Issues with lithiums in very cold weather?

[Naughty Cal]

10 hours ago, Tony1 said:

 

I should be realistic about what a warm air duct system can achieve, especially when the batteries are out in the engine bay with an air temp of minus 9 on some very occasional mornings. 

I think I'll give the air duct a go, as its cheap and easy ish to install, but it sounds like other measures might also be needed on the really cold mornings. 

The one good thing is that although you cant charge them below zero, you can still draw current from the lithiums at minus 10, so at least I can still have power (although I might break out the old gas kettle again on some mornings). 

 

 

 

 

We can still draw power from ours down to -20 but I doubt we will ever get that low with them.

 

If we go to colder climes the heating will be on to keep the cab and the battery warm.

 

For us it's only really when the van has been sitting on the drive unused that we have the no charging situation for a few hours.

[booke23]

9 hours ago, MtB said:

 

I doubt it. The internal resistance of LiFePO4 cells is so low that I'd have thought hardy any energy will be dissipated inside the battery during discharge. 

 

Yes that’s my experience. On a frosty morning I turned the inverter on and ran a low wattage electric kettle for 5 minutes….it was drawing about 70 amps from the LiFePO4 and the internal battery temp stubbornly remained at 3c. 

[DaveP]

10 hours ago, MtB said:

 

I doubt it. The internal resistance of LiFePO4 cells is so low that I'd have thought hardy any energy will be dissipated inside the battery during discharge. 

 

Some basic calculations would reveal the amount of heating effect. Does anyone have any values to hand for LiFePO4 cell internal resistances? 

 

 

I don't have direct references for the internal resistance, but believe it's of the order of 2-3mOhm for a 100ah prismatic cell.  One will also need the heat capacity of the battery to calculate the temperature rise.

 

My most demanding supply is my washing machine, which draws 150A at around 12.5v for 10-15min (about 400Wh), this causes the batteries (a pair of 100ah's) to warm by about 10degC.

 

Edited July 26, 2023 by DaveP

[MtB]

1 hour ago, DaveP said:

I don't have direct references for the internal resistance, but believe it's of the order of 2-3mOhm for a 100ah prismatic cell.

 

in that case given V = IR the internal voltage drop inside a cell with say, 100A discharge will be 100 x 0.003 = 0.3 Volts.

 

The power dissipated inside the cell will be 0.3V x 100A = 30 Watts.

 

This seems intuitively far too high a result so perhaps I've made a mistake. Or maybe that 3 mOhm value isn't right. Sounds to me more like the value for a LA battery. 

 

 

[DaveP]

20 minutes ago, MtB said:

 

in that case given V = IR the internal voltage drop inside a cell with say, 100A discharge will be 100 x 0.003 = 0.3 Volts.

 

The power dissipated inside the cell will be 0.3V x 100A = 30 Watts.

 

This seems intuitively far too high a result so perhaps I've made a mistake. Or maybe that 3 mOhm value isn't right. Sounds to me more like the value for a LA battery. 

 

 

Having looked a  bit more - prismatic cells ~0.3mOhm, so complete batteries somewhere round 1.2mOhm....

 

Edited July 26, 2023 by DaveP

[MtB]

31 minutes ago, DaveP said:

Having looked a  bit more - prismatic cells ~0.3mOhm, so complete batteries somewhere round 1.2mOhm....

 

 

Ok so closer to 3W per cell, and 12W per 15 or 20kg battery. Without knowing the Specific Heat of cell material we can't calculate the rate of temperature rise, but intuitively it isn't going to be anything useful in terms of quickly warming up a LiFePO4 leisure battery in preparation for charging.

 

 

 

[magnetman]

What about a short burst of immersion heater to wake them up on a cold day?

I mean for the water and the high discharge in theory warms the battery. 

 

@DaveP data earlier was interesting. 

[MtB]

17 minutes ago, magnetman said:

What about a short burst of immersion heater to wake them up on a cold day?

I mean for the water and the high discharge in theory warms the battery. 

 

 

But we've just been through all that in the posts immediately above. It doesn't!

[magnetman]

According to the data published up thread by DaveP the battery temperature went up around 10 degrees C. 

"My most demanding supply is my washing machine, which draws 150A at around 12.5v for 10-15min (about 400Wh), this causes the batteries (a pair of 100ah's) to warm by about 10degC."

