Need help with relays

[Stephen Durr]

Can someone recommend a reliable 15V DC relay that can handle up to 20 amps continuously for periods lasting up to several hours?

I'm having trouble with the reliability of some relays in a custom race car application.  There are several circuits that have high current loads that are supplied by relays that are controlled by SPST toggle switches.  When I flip a toggle switch ON it closes the relay's control circuit energizing it's coil which closes the relay's internal output circuit supplying the load.  When I flip the toggle switch OFF it de-energizes the relay coil and opens the output circuit turning off power to the load.

All of the loads in this car are designed to operate with input voltages in a range between 12 and 16 volts DC.  The car's electrical system consists of a 12 V battery and an internally regulated alternator that supplies between 13.3 V DC and 14.75 V DC depending on rotational speed (RPM).

Some of the high current circuits in this car have loads that are used infrequently or only periodically and are in operation for short periods of time.  Other circuits have loads that are in operation continuously, meaning 100% of the time that the cars engine is running.  I have had trouble with the relays for the continuous duty circuits; they seem to go bad after some time.  I don't have the same problem with the circuits that are operated infrequently or only periodically; those relays don't seem to go bad.  All of these circuits are using the same style of relay from the same manufacturer.

Some examples of both types of circuits are listed below:

Examples of the infrequent or periodically operated circuits (these relays don't go bad):

• Starter motor circuit, operates only while cranking/starting the engine, can draw up to 50 amps for a few seconds
• Cooling fan circuit, operates only when engine coolant temp is higher than set-point, can draw up to 40 amps for periods usually lasting less than two minutes but sometimes up to 15 minutes or more
• Transmission brake circuit, operated very infrequently, can draw up to 5 amps for periods usually lasting a minute or two
• Line lock circuit, operated very infrequently, can draw up to 5 amps for periods usually lasting between 30 and 60 seconds

Examples of the continuous duty circuits (these relays go bad after some time):

• Fuel pump circuit, in operation 100% of the time that the engine is running, usually draws about 10 amps but can draw up to 30 amps for short periods depending on fuel demands
• ECU circuit (engine control unit), in operation 100% of the time the engine is running, draws anywhere from 1 amp to 16 amps depending on operational demands
• Ignition controller, in operation 100% of the time the engine is running, usually draws about 8 amps but can draw up to 13 amps for short periods depending on operational demands

As I said, I have used the same part for all of these circuits.  The part is an integrated power relay module manufactured by Cooper Bussman.  The datasheet is here:
http://www.waytekwire.com/datasheet/46094.pdf

The power relay module incorporates a relay manufactured by Song Chuan.  The datasheet for the relay is here:
http://www.songchuanusa.com/wp-content/uploads/897.pdf

On the datasheet for the relay they specify expected life in terms of number of operations vs current.  To me that indicates that the life of the relay is related to the number of times it is switched on/off and the amount of current supplied.  In my application the relays that go bad are switched on/off very infrequently.  Basically I turn them on once before starting the engine and they stay on for the entire time the engine is running until I turn the engine off.  So I am not subjecting these relays to lots of switching cycles.

When one of these relays starts going bad it's output voltage begins fluctuating or dropping and then eventually it just doesn't provide any output voltage.  This usually results in the engine dying, which can be quite dangerous if the vehicle is at speed and surrounded by other vehicles that are also at speed!

Do I have to just accept that relays go bad after some time if operated continuously?  Or does someone manufacture a relay designed for this type of operation?

[woodchips]

First comment is that the link supplied for the relay isn't a solid state relay, it is a simple mechanical relay.

That said I would have expected the relays used to be quite happy carrying 20A at 12V indefinitely. Note that the 12V version is rated at 70A but the 24V version is only 25A, I would have said that the 25A is believable, the 70A isn't. All the current is passing through one 1/4" crimp terminal, they simply won't carry that sort of current. There is no reason for the contact rating to be downgraded with contact voltage, the losses are current, not voltage.

When you dismantle one of the bad relays what has gone wrong with it? Presumably the contacts are all burnt up?

Seems to me that the relays are being used within a sensible current limit, so must be something else. How hot do they get? Are they positioned right next the radiator, exhaust etc because high temperature increases resistance, might make the contact pressure reduce due to inadequate coil current which would start to burn the contacts. Vibration might cause a similar effect, particularly the crimp connection to the relay terminals.

