[mudder]

This thread is for any and all questions related to LiBCM during the Open Beta.
After browsing www.linsight.org, if you still have questions, please post them here.

I will answer questions posted in this thread by updating the FAQ/Instructions/Documenation/etc at www.linsight.org.
I will probably respond to questions posted in this thread by replying with a hyperlink to the answer (which I may have just written).

 

[mudder]

Here is a video showing the parts you don't need:

You need everything else in the IMA bay

 

[mudder]

Bumblebee does not offer the 12S+ module separately.

Related: I won't speak for Eli, but I understand BB will eventually start selling the 18S- cells individually.

Uncompressed cells are overwhelmingly damaged during charge/discharge cycles and heat fluctuation. If you leave the cells uncompressed in a climate controlled room, they will only age just slightly faster than compressed cells would. So maybe 9.8 years instead of 10.0 years for SoC(actual) to drop to 80% of SoC(nom). Here's a paper related to your question.

You would certainly want to compress the cells prior to charging them, no matter how low the current is. Before you take the module apart, you'll want to discharge up/down to 50% SoC (3.72 volts resting).

 

[mudder]

Good question. I've added it to the FAQ.

Yes, I always recommend cutting the WHT/GRN wire heading to the DCDC converter. This forces the DCDC to output 14.x volts (like a normal car's alternator would). Cutting the WHT/GRN wire still allows the MCM to disable the DCDC entirely (which is controlled by GRN/BLK).

FYI: I haven't discussed this previously, but LiBCM supports the ability to control the GRN/BLK wire (instead of the MCM). That allows LiBCM to more intelligently control when the DCDC is active. Once I've announced how to do this, the primary reason to do so will be to allow the DCDC to charge the 12 volt system whenever the key is in the 'ON' position... even if the car isn't started.

 

[mudder]

Cutting WHT/GRN is for the benefit of your 12 volt battery. The "hit to fuel economy" issue - if it exists - is a false economy: killing the 12 volt battery every few years by chronically undercharging it is more environmentally detrimental. I don't hypermile, so I have no data on any supposed fuel economy penalty.

LiBCM and cutting the WHT/GRN wire are mutually exclusive things. You can cut the WHT/GRN wire with or without LiBCM.

 

[mudder]

As long as you use edge trim or grommets it'll be safe. @Natalya has another cool idea to install a waterproof, external 5-15P plug near the license plate. I leave these modifications to the community at large.

 

[mudder]

Yes, the contactor needs enough voltage to kick over... otherwise the DCDC isn't activated. However, I often want to turn the radio/headlights/whatever on without first starting the car. But yeah, all bets are off if the contactor won't close.

The more important reason I'm adding this support to LiBCM is so that you can remove the lead acid 12 volt battery entirely, and replace it with the QTY6 leftover lithium cells from an 18S->12S conversion. With a lithium battery in place, LiBCM needs to be able to turn the DCDC converter off entirely as a safety backup, to prevent over-charging the 12 volt lithium battery. Obviously this also turns the dashboard battery and brake lights on, but that shouldn't happen in practice; LiBCM will first pull WHT/GRN low, which sets DCDC to 12.x mode. I haven't spent more than twenty minutes thinking about this feature, but it should work eventually.

Taking a step back, probably the #1 reason I want the DCDC to charge whenever the key is on is so the 12 volt battery doesn't go flat while grid charging (a NiMH pack). But that's no longer an issue, since LiBCM no longer needs 12 volts to run the fans.

 

[mudder]

1. That's the goal, but haven't done enough research to see if that's actually possible. For sure the stock fuse (from DCDC to engine fuse panel) would need to be larger. All TBD.

2a: We're still working the details out on this. But yes, LiBCM can technically allow the IMA to start the engine at any temperature... I just still need to do more research on the lowest temperature that can safely start the engine without damaging the lithium battery.

