Erratic reading from current shunt

Darryl

4/15/23, 4:30 PM

Well, that was short lived. Readings are going haywire again. Victron battery monitor shows +2.8A which is in line with the equipment currently running and limited sunlight fom solar today. I'm thinking that either a need a new shunt or the BMS is screwing with me...

Darryl

4/15/23, 4:47 PM

And now it's back to normal. 🙄

I did install a short length of AC wire in the same vicinity (6-10") as the shunt sense wires yesterday afternoon but the log shows the initial issues coming up in the AM, prior to that... Any chance the AC current in the wire could be causing an induced current in the sense wires resulting in an erratic reading? Is this possible even if there is nothing connected to the AC outlet?

I notice that the shunt sense wires included in the standard harness are not a twisted pair, maybe they're shielded? Should I be shortening them and twisting them tightly to protect against this interference (If that's actually an issue)?

Richard OGM Support Staff

4/22/23, 12:50 PM

This sounds like a wiring problem. This is a millivolt reading so it is very sensitive to any poor connection.

Ideas:

Check to make sure all crimps are clean and tight
clean the connection surfaces on the shunt and the fork terminals
check the pins on the AMP plug at the BMS
Make sure the shield is only connected at the BMS end

Hope that helps

Richard OGM Support Staff

4/22/23, 12:53 PM

and just caught your reference to a local A/C line. Because the shunt is such a sensitive reading there is a chance that induced AC power may be affecting the reading. Make sure the 2 wires are as far away as possible and never wrapped together

Edward

5/20/23, 8:15 PM

@Darryl

I have just had a similar occurance. I am running a REC Active BMS with a REC WiFi module attached. The system is destined for my sailboat, but is currently prototyped on the benchtop in my lab. Configuration is:

REC Active: HW v2.1 FW v2.6
REC WiFi: HW v1.2 FW v1.5.4

I have a simple circuit setup utilizing the premade cable, a charge contactor, a load contactor and a 500 Amp current shunt, all purchased from OGM, with 560 AH of cells attached.

I have configured the BMS via the WiFi, and have performed several tests, including a full discharge to cutoff, a full charge to cutoff, and a low temperature cutoff. All worked perfectly.

With the pack discharged to 50% SOC, I disconnected all loads and the charger, and began to try and integrate the CAN BUS to a VENUS OS system. At some point in the process, I noticed that the BMS was suddenly reporting a load of -55.1 Amps with no loads connected! The wiring was unchanged, other than plugging in the CAN connector. Removing it changed nothing. I rechecked all connections, and rebooted several times. Problem remains.

I'm with you on this one. When I'm off cruising in the out islands of the Bahamas, I can't afford to have the whole DC system malfunction.I need a solid provable answer to this one.

Has anybody else seen (or solved) this incorrect behavior?

Rick OGM Support Staff

5/21/23, 3:01 PM

Edward,

This is an important issue and we need to get to the bottom of it. I have not been able to replicate your issue on our test board nor seen it happen personally but it really seems like an electrical connections issue. When we didnt hear back from Darryl we assumed he had found an electrical problem and resolved it. Would you please review my recommendations earlier in this thread and see if you can find an electrical issue and let us know how you make out. I also have a message into the REc dev team to get their feedback. Remember, you are dealing with a millivolt signal here so all connections need to be very clean and tight and well crimped where applicable.

Edward

5/21/23, 9:40 PM

Rick,

Thanks for getting back to me so quickly.

For completeness, I went thru your list. All connections were clean and tight. The problem persisted. Note that the lugs on the premade cable are attached directly to the shunt, so there is at most 1" of unshielded wire, with nothing in the vicinity, and no second shield connection.

As a last step, with the power off to both the BMS and the WiFi module, I pulled the BMS connector and inspected the pins and sockets. All straight and pristine. But as I plugged the connector back in, I noticed that the red LED on the BMS flickered once! And when I powered up the WiFi module and the BMS, all was back to normal, showing only the small current from the two contactors! It seems pretty likely (to me at least) that the BMS had gotten into some "bad" state that was being maintained by the voltage from the cell sense leads, and was only reset back to sanity by disconnecting the sense leads.

