Relay Question

Go To Last Post
20 posts / 0 new
Author
Message
#1
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I'm working with a relay circuit that switches a line between 18V and essentially floating when the coil is energized. I am having issues with the relay output getting stuck in the 'non-rest' state (connected to 18V) and essentially becoming non-functional.

Two relay footprints are used, one or the other is installed at any given point. One relay works fine, the other does not.

The good relay is the IMC03CGR.
http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=PB1249CT-ND

The bad relay is the V23026D1021B201
http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=PB285CT-ND

I have attached the circuit.
Thanks in advance for any help.

Jeff

Attachment(s): 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Is there enough drive current provided by Q1?
Have you been supplied the latching version in error ?

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

We definitely have the 'D1' model (SMT, monostable), or at least that is what is labeled on the part. From my calculations I think I should have enough driving current (3.3 - 0.7)/1200 = 2.1mA : 5V/313 = 16mA : By rule of thumb 16mA/10 = 1.6mA < 2.1mA, plus if there wasn't enough current, it would have a hard time switching to the "non-rest" 18V state. At this point I am always at the 18V state. I had tried lowering the base resistor to 600 Ohms too just be be safe, and that did not help.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Two thoughts.

Are you sure the I/O pin is configured correctly as an output? If not, and the internalpull up resistor is enabled, there will be some base current flowing which could be enough to latch up the relay in the on state.

Second: Is the micro's 5V coming from the same supply as the 5V driving the relays? Is there a big switching transient when the relay closes? This could trash the micro's program. Same question about the 18V, is the micro powered from the same supply, any big switching transients when the 18V is switched?

There are several approaches to dealing with this, but I'd rather hear about how they are ocnnected before discussing it further.

JC

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

The other question I'd be asking is how much current is the load drawing? The Good relay is designed to handle 2A on the contacts, while the bad relay can only handle 1A. Is there any chance that your load current on the 18V side is more than 1A?
Also, is it an inductive load? It's possible that the contacts are being welded together by arcing that happens when you try to interrupt too large a current in an inductive load.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I do not think this is a microcontroller issue. The uC runs off a 3.3V supply, seperate from the 5V and the 18V supply. And the uC is not locked up.

I checked the base-emitter current and the collector-emitter current, and they were 2.1mA and 17.4mA respectively when CALRELAY goes high and zero when it goes low, which means that the uC/transistor side of things is working as well as expected.

On one side of the 2-pole switch is rectified voltage coming from some current transformers. The other pole is just the 18Volt supply. I'm not convinced it is a welding issue either though, as it seems to want to actuate properly sometimes, but not all the time.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I have seen this before with some larger relays. Some residual magnetization keeps the relay in the "pulled in" position when the coil current is zero. If tapping on the relay frees it, that's probably the problem.

If there is a way to adjust the armature restoring force (a spring for instance) you may be able to "adjust" it so that it overcomes the residual magnetic force.

Greg

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

dalpilot wrote:
I have seen this before with some larger relays. Some residual magnetization keeps the relay in the "pulled in" position when the coil current is zero. If tapping on the relay frees it, that's probably the problem.

If there is a way to adjust the armature restoring force (a spring for instance) you may be able to "adjust" it so that it overcomes the residual magnetic force.

Greg


That's more likely what Chris-Mouse explained, and I'll also look in that direction.

Is there a possibility to unsolder the switching transistor and replace it with a simple switch for testing purpose only?

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

It looks like you are charging a 2.2uf cap through the relay contacts. This can cause the contacts to stick (weld).

Tom Pappano
Tulsa, Oklahoma

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Simple trick disconnecting de-energized relay shall prove the cause. IMHO welded contacts. Snubber could be a cure for that; the rest of 18V circuit may be useful for telling exactly, though.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Why would having a 2.2uF cap on the relay contacts cause welding? I bet that is the problem, but I'm not really understanding why.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

jcrollman wrote:
Why would having a 2.2uF cap on the relay contacts cause welding? I bet that is the problem, but I'm not really understanding why.

The contacts don't close clean, they bounce, and with the large capacitor inrush current this causes arcing. Then the hot contact points are held firmly together while they cool. It only requires one tiny spot to get hot enough to soften.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Is 2.2mF only capacitor on 18V line? I would have suspicions there's something else.
Welding is caused by energy (squared voltage*capacitance/by two). Best monitored by inserting small resistor in series good scope.
For tiny contacts this may be enough, especially for those exhibited to such an abuse before (electro-corrosion).

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I've attached a larger version of the schematic, and I can change the location of the 2.2uF to the line called 'current+' for the same desired effect. I wonder if the other 'good' relay is just built better to withstand the inrush current.

Attachment(s): 

Last Edited: Wed. Jun 22, 2011 - 12:39 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:
Welding is caused by energy (squared voltage*capacitance/by two).

Energy in itself is meaningless. That amount could be transferred in the course of an hour, which would be harmless, or in a few microseconds (small current versus a large one).

Depending on all resistances in the path and the ESR of the capacitor this the inrush current could be quite significant.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

@jayjay1974:
energy of discharge, let me correct that and thanks for pointing at. Closer to microseconds in terms of duration.
@ OP:
Is there only me experiencing issues with "RelsyIssue2.jpg"?

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Obviously the easiest way to check if contact welding is the issue is to DISconnect the +18V and see if the relay works properly without the 18V connected.

If so, this would help confirm the contact arcing/welding as the cause of the problem.

JC

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

DocJC suggestion is a good one to confirm or eliminate the sparking scenario!
Is the value of the capacitance critical? Could you have say 100 Ohm's in series with it to reduce the inrush currents?

For D11 you show a zener diode. What is its break down voltage? As far as I see it, you only need to use a 1N4148. The advantage is that they are cheaper (with a dis-advantage of slightly increased release time).
Zener diodes are usually used from collector to emitter with a break down voltage a few volts greater than Vcc. (The advantage is that you don't affect the release time of the relay)
:idea:
What is the Vol(max) for the silicon that is driving this circuit? IIRC a Xmega can have a Vol(max) of 0.76 V, which may/would be sufficient with the low value of base resistance and a gainy transistor, to keep the relay operated, once it is operated. The "bad" relay may have a lower hold current.
I would suggest that a 10K resistor from base to emitter will ensure that the transistor is fully turned off.

:idea:
Other weird things can happen. On release you could be exciting a parasitic oscillation which then keeps the relay operated? This condition is quite likely if the transistor is not fully turned off (as described above). The condition might be critical dependent on the load inductance of the "bad" relay!

Charles Darwin, Lord Kelvin & Murphy are always lurking about!
Lee -.-
Riddle me this...How did the serpent move around before the fall?

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I fixed RelayIssue2.gif.

A 100 Ohm resistor in between the relay and the capacitor is a possibility. Moving the capacitor to the line 'current+' is also a possibility.

@ LDEVRIES - D11 is a Zener, and it would probably make sense to use something cheaper here. The driving silicon is an XMEGA. I could increase the base resistor to see if it changes anything, but for now I have moved the capacitor in question to see if I still get the issues we have been discussing.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

A good practice do is to have base of switching transistor drained to GND via a resistor in order to close it reliably.
With everything Lee suggests it shall work.
Me personally will look for relays with good overhead in terms of current - it's always better eliminate possible source of issues.
Looks unlikely that it's welding, indeed.