2b: Three separate computers control whether the backup starter is used. The BCM (replaced by LiBCM) gives "all systems go" if it thinks the IMA battery is in good enough shape to IMA start. The MCM listens to the BCM's status. If the BCM says 'go', then the MCM does its own checks, too. If all is good in IMA land, then the MCM tells the ECM "all systems go". If the ECM gets a "go" signal from the MCM, then it will wait five seconds for the IMA system to start the engine. if the engine still isn't spinning after five seconds, then the ECM will activate the backup starter.

So what this means is that yes, LiBCM can 'trick' the car into starting at ANY temperature, as long as the IMA system isn't throwing any P-codes (which would make the MCM abort the IMA start).

 

[mudder]

I've received a couple PMs regarding cell balancing.

Q: Is cell balancing safe right now?
A: Yes, LiBCM's existing balancing algorithm is very safe, but not terribly efficient. The OG Beta Alpha crew and I have been using it for several months without issue. It's well tested.

Q: The cells in two different modules (that I'm going to use for LiBCM) are at vastly different voltages. How long will it take LiBCM to initially balance the cells?
A: The existing balancing algorithm is very slow, and requires that the grid charger be plugged in (i.e. balancing does not occur when the car is off).
A: The next version of the algorithm will balance around twice as fast, and will do so whenever the key is off (whether or not the grid charger is plugged in).

Q: When will the new balancing algorithm be released?
A: This is my next firmware task.
A: My goal is to write this firmware tomorrow, but then it will take a day to run through the safety process (before I push the code up to y'all).
A: Obviously if any major issues are reported by the Beta crew, the balancing algorithm will be put on hold until I fix said issues.

Q: Can I drive my car with an unbalanced pack?
A: Yes, but the SoC range will be restricted. For example, if one 18S module is at 30% SoC, and another is at 70% SoC, then LiBCM is only going to use 35% of the SoC range (to prevent cell over/under voltage).
A: If your pack is unbalanced, you will want to download the above-mentioned firmware update (once I write and release it). Otherwise, if you use the latest code as of today, you will need to leave the grid charger plugged in - whenever you aren't driving - until the cells are balanced (which could take days).

 

[mudder]

Wow I didn't know it was possible to over tighten the #3 Phillips on the IMA switch. Were you using power tools?

 

[mudder]

Peter:
Now that I've seen your notch, I don't have any concerns. Note that if you follow the instructions on linsight.org, then when you install the 12S module - and connect the cables to the junction board - the maximum possible voltage is only 49.2 volts (less than 60 VDC safety limit). So the shock hazard is near zero. However, your way works, too... just unnecessary steps IMO. No harm in doing that, so I have no issues with it.

In regards to your fan issue, you're probably correct that the BMS ribbon cable is pressing into the fan. Use an insulated tip to move it out of the way (as shown in the installation video).

You now have more assembled LiBCM units than I ever have .

...

@Crimpcap
Based on the pictures provided:
-It looks like one of your modules is at a different state of charge than the others.
-Peter's cells are all from the same batch, so they're all at the same SoC.
-It's ok for the delta to increase during the initial bulk charge (i.e. until you get to 3.9 volts for the first time).
-Once you get to 3.9 volts, the balancing circuit will take over, discharging the cell(s) with the highest voltage.
-With the existing balancing firmware, it's going to take around 16 hours to balance your pack. Once I update the balancing algorithm, the balancing algorithm will be more than twice as fast.

My recommendation is to keep grid charging unless the high cell voltage (top left corner on the LCD display) exceeds 4.2 volts... as long as that doesn't happen, there's no safety issue. Note that while grid charging, the high cell voltage shouldn't ever get above 3.900 volts.

Overall, I'm not worried.
However, until you're comfortable the grid charger is working properly, don't leave it plugged in overnight.

...

I'm just finishing up applying comformal coating to QTY15 PCBs... so they can ship tomorrow (the coating takes hours to dry). So now I'll switch over to rewriting that grid charger firmware (to make it faster).