Now, I'm not a big believer in coincidences. The BMS, which had been performing flawlessly for many hours before this, suddenly started misbehaving shortly after I started trying to use the CAN BUS. In my experience, its usually the last thing you touched that caused the problem.

I had used the WiFi control panel to send an RS485 "CANF?" command, and gotten back an answer of "250". I then used an Inno USB2CAN converter between the REC Active and my laptop (with terminators), and seen messages coming out of the REC at 250 kbps. I then hooked the USB2CAN up to the RPi that I am running Venus OS on, and just as I was about to adjust it's CAN frequency, I noticed the bogus current reading.

Note that I am not running the latest firmware on either the REC or the WiFi. Was there any problems in the CAN communications in the earlier FW? I hate to go off flashing things before I have a better idea of what's going on, and whether my HW versions are compatible with the latest firmare.

BMS Prototype Shunt Connection

https://rec-bms-na.com/product-manuals/#7-8-wpfd-rec-active-bms

Rick OGM Support Staff

5/22/23, 1:51 AM

REC doesnt share all of the details of their updates but it can't hurt to start there. You can download the firmware from here:

and you will want to update your wifi module as well

https://rec-bms-na.com/product-manuals/#7-12-wpfd-rec-wi-fi-module

And we have video tutorials here:

https://rec-bms-na.com/tutorials/

Edward

5/22/23, 10:46 AM

Thanks Rick, but I'm going to hold off updating and try to find a way to reproduce the problem. This way, when I update, I'll have a way to determine if its been fixed or not.

Rick OGM Support Staff

5/26/23, 11:27 AM

Here is some feedback from Tine at REC head office.

Paraphrased

Looks like a wiring problem. If the problem was in the burned current protection resistors the current would always stay round -2 to 3 A -> open leads. Is the original cable used for the current measurement?

All the power wires should be traced 90 degrees with the shunt since values in the uV (microvolt) range are measured.

Edward

5/30/23, 11:04 AM

I am using the OEM supplied cable. And, due to the prototype nature of my setup, all cables are short. pristine, and easily inspectable. I am 100% sure that this is NOT a wiring issue, since it cleared by removing all power from the BMS without any wiring changes. My gut and 40+ years of embedded software development tells me that this is a firmware problem, but I have no evidence to back up this claim. :(

After over a week of running the system through charge and discharge, and exercising the WiFi and CAN bus interfaces, I have been unable to reproduce the problem. :(( Looks like I'll just have to upgrade the BMS from v2.6 -> v2.9 and the WiFi from v1.5.4 -> v1.6 and hope for the best.

Thanks for all your continuing help.

Neil

6/3/23, 5:17 PM

have had the same thing happen to my system recently suddenly strange ghost loads appearing i had one over noght that too the 4s pack down to what it thought was 8% capacity but the voltage was sat at 13.3V.

the pack has been running for over 2 years al be it i decommisoned it for the last 6 months but apart from firing it back up the only major variable that has changed on the system is i have added the WIFI module.

will check my firmware version but i have not updated for at least 18 months so definatley not on the latest

Neil

6/5/23, 4:59 PM

Here is an example captured tonight, the boat is pretty much idle plugged into shore power with an actual current draw of 184W the only loads are just 2 fridges wifi router, and a victron Cerbo total max about 15A but the REC is reporting a draw of 57 amps 43 amps on the battery.

State of Charge is now gon down to 1% however the battery voltage is full at 13.99V

Rick OGM Support Staff

6/5/23, 5:24 PM

I have pushed this back to Tine at REC for his review. As soon as I hear back I will post it here.

tine

6/9/23, 6:11 AM

Does the ABMS show correct current (0 A) when all is disconnected? IS the shunt connected directly to cell 1- pin or via fuse or ON/OFF power switch? What avout the cable from the ABMS to the shunt? Is it the original one with the pair waire and shield? Loops?

It can accept -0.3 to 5V on sense- and sense+ pins.

Darryl

12/12/23, 1:48 AM

@Edward

Its been a long time since I started this thread and that's because I thought my issues had gone away. Turns out I was wrong. Should have known that things don't fix themselves.