I may be getting totally the wrong end of the stick but it seems to me that one could set a 150a discharge (1.5kw immersion heater?) to come on via a timer for a short time before charging commences and it would warm the batteries up. 

 

If it was an immersion heater it would be silent so not intrusive early in the morning..

 

 

 

 

Of course there won't be excess solar during very cold weather so it might not be worth it. 

[MtB]

17 minutes ago, magnetman said:

According to the data published up thread by DaveP the battery temperature went up around 10 degrees C. 

"My most demanding supply is my washing machine, which draws 150A at around 12.5v for 10-15min (about 400Wh), this causes the batteries (a pair of 100ah's) to warm by about 10degC."

I may be getting totally the wrong end of the stick but it seems to me that one could set a 150a discharge (1.5kw immersion heater?) to come on via a timer for a short time before charging commences and it would warm the batteries up. 

 

 

Ah I see. My apologies. I didn't look closely at that post as I'm accustomed to skipping past reams of graphs and stuff as people (not DaveP) tend to post screenshots with loads of irrelevant data as well as the one thing we need to look at. 

 

 

[Tony1]

 

A quick query for the electrical wizards, if I may: 

 

I use BMV712 units to monitor battery voltage, current, SOC etc. At the moment the positive wires from the BMV shunts are placed on the positive terminal of the battery bank. (I think its referred to as the 'most positive' terminal as its the first battery in the row, and where the pos cable is connected.) 

 

When the batteries are moved to the engine bay, my first thought was I'd have to move the tiny pos wires from the BMV shunts out there too.

And to reach out there, I'd have to add 3 or metres to their length.

 

So here's the question- do these BMV shunt wires really have to be be sited on the pos battery terminal?

Or could I place a bus bar (or post) on the end of the main pos cable that leads out to the batteries, and site the BMV wires on that? 

Would the readings they make still be usable, if they are sited at the end of a 4 metre run of say 50mm sq cable coming in from the battery positive?

 

Would they show voltage readings that were lower than the true reading at the actual battery terminals, and if so, would it be a consistent drop that I could sort of compensate for? 

And if I run them out into the engine bay, would lengthening them by say 3 metres also make the voltage readings artificially low in some way? 

I've got to lengthen the temp sensor wires anyway, but they are more simple resistance based, so I'm hoping their accuracy wont be affected too much by adding length to their wires. 

 

 

[nicknorman]

21 minutes ago, Tony1 said:

 

A quick query for the electrical wizards, if I may: 

 

I use BMV712 units to monitor battery voltage, current, SOC etc. At the moment the positive wires from the BMV shunts are placed on the positive terminal of the battery bank. (I think its referred to as the 'most positive' terminal as its the first battery in the row, and where the pos cable is connected.) 

 

When the batteries are moved to the engine bay, my first thought was I'd have to move the tiny pos wires from the BMV shunts out there too.

And to reach out there, I'd have to add 3 or metres to their length.

 

So here's the question- do these BMV shunt wires really have to be be sited on the pos battery terminal?

Or could I place a bus bar (or post) on the end of the main pos cable that leads out to the batteries, and site the BMV wires on that? 

Would the readings they make still be usable, if they are sited at the end of a 4 metre run of say 50mm sq cable coming in from the battery positive?

 

Would they show voltage readings that were lower than the true reading at the actual battery terminals, and if so, would it be a consistent drop that I could sort of compensate for? 

And if I run them out into the engine bay, would lengthening them by say 3 metres also make the voltage readings artificially low in some way? 

I've got to lengthen the temp sensor wires anyway, but they are more simple resistance based, so I'm hoping their accuracy wont be affected too much by adding length to their wires. 

The fused positive wire on the 712 has two purposes, to power the device and to allow it to display battery voltage. Battery voltage doesn’t have any bearing in SoC. However taking the positive supply remotely from the battery will mean that there is some voltage drop when current is flowing, and so the voltage will read a bit low during discharge (faster discharge = more error) and a bit high during charge. Provided the wiring is fairly chunky I don’t think it will be an issue, and the voltage will read correctly when there is no heavy load or charge on.

[IanD]

6 minutes ago, nicknorman said:

The fused positive wire on the 712 has two purposes, to power the device and to allow it to display battery voltage. Battery voltage doesn’t have any bearing in SoC. However taking the positive supply remotely from the battery will mean that there is some voltage drop when current is flowing, and so the voltage will read a bit low during discharge (faster discharge = more error) and a bit high during charge. Provided the wiring is fairly chunky I don’t think it will be an issue, and the voltage will read correctly when there is no heavy load or charge on.

