How to Protect Relay Contacts

First, make sure that voltages and currents have actually the values you assumed. Don't rely just on expected values.

If everything is OK, and the circuit being commanded carries only DC, an antiparallel (i.e., oriented against the normal line voltage) diode across the relay contacts should suppress the arcing, which in most cases is responsible for contact damages. The problem you are experiencing sounds very much as due to arcing produced by a heavy inductive load.

If your relays command loads located outside the PCB, it shouldn't be difficult to put a diode somewhere.

As you suggest that these relays may be involved in safety issues, take into account that a so-placed diode, while suppressing the arcing, may slightly delay the off-time of the circuit being commanded, depending on its inductance (usually not more than a fraction of a second). So, test that delay before adopting a final solution. If the delay turns out to be unacceptable, insert a resistor in series with the diode. It will dissipate more rapidly the energy stored in the circuit at expense of a greater shut-off current (You are supposed to find empirically a reasonable value, which depends on load inductance and expected response time).

Finally, use diodes that fairly exceed the expected maximum current (perhaps, one rated for 6A).

Many thanks fernandotasso, I'll try this out.

Be sure you understand about the diode. It will be of little value across the contacts. It must be placed across the coil of the load.

The other thing I note is you state your relay switching capacity as rated right at the load it is switching. This gives you no margin. If I were designing for a safety application I would probably use a 2 amp 100 vdc contact rating to ensure longer life.

The one last thing you can do, and especially if you can't put a diode across the coil of the load, is to put a snubber across the contacts. It won't do you any good on contact closure, only over-rated current capacity can address that. But it will reduce arcing on opening. A snubber is merely a low value resistance and a capacitor of adequate voltage rating with enough capacitance to slow the rise in voltage across the contacts while they are opening.

correct

Arc suppression on the relay contacts:

http://www.multi-contact.com.sg/elesta/images/Technical%20Information.pdf

Oops! I just reread this thread and realized my BIG MISTAKE. You're right, mine was an unforgivable slip. So I rated it as off-topic. GA for you. My answer, aiming to a local solution, should've been placing a diode between the positive contact of the relay and the supply's positive, usually available at the PCB.

What really puzzles me is how that mistake could have been rated as a GA! Please, anyone who read this help me rate that answer as off-topic, so it cannot be misleading!

so why did u show L in series with R and not C ?

Bung a flywheel diode in reverse polarity across the load relay coils. 1N4001 is made to do this job, and they are cheaper than the chips in Broadmayne's fish shop (spare us some coppers for a cupper, Guv?). <splutter, cough>

Just don't go cheap like I did many years ago:

My church had the same problem with the key contacts in an electro-mechanical pipe organ.

I knew about diodes from a college computer interface project, and I knew I needed a bunch of them, so I bought some packs of little, glass switching diodes. Yes, of course, the 50-pack from R.S.

The crazy thing was -- and I cannot remember how the circuit would do this -- when a diode would fail, the note would continue to sound.

I was a little scared of the voltage, so my upgrade diodes were something like 1N4005 or 6 to get higher reverse voltage capability. That way, for not much more cost, the diode was good for whatever the inductor wanted to throw at it.

If the note continued to sound, the relay was probably welded shut. That must have really ruined the organist's Sunday.

The note continued to sound because the diode in parallel with the contacts (and in series with the solenoid) failed short.

Sounds like you need to use a Transorb (2 high current diodes common cathode connection) type of ac suppression across contacts or even better would be an H bridge contact configuration.

Solid state relays are the best choice since the have CEMF diodes inherent in their construction and the MOSFET(s) used provide excellent stand off isolation from inductive spikes.

Any coil type relay must have CEMF diode (Cathode connected to the + coil lead & Anode connected to - coil lead). I wound also use one for the "coil" of a solid state relay as well since it only needs to be a fast signal type diode like a 1N4148/1N914B or a low current schottky diode. CEMF stands for (Counter Electro-Motive Force).

