"welding" of relay contacts problem

[archielim22]

Why does my relays stop "switching" after some time of use. I am currently working on controlling a 24v 5A DC motor using a microcontroller and a uln2003a darlington array. Also using a 5v coil rating, 28v contact rating, 10A relay to control the motor. Everything is working perfectly except that after a few minutes of use, each relays stops switching....

i read somthing about using supression circuits, but the uln2003a has, and should have a built in protection... might need to protect the contact too? how? any help would be greatly appreciated.

also, does using an other types of relay with different material helps? i heard about expensive relays which supposedly does not weld... is it worth it?

 

[cbarn24050]

Hi, when you switch on a dc motor it is allmost a short untill it starts moving. This causes a large current to flow which can weld relay contacts and also weaken the field magnets of the motor. You need some current limiting in the circuit, a low value resistor might work or some active element such as a transistor or surge limiting thermistor.

 

[electricpete]

yes it's a good point by cbarn. I certainly don't disagree with the possibility of contact welding.

I don't understand what is the difference between switching fast and switching slow. Are you talking about the longer time between turning the motor off and then turning it back on?

 

[archielim22]

cbarn: i will try thatÜ any recommendatin on the resistor value?

pete: yes switching fast means turning the motor on and off... as far as i know, DC motors have high transient currents, causing the 'melt'

i just read about the double break contacts, thankx to jbartos and crowbar, will also try to luk into thatÜ

 

[karilaakso]

I would put a truck headlight relay there. It is made for 24 V and 30 A.

 

[cbarn24050]

Hi, you will need to experiment with the value, get a pack of 1/2ohm 10W resistors.

 

[Shortstub]

archielim22,

1) The driver, ULN2003A, does have a suppression circuit. However, unless that suppression circuit has been provided externally, pin 9 must be connected to the positive side of the relay.

2) Much has been said about the relay's output, but what is the relay's coil resistance?

3) Much has been said about the microcontroller driving the ULN, but what is the voltage at the ULN's input pin?

4) Are any of the other drivers in the ULN being used? If so, what is the combined coincident load, in mA?

5) What is your "switching-rate?"

6) If the other drivers aren't being used, then, parallel 1 or 2 others to the one being used!

 

[archielim22]

cbarn: ok ill do that. problem is the relays might already be busted, and i have to keep busting relays to see what value worksÜ
But i have to do it, got no other choice...


Shortstub:
1.) Yes, pin 9 is connected to the positive side of the relay.
2.) I think coil resistance is 63ohms
3.) Voltage at input is also 5V, using the darlington array to increase current.
4.) Although 4 of the arrays ports are being used, only one port is being used at a time.
5.) Switching rate varies depending on the user, but the maximum rate would be around 5 switches per second.
6.) Connecting them in parallel would increase my driving/sinking current right? And this would increase the 'power' to unweld the contacts? or yuo have a different reason in suggesting thisÜ pls explain if yes, thankx!

Any relay experts out there?Ü pls suggest what relay to use, including manufacturer and model, and why Ü

 

[Shortstub]

Archielim22,

Reur pt 2. Sinking current is 380 mA, which is within the chip's rating.

Reur pt 6. That is correct. Two drivers will reduce the current to 1/2 of the above, or 190 mA!

Based on your measurements, I don't believe the problem is the chip, but instead, the switching frequency. Contact welding is not usually a problem of "making" DC current. However, it is a severe problem for "breaking" the current. Some visible arcing is normal, but is there considerable "visible" arcing? Arcing is more of a problem if the current that is being interrupted is long in duration, i.e., high L/R. Double-break contacts will increase the interrupting capacity by about 3 times over single-break. But, in the end, contact "life" is very, very, dependent on the L/R of the "load!"

The automotive-type may be worth a try. You might even try "breaking" both the neg and pos lead, but frankly, I believe that the application is much to severe for the relay you have chosen.

Relay selection is very much an "art", therefore I suggest you contact a relay "specialist!"

 

[cbarn24050]

Hi shortstub, could you explain how a contact can weld as it breaks.

 

[Shortstub]

Archielim22,

Am I wrong? Or, are you using this circuit for PWM control of the DC motors' speed?

If so, you'd be far better of using a power semiconductor, like a MOSFET, in place of the relay! An addtional benefit it can be driven by the microcontroller, eliminating the ULN2003A!

 

[Shortstub]

Cbarn24050,

If I interpreted Archielim22's comments correctly, then the relay is being energized to operate the motor thru "making" of its contact(s). The coil is not being energized to "break" the motor circuit. I'm not suggesting contact "welding" occurs only because of "breaking" of the current circuit, but, it is the greatest contributing factor. The L/R ratio determines the duration of the "arc!" Extinquishing of the arc is a function of the speed of separation of the contacts. The required speed is not possible with a standard relay.

