TVS selection criteria

Go To Last Post
17 posts / 0 new
Author
Message
#1
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Morning all,

I'm suffering the usual weekend brain death!

What are the considerations when selecting tvs diodes? Is the thinking the same as selecting a regular zener?

The answer is probably painfully obvious, but I'm struggling to see it.

Example: say I'm switching a 5 volt coil relay with 100ma turn-on current via a transistor from the mcu and the transistor is switching with 10ma. The load the relay is switching is for example 240 volts/ 8 amps. To protect the mcu I would like to add tvs diode from the mcu signal to ground. What value/ratings should I be considering?

--greg
Still learning, don't shout at me, educate me.
Starting the fire is easy; the hardest part is learning how to keep the flame!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Are you suggesting putting a tvs diode across the relay coil? The only advantage might be slightly faster relay release time versus using a standard diode. One would normally use a tvs diode on signals to go out to the real world.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:

Are you suggesting putting a tvs diode across the relay coil?
No. To protect the mcu I would like to add tvs diode from the mcu signal to ground.
My thinking is, if the relay fails, 240V shoots (potentially) through the transistor and thus into the mcu.

In the example (ignoring all values) it's D2 as a TVS Diode I'm wondering about.

Attachment(s): 

--greg
Still learning, don't shout at me, educate me.
Starting the fire is easy; the hardest part is learning how to keep the flame!

Last Edited: Sun. Oct 16, 2011 - 02:11 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Protecting against flyback from a relay is best done with a normal diode. The flyback voltage can be on the order of several hundred volts and appears when the coil's magnetic field collapses. In other words, when you turn it off. The voltage spike is a negative going spike relative to the coil energizing voltage. The diode is connected so that the anode goes to gnd and the cathode to the high side of the relay coil. While the relay is energized no current flows through the diode as it has a high reverse breakdown voltage. When the coil voltage is turned off, the negative going spike is clamped to within a diode drop as the diode conducts in the forward direction.

A TVS diode works by exploiting the zener/avalanche breakdown phenomenon. You select a TVS diode based on the input/output range of the port you want to protect, and connect it just like the relay protection diode. For example, if you want to protect a port with a range of 0 -5 volts you would select a TVS with a breakdown voltage of about 5 volts. When a spike exceeds that value in the positive direction the TVS diode breaks down and conducts in the reverse direction clamping the input to about 5 volts. If the spike goes below gnd then the protection afforded is just like a normal diode.

You have to be careful if you use a TVS to protect against flyback from a relay. If your relay energizing voltage is say, 12V, and you try to use a TVS with a breakdown rating less than that it will conduct when you try to energize the relay. That's not what you want and you'll probably kill the TVS or fail to turn on the relay. You need to use a TVS with a breakdown voltage greater than the relay energizing voltage so that the TVS diode does not conduct in the reverse direction (like it's designed to do) with the relay energized. When you turn the relay off, the negative going spike is caught and clamped by the TVS as it conducts in the forward direction just like an ordinary diode.

So, you can use a TVS diode to protect against flyback, but the protection it gives you is the same as an ordinary diode. And perversely, you have to protect your circuit against the special characteristics that make a TVS diode valuable in the first place, namely a well characterized reverse breakdown specification, which you must be sure to avoid triggering!

Greg

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

dalpilot: Thanks for that, but its not flyback I'm worried about. In my example I'm switching mains voltage, and concerned about a failure of the relay, not flyback per'se.
I've added a little diagram to explain where the TVS would be placed. (D2) forget the values in the schematic, I just knocked this up quickly to explain what I'm asking. The flyback diode (D1) is already in place in my design.

According to your description (thanks) D2 should breakdown at 5 volts? (or slightly higher?), but what would the power rating be, given I want to protect against a relay failure (potentially 240V @ 8 amps)... somewhere in the region of 2kW?

--greg
Still learning, don't shout at me, educate me.
Starting the fire is easy; the hardest part is learning how to keep the flame!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

TVS diodes are designed to protect against transient events. There's no way a TVS will protect your circuit if your relay fails in some way that puts 240 volts across the TVS protected port for more than a few milliseconds at most. The TVS will fail as it can't handle that much energy. If there's a big enough resistor between the TVS and the 240 volts it might survive, but I'm not sure how you calculate the length of time it could withstand it. Typically I've seen specs which show the max pwr the TVS can survive given a 2s pulse. The ones I'm using specify 300W for 2s. But, I've not seen data that characterizes their steady state power handling ability, which I suspect is tiny.

I don't know a whole lot about relay failure mechanisms, but do they typically fail so that the switched voltage gets applied to the coil? I'd consider that pretty rare.

Greg

Last Edited: Sun. Oct 16, 2011 - 03:12 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:
I'd consider that pretty rare.
So D2 in my example is pretty useless then, given that D1 already handles the flyback. Perhaps I'm trying to implement the wrong tech? Would an opto-isolator afford me any greater protection in this scenario?
As already noted, I'm concerned about protecting the mcu, as it's (ultimately) going to be driving some seriously heavy gear. I need to effect a complete shutdown of the attached hardware if any fault develops. As much as is possible, I'd like to implement protection in the circuits carrying signal data, such that the mcu can report (for example) 'Relay 1 failure...', rather than a complete meltdown of the controller.