Actually since he indicates he is switching DC he can use a DC (single diode) Transorb.

True, good point.

I assume the welding of the contacts is happening during the make and not during the break of the contacts. That being the case, the contact material of choice would be one of several AgCuNi alloys. They should be either self burnishing or offset cupped configuration.

The use of a non-linear resistance like that in a simple tungsten filament light bulb can be used to limit the dI/dT transient that is welding the contacts closed. This can be a very simple solution that will only work if one knows the maximum expected peak current for a short duration of time. "Quench-arcs" also work well and are typically rated in joules which is at least the right units for a dI/dT transient.

Take Fernandotasso's lead and check that the current you actually have is not in excess of the rated current of the contacts. Look at the rating/ spec sheet for the relays. You may well find they are 1A contacts for resistive loads but only a fraction (<10%) of that for inductive loads.

For the contacts to burn off completely it means there is quite a lot of energy there.

Do you know what the actual load is? Is there a possibility that it is a small servomotor? If so then the energy involved is related to the inertia of whatever is being moved. Can you give us a diagram of the circuit?

Regards

Chas

I used to help manufacture switches, both DC and AC, when Eagle Electric was still in business in the US.

AC switches have it easy. The current changes direction and goes though zero 120 times a second and engineers use that momentary lull, to help terminate the power with modestly sized contacts.

DC switches don't have the benefit of that. The "push" is constant and the arcing that is welding the points is evidence of it's resistance to termination.

Often the problem in a properly designed DC system is that old relays have grown sluggish and take longer to open than they should. This makes an arc of longer duration, which of course elevates the temps at the contacts making them more prone to melting and welding.

First thing I'd check is to insure that the power that energizes the relays is adequate. Voltage drops from loose connections, corrosion, fading power supply, etc will effect what's available at the relay and have an immediate effect on the speed with which it opens.

I agree with RCAPPER: The use of a properly sized capacitor will go a long way towards absorbing the energy that is melting the contacts.

There is nothing novel about this solution. Capacitors were used in exactly this manner to protect the points on the old Kettering-Coil ignition systems of decades past. And, they were very effective too.

I'm assuming, of course, that some other additonal device hasn't added to the load and expanded the current flow beyond the rated current capacity of the relays. I suggest you test that first.

Don't simply treat the symptoms. Find the cause.

Good luck and please let us know how it turns out.

L.J.

I`ve extended Qqbercis drawing for explanation. Be sure to use a solution like the following:

The diode protecting the coil and contacts is one thing, the other is to protect the contacts from burning which is done by correctly dimensioning the R1C1 snubber network.

Assumed your voltage Vs is 24V DC and the DC resistance of your inductive load is 100 Ohm then your current with a closed contact is 0,24 A DC.

C1 is dimensioned with a rule of thumb as about 0.1µF per Ampere of Load current which is 0,024µF in our example or, with an available value, 0,027µF or 27nF.

Lets further assume your relais allowed discharge current is 1 A you get R1=24V/1A=24 Ohm or, from E12/E24-Norm-Row, 27 Ohm.

Regards Uwe

Fortunately there are several choices, the mentioned diode is the most commonly used, another inexpensive option is to add a 0.0047 micro Faraday 500 volt disc capacitor in parallel to the contacts, this can be used in addition to the diode an will add protection to both the contact and the diode in both pole directions.

This suggestion does not apply to very light loads (obviously this is not the case) because it might couple enough energy to trigger electronic relays with high impedance input.

Since there are safety issues involved it is strongly recommended to use the service of an experienced engineer, he will probably replace the control system with a PLC or heavy enough solid state relays, life worth's it.

RF

In doing plant start-ups there were sometimes 90 relays in a control panel, after a year of intermittent failures, the manufacturer did a recall and replace of the relays. They were AC coils.

Another 125VDC control panel came without the diodes across the relay coils, had to add them in the field to stop failures.