 

[mudder]

In the first video, the LEDs are working as expected.
In the second video, the voltages shown on the LCD are not good! Fortunately, they're probably not accurately displaying the cell voltages either, as 5.016 volts would almost certainly have caused a fire by now. My guess is something isn't plugged in correctly. Check the following:
-Is the RED 75 A Anderson connector plugged in? Note that if the grid charger is plugged in while the RED Anderson connector is disconnected, then that will permanently damage a component on the LiBCM PCB.
-Are all three BMS connectors plugged in?
-What are the individual cell voltages? (see below)

Video showing how to see all QTY48 cell voltages:

FYI: The reason the screen is flashing (and the buzzer is beeping) is that LiBCM thinks a cell voltage is too high... so it's trying to get your attention.

 

[mudder]

The screen will flash (and the buzzer will beep) if any cell voltage exceeds 4.200 volts. In the picture you posted, the highest logged voltage (during this keyON event) is only 4.061 volts... so is that picture representative, or actually just after the event you describe occurred?

In general, as long as the beeping stops when you disable regen (e.g. Calpod), there's no safety risk. FYI: The maximum 'safe' cell voltage the EHW5 module can tolerate is 4.25 volts.

Please make sure you're using the latest firmware (0.6.2), which has much improved SoC limiting over the previous versions.

 

[mudder]

Thanks for the response. I'll add it to our dataset... better code can fix this more gracefully (in a future firmware update).

 

[mudder]

@Crimpcap
Any updates on your LiBCM issue?

 

[mudder]

You said that parts get destroyed on the PCB is the grid charger is on and the Anderson connector gets unplugged. Would the same parts get destroyed if I switched the pack off before I unplugged the grid charger?

Turning off the IMA switch will not harm anything. It's only that cells 1:36 must remain connected, due to how the LTC6804 ICs are configured. The IMA switch is placed between cells 36 & 37, so no harm will result if the IMA switch is turned off.

This is ultimately a limitation of the parts available... I used the LTC6804 (which monitors QTY12 cells), whereas I wanted to use the LTC6813 (which monitors QTY18 cell)... alas, the LTC6813 has been out of stock for over a year now.

...

Looking at your serial data:
cells 42 & 43 are intermittently reporting erroneous voltages. For example:
-in frame02, #[email protected] volts & #[email protected] volts
-in frame03, #[email protected] volts & #[email protected] volts
-in frame04, #[email protected] volts & #[email protected] volts
-in frame05, #[email protected] volts & #[email protected] volts

Note that the sum of cells 42 & 43 is always ~7.5. If you divide that by two, then each cell is actually at ~3.75 volts, which is approximately what all the other cells are.

So then it would appear the the battery sense lead connected between cells 42 & 43 is floating. This could be caused by several different issues. Starting from the PCB and working back towards the battery module, there could be:
-a loose solder joint on the PCB itself (e.g. the LTC6804, etc).
-a loose pin on the BMS adapter that connects the PCB to the battery.
-a loose wire on the battery BMS harness.

If you're up for troubleshooting, here's a video showing my approach:

If you don't feel comfortable troubleshooting, I can send you a replacement PCB and 12S adapter cable tomorrow (you'll need to send me your existing PCB and 12S adapter cable back). Let me know.

 

[mudder]

I've seen this sort of thing when there is a bad connection from one of the LTC taps to a cell. The hint: one voltage is high and the other low; add them and divide by two and you are near the cell voltage. This is NOT a diagnosis, just an observation. Please wait for and rely on whatever guidance @mudder provides.

After posting the above, I see that's Sean's root cause suggestion is identical to mine.

 

[mudder]

Well I'm glad you were able to figure it out!

Can you clarify where the wire was broken?
-Was it the wire attached to the battery, or;
-was it the wire on the adapter cable (included with LiBCM)?

If it was on the adapter cable I made, I'll need to figure out what I'm doing wrong when crimping them... and hopefully it was a one-off issue; otherwise I'm going to have to remake a bunch of adapter cables!

 

[mudder]

Boo... If I send you a new 12S adapter cable, can you send me the old one back? I'd like to see what happened at the crimp interface.

 

[mudder]

Yes. Let me know if you need my address.

Only if it's changed in the last week.