Victron Question

[WorldwideDave]

The Cerbo temp sensor relies on conduction from the metallic ring terminal. You are sampling the air above the battery, not the surface. It needs to be firmly clamped to a surface, i.e., a battery terminal.

I think we know how the last sensor I attached to a terminal worked out. You suggesting I put this exactly where the one I just removed was? If I attach to any lug, I would also want Ruuvi to be in similar situation.

I am charging now at about 60 amps - not a lot - won't likely get to 100 Amps today with cloudy weather. But I will reach 100% SOC.

Plan is to get it to 100% SOC then do a full 100A discharge for 2 hours straight or until I get some alert or worse. it will be my first discharge test without that shunt-based heat temp sensor attached at 100A since new BMS.

I am not likely to move temp sensors until the battery reached 100% SOC.

I really wish the victron setup would tell me how much time remaining until 100% SOC is reached when charging, not just telling me time remaining on discharge. My new BMS will tell me this but I have to walk 150 feet away and get close to it via bluetooth to see that.

Current, PV, temps, etc.:

 

[sunshine_eggo]

I think we know how the last sensor I attached to a terminal worked out. You suggesting I put this exactly where the one I just removed was? If I attach to any lug, I would also want Ruuvi to be in similar situation.

Nope. I'm saying you shouldn't use the temperature sensor in a way it's not designed to work. You could connect it to the (-) terminal, but that would be worthless too. It doesn't HAVE to be bolted to a terminal, but it needs to be firmly in contact with what you want to measure.

This entire process has determined that terminal mounted temp sensors on "canned" batteries are pretty much worthless. They are not representative of the temperature of the battery but of the heat generated by connection resistance. BOTH battery terminals are connected to the cells via wire, so there's no direct connection to the cell.

Now if you were to DIY a battery and connect it directly to one of the cell terminals, yes. It would be decent.

 

[WorldwideDave]

Nope. I'm saying you shouldn't use the temperature sensor in a way it's not designed to work. You could connect it to the (-) terminal, but that would be worthless too. It doesn't HAVE to be bolted to a terminal, but it needs to be firmly in contact with what you want to measure.

This entire process has determined that terminal mounted temp sensors on "canned" batteries are pretty much worthless. They are not representative of the temperature of the battery but of the heat generated by connection resistance. BOTH battery terminals are connected to the cells via wire, so there's no direct connection to the cell.

Now if you were to DIY a battery and connect it directly to one of the cell terminals, yes. It would be decent.

I went through the 'post your battery thread - no comments' that will had started in july/august of 2024. I clicked on every photo last night. Every. Single. One. Out of the HUNDREDS of photos, I saw 4 that appeared to have a victron battery temp sensor attached. 4. One was mine. I contacted the 3 posters of the others. I heard back from one. So I don't have a large sampling to go with of people that even use this sensor. But I did see 3 others who did have it mounted on a positive battery post. In ever case they were NOT DIY batteries - they were closed top plastic ones like my old CHINS was.
My take away is that they aren't used in LiFePO4, or people test the BMS overtemp protection cutoff themselves somehow, or they don't use victron, or they do use victron but use another method (like Ruuvi - heard that from way more people, which is why I bought to play with).
I will sit there on my duff with a flir - my next major expense - and do readings that way. But for today, when I do the 100A discharge for 2 hours after battery at 100% SOC, I will just use my infrared and the ones where they are. I will do another test with them in a different location on either wednesday night or thursday, because SOC will be 0% today and I will need tomorrow to re-charge them. Very cloudy.

 

[HRTKD]

Now if you were to DIY a battery and connect it directly to one of the cell terminals, yes. It would be decent.

My situations that I write about were DIY, so I have direct access to the cell's surface as well as the cell terminals.

This entire process has determined that terminal mounted temp sensors on "canned" batteries are pretty much worthless.

I agree with this. It's kind of like "Six Degrees of Kevin Bacon". The terminal on an off-the-shelf battery is a full degree away from the actual cell(s). The battery case is at perhaps two degrees away from the cells.

