Solid State Relay Puzzle

[Dockhead]

Following Dave's (Goboatingnow) advice, I bought a solid state relay for my autopilot, a Celduc SOM40100.

Now I am puzzled, because the datasheet (there are no other instructions) shows a diode and a capacitor which appear to be connected across negative and positive of both power and control circuits. WTF? There is a cryptic reference in the data sheet to "See Instruction sheet to select the right protective components." But there is no instruction sheet; the data sheet (here: http://www.farnell.com/datasheets/56391.pdf) is all I have.

Has anyone messed with these MOSFET solid state relays? Are these diodes and capacitors shown in the diagram the referenced "protective components"? What do they do? Can I use the relay without them?

[Paul Elliott]

The diode and/or capacitor is used to protect the relay from "inductive kickback" that occurs when the current through an inductor (coil) is interrupted. Depending on the coil, the current, and the speed of interruption this can be hundreds or even thousands of volts. Whether you need protection depends on what you are switching.

[Dockhead]

Originally Posted by Paul Elliott

The diode and/or capacitor is used to protect the relay from "inductive kickback" that occurs when the current through an inductor (coil) is interrupted. Depending on the coil, the current, and the speed of interruption this can be hundreds or even thousands of volts. Whether you need protection depends on what you are switching.

Thanks for that.

I'm switching the bypass valve in my autopilot ram. It's the "engage/disengage" function of the pilot. My Simrad/B&G AC42 pilot computer claims to be overloaded by my bypass valve, although I have measured the current and it seems to be within spec. Here's the thread on it: http://www.cruisersforum.com/forums/...ms-101501.html

[Paul Elliott]

I've never actually designed transient (kickback) suppression at this current level, but the two methods I've used at lower currents are a diode in parallel with the coil (opposite polarity, with the cathode connected to the positive terminal), or a resistor and capacitor in series, connected across the coil. Using the resistor/capacitor design, I would choose a resistor of about 4.7 Ohms (more or less, it's not critical), which, since your coil has a DC resistance of about 6 Ohms, should limit the voltage spike to less than 24V. The power rating could be 1/2 watt (or greater), since the spike is very brief. The capacitor could be 0.1uF, although I'm kind of guessing about that. I would use a big ceramic disc, or non-polarized film capacitor, with a 100V rating. 50V might be OK.

For a diode, you would use something with a 4A rating or better, probably 100V rating.

This assumes a 12V system. Double the voltage and power ratings for 24V. I would welcome hearing from anyone with practical experience with this!

[tbodine88]

The diode could be something like this
40 Amp 600 Volt Stud Blocking Diode - Amazon.com

I picked the 40 amp diode since it meets the capability of the switch.

For the capacitor, I'd use a 1 uf 100v like this

Capacitor 100v 1UF Axial Elec. Nonpolar Crossover | 31-1295 (311295) | Distributed By MCM

I'm guessing for this, since its supposed to pass the high frequency portion this might be to big.

[Paul Elliott]

One reason it's hard to guess a capacitor value is that we don't know the inductance of the coil, and so we don't know the time-constant / stored power of the coil. I'm sure there is a typical value, but I don't know it. I suspect that tbodine88's recommendations would work just fine. My diode suggestion may be a bit wimpy -- there's no harm in upping the diode specs.

[Dockhead]

Christ, if I had known I would have to design a whole circuit in order to use the bloody thing, I would have bought a banal car-type mechanical relay.

I have just discovered that I grossly overspecced the relay -- it's a 40 amp one, when the current required is only 2 or 3 amps (obviously inrush current will be more).

Maybe the overspecced-ness of the relay will be enough protection?

[s/v Jedi]

Originally Posted by Dockhead

Christ, if I had known I would have to design a whole circuit in order to use the bloody thing, I would have bought a banal car-type mechanical relay.

I have just discovered that I grossly overspecced the relay -- it's a 40 amp one, when the current required is only 2 or 3 amps (obviously inrush current will be more).

Maybe the overspecced-ness of the relay will be enough protection?

