Ground Check not working as expected

[Scott Rubin]

Hi All,
I have been testing my open EVSE on the bench.(Ok well dining room table)
Connected to 120 Vac single phase, the display indicates L1 mode.
If I disconnect the ground wire, when I power it up the display indicates L2 mode.
I was expecting the no ground error to appear.
Has anybody else seen this behavior?

Also what is supposed to happen state machine wise after a stuck relay or open ground fault?
I was expecting the error to prevent charging, but what clears the error? Is a power reset needed or does the S/W keep scanning the error conditions and clear the error when they go away.

~~ Scott

[Greg Scanlon]

I can't explain the behaviour you're seeing, but the theory of operation including a startup flow chart is on the wiki now.  

http://code.google.com/p/open-evse/wiki/theory_of_operation

[Chris]

I have a theory... I think what is happening is it is not seeing anything on either Mid400 and is using the default service level. If you enter the charge state I bet you will get the Error bad ground...

 

In previous versions of code we were testing for ground continuously while charging, that will work on L1 but for L2 you need to test ground independently with 2 relays.

[Scott Rubin]

I am wondering if the current is feeding thru both MID400s since the gnd is open. Maybe need to increase the resistors to reduce current when both are connected in series.

[lincomatic]

Stuck relay will automatically clear if the status changes.

If a GFI fault is detected, the EVSE waits 15 minutes, and then checks it again.  After 3 tries, it gives up and stays EVSE_STATE_NO_GROUND until it’s power cycled.

-Sam

[lincomatic]

Sorry, I should have mentioned.. there are two different ground checks.. hardware GFI, and software ground checking.

 

GFI fault (EVSE_STATE_GFCI_FAULT) is triggered by hardware interrupt from the GFCI circuit and will retry @ 15 minutes (GFI_TIMEOUT) for 3x GFI_RETRY_COUNT), then stay in EVSE_STATE_GFCI_FAULT.

 

No Ground (EVSE_STATE_NO_GROUND) state is possible only with ADVPWR turned on, and is tested via software only when the relay is closed (car charging).  It will also wait for GFI_TIMEOUT before retrying up to GFI_RETRY_COUNT times, after which it stays in EVSE_STATE_NO_GROUND.

 

From: lincomatic [mailto:linco...@gmail.com]
Sent: Wednesday, June 26, 2013 5:00 PM
To: 'open...@googlegroups.com'
Subject: RE: Ground Check not working as expected

 

Stuck relay will automatically clear if the status changes.

If a GFI fault is detected, the EVSE waits 15 minutes, and then checks it again.  After 3 tries, it gives up and stays EVSE_STATE_NO_GROUND until it’s power cycled.

-Sam

 

[Scott Rubin]

Here is the problem: If the gnd is open then both MID400s are in series with  44k (x2 22K) resistors across the 120 Vac line. So at 120 Vac / 44K = 2.72 mA. According to the MID400 app note:

Off-State Mode
When the input RMS AC input current is below 0.15mA the MID400 output will be in the high state as per specifications.

So both MID400s appear to turned on with a missing ground since 2.7 mA is greater than the turn off  0.15 mA. this is why the charger shows it is at L2 state.

~~ Scott