Arc duration contributes to the melting of the contact material, contact pitting, and distribution of the arc products. All play a part in the "welding" phenomena. Contact bounce upon "making" could also be part of the problem, as well as "chatter!" However, the latter is more of a problem with AC circuits, but that's a topic for aother thread!

Of course, a sluggish movement could also be a contributing factor. Proof of the latter would be arcing upon "make!" In any case, 5 contact operations per second is excessive. I am certainly not suggesting the use of high-speed photograpy, although I did use a fast-frame cam-corder to analyze a similar mechanical function.

Contact "welding" is never the result of a single break/close cycle. For the given "switching-rate" I am sure if the contacts were examined periodically, then, the worsening condition would be observed. Anyway, in my experience, the problem is not the electronics, but the electromechanics. Typically, contact "life-expectancy" is expressed in millions of operations. Of course, I'm unaware of the actual number in Archielim22's case, but, it could be as high as 400,000 operations per day!

 

[cbarn24050]

Hi shortstub, am I to take it that you cant explain it. I cant say i'm suprised.

 

[Shortstub]

Cbarn20405,
And what part wasn't explained to your satisfaction?

 

[jbartos]

Suggestion/question: Is ULN2003A properly applied? Have you tried to replace it by another one, known to be good one?

 

[cbarn24050]

hi shortstub, you didnt explain how a contact can weld as it breaks.

 

[Shortstub]

Cbarn24050,

Now I understand where you're coming from... the resistors! I wondered why you were so enraged at my reponse that you decided, albeit, anonomously, to throw a "barb!" Now, don't get me wrong, it's your perogative to question anything I provide as a technical view. But do it in professional manner. Say, it's B... S..., and then explain your position! This forum, afterall, is supposed to be about "give & take!" But, more on this later.

I agree that "welding" can occur on "make!" Obviously, it can only occur while the contacts are "mated." It's a DC motor, about 1/6 Hp, and it is reasonable to assume that inrush can be about 2 to 3 times FLA. This value is not too large given the rating of the contacts, i.e., 10 Amp. But, before mating surfaces can fuse, there must be a large current-density, i.e., current thru a small area of contact. Archielim22's observation was that the contacts' operated for some time, so its also reasonable to assume that the contacts were"good" at the start. Further, it was my position, that the degree of stress on "make" was far less than that at "break!" I didn't say the contacts fused, while parting. Also, Archielim22 did not observe a drop in the coil voltage, which could have led to "chattering!" Furthermore, I've "played" with quite a number of ULN's, although 2001's instead of the 2003's, for many years!

That said, then, as far as this thread is concerned, it is clear to me the problem is the relay switching-rate, at least until proven otherwise.

Now, cbarn, if you're questioning my credentials I'm more than happy to provide them to you, or anyone...on list... or off! My real "gripe" with you, cbarn, is, that I perceived the "barb" as questioning my veracity.

 

[archielim22]

Wow so much replies since i last read this....

1.) Shortstub maybe right that switching rate is a problem since if i 'switch' the relay at lower rate, evreything works okay....

2.) i dont know if this will help, im using a SPDT relay, and looking at my schematic above, the motor's EMF during break sinks current at the NC pin.... which is a very high current, which can lead to a melt, and consiquently, a weld. i have already provided a diode at this pin for supression.

3.) For the 'make' part, im providing a 3.3OHM series resistance for the initial moment that the NC and NO are shorted, producing high current at startup.

4.) i will now be testing a power PCB relay with a 16A 125VDC contact rating. If this does not work, i dont know what willÜ

5.) A side note, the OEG relays lasted for about a minute of high rate switching before welding, the OMRON lasted for about 3 minsÜ i believe the MATUSHITA ralys are considered the best, but they are so f*@&$ expensive..Ü

 

[cbarn24050]

Hi shortstub, I think you need to read again what you posted, if your retacting it now thats fine. You seem to be 1 of those hotheads who posts before reading, may I suggest more use of the "preview post" button.

 

[Shortstub]

Archielim22,

Now that the mechanism of contact "welding" has been explained, at least to the satisfaction of one, would you like to pursue the "art" of contact protection?

 

[archielim22]

Shortstub,

i think i have solved my problem, at least i think so... i did the following:

1.) supplied a diode across NC and Common contact to supress incoming EMF current during 'turn-off'

2.) connected a 1 ohm resistor between relay and motor to limit the current during startup

3.) decreased switching rate

4.) supplied a relatively expensive relay for robustness of contacts