Relay will not open (but will close)

Hi Folks-

I'm hoping some experienced EEs in the group can help me with a problem I'm having with a circuit using Kemet EC2-12NU miniature signal relays, a low voltage relay. The switch will reliably close when the coil is energized. The switch will open about 50% of time when the coil is de-energized. There is no delay effect where it might open shortly after the coil is de-energized. If it won't open, it will never open until I energize the coil and then de-energize again, at which point it might open.

We have three of these relays on our board all of which are under microcontroller control. The failing one is different from the others because it switches the 12V supply voltage to a 12V->500V DC power supply. The 500V power supply consumes up to 650 mA under full load. My testing is unloaded so I don't know what the no-load current is. The relay behaves the same whether the 500V power supply is loaded or not. The contacts are rated at 2A.

The relay coil drives an internal DPDT switch. I can clearly hear the switch click a little muffled when it won't switch. I presume only half of the internal switching hardware is slamming when it fails to switch.

I believe there is some sort of external interaction that is causing the relay switch to refuse to open. The relay functions 100% reliably when it's removed from the circuit. The other relays in the circuit work reliably.

Can anyone offer any possible reasons why I'm seeing the relay fail to reliably open?

Thanks - JJS

Reply to
John Speth
Loading thread data ...

Maybe you're driving a capacitive load? If it's a capacitive load then the contacts may be welding on make (because of the current surge). Then not releasing on break because -- well, they've been welded.

There's a slight chance that the load is heavily inductive and the contacts are getting crapped up on break (because of sparking), but I'm much less sure of that.

A far less likely chance is that there's a permanent magnet lurking near that one relay, that's "helping" it stay closed.

If it's welding or sparking then a teardown of the relay, and of a brand new one for comparison, should show some significant differences in the contacts.

--
Tim Wescott 
Wescott Design Services 
 Click to see the full signature
Reply to
Tim Wescott

I'm wondering about damaging the coil driver circuit; some of those relay coils can have a ton of inductive kickback. How are you driving the coil, and do you have an explicit clamp?

--
Rob Gaddi, Highland Technology -- www.highlandtechnology.com
 
Email address domain is currently out of order.  See above to fix.
Reply to
Rob Gaddi

You seem to be focusing on the contact side while the coil has not be analyzed so much. A relay coil has make and break specs. To energize the relay the minimum Pull-In Voltage must be met. For the relay to de-energize the maximum Drop-Out Voltage must be met. If your driver circuit has any leakage current it might be enough to keep the relay energized.

This is a 12 volt coil relay. What is your driving circuit?

--

Rick C
Reply to
rickman

The Kemet EC2-12NU relay seems to have a polarised coil. Are you driving with the correct polarity?

Yes I know its a non-latching type but have experience when you drive some relays the wrong way that things didn't work out well!

That might not be the cause, in which case I'm also wondering if you're switching a capacitive load?

--
Mike Perkins 
Video Solutions Ltd 
 Click to see the full signature
Reply to
Mike Perkins

I can hear the coil slapping the contacts so I assume the coil is actuating adequately. The two other relays in the circuit are behaving nicely.

The driver is a FQD13N10L N-Channel MOSFET. We have a clamping diode on all relay coils. The gate is driven by a 3.3V push-pull digital output from a National Instruments uSB-6001 DAQ card. The coil high side is connected to the 12V rail. The MOSFET is swithing on the low side.

JJS

Reply to
John Speth

I double checked that. I even wired one on a bench breadboard to confirm our circuit application of the part.

I'm wondering about the same thing. The HV power supply is a Matsusada RD6-0.5P-LU. The data sheet is kinda lacking in characterization data for extreme and transient conditions. It's all typical steady state data.

JJS

Reply to
John Speth

You could be right about that but the fact that there are two other relays being driven identically makes me doubt that the coil could could be the culprit. The relay data sheet makes a vague statement that holding the coil energized can heat the coil and emit an organic gas that can destroy the contacts. There was no quantifiable information the accompany the statement.

My conclusion: Don't ignore the possibility of the coil being part of the problem.

JJS

Reply to
John Speth

Tim, I believe you might have given me the next line of investigation. I need to address the capacitive load possibility next. I added a clamp diode across the switch "output" to ground (cathode at 12V, anode to ground) believing it would clamp turn off inductive transients. It didn't help but it wasn't incorrect for the inductive load possibility. I need to get a current probe. My senior EE associate will take over the problem soon.

JJS

Reply to
John Speth

I humbly thank all respondents for their help. There is certainly a wealth of hard-earned knowledge here considering how small the news group has gotten.

In a nutshell, I think the contacts are welding shut due to arcing. That's because we have little engineering information about the load that's being switched. We will characterize the power supply and take steps to cor3ect deficiencies.

JJS

Reply to
John Speth

Did you ever figure this out? I've got an index card in my head with the "answer" section blank

--
www.wescottdesign.com 


----Android NewsGroup Reader---- 
 Click to see the full signature
Reply to
Tim Wescott

We have no root cause and we have no budget to get it. A circuit workaround was implemented instead.

We wrecked three relays to various degrees. Each was opened. Slight contact debris was noted. Since it's destructive to get a view of the contacts, we can't say if the debris is significant or not.

We think the load is capacitive enough to weld the contact when it closes. Then it's stays welded until something jars it loose. The load is a 500V power supply. The power supply manufacturer was not helpful in providing any information other than what's in the data sheet.

JJS

Reply to
John Speth

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.