You may also want to consider a "Solid State Relay". Since an SSR does not actually have contacts it should eliminate the problem.

http://www.newark.com/magnecraft/861ssr115-dd/solid-state-relay/dp/30M2664

Hello Gaz3110,

May I ask is this the first time you have found this problem, either in the same building or others?

Can I also ask if you work for the Elevator manufacturers? As it would seem they have done something wrong in the installation for any induction coil to be able to work like that?

It seems clear that the induced current is too much. But you have been given a potential remedy so good luck!

bb

We have elevator security systems in numerous buildings across a variety of elevator companies but have experienced the relay contact welding issue in only a few. The problem seems to lie with what we are given from the elevator company. Generally it is one 'common' per elevator and an individual switch wire per floor. A closed contact on our control relay 'secures' the desired floor. An open contact on our relay provides normal access. This scenario can change with elevator manufacturer. Our relays are part of a multi purpose panel with no choice of changing design or relay (apart from replacing failed relays as they are socket mounted on the pcb) and only a little room for adding additional components across the contact terminals if diodes or caps etc. are required.

Hello Gaz3110,

Just to say it must be to must power getting through which is welding the contacts together?

Though you want the contact to work you do not want them to work THAT well right? I cannot say what but, it sounds like you need something to prevent the power surge. I cannot recall whether this has been discussed in other posts?

This is something I would have thought should have been sorted over the years elevators have been used.

bb

In my opinion, (which may not mean anything to you, even though I've been in the field 30 successful years)

this is way out of your league. Contact Honeywell, Grinnel or an elevator servicer such as Thyssen

Hello Guest,

I was wondering why the OP was messing with an apparently ready fitted system IE the elevator, and trying to right a problem which should be the elevator engineers job to sort out?

Take care.....

bb

You didn't say but it sound like this is an existing installation that has been in operation for awhile. If this is a new phenomenon then check the wiring and the load because that would most likely the source of the problem. To literally burn off a contact requires a significant overload, even for a 1A contact at 24 VDC.

But more to the point, if the damage is because of an inductive load issue and the arcing that occurs when the load is turned off, you might put a diode across the load with the cathode of the diode connected to the positive side of the load and the anode connected to the negative side of the load (I don't know if the load is switched to common or switched to the positive supply.)

Hi RcEe,

In this particular installation it would appear that our security system relay in turn switches a control relay in the elevator motor room. All we receive from the elevator company is a pair of wires that wire into their elevator controller. Different elevator systems require different switching. In some systems we use our relay to break the feed to the floor button so that it can't be accessed while on 'security'. Other systems require us to switch another relay in the elevator motor room which is what this one appears to be. Voltages that we switch also vary, in some installations it is as high as 110vdc but most are 24vdc. This particular job is 24vdc which we receive by way of one 'switch' wire and individual 'floor' wires per landing. In this example it is a 20 floor building with 6 elevators. Each level has a control relay and the majority of the floors are on security 24hrs. This means that with users with access cards traveling in the elevators cause the floor relays to change state numerous times during the day. The failure of the relays is more evident in this building because the floors are meant to be on security all the time (government tenant) and a non secured floor is reported quickly. Most buildings on the other hand only secure their floors at night when there is minimal use of the elevators and as such, there is minimal switching of the relays. Repetitive failure of 120 or so relays is in need of a reliable fix which so far points to fitting a diode across the elevator common and switch wires at our relay.

Gas3110;

Am I correct that with 6 elevators and 20 floors all on security would amount to 120 relays driven by your single 1A relay contact? 120 relays, even at a modest 30mA @ 24VDC, amounts to 3.6A DC. Most, but not all relays, require the DC ratings to be reduced by a significant factor 5 or 10 to 1, from their AC rating. If this is the case then I would opt to have your control relay drive an AC or DC relay that had 10A rated contacts or go to a solid state DC relay that can handle those types of loads as an interface between your control relay and that security relays for all of the other elevators and floors.

Take care

RcEe

RcEe,

To clarify, we have one relay for each floor landing we have to control. So in effect we have 120 'security' relays switching 120 'elevator' relays.