Stuck relay / contactor issue, round 2

[Andre Eisenbach]

I posted a while ago about my contactor issues (https://groups.google.com/d/msg/openevse/hVj54uIdy5o/dNlk7YgHyNgJ), but the conversation got distracted by a firmware issue. Now that the firmware issue is resolved, I'm still seeing a problem on long charges:

I'm getting a "stuck relay" error after charging or a while. I can't reproduce it by charging "a little" or repeatedly charging/stopping. It seems that only prolonged charging cycles (hours+) cause this. I'm pretty sure it's temperature related.

I've already ordered a new contactor just in case, but any other idea what might be happening?

Any ideas for debugging/testing?

Thx

[lincomatic]

You mean you get stuck relay error when the charging terminates, right? Does it just terminate into stuck relay, or does it enter stuck relay state a while after charging terminates?

--
You received this message because you are subscribed to the Google Groups "OpenEVSE" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openevse+u...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Andre Eisenbach]

Since it happens when I'm not watching, I'm not entirely sure. Will have to add some logging. I think it terminates into stuck relay state.

Mind you it seems like the contactor actually closes. But again, more work required....

There's no cap or anything on the AC sense pins, right?

[lincomatic]

No cap that I know of. But if you have the diodes installed, then you have to sample them. You would know, though, because it wouldn’t work at all w/o sampling.

 

You can easily check to see if the contactor is stuck by testing for voltage on the output. If it’s not stuck then it’s probably a timing issue. Do you give a debounce interval before you test for stuck relay after it opens?

[Andre Eisenbach]

I do sample the pins and do have a 150ms debounce delay before testing... :(

[lincomatic]

Try v3.8.4 and see if you get the same error?

https://github.com/lincomatic/open_evse/releases/download/v3.8.4/open_evse-384.hex

[Danny ter Haar]

On Thursday, May 28, 2015 at 10:32:18 AM UTC-7, lincomatic wrote:

Try v3.8.4 and see if you get the same error?

https://github.com/lincomatic/open_evse/releases/download/v3.8.4/open_evse-384.hex

Hahaha ,,,, 

[Andre Eisenbach]

Looks like the OpenEVSE firmware waits 1s.

I'll try that instead...

On Thursday, May 28, 2015 at 10:32:18 AM UTC-7, lincomatic wrote:

[lincomatic]

I just meant for testing. If it fails, then it's highly likely that it's hardware. 

Sent from my iCrap

--

[lincomatic]

I think you found the solution. I vaguely remember making it extra long because of this issue, now that you mention it.

[Nick Sayer]

On Thursday, May 28, 2015 at 9:32:13 AM UTC-7, Andre Eisenbach wrote:

Since it happens when I'm not watching, I'm not entirely sure. Will have to add some logging. I think it terminates into stuck relay state.

Mind you it seems like the contactor actually closes. But again, more work required....

There's no cap or anything on the AC sense pins, right?

The MID400 is "slow" so that it can (normally) slide over zero-crossings.

[Andre Eisenbach]

More data from tonight:

- Charged the car for 4 hours (23+kWh)

- Temperature around 44C max

- Increased the relay stuck check delay to 500ms

Here's what happened:

- Car went into state B

- Stuck relay error nearly immediately

- Reset the EVSE remotely via Bluetooth, comes up "stuck relay"

- Walk into the garage, unplug the car and measure charge cable; 0V...

- Reset the EVSE via settings menu, comes up "stuck relay"

- Open up the EVSE (live... yikes...) measure mains 240V, contactor output 0; coil voltage 0; AC sense line 0 (earth referenced)

- Reset again via settings menu

- EVSE comes up no error

So, the stuck relay state persisted for many seconds. At least enough time for me to walk downstairs, measure voltages and reset the EVSE twice.

Then it cleared itself.

Maybe the temperature went down, maybe measuring did something (though previously it also auto-cleared after a while). Maybe something else.... no idea

Ideas?

Thx

[Craig Kirkpatrick]

In your code it is important to check for the stuck closed condition. Maybe you are erroring with stuck open which really isn't a safety error. And maybe just giving the contactor some more milliseconds to close is all that's needed. Otherwise I'd double check the crimp connections on the relay AC coil and the relay test.