[eetech00]

First comment is that the link supplied for the relay isn't a solid state relay, it is a simple mechanical relay.

That said I would have expected the relays used to be quite happy carrying 20A at 12V indefinitely. Note that the 12V version is rated at 70A but the 24V version is only 25A, I would have said that the 25A is believable, the 70A isn't. All the current is passing through one 1/4" crimp terminal, they simply won't carry that sort of current. There is no reason for the contact rating to be downgraded with contact voltage, the losses are current, not voltage.

When you dismantle one of the bad relays what has gone wrong with it? Presumably the contacts are all burnt up?

Seems to me that the relays are being used within a sensible current limit, so must be something else. How hot do they get? Are they positioned right next the radiator, exhaust etc because high temperature increases resistance, might make the contact pressure reduce due to inadequate coil current which would start to burn the contacts. Vibration might cause a similar effect, particularly the crimp connection to the relay terminals.

I agree. The relay doesn't appear to be solid state as shown on the Relay spec sheet.
But...is the coil failing or the contacts?

Need to inspect/test a bad one to find out..

Also, if it is a standard relay, ensure you are using a diode (1N4007) snub around the relay coil.

eT

[croyleje]

Hello Crydom is probably the biggest out there and makes great products also easily available from most supply houses.  They are a little pricey but never had one go bad on me also if your having that much trouble with solid state relays make sure there isn't some overload happening.  And remember if there is any other coils in the circuit @relays, inductors, ignition coils, ect. when those power off and the magnetic field collapses they can induce a high voltage in the line consider some type  of protection. 

Jason

[theatrus]

I'd imagine any automotive relay would be up to the task. Sticking on is usually a sign of the contacts fusing due to arcing or overload. DC is more demanding as there isn't a low current spot to switch the relay as in AC, so make sure the relays are rated for DC current.

[Stephen Durr]

Thanks for correcting me, I guess these are not solid state relays, just regular electro-mechanical ones (I removed that reference from my original post so as not to confuse others).  When they turn on/off you can hear a solid clicking noise.  Unfortunately I never cut one open for inspection, I just tossed the bad ones out, but if another one goes bad I will save it and cut it open.  These relays are not in the engine compartment they are in the passenger compartment underneath the dash, so they are not subjected to heat or moisture.

[max_torque]

You definitely need to inspect one of the failed relays to establish the root cause of the failures!

Generally, relays (and other switches) that don't get "switched" very often suffer from contact corrosion and brinnelling.  For circuits without in-rush protection, like fuel pumps etc, you can get carbon deposits build up and increase the contact resistance.

Also, high frequency vibration in an automotive environment can cause a lot of issues with contact damage at a macro level.


I used to use BOSCH normal 25A relays for just about everything in numerous competition cars (before i went fully solid state) and never had a problem with those.

Another good idea is to buy those little stick on "temp tabs" that will tell you the maximum temperature a component has reached.

None of your loads should be enough to cause over heated contacts in those relays, but it's worth checking the new / used contact resistance to work out the rough thermal load into the devices.

[Dave Turner]

Check that that your power supply can supply enough 'juice' in the worst case for all circuits. It's possible that under full load there isn't enough to keep the relays fully on which together with vibration can cause contact burn out particularly in those that are meant to be on continuously.

As suggested previously you've got to take 'em apart to see what's wrong.

[Kjelt]

Also measure the contact resistance, at 20A each 10mOhms is a 0,2V drop , I would seriously look into a power DC SSR if I were you.

[Swake]

Use normal car relays. Preferably use quality stuff of a well known brand such as Bosch. And respect the minimum cabling gauge for the amps you draw. same goes for the plugs.

You'll find these relays mainly in 4 different physical flavors:

Micro Relay: small rectangular, about half the size of 'normal relay'
Normal relay: Typical cubic version.
High current: Same cubic shape but with thicker/larger connectors. This is the one your need.
Very high current: much bigger, not used very often.

It is a standard in size and pin-out, many brands sell them and probably all car mfg's use them. Some with build in suppression diodes and/or signal LED's. Holders are available in all formats and sizes, also waterproof.

Search google for 'car relay'

Eventually salvage them from German brand cars as I do.

I regularly drive 4X4's off road in harsh conditions and completely rebuild the electrics of my own cars to my own specs.