The problem is that the BMS will then get the SoC wrong with any loads on. Lithiums have a very flat voltage vs. SoC curve so a small drop in voltage will be interpreted as a big drop in SoC.

 

Assuming there is a BMS which indicates SoC, and it gets its voltage reading from the BMV712, and the batteries are LFP... 😉

[Tony1]

1 hour ago, nicknorman said:

The fused positive wire on the 712 has two purposes, to power the device and to allow it to display battery voltage. Battery voltage doesn’t have any bearing in SoC. However taking the positive supply remotely from the battery will mean that there is some voltage drop when current is flowing, and so the voltage will read a bit low during discharge (faster discharge = more error) and a bit high during charge. Provided the wiring is fairly chunky I don’t think it will be an issue, and the voltage will read correctly when there is no heavy load or charge on.

 

Cheers Nick, I'll give it a go and see how the voltage varies in different scenarios. 

I already have a conundrum anyway, in that I have three BMVs wired in sequence, and they all give slightly different voltages and SoC readings etc.

Understandably I guess, because each BMV and shunt will take a small amount of current itself. 

For day to day use, I use the BMV whose values compare closest to the 'true' readings that you get when linking the batteries to the PC. 

 

 

50 minutes ago, IanD said:

The problem is that the BMS will then get the SoC wrong with any loads on. Lithiums have a very flat voltage vs. SoC curve so a small drop in voltage will be interpreted as a big drop in SoC.

 

Assuming there is a BMS which indicates SoC, and it gets its voltage reading from the BMV712, and the batteries are LFP... 😉

 

No problem in my case, as my BMS is basically the BMV712s, combined with disconnect switches! 

I understand that the BMVs dont use voltage when displaying SoC, as Nick says, so that shouldn't be a complication in this case.

I dont love the fact that because the voltage curve is so flat, I might voltage readings that dont seem to correspond well with the calculated SoC, but hopefully over time I'll get a feel for any innaccuracies. 

When checking the batteries' SoC, I always read their voltage at the same time. 

So if I read a SoC of say 70%, I'm looking to see a voltage of about 13.20v. If there is a small load, I add on 0.01 volts for every 2 amps of load, so if the fridge is on and there is a 6 amp draw, I might see a voltage of 13.17v, and I add 0.03v to compensate for the load caused by the fridge. 

If the load goes above 10 amps it becomes difficult to compensate the voltage accurately, but when possible I always corss check the voltage against the SoC, just for peace of mind. 

 

And ultimately if I'm not happy with the voltage accuracy I'm getting, I'll have to re-site the BMV wires onto the battery terminal itself, but fingers crossed I can get decent readings at the end of a long 50mm sq cable...

 

Edited July 26, 2023 by Tony1

[Alan de Enfield]

14 minutes ago, Tony1 said:

And ultimately if I'm not happy with the voltage accuracy I'm getting, I'll have to re-site the BMV wires onto the battery terminal itself, but fingers crossed I can get decent readings at the end of a long 50mm sq cable...

 

Can you fit a 50mm2 cable into the BMV terminals ?

[Tony1]

Just now, Alan de Enfield said:

 

Can you fit a 50mm2 cable into the BMV terminals ?

 

My plan is that the 50mm cable will run from the main fuse to a bus bar, then out to the engine bay and onto the pos battery terminal. The BMV pos control wires are tiny things, maybe 18AWG. They run from the BMV shunt (which sits on the main negative cable run), onto either the pos battery terminal (M8 in my case), OR if I can get away with it. they'll sit  on the same bus bar or post as the 50mm pos cable.

I'm afraid I'm not describing it very well! 

 

 

[Alan de Enfield]

10 minutes ago, Tony1 said:

 

My plan is that the 50mm cable will run from the main fuse to a bus bar, then out to the engine bay and onto the pos battery terminal. The BMV pos control wires are tiny things, maybe 18AWG. They run from the BMV shunt (which sits on the main negative cable run), onto either the pos battery terminal (M8 in my case), OR if I can get away with it. they'll sit  on the same bus bar or post as the 50mm pos cable.

I'm afraid I'm not describing it very well! 

 

 

 

That is what I thought. When I fitted one in the 'Cat' it was very fine wire, like 'speaker cable'

 

Don't know what 18 AWG is.

 

Looking it up I think it is much smaller than18AWG.