 

[WorldwideDave]

My situations that I write about were DIY, so I have direct access to the cell's surface as well as the cell terminals.



I agree with this. It's kind of like "Six Degrees of Kevin Bacon". The terminal on an off-the-shelf battery is a full degree away from the actual cell(s). The battery case is at perhaps two degrees away from the cells.

How do you recommend I attach the victron and ruuvi to the raw cells?

 

[HRTKD]

How do you recommend I attach the victron and ruuvi to the raw cells?

Unfortunately, I think you have to pick the least worst option.

With your Flir camera, find the hot spot on the battery's surface while conducting a stress test of the battery. That's where you want to place the sensor.

You've already opened the battery box, so you're ahead of just about everybody else with an off-the-shelf battery. Direct contact with one of the two middle cells is where the sensor should be. Slide the sensor in from the end between the plastic cover and the top of the cells.

 

[Dooug]

I found my 3 temperature sensors where all slightly out and was able to get them all together by adjust the offset setting in the cerb temp sensor config. You can't move it far, but I might be helpful.

I have 3 monitoring air temps in different plumbing areas of my camper basement and one mounted beside my batteries to see how cold or hot the compartment is.

 

[WorldwideDave]

Unfortunately, I think you have to pick the least worst option.

With your Flir camera, find the hot spot on the battery's surface while conducting a stress test of the battery. That's where you want to place the sensor.

You've already opened the battery box, so you're ahead of just about everybody else with an off-the-shelf battery. Direct contact with one of the two middle cells is where the sensor should be. Slide the sensor in from the end between the plastic cover and the top of the cells.

You aren't suggesting I cut the Kapton tape holding the cells together, slide the lug between the blue-wrapped cells, and tape it back up, correct?
You want me to use kapton tape to hold it on the top, similar to how the kaptop tape is holding the BMS' sensor now? I can put the BMS' external sensor anywhere you tell me to. There are 8 cells; 2 series 4 parallel, and I can put both on top resting and tape there in the middle.

 

[WorldwideDave]

I found my 3 temperature sensors where all slightly out and was able to get them all together by adjust the offset setting in the cerb temp sensor config. You can't move it far, but I might be helpful.

I have 3 monitoring air temps in different plumbing areas of my camper basement and one mounted beside my batteries to see how cold or hot the compartment is.

I presume to ensure you don't freeze the pipes/water, but with Pex, is that an issue in an RV?
Put differently, I assume you are using the cerbo to send you push alerts/emails when an issue is raised, but does it actually kill the inverter, the MPPTs or actually 'do' anything in your system?

 

[HRTKD]

You aren't suggesting I cut the Kapton tape holding the cells together, slide the lug between the blue-wrapped cells, and tape it back up, correct?
You want me to use kapton tape to hold it on the top, similar to how the kaptop tape is holding the BMS' sensor now? I can put the BMS' external sensor anywhere you tell me to. There are 8 cells; 2 series 4 parallel, and I can put both on top resting and tape there in the middle.

Based on the picture in post #96, slide the Victron sensor under the yellow plastic that is held in place with the black cable ties.

 

[WorldwideDave]

Based on the picture in post #96, slide the Victron sensor under the yellow plastic that is held in place with the black cable ties.

If I cut those cable ties, I can put the victron sensor and ruuvi anywhere, and can use Kapton tape or glue if someone wants to tell me how and where for sure. I will be happy to cut the ties and shove it under there, but without it not being attached, I guess the concern is I will get the 'ambient air temp' vs. the cell temps I'm trying to measure.

 

[sunshine_eggo]

My situations that I write about were DIY, so I have direct access to the cell's surface as well as the cell terminals.



I agree with this. It's kind of like "Six Degrees of Kevin Bacon". The terminal on an off-the-shelf battery is a full degree away from the actual cell(s). The battery case is at perhaps two degrees away from the cells.

This is a strong argument that batteries without built-in temperature protections shouldn't be purchased as there is no reliable way to externally measure their temperature - even for something like Ruuvi or SBS attached to the outside, there is likely an air gap between the case and cells. If you want an actual measurement, then this is a further argument for only buying batteries with a smart BMS that reports temperature as well.