I would swap it for a small one and check details for installation first, i.e. choose one that has some protection already or is hardened enough to not need it. Your application is very low power.

[Dockhead]

Originally Posted by tbodine88

The diode could be something like this
40 Amp 600 Volt Stud Blocking Diode - Amazon.com

I picked the 40 amp diode since it meets the capability of the switch.

For the capacitor, I'd use a 1 uf 100v like this

Capacitor 100v 1UF Axial Elec. Nonpolar Crossover | 31-1295 (311295) | Distributed By MCM

I'm guessing for this, since its supposed to pass the high frequency portion this might be to big.

Thanks very much!

So maybe something like this for the diode: http://www.farnell.com/datasheets/3247.pdf
?

[s/v Jedi]

Originally Posted by Dockhead

Thanks very much!

So maybe something like this for the diode: http://www.farnell.com/datasheets/3247.pdf
?

I'm sorry but a 40A diode is crazy. Who calculated that and using which math?

[Maine Sail]

Perhaps I missed something about why the solid state relay was used? Why not just use a simple contactor/relay? You can find them with a simple diode across the + & -...

I have used relays like this in a number of installations where the course computer is undersized for a newly installed drive unit but the CC is not worth getting rid of. Works like a charm and is quite inexpensive to do...



Cole Hersee also makes some nice 70A continuous duty rated relays. If you don't mind a blade terminal connection they work great and last a long time. They can also be had with a diode across + & - of the coil.....

[Paul Elliott]

A solid state relay probably has a lower current-draw than an electromagnetic relay. This may be an advantage, or it might not matter.

[Paul Elliott]

Originally Posted by Dockhead

Maybe the overspecced-ness of the relay will be enough protection?

The relay is spec'd for 60V open-circuit voltage, and your un-suppressed AP solenoid will probably put out a spike greater than that. The relay can take a non-repetitive transient of 100V, but you don't want to stress that spec.

I do see that the relay includes a 56V transient suppression diode (D1 in the relay schematic) -- I wonder how they can spec 60V operation with a 56V overvoltage clamp diode? In any case, with your 2A current in the AP coil, and 12V voltage, I suspect that the built-in clamp diode will do the trick.

[Dockhead]

Originally Posted by Paul Elliott

The relay is spec'd for 60V open-circuit voltage, and your un-suppressed AP solenoid will probably put out a spike greater than that. The relay can take a non-repetitive transient of 100V, but you don't want to stress that spec.

I do see that the relay includes a 56V transient suppression diode (D1 in the relay schematic) -- I wonder how they can spec 60V operation with a 56V overvoltage clamp diode? In any case, with your 2A current in the AP coil, and 12V voltage, I suspect that the built-in clamp diode will do the trick.

It's all 24v nominal, not 12v.

So you think I can just wire it up as it is?

I have wasted a lot of time reading other manufacturer's literature. Some of it says that for an inductive load, you just need a "freewheel diode" -- so, a diode which will allow the pulse or surge or whatever it is to circulate the current after the inductive device is shut down.

Here's something I found: Flyback diode - Wikipedia, the free encyclopedia

Maybe I could just put in one diode and be done? Do I really need the capacitors?

[s/v Jedi]

Originally Posted by Dockhead

It's all 24v nominal, not 12v.

So you think I can just wire it up as it is?

I have wasted a lot of time reading other manufacturer's literature. Some of it says that for an inductive load, you just need a "freewheel diode" -- so, a diode which will allow the pulse or surge or whatever it is to circulate the current after the inductive device is shut down.

Here's something I found: Flyback diode - Wikipedia, the free encyclopedia

Maybe I could just put in one diode and be done? Do I really need the capacitors?

Yes, that's all you need. I'm pretty sure a 5A rectifier diode will be fine. Like you have in spare for the solar panel bypass?

For schematic: just put it parallel to the load (that AP solenoid) with the cathode (the pointy end marked with a band) pointed towards the positive side of the voltage source.

EDIT: but as near to the relay as possible, not at the location of the solenoid.