18AWG is 0.82mm2 (and rated at 14 / 15 amps) which is bit bigger that typical mains cable (0.75mm2) and half the size of a 1.5mm2 ring main

[Tony1]

14 minutes ago, Alan de Enfield said:

 

That is what I thought. When I fitted one in the 'Cat' it was very fine wire, like 'speaker cable'

 

Don't know what 18 AWG is.

 

Looking it up I think it is much smaller than18AWG.

18AWG is 0.82mm2 (and rated at 14 / 15 amps) which is bit bigger that typical mains cable (0.75mm2) and half the size of a 1.5mm2 ring main

 

Yes, I think you're right, the original BMV wire is thinner than 18AWG. 

If I do have to extend them out into the engine bay, my thinking is that by using a thicker wire, it would reduce the voltage drop. 

 

(But to be honest, the idea of me planning electrical things is akin to a 12th century monk trying to plan the space station.) 

 

 

Edited July 26, 2023 by Tony1

[nicknorman]

1 hour ago, IanD said:

The problem is that the BMS will then get the SoC wrong with any loads on. Lithiums have a very flat voltage vs. SoC curve so a small drop in voltage will be interpreted as a big drop in SoC.

 

Assuming there is a BMS which indicates SoC, and it gets its voltage reading from the BMV712, and the batteries are LFP... 😉

Tony’s batteries have internal BMS but the SoC is only accessible via PC. The BMV 712 SoC won’t be affected by a slight reduction in voltage, it is charge counting only.

[cuthound]

On 24/07/2023 at 21:51, Tony1 said:

 

I dont know whether you're trolling or whatever by placing that there. If you are then well done and give yourself a pat on the back.

But it doesnt change the fact that LiFeP04 batteries are very safe. 

One recall incident does not a BSS requirement make. 

 

 

 

Ditchcrawler is definitely not a troll! Still at least you apologised later in the thread.

 

No company issues a product recall unless they have to, because of the cost and perceived reputational damage. Often they are forced to by consumer protection groups.

 

The recall is always based on a number of the items failing in a potentially damaging way  which results in the product call.

 

Whilst I agree LifePo batteries are safer than most Li-ion technologies, poor design or poor manufacturing techniques and/or lack effective quality control can result in any product failing in a potentially damaging way.

[MtB]

5 minutes ago, cuthound said:

 

Whilst I agree LifePo batteries are safer than most Li-ion technologies, poor design or poor manufacturing techniques and/or lack effective quality control can result in any product failing in a potentially damaging way.

 

Even so, AUIU, LiFePO4 batteries in combustion are not self-sustaining (containing both fuel and oxygen) like some other lithium ion chemistries.

[cuthound]

5 minutes ago, MtB said:

 

Even so, AUIU, LiFePO4 batteries in combustion are not self-sustaining (containing both fuel and oxygen) like some other lithium ion chemistries.

 

That is what I've been told too, but there must have been several serious failures of some kind of the type of battery that Ditchcrawler referred to, or the recall notice wouldn't have been issued.

 

Perhaps the failures resulted in the battery overheating with the potential to ignite nearby combustible materials, rather than a self-sustaining battery fire?

[MtB]

4 hours ago, cuthound said:

 

That is what I've been told too, but there must have been several serious failures of some kind of the type of battery that Ditchcrawler referred to, or the recall notice wouldn't have been issued.

 

Perhaps the failures resulted in the battery overheating with the potential to ignite nearby combustible materials, rather than a self-sustaining battery fire?

 

Possibly. They certainly overheat during over-charging, which is why all decent BMS have cell high temperature monitoring, to cut off charging when cell temp gets to (typically) 50c as a crucial backup safety measure.

 

Where the cells in DC's recall notice LiFePO4s then? I thought not.

[cuthound]

50 minutes ago, MtB said:

 

Possibly. They certainly overheat during over-charging, which is why all decent BMS have cell high temperature monitoring, to cut off charging when cell temp gets to (typically) 50c as a crucial backup safety measure.

 

Where the cells in DC's recall notice LiFePO4s then? I thought not.

 

Yes they were LiFePo4's, see the image copied from DC's post below:

 

[MtB]

5 minutes ago, cuthound said:

 

Yes they were LiFePo4's, see the image copied from DC's post below:

 

 

 

Thanks. I totally missed that headline! 

 

My guess would be they are being sold with no BMS at all then, overheating and the case catching fire when relentlessly overcharged.