Automotive relays

[raptor1956]

We all know that DC is a harder beast to turn off with a switch than AC is given the 100 or 120 times a second that AC turns itself off and the tendency for DC to arc and destroy contacts.  So, my question is -- how do the relays in cars and trucks handle the sometimes pretty high current switching that can be well past 25A and sometimes much higher?  How do automotive relays do what other relays can't?


Brian

[ataradov]

All the automotive relays I disassembled had nothing special in their design. Contacts are made out of copper and that's mostly it.

Some industrial high energy relays use a spare set of contacts that connect first and disconnect last, and are allowed to ark. The main set of contacts actually carries the current.

[Paul Moir]

Don't forget they only have to survive for something like 5000 operating hours.  The actual number of operations on something frequent like a cooling fan motor relay is not all that high.

[raptor1956]

OK, I have a project using a digital temperate control board with relay output to control a resistance heating circuit of about 4A and 12VDC in an automotive application so the voltage could actually be a bit over 14VDC.  I can program in enough hysteresis to reduce the switching frequency to, perhaps, once every few minutes or perhaps even longer and I doubt the lifetime would need to exceed more than 2000 hours and probably as little as 500 hours.  Basically making a LiPo battery heater to keep them in the range of 25C-35C for use in the winter.  I just need to keep them warm enough until they are needed and then the power draw in use should keep them well above 20C on all but the coldest (<-20C) of days.  Found a cheap control board that runs from 12VDC with a relay rated at 20A for both AC and DC, which is a bit dubious, but given the relatively low cyclic rate for the relay and relatively low DC resistance current I feel safe to avoid adding a n-channel MOSFET to drive it -- the relay should be sufficient.


Brian

[max_torque]

I lot of automotive loads are, or certainly used to be, mainly resistive (filament lamps, resistive heaters etc), meaning dumb relays worked ok.

As more and more loads become inductive then relays are getting replaced with solid state switching.

Automotive relays generally have a limited life time (number of switches) but carry enough current that contact damage can be self healing to a fair degree.  Breaking inductive loads does result in arc damage to the contacts of course.  One option, if you know you will be doing a lot of switching is to bypass the contact with a suitable resistor and/or TVS etc.  If you choose a breakdown voltage for your TVS above the supply voltage then it will just kick in to relieve the worst inductive kick, and limit arcing across the contacts.  Of course, this does slow the turn off of the load, because current can continue to flow.  This may or may not matter for your load (if it's a heater, then it probably doesn't)   A large value resistor can also be used in series with the TVS, or even an inductor.  Plenty of options exist, and it's pretty simple to experiment or simulate to find a good solution

[BradC]

Breaking inductive loads does result in arc damage to the contacts of course.  One option, if you know you will be doing a lot of switching is to bypass the contact with a suitable resistor and/or TVS etc.  If you choose a breakdown voltage for your TVS above the supply voltage then it will just kick in to relieve the worst inductive kick, and limit arcing across the contacts.

Another option is a ceramic capacitor across the contacts. I still use that one with troublesome inductive loads in cars (relays, fans, A/C clutch coils). A decent back EMF diode and a small cap can do wonders.

[calexanian]

I have asked this very question of the relay manufacturer we use. btw relay manufacturer is a misnomer. most use the same factories in china and create their own combinations, but run on the same machines their competitors use, but that's another conversation. There are some differences. for a given relay configuration automotive relays will have a stronger spring to separate potentially wetted, or stuck contacts. Also the larger ones have longer arc break distances. The primary difference is in the contact material though. Since they are not to worried about low current operation they can use harder alloy contacts. Some use a copper chrome type alloy that is also used in resistance welding jaws and such. Every "Manufacturer" has their special magic blend.

Other than that to get the qualifications for automotive road use there are additional tests that must be performed and there are a few footprints that are only found in automotive use to further differentiate things.
But in the end it just has to make and break.

Hope that helps.

[electrolust]

related story.

i had a car (race car) that had a flaky master switch.  at the same time, the master wouldn't actually disconnect power to the ignition system (bad diode on alternator).  so the car actually ran, until it didn't.  you didn't actually notice the switch flakiness, what you noticed was that the fuel relay would burn out very quickly (tens of minutes).  went through quite a number of relays and hours of diagnostics to ultimately find the master switch -- something you normally consider to have ultimate reliability -- to be faulty.

portable scope would have found it.  the switch would only act up under the mechanical vibration of on-track activity, so it was undiagnosable in the garage or paddock.

