Juicebox not charging my Leaf

*** UPDATE: THE PROBLEM WAS THAT MY J1772 CONNECTOR ONLY WORKS RELIABLY WHEN FIRMLY PUSHED IN TO THE LEAF ***

I have a 40A refurb juicebox (non-Pro) from eMotorworks. I got about 100 charges out of it, then I woke up last night hearing the car beep every few seconds as charging restarted.

Today, charging ran for 6 minutes, then aborted. After the 6 minutes, it returned to last night's pattern: charging starts, runs for about 5-10 seconds, then I hear the EVSE relays click. Then charging restarts.

From what I've read (like the Juicebox kit instructions), it sounds like GFCI nuisance trip is the most likely culprit.

i tested lowering charging amps to avoid a possible GFCI trip:


- Turning Juicebox trimpot R20 to "9 o'clock" yielded a charge rate of about 1.6kw, and that worked for a 2 hour test charge.
- "12 o'clock" gave about 5kw
- "10:30 o'clock" gave about 3kw.
- "9 o'clock" gave about 2.75kw

Is this an internal GFIC? Or external like installed in the breaker panel?

Oilpan4 said:

Is this an internal GFIC? Or external like installed in the breaker panel?

Click to expand...

It's the wire-lead current transformer inside the Juicebox. The original Juicebox (sold as a kit) has a procedure for adjusting this (you push/pull the A/C leads where they pass through the CT core, until the number of millivolts returned by the CT's secondary is minimized; then glue the wires in place with electrician's silicone).

To measure the CFCI output, you measure the millivolts between Arduino pin A1 and ground. In my case, I got a reading of "355 mV" at 6A of charging current, which doesn't seem to match the manual.

*** UPDATE, I got the text output from the Juicebox's Arduino microcontroller board ***

I connected a laptop with Arduino IDE to the Juicebox controller, and now I can see that it is throwing Failed Diode Check errors:

JuiceBox JuiceBox Classic:: HW:081401:: FW:100107:: Sup FW:04:: JNet Ver:07
Jfirmware MFG Date:Oct 25 2018:: Time:13:44:19
NFO:GFI Circuit Line Voltage Sensed at 275
0814011001070449459318990402BG,103iode Check Fail:
0814011001070449459318990402BG,603:CLR Diode Check Fail:
0814011001070449459318990402BG,103iode Check Fail:
0814011001070449459318990402BG,603:CLR Diode Check Fail:
...

Click to expand...

(for posterity: I connected an Arduino FTDI's RX pin to juicebox TX pin, FTDI TX to Juicebox RX, ground to ground; and FTDI DTR to juicebox ground. The FTDI's CTS and +5V were left unconnected. I set the Serial Monitor app in Arduino IDE to "9600,8,N,1")

After seeing this Diode Check Fail error, I checked resistance from the Leaf's J1772 pilot pin to ground pin, and vice-versa. My MS8239D multimeter read 12 mega-ohms from pilot to ground, and infinite from ground to pilot. A second (even cheaper) multimeter read 3 meg-ohms pilot to ground and infinite from ground to pilot. This is odd, since YouTube claims that I should be reading K-ohms from pilot to ground, and Mega-ohms from ground to pilot ( https://www.youtube.com/watch?v=5R4dGg8nIUs ). But I checked both meters using a pile of 10K resistors, and on every resistor, both meters agreed to within 0.2%. I have no idea why they read an order-of-magnitude different from each other when checking for a working diode on the Leaf.

I've tested L2 charging the Leaf on public chargers, and it worked fine on Chargepoint, Siemens and Aerovironment. So I think the diode is fine. My takeaway is: don't expect that can reproduce the readings in that YouTube video for testing the Leaf diode.

*** PROBLEM DIAGNOSED ***

I think the "failed diode check" message is a "red herring". The real problem appears to be that my Juicebox's J1772 connector has an iffy connection to the pilot pin on my Leaf. If I connect the J1772, then tug on it to be sure it is well connected, then usually the Leaf fails to charge, with the Juicebox "diode check" error message, and the Juicebox restarting charging every 10 seconds.

But if I connect the J1772, and firmly push it into the car, without ever tugging or wiggling it at all, then the Leaf charges reliably.

This is different from my Nissan trickle charger: I used to tug on that J1772 handle to verify it was fully hooked to the Leaf, every time that I connected it. Also, I have never had a problem with more than a dozen public L2 EVSEs. Maybe my Juicebox J1772 has a bit of play in the plastic moulding. Or maybe the pilot pin in my Leaf J1772 port is shorter than it is supposed to be.