--greg
Still learning, don't shout at me, educate me.
Starting the fire is easy; the hardest part is learning how to keep the flame!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:
So D2 in my example is pretty useless then, given that D1 already handles the flyback.

Not at all. It's a good idea to have it, as there certainly other transients which the TVS will protect against. I just don't think it can protect against the 240 volt switched circuit getting into to it unless R1 is "big enough." And I don't know how to figure out what "big enough" is since the steady state pwr handling capability of the TVS is not stated, or at least I couldn't find it with the devices I'm using.

Greg

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

You could make the avr completely opto-isolated from the power stages. If the relay arcs and welds together, there will be horrenous ground current for a few ms till some fuse blows. This might raise the ground above vcc and zap the mcu. I think lots of motor controllers and inverters have opto isolated gate drives. Good idea, but a little more expensive and complicated. Extra parts.

Imagecraft compiler user

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

dapilot: Thanks again... I'll keep it in. I'll bet on 2kW to be safe (granted, it won't handle 240v for very long!) Incidentally, the relay is switching a 'contactor' (is that the right word?) which will itself be switching 3 phase at a far higher potential).

Bob: opto-isolating the drive stages seems pretty sensible also. I'll incorporate this after the tvs. Extra parts is not a concern... I'm not paying for the components, just the design and programming. There are only 50 - 100 of these going out each year, and the total cost of the controller will be a minor fraction of the completed 'unit'. (This assumes I can design a system 'robust' enough to handle stupid operators, any and all foreseeable screw-ups, let alone the ones I can't foresee!). Isolation, Isolation, Isolation!!!

--greg
Still learning, don't shout at me, educate me.
Starting the fire is easy; the hardest part is learning how to keep the flame!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

It seems to me that IF one were to use the TVS then it would be micro, reverse polarity protection diode, zener (TVS), resistor, FUSE, transistor.

When the Zener (or TVS) conducts, the (fast blow) Fuse blows, limiting the power which needs to be dissapated by the TVS.

The opto is a much better solution.

JC

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I've seen faults melt an opto and zap the low voltage side, but I've yet to see a relay fail the same way. Contactors usually have greater creepage distances. Another thing of note, DC coil contactors are usually more expensive than their AC coil counterparts so it may be more economical to use a smaller relay controlled by the micro to switch the AC to the contactor. Two levels of isolation to boot. Another advantage is that you can size the contactor to suit.

If the relay did fail in the manner that you suggest, I dare say there would be more problems than a TVS diode alone could solve.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:
so it may be more economical to use a smaller relay controlled by the micro to switch the AC to the contactor.
Exactly what I'm planning.

Thanks JC, I'll look into that.

--greg
Still learning, don't shout at me, educate me.
Starting the fire is easy; the hardest part is learning how to keep the flame!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I would never try to use a TVS as a relay coil snubber. It is simply not the right tool for the job. Consider what you need to happen:

1) Relay coil energized. The snubber has to take little or no power in this direction. A normal diode is easy. You simply connect it so that it is reverse biased. What would you do with a TVS? If it is a bipolar device, then you would need to choose one with a breakdown voltage greater than the supply voltage. If it is a unipolar one, you have to connect it in breakdown direction, and the voltage, again, has to be larger than the supply voltage.

2) Relay coil is de-energized. You need to prevent the relay flyback voltage from rising so high that it damages the coil control switch (often a transistor or FET). With a normal diode, it simply forward conducts, limiting the flyback voltage to less than Vcc + 1V. With a bipolar TVS, the flyback voltage is Vcc + Vb (Vb = Vbreakdown for the TVS). This will likely be more than 2*Vcc (since Vb > Vcc, from the first requirement) and may require un-necessarily large breakdown voltage requirement on the switch. If it is unipolar TVS, then it forward conducts, just like a normal diode.

So, with a normal diode, you pay for a normal diode and get a flyback voltage that is, at worst, only slightly higher than Vcc. With a bipolar TVS, the flyback voltage will be more than 2*Vcc (since Vb > Vcc to avoid conduction when the relay is activated); this will give a controlled max, but much larger than necessary. With a unipolar TVS, you gain nothing over a standard diode, and pay more for the privilege.

Jim

Jim Wagner Oregon Research Electronics, Consulting Div. Tangent, OR, USA http://www.orelectronics.net

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

What is this fixation with me using a TVS as a flyback diode? I've said repeatedly that that is not the intention. The TVS is D2 in my diagram.
Thanks for your explanation Jim, but I fully understand how to implement the relay protection diode. Go back over the thread again to see what it is that I'm concerned about.

--greg
Still learning, don't shout at me, educate me.
Starting the fire is easy; the hardest part is learning how to keep the flame!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

OK, my error.

I would put a series resistor, say a few hundred ohms, ahead of D2. Now, the power/energy requirements on D2 are MUCH lower. Yes, the resistor may have significant power dissipation needs, but that is a lot cheaper than diodes.

Further, you don't have to worry about the elevated breakdown voltage at high currents, because there are no high currents.

This gives you a more controlled situation, and everything will be happier when the fault hits.

Jim

Jim Wagner Oregon Research Electronics, Consulting Div. Tangent, OR, USA http://www.orelectronics.net

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Thanks Jim.

--greg
Still learning, don't shout at me, educate me.
Starting the fire is easy; the hardest part is learning how to keep the flame!