Does this show up up with OpenEVSE firmware? That would just be an informative data point.

[lincomatic]

The fact that the charge cable measured 0V even though you were getting stuck relay proves that the contactor is not the culprit. That leaves either the sense circuit or a bug in the code that handles them.

Check the wiring and make sure the terminal blocks are screwed down very tight.

One thing I’ve found very helpful for debugging is to display the live values of the AC pins on the LCD during a fault.

 

From: open...@googlegroups.com [mailto:open...@googlegroups.com] On Behalf Of Andre Eisenbach

Sent: Friday, May 29, 2015 12:41 AM
To: open...@googlegroups.com

[Chris]

In no Spark did you set the pullup or is your digital line floating?

[Andre Eisenbach]

Yes on the pull-up.

[Cory Clemmer]

What about two led or neon 240vac indicator lights on the side of the unit, one connected to the coil and the other to the output to aid in debugging.

[Andre Eisenbach]

Ok, next round of debugging:

- Tried again with larger delay

- Unplugged after about an hour of charging; stuck relay error

- Measured all the voltages, all 0 (sense, coil, output); Well, sense wire had ~0.4 VAC is that normal?

- Resetting multiple times got instant stuck relay error

- Debug print showed both sense pins read low (according to my code)

- Downloaded the official firmware

- Did a few shorter charging cycles and all good (just like with NoSpark).

- Then did a long charge and got "service required; stuck relay" error after unplugging

So, it's not the contactor as Chris said.

But I still do think it's maybe heat related?

No idea though....

[Craig Kirkpatrick]

I read and interpret what you give as observations. Then I'm befuddled when you say it is not the contactor. From your observations and what you report I say it is the $12 contactor.   Am I misinterpreting what you say?

Best Wishes,

Craig K

--

[lincomatic]

He said the contactor output is 0V. Yet the sense pins are both low (voltage present).

Something’s wrong w/ the sense circuit.

[Andre Eisenbach]

Exactly.

Btw. can somebody measure the sense pin in normal idle condition (in a full circuit)?

Is the 0.4 VAC normal? I'd assume the optocoupler needs much more to sense a voltage?

[Andre Eisenbach]

Assuming it's not the optocoupler (might be a bad assumption, but I figure it's probably fine), could there be anything wrong with my wiring?

Maybe having the sense wire crossing the hot lines is causing the ~0.4VAC and maybe even some issues?

#claspingatstraws

Here's the wiring again:

https://lh3.googleusercontent.com/-DXtaPczBF3c/VTMk6cShFCI/AAAAAAAAUzc/S4CzJOaESkU/s1600/image%282%29.jpg

Any tips would be appreciated.

[chris1howell .]

There really is not much that can go wrong. On the line side you have hot resistor mid400 ground. On the other you have input resistor mid400. There could possibly be a soldering issue but other than that the most probable hardware culprits Mid400 and microprocessor. I have seen both exactly once each.

I have a top secret (for a couple more weeks) board to send to you to play with so I'll send a Mid400 with it as well...

[Craig Kirkpatrick]

I measured the AC test line on a v3 on my bench powered by 120V.  Clearly there is some 60Hz noise signal showing up on the AC test node.  It is pretty small amplitude, measuring 0.6 Vacrms.

Basically I'd say I'm seeing the same low level AC on that signal while the EVSE is idle that you are seeing.  It is hard to say where it is coming from and could even be coming from the EVSE controller board itself.  

I think you can rule out the small noise signal as the problem since I'm seeing it too and using your code with an EV simulator and not seeing the same trouble.  The only difference is I have the relays on this one instead of the contactor and the contactor will take more de-bounce time I imagine.

Depending on other priorities today I'll put NoSpark on my v3 contactor and charge my Ford Focus to see if I run into the problem you are seeing.  I think enough folks are already using NoSpark and have already proven that they don't see the problem.  Hopefully the MID400 component change in a few days will remedy the problem for you.  Let me know how I can help.

b

[Andre Eisenbach]

Thanks for checking!

Can't wait for the mystery board delivery now :D