GEVCU 6.2c no CAN traffic

91 views
Skip to first unread message

Lars Rengersen

unread,
Apr 14, 2021, 3:55:01 PM4/14/21
to GEVCU-Development
Hope this is the place to be for some help with my GEVCU. I have a 6.2-c and have flashed the DEV branch of Collins GEVCU7 repository (including the dev version of Adafruit BLE).

In the serial monitor I can see it is trying to send messages.

However, both the car CAN and the DMOC can seem corrupt. As soon as I connect them the car CAN breaks and the DMOC goes into error.
Even on the bench with just my Peak CAN adapter connected I cannot sniff any traffic. As soon as I connect the bus cables, the led on the Peak adapter flashes red and there is not any baudrate that helps in seeing any data on either bus.

I have done some hardware checks, the PCB has 3.3V, 5V and 12V.

Since I thought, bottomline the GEVCU is perhaps the same as the Due I flashed the CAN_Echotest from the can_due library. Unfortunately no luck. No CAN messages (on in that case 250 kbps) and not even a serial connection.

Hope you have some hints since I hope to spin my Siemens motor using the DMOC645 soon.

Collin Kidder

unread,
Apr 15, 2021, 10:36:07 AM4/15/21
to gevcu-development
Try to see if there is resistance between the two CAN wires. Between H
and L on a given set of wires you should get 120 ohms resistance. It
is possible that one or both of the wires might not be making good
connection at the harness.

Otherwise, yes, you should be able to use Due CAN sketches for
testing. The CAN hardware in GEVCU does not need an enable pin or
anything special to work. So, things like CAN_Echotest should work.
> --
> You received this message because you are subscribed to the Google Groups "GEVCU-Development" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to gevcu-developm...@googlegroups.com.
> To view this discussion on the web visit https://groups.google.com/d/msgid/gevcu-development/a542a8dc-4cee-4142-9d35-33034c5fb852n%40googlegroups.com.

Lars Rengersen

unread,
Apr 15, 2021, 11:58:34 AM4/15/21
to GEVCU-Development
Thanks Collin,
As long as the GEVCU6 is turned off, on the bench with noting else connected I measure 120 Ohm across the CAN-H and CAN-L.
But as soon as I turn it on (12V+ to C1 and GND on A3) then strange things happen.
Then on the car CAN my multimeter cannot measure a value and on DMOC CAN is is 4 MOhm.
I have checked with a scope and it does appear I see some pulses on Tx so the SAM3X is trying to send.
That plus the fact that even the CAN_Echotest does not work makes me think I have a broken unit.

Is there any documentation on what the DMOC needs in terms of CAN messages and flow/sequence?

Regards,
Lars

Op donderdag 15 april 2021 om 16:36:07 UTC+2 schreef Collin Kidder:

Angelum Band

unread,
Apr 15, 2021, 2:32:33 PM4/15/21
to gevcu-de...@googlegroups.com
Hi question for both: I have a UQM Powerphase 145 and in the manual they recommend to have CAN Grnd, Serial Grnd, General GRND all independent and separated from each other. Would that be the case for the DMOC? 

Lars Rengersen

unread,
Apr 15, 2021, 3:28:14 PM4/15/21
to GEVCU-Development
As far as I know (but do not have a full pinout diagram) the DMOC only has common ground and RS232 ground, not a separate CAN ground. But the GEVCU does not have CAN-ground output or a CAN ground plane on the PCB.

Op donderdag 15 april 2021 om 20:32:33 UTC+2 schreef angel...@gmail.com:

steven Mathews HM

unread,
Apr 18, 2021, 6:09:28 PM4/18/21
to gevcu-de...@googlegroups.com

Hi all

Further to Lars comment.

I finally /  just about to wire up UQM Coda to Gevcu. I have noted in documentation it is clear that UQM has Can H, L & Com //Ground. But as mentioned, I also note no Gevcu CAN-Ground.

Can anyone advise if there is a particular ground that should be used for CAN-Ground for the UQM on the GEVCU.

 

 

 

I look forward to your reply.

steve

image001.png
image002.png

Angelum Band

unread,
Apr 18, 2021, 8:59:35 PM4/18/21
to gevcu-de...@googlegroups.com
I'll try to see if I can find said diagram. Do you have the version, exact model and year? 


Collin Kidder

unread,
Apr 18, 2021, 9:18:42 PM4/18/21
to gevcu-development
On Thu, Apr 15, 2021 at 11:58 AM Lars Rengersen <la...@larsrengersen.nl> wrote:
>
> Thanks Collin,
> As long as the GEVCU6 is turned off, on the bench with noting else connected I measure 120 Ohm across the CAN-H and CAN-L.
> But as soon as I turn it on (12V+ to C1 and GND on A3) then strange things happen.
> Then on the car CAN my multimeter cannot measure a value and on DMOC CAN is is 4 MOhm.
> I have checked with a scope and it does appear I see some pulses on Tx so the SAM3X is trying to send.
> That plus the fact that even the CAN_Echotest does not work makes me think I have a broken unit.
>
> Is there any documentation on what the DMOC needs in terms of CAN messages and flow/sequence?
>
> Regards,
> Lars
>

It is possible that the transceiver has gone bad. GEVCU uses special
isolated transceivers which should be very reliable but anything is
possible. You could try scoping the CAN output wires and see what it
looks like.

