolated input. The control range is 3 to 15 V at 15 mA (300 Ω input impedance) for the CX240D5. The control range is 3 to 32 volts (1500 Ω input impedance) for the MPDCD3.

I do not understand how placing a capacitor on the output of a CX240d5 would in any way "filter" the output, nor do I understand how the CX240d5 would even work with a DC power source. But if the filter capacitor does work, and you do have a solid state relay, I wouldn't worry about the "relay" cycling wearing it out. On the other hand, if you can actually hear contacts on YOUR relay clicking on and off, then that is a mechanical relay and your concern is justified.

If you are worried about the pulses cycling actual mechanical contacts on your relay, try @KrisBlueNZ recommendation above to use a diode and a capacitor to smooth out the pulse width-modulated input. A smaller capacitor will be needed for the MPDCD3 because it has a higher input impedance than the CX240D5. Moreover, some experimentation with the capacitance value may be necessary to ensure the capacitor charges up sufficiently between pulses to keep the switch "ON" at the lowest duty cycle.

 

[Gryd3]

http://www.allaboutcircuits.com/vol_3/chpt_3/5.html

Is the relay even required? I've seen another guy on here make a peak Voltage adaptor to be used on a standard DC multi-meter. Would something like that work?
Diode, capacitor and resistor. Assuming the monitoring device has a very high input impedance, the capacitor should remain charged sufficiently.

 

[hevans1944]

http://www.allaboutcircuits.com/vol_3/chpt_3/5.html

Is the relay even required? I've seen another guy on here make a peak Voltage adaptor to be used on a standard DC multi-meter. Would something like that work? Yes! I don't see why it wouldn't. Good observation! Of course you would place the resistor in parallel with the capacitor to ensure the capacitor discharged when the pulses went away.
Diode, capacitor and resistor. Assuming the monitoring device has a very high input impedance, the capacitor should remain charged sufficiently. That would depend on the minimum duty cycle of the pulses, whether or not the capacitor charged sufficiently. Should still work though for some combination of R and C.

The OP didn't mention what the "monitoring equipmnt" actually was, implying however that several things were monitored and only this one thing was giving problems. Anyway, in lieu of further comments from the OP, I think we are done here. I know I am.