[brainwash]

I've written in another topic on this forum, there are several replacements for relays and they cost as much, or less. The automotive manufacturers have been using them almost exclusively since the 2000's.

On the parts side, I have quite a few SMD automotive high-side smart MOSFETs (relay replacements) in my parts bin:  BTS6163D, BTS5012SDA, BTS6143, BTS4443, VN750, BTS247Z, ... I also have a couple of those with current sensing (AUIPS7145R, VND5E025MK-E, VN7140, ...).

Their prices range from 0.7$ to 2.5$. They all have current limiting and back-EMF protection and can be paralleled.
Above that level (200A) there are the cranking solenoids, (still) cheaper to use than solid-state solutions.

BTW, 4A@12V is perfectly fine for a relay, thousands of actuations. Omron ( https://www.omron.com/ecb/products/search/?cat=1&did=4&prd=dcpower&lang=en ) quotes 10k actuations for their smallest power relay at 15A. Assuming 8 actuations per hour, 8h usage per day, 50 usage days per year, you would get at least 3 years out of one relay if you drive it at its rated maximum (15A). Much more if you drive it at 4A.

[calexanian]

For this application I would use an off the shelf automotive horn relay or headlight relay. They are made for this kind of service and are cheap as online.

[raptor1956]

I've written in another topic on this forum, there are several replacements for relays and they cost as much, or less. The automotive manufacturers have been using them almost exclusively since the 2000's.

On the parts side, I have quite a few SMD automotive high-side smart MOSFETs (relay replacements) in my parts bin:  BTS6163D, BTS5012SDA, BTS6143, BTS4443, VN750, BTS247Z, ... I also have a couple of those with current sensing (AUIPS7145R, VND5E025MK-E, VN7140, ...).

Their prices range from 0.7$ to 2.5$. They all have current limiting and back-EMF protection and can be paralleled.
Above that level (200A) there are the cranking solenoids, (still) cheaper to use than solid-state solutions.

BTW, 4A@12V is perfectly fine for a relay, thousands of actuations. Omron ( https://www.omron.com/ecb/products/search/?cat=1&did=4&prd=dcpower&lang=en ) quotes 10k actuations for their smallest power relay at 15A. Assuming 8 actuations per hour, 8h usage per day, 50 usage days per year, you would get at least 3 years out of one relay if you drive it at its rated maximum (15A). Much more if you drive it at 4A.

Thanks for the feedback everyone, in my use case with relatively low resistive loading and the ability, owing to the huge thermal inertia, to have a low cyclic rate, the lifespan of the relay should be fine as is.  I thought about having the relay control a mosfet but that's not needed in this case.  But, the idea made me question the way automotive relays function and how they differ from general purpose relays.  It make sense that solid state devices will begin to take over and thus eliminate the arcing problem altogether. 

Again, thanks for the feedback.


Brian

[arclight]

You can also add an MOV of an appropriate voltage in parallel with the output contacts.  That will help control the back EMF and extend the contact life on an inductive load.  I have seen this used on alarm/access control systems.

Arclight

[brainwash]

Thanks for the feedback everyone, in my use case with relatively low resistive loading and the ability, owing to the huge thermal inertia, to have a low cyclic rate, the lifespan of the relay should be fine as is.  I thought about having the relay control a mosfet but that's not needed in this case.  But, the idea made me question the way automotive relays function and how they differ from general purpose relays.  It make sense that solid state devices will begin to take over and thus eliminate the arcing problem altogether. 

In the old cars (have had access to quite a few) the relays were pretty standard. They could have been working for 20+ years or could have failed after 5 years. A few of them had double contacts, but otherwise nothing special.
Industrial relays might be a different beast. For example the ones used by Caterpillar are rated to millions of actuations, and so are the switches. But it's just a matter of over-speccing and tighter tolerances (quality control).

The automotive industry is mostly cost-driven and has only a baseline reliability rating, so they will use at one time whatever provides the most bang for the buck. Right now it seems solid-state is the way to go, at those quantities. The switches I enumerated are actually used (mostly) in place of headlight relays. The downside is that now everything is integrated and a module performs a lot of functions, including diagnostics. And then you would need to re-code a replacement module. Relay changes could be done by anyone but at that time it was still people like us that did it, so overall not much has changed, the skill pool is limited just the same.