I don't know that there is anything written up to document the DMOC
CAN messages. One thing I do remember is that the DMOC is extremely
picky when it comes to CAN messages. They need to start basically
immediately upon power up. Then CAN messages have to continue coming
reliably or it faults. The messages required are in the 0x232 range (I
think 0x232 through 0x234 are required) Then, there is a flow sequence
for getting into gear.

Collin Kidder

unread,
Apr 18, 2021, 9:24:04 PM4/18/21
to gevcu-development
No, CAN is supposed to be differential and you are not supposed to be
referencing it to any ground. Though, in practice most transceivers do
kind of reference the CAN signals to local ground. As such, it does
pay to have a good common ground between devices. You'll notice that
most cars have all their ECU/VCU boxes with a common ground. For what
it's worth, GEVCU does not reference CAN to any ground. The
transceivers are fully isolated. So, it doesn't care about CAN
grounding. It sounds like the UQM does care. GEVCU only has one common
ground but it has a lot of ground wires coming out of it. You'll have
to connect all the UQM grounds together.
> To view this discussion on the web visit https://groups.google.com/d/msgid/gevcu-development/CAOD4C3CBx4fSixeNJS-3HvFSVPecsY7%3DaQh87FO9KCAUHa-_FA%40mail.gmail.com.

Lars Rengersen

unread,
Apr 19, 2021, 2:09:22 AM4/19/21
to GEVCU-Development
Thanks, yes, I think the transceivers are bad. I've checked with a scope and I do see pulses on the Tx pin (so the microcontroller side) of the transceiver.
So it is actually trying to send something. But nothing on the CAN lines.
Will try and see if I can replace them.

The good news is that my motor spins!
I also had a GEVCU 4.2 but that did not work either. After making a .dbc and sniffing the CAN I saw the inverter was reporting a temperature of 100 degrees. So I have set the allowed heat sink delta temp to 200 degrees as a hack/workaround and now the 4.2 does control the motor.

Op maandag 19 april 2021 om 03:18:42 UTC+2 schreef Collin Kidder:

Lars Rengersen

unread,
Apr 22, 2021, 3:05:14 AM4/22/21
to GEVCU-Development
A follow up on this one: the CAN transceivers were indeed faulty.
After replacing one of them the resistor value remained 120 Ohm both when powered on and off.
The bad news is that while removing the old one I damaged a PCB pad so the Rx pin is not properly connected.
So I have an expensive wall ornament in the end...
Now just have a GEVCU 4.2.
Will reply in the other topic on my options / ideas.

Op maandag 19 april 2021 om 08:09:22 UTC+2 schreef Lars Rengersen:

Michael Brown

unread,
Apr 22, 2021, 5:03:30 AM4/22/21
to gevcu-de...@googlegroups.com
Lars, I may be able to resurrect the damaged GEVCU 6 for you, if you're willing to send it to me here in Thailand.  Can you take a very clear picture of the damage and send it to me at mwbr...@gmail.com and I'll see if I have a chance at a repair.  Maybe one of the other guys more local could try for you also.

--

Mike Brown

"Driving a gasoline sports car is going to feel like a steam engine with a side of quiche."  Elon Musk


--
You received this message because you are subscribed to the Google Groups "GEVCU-Development" group.
To unsubscribe from this group and stop receiving emails from it, send an email to gevcu-developm...@googlegroups.com.

Lars Rengersen

unread,
Apr 22, 2021, 5:14:04 AM4/22/21
to GEVCU-Development
Thanks for your offer Mike.
The problem is I was too impatient and/or did not set the temperature of my hot air soldering station high enough.
I slightly lifted the board using the chip and assumed it would come loose as soon as the solder melted.
However after 6 pins melted loose, two other were ripped.
See attached photo of the chip with the remains. Pad for pin 1 is also a bit damaged.
Unfortunately I do not have a photo of the PCB.
One pin is not used but the other is Rx.
I scraped a bit of trace and have put solder on in an attempt to repair it but it did not work out.
In my view this board will never be reliable enough again to be used in a car.
So if someone wants to do an attempt to repair it or have it repaired by you, make me an offer.
Am willing to sell for cheap, I am likely not going to use it and/or trying to repair it for further testing.

Regards,
Lars

Op donderdag 22 april 2021 om 11:03:30 UTC+2 schreef mwbr...@gmail.com:
IMG_0186.jpeg
Reply all
Reply to author
Forward
0 new messages