My system is reading +50 amps in while actually receiving something like 2-3A from shore power. Its 11Pm here so no solar.

I will try to troubleshoot tomorrow but likely it will need to wait til the weekend as long as it's not giving me grief.

Did upgrading the firmware make a difference in your symptoms?

Darryl

12/12/23, 1:50 AM

@Neil did you have any luck solving your issues? Mine have cropped up again. I will be carefully checking wiring but it sounds from Edward's experience that this is more likely firmware related...

tine

12/12/23, 5:12 AM

Disconnect everything from the shunt consumption. The current should show about 0.0XX or negative near zero. I do not see clearly if this is 500A/50mV shunt or 300A/50mV shunt. First check/set the Voltage to current coefficient of the shunt. For the 500A/50mV it is 0.01953125, frot 300 it is 0.01171875.

Then check IOFF offset current. Shouold be zero by default.

If the current is not near zero and fluctrating then disconnect the sensing wires from the shunt, open the ABMS and check the continuity to R46 and R47 right side next to common mode choce L2 on the circuit. This will eliminate wiring. Then observe/measure the resistance of the R46 and R47. Usually they are burned and in open circuit and the input of the current measurement ADC is floating showing wird values.

Rick OGM Support Staff

12/12/23, 8:19 AM

Darryl,

If after going through Tine's recommendations you do not find any issues please contact me at support@ogm-energy.com with your current shipping address and we will send you a new current shunt wiring pigtail. You will need to push out the pins on the Amp connector and click this one in.

Thanks and best of the season!

Rick

Edward

3/12/24, 8:59 PM

@Darryl

I only recently upgraded to FW 2.92-2. In the middle of trying to make sense of all the changes to the use of the relays and optos, I noticed once again that the WiFi homepage was reporting about -55A drain with no loads attached! And once again, the only way to clear this "ghost" load was to disconnect the big BMS connector for about 10 sec.

Neil

4/29/24, 6:57 AM

Has anyone come up with a reliable solution to this problem? I went out for a week and can't use or trust the capacity battery percentage anymore.

super frustrating!

Rick OGM Support Staff

4/29/24, 7:38 AM

Did you go through Tine's troubleshooting tests above?

And verify the pin connections at your AMP connector?

tine

5/1/24, 3:47 AM

It is not the matter of code or hardware, otherwise the > 2000 people will send me emails. It may be the thing of wiring and component placement. Like on this picture. See how the battery and the power lines make a loop? Then the shunt with the sens wires able to controll uV is right in the middle of the loop. I would move the wiring from the loop.

Any EMI ripple then translates in the sensing wires though we keep this on the minimum. a twisted pair is used with shield. In the BMS we use common mode choke and serial/differential filters and then we do the 0.3s integration of current to eliminate whatever sinus or AC current from tha actual signal. When we started with the Greenline boats we had 7R BMS using hall sensor. Without any software tweeks the current jumped +/- 30% appart due to the EMI. We have measured a sine wave in the signal and had to perform full sine wave signal integration. Otherwise we may measure on the positive peak or negative peak or somwhere on the AC sine wave that was added to the actuar measured DC. Once I have also noticed that one of the user changed only one of the sensing connectors with a different material. This unequal bi-metal petiential mede an offset with the measurement. If 500A is sensed @ 50mV it means 3.5 mV for the 35A.

Neil

5/5/24, 7:08 AM

Still pulling my hair out though I think it's the cable.....

I have opened up the BMS and have continuity between pin 24 and 25 of the main connector back to R46 and R47

R46 measure resistance of about 3314 ohms so not blown.

There is some funny brown marks on the PCB see picture I'm not sure if this is normal or not 🤷

I also observe when I move the cables around the readings fluctuate but I can't pin it down to a specific wire that is causing the issue

Rick OGM Support Staff

5/6/24, 7:12 AM

That brown mark doesnt look like an issue.

Did you ever check the shunt wire pin crimp connections?

Remember, you are dealing with millvolts and any poor connection is suseptible to induced galvanic voltages

You may even be able to solder the pin crimped connections

tine

5/6/24, 11:08 AM

R46, R47, L2, C52-C54 are the input filter for the current measurement. maybe you are getting something via burned PCB