Relay Contact Power Losses?

[AcHmed99]

I’m trying to figure out whether it’s more efficient to use two cheap IGBTs low Vcesat as an SSR or just use a relay. While power dissipation in the coil is easily calculated I cant seem to find anything on the actual contacts other then the initial resistance which ranges anywhere between 0.050 to a little over 0.1 oHms.

I’m just using it as a bypass for an inrush limiting resistor. The load is worst case 8.5Arms a little over 750W resistive.

With two low Vcesat IGBT’s (placed after the rectifier) using spice to crunch the numbers I’m getting a little over 4W total conduction losses.

If I assume a 0.1oHm contact R for a relay losses are in excess of 7W.

Do they not spec actually contact resistance at working loads that the relay is rated for?

[Kremmen]

The efficiency is similar, there won't be a dramatic difference between a good relay and a good FET. I have made several kinds of inrush limiters over time, and my choice is the relay. I have had good luck with this one and its relatives: http://www.digikey.com/product-detail/en/G8P-1A4P-V%20DC12/Z3309-ND/1813598.
My reasons are that there is no great difference in efficiency; as long as you keep to the max ratings of the relay, you don't need to worry about direct cooling because it is not done for relays; a relay is basically more robust component than a FET. given that a FET can take lot of punishment before dying, as well; relay contacts are potential free and insensitive to polarity; driving circuit is not associated with power circuit in any way so coil driver can be easily implemented from galvanically isolated circuits.
Not all reasons may be significant in every case and not for you. I have found it easiest for myself to use the tried and working circuit for all cases. I don't use NTCs for instance due to the issues associated with them.

[SeanB]

Relay contact losses at 10A for a 16A or 20A relay are pretty low, I have not seen undue heating from current from the contacts, so definitely they are pretty low, provided you use a relay rated for mains use.

[Kremmen]

Have you had any issue for say using them for a wide load? The load may be in standby or draw only a couple of watts for extended periods. So I was wondering if that could be an issue with oxidation build-up on the contacts. The lowest minimum current relays I could find are 0.1A that could be 10-14W and sometimes the load may not draw that.

I agree with the NTC’s issues it’s why I am not using them.

Low current can indeed be problematic for power relays due to the contacts not contacting properly. Personally however,  i have never had issues with the relays i linked. Omron is a quality mfg and that series like many others has AgCdO i.e. silver cadmium oxide contacts. Provided they see high current use reasonably regularly, i don't think you can do better than AgCdO. For some cases where (separate) contacts need to carry low and high currents in one relay, there may be an option to use gold plate on the contacts (plate, not wash). At least several mils. The plating will burn off from the high current contacts exposing the material below - hopefully AgCdO - while the low current contacts make solid connections over the gold plating. Unfortunately this scheme does not work if you alternate between high and low currents. Then the cadmium is your best bet.

[SeanB]

Using a relay in an inrush limiter it will always be switching a low current and not handling inrush at all, and will not be handling the switch off arcing at all. Thus a simple silver contact in a sealed or semi sealed relay will essentially last almost forever, and will not suffer arcing or excessive wear. Even the Omron linked to will work provided you have a 12V source to power the relay.

I have used low power signal relays for this with no problems, a gold flashed 4 pole unit with all 4 poles in parallel, and it has been running for decades. It was the only 48V unit I had to hand, and was rated at 2A per contact.

[penfold]

If you're trying to figure out from the perspective of efficiency whether to use a device which dissipates 7W or one which dissipates 4W, in a system consuming 750W, being a difference of 0.3% - hats off to you're measurement resolution for efficiency.

Assuming that the system won't spend its entire life at maximum power, assuming a resistive model for the contact losses from the relay then these will be much lower at lower current, whereas assuming constant voltage drop on the IGBT then these will only drop linearly with current

[Kremmen]

Using a relay in an inrush limiter it will always be switching a low current and not handling inrush at all, and will not be handling the switch off arcing at all. Thus a simple silver contact in a sealed or semi sealed relay will essentially last almost forever, and will not suffer arcing or excessive wear. Even the Omron linked to will work provided you have a 12V source to power the relay.

I have used low power signal relays for this with no problems, a gold flashed 4 pole unit with all 4 poles in parallel, and it has been running for decades. It was the only 48V unit I had to hand, and was rated at 2A per contact.

Like i wrote i don't use NTCs because of their thermal cycling limitations. If you do use them, then repeat starts _will_ switch high inrush currents, otherwise of course not. Anyway, i use dual relays and normal resistors and in that scheme the first relay will always switch the full inrush. That's why the choice of contact material. And no, pure silver will not do anywhere as well as AgCdO, that's precisely why the latter is used. As to the Omron, i say that _especially_ that one will work. Of course anything will work for a while, but your gold flash in a power relay is useless. I promise you it is long gone if it switches any appreciable power at all. Flash will tend to rub off over time even during normal contact swiping operation, never mind an actual current over the contacts. That of course does not in itself mean that the relay cannot still work properly.

[SeanB]

Gold flash over german silver contacts, so long gone and running on silver contacts. They were gold flashed as signal switches or power switches, just you could not go from power switch to signal switch after a few cycles, but otherwise good relays.

[Kremmen]

Sure, silver is traditionally used as contact material and does work, of course. I don't mean to criticize your choice of component, just calling the facts as i know and understand them. The added cadmium in the contacts does provide less erosion and less tendency for welding. So why not specify it when it is indicated?

[SeanB]

I have Boveri relays with standard german silver contacts that have done well over the rated million cycles by at least a factor of 10, they generally do not fail because of contact problems except those that are run at full power of 16A, but from the cores themselves failing mechanically.

[Kremmen]

I have Boveri relays with standard german silver contacts that have done well over the rated million cycles by at least a factor of 10, they generally do not fail because of contact problems except those that are run at full power of 16A, but from the cores themselves failing mechanically.

Heh, it is a small world. I used to work for the Finnish Stromberg, now part of Asea Brown Boveri or ABB for short. Stromberg was and ABB still is a major mfg of power electronics, switchgear, relays and contactors of all kinds. Most of the current contactor types seem to be descended from original Stromberg designs but surely there is some Boveri there too. All quality stuff anyway. In my job we used to compete with Brown Boveri for paper machine drive systems. In my time we used to win most cases easily but the world is a different place now. Probably the Chinese take most of the deals these days...

[SeanB]

Relays almost always give a worst case value, often it is well below this. Only switchgear I have that heats up is handling 30A or more in a distibution board, and that only has a 30 degree rise in ant case.