[exxon]

A race car with mechanical relays? So strange...

Race cars demand for special SSR like the one in the picture. We developed it for FIA approved ECUs few years ago. It has inside a BTS555 in TO218AB-5 case and few other components. We tested this product at 150 A continuous current load with practically zero defect under really harsh conditions.

In my opinion, the reason of the failures you experience is vibrations. In race cars moving contacts are... simply forbidden. Even push buttons and toggle switches have to be SS.

[woodchips]

The problem with solid state relays is voltage drop, which will always be more than from a simple mechanical relay.

At 150A a transistor is dropping about 0.2V collector to emitter, much worse if a darlington. That is an awful lot of watts which means that the SSR must have a heat sink, which in a body shell made from steel isn't so easy. Yes, vibration could be a problem but seems easier to fix, rubber mount, that the heat losses.

Can't accurately scale the photo, but possibly only 30mm wide?

[max_torque]

I suspect that simply using proper quality relays will remove the issue......... ;-)


(also, using lower "rated" relays will help with contact cleaning, like a typical 20A Bosch relay for example)

[jlmoon]

Have to agree with Max-Torque, used those 4 or 5 terminal Bosch relays for years in custom applications.  They can't be beat for automotive / 12VDC applications.  Watch out for the knock-offs though they're out there.. beware!

[exxon]

I don’t state mechanical relays don’t work. I say they are definitely not the best choice for a race car. The only advantage a mechanical relay has over a SSR like the one in the picture is the lowest cost (not an issue for a race car). When you are on the ring, fighting for 1/1000 of a second, the last thing you want to worry about is a failing relay. I assure you are happy to spend few tens of bucks more.

Technically speaking, that beast has around 3 mOhm of Ron, lower than the most part of mechanical relays (especially after some tens of operations…), it is lighter and a lot smaller than mechanical counterparts (see the drawing). No arching when opening inductive loads, no wearing out. It is assembled on a heatsink to have a rugged case to be screwed on the chassis, but runs few °C over the ambient temperature (not at 150 A, obvious…). The aluminum case is capable to dissipate the big power converted to heat in case of short circuit without any damage to the electronics: yes, it doesn't die if shorted…

[woodchips]

Ah, the SSR is using a FET, so lower Ron. Looked up the data sheet and what intrigues me is how it carries 150A on two leads each 0.5mm by 1.5mm, 1.5sq mm, normal current capacity about 20A or so? Other data sheet limitation is the need to know the circuit inductance for breaking the current, sure that could catch you out!

An impressive device, must have a use for it somewhere.

[max_torque]

Typically, the limit current for a typical TO package lead is around 75Amps, assuming that lead is nicely coupled to suitably big pcb tracks.  With this sort of device i've always found making big enough tracks is the issue, and not the short power silicon leads themselves (anything over about 30A and you're gonna end up with some form of busbar to beef up your tracks)

Those smart ProFets are quite good, with built in thermal, over current, under voltage limits etc (although the over current limit cannot be relied on completely, as a dead short will eventually damage the device).  They are however quite expensive in low volumes, but as mentioned that is not much of an issue for pukka race cars ;-) The fact they also have a integral current monitor output is handy too.

The latest 7 lead or leadless SMC Mosfets can carry massive currents.  I've used only 4 in parallel for a 12v 500Arms inverter!

[exxon]

Most of us calculates the electronic components pin maximum current capability, the same way electricians calculate wire size for house electric wiring (say few amps per square mm). On the other hand, an uninsulated long round 1 square mm wire suspended in free air can carry more than 100 A before melting. For an electronic component pin, the max current density is between these two limits.

In the case of this SSR, the tiny rectangular section pin connecting the electronic component to the PCB has a very complex thermal behavior. It is so short that the draining of heat from the two sides plays a bigger role than surface dissipation. So, the most important role is played by the heatsink capability of the landing pads on the PCB.

In our design, the substrate is alumina and the component has its pins preformed in a way to allow surface mounting on a very large metal area, capable to spread the heat on the substrate. The whole assembling is then vacuum potted using high thermal conducting epoxy, used to remove heat from the alumina substrate and transfer it to the aluminum heatsink.

In facts, our biggest effort in the design of this system has been to squeeze out the maximum from the BTS555 by a smart thermal design. It paid.