View Full Version : Electronics question: sizing arc-suppression diodes

07-12-2008, 04:51 AM
The coils used in electric clutches generate a high-voltage spike when the power is removed. The arcing is bad for the switches. I tried putting a 1N4001 diode (I just happened to have one on hand.) across the contacts, but it doesn't kill the spark.

I also tried some low-wattage 12-volt light bulbs. They work better than the 1N4001, but there's still some visible arcing. If I get the wattage too high the clutch won't release, since the hold-in current is quite a bit less than the pull-in current.

Is there a guide somewhere for picking the right diode? I don't have any specs on the coil, but I know it draws about three amps at 12 volts. The 1N4001 has a PIV of 50 and a surge capacity of 30 amps. Any suggestions for what I should use here? Will I be able to get rid of the arc completely with the right diode?


07-12-2008, 05:54 AM
Use the diode across the clutch coil, but also add a capacitor across the switch contacts. A value in the range of about .1 uf to .47 uf should be right. Try a .22 uf capacitor rated at 50 volts or more, that should be about right to suppress switch arcing when it opens.

07-12-2008, 07:44 AM
Use a fast recovery diode with a low forward drop such as a schottky diode from a computer power supply.

They look like this (most often)


Weston Bye
07-12-2008, 08:16 AM
The 1N4004 is typically what is used for most solenoids where I work, and also for some clutch coils. The back EMF negative spike typically exceeds 150 volts. A diode is most effective in shunting and absorbing the energy. An ordinary resistor in the 680 Ohm is used in some applications, but is not as effective as the diode.

I suspect that the 4001 was breaking down as it was not man enough to handle the back EMF spike. Go with a higher working voltage. If you are cycling the clutch very frequently, a physically larger, higher wattage diode might be necessary. However, I doubt it. In one application that I'm familiar with the coil was being PWM modulated at a 2 Hz. rate, so twice a second the back EMF spike energy from the coil was being dumped into the diode to be dissipated in the form of heat. The diode, a surface mount package, got hot enough to desolder itself from the circuit board - but the diode itself never failed. The solution (a silly band-aid solution, I think) was to use high-temp solder.

J Tiers
07-12-2008, 09:59 AM
You need to understand WHAT is happening.......

The coil inductance is storing energy. That energy is in the form of current flowing in the inductor.

Energy storage (or release) in an inductor occurs from "volt-seconds". When a voltage is applied across it, the current starts to rise. It continues to rise until limited by circuit resistance, or until the voltage is removed. Higher voltages cause a faster rise of current.

The converse occurs when 'removing" energy. Higher voltages remove the energy faster, causing the current to drop faster..


When the switch on your clutch opens, the current in the inductor is interrupted. But, the only way the inductor can get RID of the stored energy is for the current to continue.

BTW, while it continues, the clutch is still holding closed...... Only when the current drops below the holding current will it open. So if you need quick opening, you need to dump the energy fast.

So, the voltage across the inductor "reverses", and the terminal voltage increases until current will flow, via a spark, or current in a varistor, etc. If the coil was from +12 to ground, with switch to ground, when the switch opens, the "free end" of the coil goes POSITIVE.

Therefore a diode across the switch does nothing. It is "pointing" the wrong direction, unless it is a zener.

You need to provide a way for the current to flow until the energy is dissipated, WITHOUT creating a high voltage spark. The higher the voltage you can tolerate, the shorter time the current flows and the faster the clutch will open.


1) diode across coil, poled "bar end" to plus. This works well, as long as the diode can handle the current pulse. it DOES have the longest "hold time", of all methods, since the terminal voltage is the lowest, so the volt seconds are "short on volts", and energy is not dissipated fast. Least stress on the switch

2) same, with resistor in series added. This allows you to "tailor" the voltage and the opening time to whatever you need. Works pretty well, especially if you can make the resistor bigger so voltage goes higher. Variable sress on switch.

3) a varistor. Typically a higher voltage, but quick dissipation because the volt seconds are removed fastest with a higher voltage. More stress on the switch, usually.

4) other schemes with zeners, resistors, capacitors etc. usually are not much better, if any, than the previous, and use more parts.

5) capacitor across switch. Can be OK, but larger capacitors, which limit the voltage better, may cause a spark when CLOSING the switch. Usually used with a series resistor to do the dissipation, and generally require some calculations to get the values right.

Overall, I'd recommend the diode and resistor, starting with a resistor of about 1/2 the coil resistance, and varying from there.

larger resistor increases sparking, but dumps energy faster. Lower, or zero resistor reduces sparking, but delays opening of the clutch or relay

Picking diode?

The conservative method is to pick a diode rated for the actual coil current. That will be bigger than really necessary.

The optimum diode is rated to carry the coil current for the time of current decay...... Most 1N4001 etc of good manufacturers will take a 30A single cycle pulse, so they are OK for many coils up to 3 or 4 amps, anyway. But any more, 1n4001 series parts are "generic" diodes, and may not meet the "old school" specs.

07-12-2008, 02:54 PM
Thanks fro all the replies and info. I was able to find a 1N4004 as Wes suggested, and it gets the spark down to where it's just barely visible. That's a BIG improvement, and I'm going to go with it.

I wasn't able to notice any difference in the operation of the clutch. It seems to be releasing the same as it did with no arc suppression at all.


doctor demo
07-12-2008, 04:04 PM
Roger, I'm not electronicly savy, and this thread is no different however I think I understand what You are trying to acomplish .(get rid of arc to save switch life) so I wonder if you could use good old fashon mechanical mercury switches ? Instead of all that fancy lectronic stuff.

J Tiers
07-12-2008, 07:23 PM
Do you have the diode across the switch or the coil? You originally said the switch.......

07-12-2008, 09:17 PM
That would make a difference, wouldn't it.

The arc suppression obviously needs to be across the coil, not the switch. I'll fix it in the morning and report back. Thanks for pointing out the error.


07-12-2008, 11:49 PM
That would make a difference, wouldn't it.

The arc suppression obviously needs to be across the coil, not the switch. I'll fix it in the morning and report back. Thanks for pointing out the error.


Actually, a capacitor across the switch and a diode across the coil can be a good combination. The arcing is hi-freq and the capacitor shunts that quite well. If current in the capacitor is a problem you can put in two capacitors, one with a 10x value of the other and with a resistor in series (such as a .1 and a .01 mfd). This will limit current in the larger cap and the smaller cap will still be an effective shunt. But if the diode is doing a good job it won't take much capacitance to arrest any residual spark.

07-13-2008, 12:47 PM
OK, now I've got the 1N4004 diode across the coil, and I have to really strain to see the tiny little arc when the contacts open and close. That's a huge improvement from what I saw with no suppression at all.

I think that's going to be good enough.

Thanks for the help.

07-13-2008, 01:13 PM
Just a cautionary note about reducing the arcing to much.
In the late sixties I and lots of others were into making electronic ignition for our cars which was great except that if you went to far the current through the points was to low to 'clean' them so oxide built up with the consequent missfiring. All easily cured by letting a bit more current through, these days all of no consequence due to optical or hall effect switching.
Just thought I would let you know.