 

[HRTKD]

If I cut those cable ties, I can put the victron sensor and ruuvi anywhere, and can use Kapton tape or glue if someone wants to tell me how and where for sure. I will be happy to cut the ties and shove it under there, but without it not being attached, I guess the concern is I will get the 'ambient air temp' vs. the cell temps I'm trying to measure.

As a non-destructive test, cram the sensor between the plastic and the cells. Then test the temperature reading.

 

[sunshine_eggo]

As a non-destructive test, cram the sensor between the plastic and the cells. Then test the temperature reading.

And don't be surprised when it's different from the Ruuvi.

 

[robbob2112]

If I cut those cable ties, I can put the victron sensor and ruuvi anywhere, and can use Kapton tape or glue if someone wants to tell me how and where for sure. I will be happy to cut the ties and shove it under there, but without it not being attached, I guess the concern is I will get the 'ambient air temp' vs. the cell temps I'm trying to measure.

Seems like a lot of the thermistors are attached in place with the white thermal glue/caulk/goop/adhesive -- it is not straight white RTV because the gasses put off as it dries corrodes electronics...

 

[WorldwideDave]

Hit 100% SOC, so ready for tomorrow's test. I do have 90 more minutes of sunshine at the most, so will let things cool off a bit outside, then go out and make the changes recommended above (the cram between test), no glue,

 

[WorldwideDave]

Okay friends, testing completed earlier today.
Questions I'd like answers to:

1.) Why the difference in SOC between the battery BMS and the Victron shunt.

2.) Why the difference between current readings between the BMS and the Victron shunt

3.) Recommendations on configuration of the JBD BMS to make it more accurate. Happy to use JBD app or Overkill Solar app.

4.) How to change things up so that the low SOC and low voltage coincide with the inverter shutting off. When discharging at 100 A, I have 4 audible alerts, and about 14 minutes max from the time I get a Low SOC alert and inverter dies. inverter not configurable (Giandel 12V).

Here's my findings:

Current differences:
There was always a current difference between the BMS and the Victron. While the current would fluctuate, the load is a pool pump running at 3120 RPMs through inverter through fuse through bus bar through kill switch through T-class to battery. Negative side is just from inverter through inverter shunt (for testing) through bus bar through battery shunt (measurements I used) through to the battery.



State of charge differences:
Seems the differences get more and more the lower the SOC is. Notice that even before starting, the victron thought SOC was lower by 7.8%, and before inverter shutoff it was 27.2% difference. That's not great. Over the course of the hour and a half test, the average difference was 15.5%. Here's the measurements:



Maybe there's some sync button in the BMS for low SOC? I know there is one for the 100% SOC in the victron and I presume in the BMS as well. Was pretty sure I had done that already when at full SOC over the weekend.

Temperature Differences:
Temp measured with infrared by pointing at battery conductor lug at T-Class fuse. Pointed at wire, at lug, and shrink wrapping; number posted is max number I got. I know it is a reflective surface, but still, not getting 70+ degrees C like I was with old BMS and conductors.
The BMS's internal temp sensor (wire taped to top of cells with Kapton tape) seems to be closest in number to the victron temp sensor I have on the top of the cells about 1/4" away from the taped one. The biggest difference is between the Ruuvi and the other two temp sensors about a 2.55 degrees C difference. I'm okay with that, but seems that the Ruuvi always will report 'colder' than something taped directly to the cells.

Here's the temp measurements:



The highest reading I got was 93 degrees F or 34 degrees C, and when I touched the wires - positive - they were warm but not hot. It almost seemed like the rubber casing around the conductors were 'loose' - like a wire labeled as 6 AWG was actually 8 AWG inside (using a guess / making an analogy / not facts / trying to explain). Just seemed like air was between the hotter wires inside compared to shething outside for those conductors coming off battery. But again, 93 degrees F for wire rated probably for 200 degrees F, not sure I need to worry? Remember my old conductors were getting 160 degrees F before, and in 1/5 the time (much faster - in 15-20 minutes, very hot).

