Transorb, Zener or Resistor for Back EMF Suppression of Relay?

[¡MR.AWESOME!]

I searched practically all of the nets and webs and still my questions are unanswered, so I decided to consult y'all. I have previously asked you guys about diode sizing to prevent back emf on a relay and you were very helpful. I am once again calling upon your help.

I have some relays in my design of a pushbutton starter and they need back emf suppression. Right now, they each have a simple diode running inverse and parallel to the coil. I did some more reading, and apparently, that isn't the best solution. The single diode causes the opening time of the contact to increase and causes prolonged arcing which reduces contact life. I read (http://www.schrackrelays.com/appnotes/app_pdfs/13c3264.pdf") that a zener run in series to a normal diode provides the best solution. So that's just fine and dandy.

But then I came across transient voltage suppression diodes, aka transorbs. They are designed for this, right? So can I just replace my single diode setup with a transorb and be done with it, or is it still ideal, when using transorbs, to also have a normal diode in series with it? On the topic of sizing a transorb, I ran across http://www.microsemi.com/micnotes/125.pdf" article. It seems informative, but I can only understand about half of it. Any help on how to size it would be, well, helpful.

I also have one relay that will be getting current going both ways through it's coil. If I can just use a transorb, than I would get a bidirectional one and that's that. If it's best to use the zener and normal setup, can/should I just double up and mirror the zener/normal circuit?

As I understand it, the other option available to me is to use a resistor instead of any diodes. That sounds good to me, but I have no freaking clue on how to size it.

A bit of topic, but while I have the experts here, I shall ask. I have some relays driving other relays. Should the coils of the driven relays be suppressed as well? My only concern is that the spike from the driven relay will cause arcing across the driving relay's contacts. Is that a legitimate concern?

To make it easier, here is what I'm asking:
1. Can I use a single transorb to replace my diodes, or should I use a zener somewhere in there as well?
1a. (If so) How do I size a transorb?
1b. (If not) Can I make a zener/normal setup bidirectional simply by mirroring it?
2. (If applicable) How do I figure out what resistance is needed if I am to use a resistor instead of any sort of diode?
3. Should I also suppress the coils of relays that are driven by other relays?

Thanks for any help that you may be able to provide.

 

[Carl Pugh]

The relay coil is an inductance.

When the power source is removed from a relay coil, the current in the relay coil continues.

The faster this current in the relay coil is reduced to zero the faster the relay operates.

The drive transistor see the coil source voltage plus the voltage across the relay coil.
With a diode across the relay coil, the drive transistor see's the coil source voltage plus about 3/4 volt.
With a diode in series with a transorb across the relay coil, the drive transistor see's the coil source voltage+ the transorb voltage + 3/4 volt.
With no surge suppression across the relay coil, the drive transistor see's the coil source voltage + the voltage across the relay coil. The voltage across the relay coil is limited by arcing (relay contacts) or breakdown (drive transistor), or the energy in the relay coil is stored in stray capacitance and there is a dampened oscillation.

 

[¡MR.AWESOME!]

I assume the drive transistor is the PIC microcontroller that I'm using. So what you're saying is that I am worse off with a transorb because it sees the coil voltage plus the transorb voltage plus 3/4 volt. You are also saying that, somehow, arcing across the relay contacts limits the voltage created by the inductance of the coil. Right? Excuse my ignorance, but that seems completely wrong to me. The first three sentences make perfect sense and are congruent with my understanding. The rest confuse me. If you would explain and possibly give some examples, you would be a lot more clear.

 

[Bob S]

If you decide to use two zener diodes, put them in series (back to back), because a zener is a diode in the forward direction. You could put a resistance in series with the two zeners. The more resistance you put in, the shorter the L/R coil current time constant. But at the same time, a shorter time constant means a higher V = L dI/dt voltage, which could damage an npn collector.

The real value of a shunt across a relay coil is to quickly dissipate the energy stored in the coil inductance; ½LI². The larger the voltage drop across the shunt (IR drop), the faster the coil discharges.

I recall getting a substantial shock (over 50 years ago!) running a dc relay at the end of a long coax cable on a 6 volt battery. It must have been over 100 volts. Also the high dI/dt couples into other circuits.

Bob S

 

[Carl Pugh]

A relay that has a transorb across it's coil will operate faster than a relay with a diode across it's coil.
A relay that has a transorb across it's coil will have a higher voltage across the drive transistor than a relay with a diode across it's coil.

WITH NO TRANSIENT PROTECTION:
If the current in a relay coil is opened with contacts that have 1000 volt breakdown, then the energy in the relay coil is transferred to stray capacitance and the circuit oscillates.

WITH NO TRANSIENT PROTECTION:
If the current in a relay coil is opened with a transistor that has a breakdown voltage of 100 volt, then the voltage across the transistor goes to 100 volt. Most likely the transistor will be damaged.

Disclaimer: The previous statements are true in theory. In practice there are many things that affect the magnitude of the transient voltage.

 

[¡MR.AWESOME!]

Alright. So now I think I get it. I had to do a lot more reading for your responses to even make sense.

So I want the current to dissipate quickly to prevent arcing on the relay contacts due to the contacts opening slowly. The quicker the current dissipates, the higher the voltage will be. However, I have to keep the voltage below the maximum of the driving transistor or else the transistor will be destroyed.

The transistor I'm using (BD681) has a max collector-emitter voltage of 100V. That means I need to keep the voltage spike below 100V, right? According to Carl, the transistor will see the source voltage+the transorb voltage+0.75V. With a 12V source I would choose a transorb with a Voltage Breakdown below 87.25V, right? I know I would really want want to keep the voltage lower for a safety margin.

Now, if I was driving the relay straight from the PIC, what would I do? The pins on the PIC can't take more than about 6 volts. Does that mean I would use a transorb that would keep the voltage under 6 volts? I assume that just isn't done and that a transistor is always used to drive a relay. Please let me know.

I have one last question. I have one relay that will be powered by an unknown, 12V power source. I do not know if it will be a transistor or a relay or something else that will switch it. Is there a generalized, safe voltage I should clamp the spike to, or should I use a method other than a transorb?

Thanks for your help so far.