Notes/Log:

Everything was doing great until around the 90 minute mark. It was acting as it should, just inaccurate readings.

At 10:45 AM, VRM reported a low state of charge, likely when reached 10% or 5% - whatever it is setup to alert me at. I expected this.

At 10:50 AM, while watching battery BMS, it jumped from SOC of 27 down to zero in seconds. At this time, the battery voltage went from 12.06 V to 12.00 V in the BMS very quickly. It was almost like the BMS is configured to be 12V=0% SOC. I'm aware of the steep curve with LiFePo4. Well at least aware of it, not an expert. Guessing config issue with BMS?

At 10:55 AM, Victron low voltage alarm, which I expected.

At 10:59 AM, non-Victron inverter died due to low voltage protection. You cannot change this voltage. I think it is set to 11.5 V; Battery BMS said 11.5V. I get four audible beeps from inverter, then it shuts down - that's very fast when discharging at 100 A.

It never hit low voltage protection in battery on this test - would not want it to, but was probably about to. It did the other day when I tested, but I think I had current up even higher (maybe to 150 A for that test).

At 11:03 AM, turned inverter back on, along with breakers, disconnects, etc and started the SCC back up and running to charge battery back up.

Victron VRM reports consumption at 2.5 kWh for this test, since midnight with the idle consumption.

Victron Connect app reports consumed Ah at 197.8 Ah (200 Ah battery) after the inverter shut off. There is still likely that 2.2 Ah left. I have done 200 Ah discharge tests before successfully.

After starting the SCC back up, it took over an hour for the shunt to report anything above zero percent state of charge. That seems strange to me, as it was charging around 15 Amps for almost an hour. About 7 A when started back up (cloudy day) and about 30 A when I just checked, so calling it 15 A on average during the hour.

Thank you all so much for reading and following along.

 

[Brucey]

Hit 100% SOC, so ready for tomorrow's test. I do have 90 more minutes of sunshine at the most, so will let things cool off a bit outside, then go out and make the changes recommended above (the cram between test), no glue,

You hit 100% based on the bms soc. Not the smartshunt?

You can see where the battery bms falls off a cliff at 29% while the smartshunt tracks better to give you a more realistic actual soc.

Try charging till your smartshunt hits 100%.

 

[WorldwideDave]

You hit 100% based on the bms soc. Not the smartshunt?

You can see where the battery bms falls off a cliff at 29% while the smartshunt tracks better to give you a more realistic actual soc.

Try charging till your smartshunt hits 100%.

It was 100% last night, lost 10% overnight, hence the difference at the start of the day.
But prior testing I did after charging it all to 100%, and I noticed the difference then as well - think that was sunday or monday test. I noted it then as well that the current display and the SOC differences then as well. This is just better documented.

 

[HRTKD]

The BMS's internal temp sensor

Be careful with that phrase. The BMS likely does have actual internal temperature sensor that is on the circuit board of the BMS. My JBD BMS had that in addition to the external sensor wire that was placed on the battery cell. There was a lot of discussion about which temperature the BMS would use for certain operations. The conclusion (as I recall) was that it took the worst case reading.

 

[WorldwideDave]

Be careful with that phrase. The BMS likely does have actual internal temperature sensor that is on the circuit board of the BMS. My JBD BMS had that in addition to the external sensor wire that was placed on the battery cell. There was a lot of discussion about which temperature the BMS would use for certain operations. The conclusion (as I recall) was that it took the worst case reading.

That is fair thank you. I will turn on all the sensors that it says it has and will send them to the BMS and see if another one shows up. I think I tried that already, but I don’t know.

 

[Brucey]

I remember it wasn’t the quickest flight across the pond, possibly handled by DHL once or twice so fingers crossed you have success.

Duppa out for delivery today, fortunately didn't get mired in small package delays as it was from Europe and not China or Hong Kong sweet mercy.

 

[740GLE]

I chatted with my coworker with a 3d printer, he said he’d gladly print me out that duppa mounting case that’s linked for $10. He said he had black fillet, I didn’t want to push it for victron Blue, lol.

Be interesting to see how it comes out.