diy solar

diy solar

LiFePO4 heating pad for cold temperatures

From one of the comments on there: " They each draw 3.5 amps at 12.3v or 43 watts" - it's a bit much. Personally I prefer something lower power so the battery temperature changes more gradually. A battery has a lot of thermal mass, and you want to make sure the cells get enough time to properly warm up. With too much heating power, this will be harder to control.
Yes I tend to agree. I have a 8S 24v pack, and 2 12V, 12W heaters in series, so a total of 24W, drawing a bit over an amp from my battery when on. I have another two of the same heaters, and was thinking if I had to I would add them in parallel to the other two, for a total of 48W.

I'm continuing to run experiments here. A few nights ago the outside temps got down to about 28°F. I put my battery box out in the driveway and turned on the logging all night.
Nov10-Nov11_graph2.JPG
The main thermostat on the batteries was set to turn on at 50°F and turn back off at 60°F. The green line shows when the heater was running, which was from basically 8:26pm to 10:31pm. (Excuse the blip in the temp measurements right as it started to heat: I had taken the lid off the box to check something and forgot about it for a few minutes. Came back to find the heater on, and I put the lid back on the box.) As you can see, the aluminum plate under the cells heated up pretty quickly (blue line) but then heated up linearly, topping out at 71°F. The top of the cells heated up linearly, and - as might be expected - continued to heat after the heater was turned off at 60°F as the heat from the bottom migrated up to the top. From 10:51pm to 7:33am the next morning (a bit over 8.5 hrs) there was no additional heat and the temp of the cells only dropped a bit over 6°F before I stopped the experiment.

So I expended about 2.5Ah as the heater ran for about 2 hours, which was enough for the whole night. This has me convinced that I don't need to add the other two heaters. 24W should be plenty.
 
Fluke 325 said all mine were within 0.5 F. Close enough for me.
I'm not worried about the heaters so much as the temp sensor in the bms. I assume it's very close to correct if not bang on so I'm not that stressed to verify it, but I'd like to be sure.
 
Yes I tend to agree. I have a 8S 24v pack, and 2 12V, 12W heaters in series, so a total of 24W, drawing a bit over an amp from my battery when on. I have another two of the same heaters, and was thinking if I had to I would add them in parallel to the other two, for a total of 48W.

I'm continuing to run experiments here. A few nights ago the outside temps got down to about 28°F. I put my battery box out in the driveway and turned on the logging all night.
View attachment 72250
The main thermostat on the batteries was set to turn on at 50°F and turn back off at 60°F. The green line shows when the heater was running, which was from basically 8:26pm to 10:31pm. (Excuse the blip in the temp measurements right as it started to heat: I had taken the lid off the box to check something and forgot about it for a few minutes. Came back to find the heater on, and I put the lid back on the box.) As you can see, the aluminum plate under the cells heated up pretty quickly (blue line) but then heated up linearly, topping out at 71°F. The top of the cells heated up linearly, and - as might be expected - continued to heat after the heater was turned off at 60°F as the heat from the bottom migrated up to the top. From 10:51pm to 7:33am the next morning (a bit over 8.5 hrs) there was no additional heat and the temp of the cells only dropped a bit over 6°F before I stopped the experiment.

So I expended about 2.5Ah as the heater ran for about 2 hours, which was enough for the whole night. This has me convinced that I don't need to add the other two heaters. 24W should be plenty.
Man I really hope I see similar. 2.5ah for 8 hours/7.5ah per day of heating is so trivial for the 2 night trips I commonly take that I won't even need to think about it.
 
The key to this is proper insulation.I just checked mine: it's -6C (21F) outside right now, and in the battery box (currently in an unheated tech building) sits at +10C (50F). All heated with a single 7W polyimide heating pad that's been running maybe 50% of the time (so taking what, 85Wh per day per battery).
 
I snowbird in a campervan and avoid freezing weather whenever possible.

My 100Ah Rebel Batteries LFP has a JBD with low-temp charging shutoff so my concerns are:
  1. battery longevity
  2. charging at ~0.2C (sometimes as high as 0.33C) in cold ambient temps
  3. accidentally getting caught in a freezing ambient temps
My better-safe-than-sorry reading of the datasheets collected by @Dzl in this post suggests that with cell temps held on the warmer side (like ~15C/59F) these charge rates are safe. ~10C/50F or ~7C/45F might also work for my use case but I have solar power and Ah excess to my needs. 15C also seems to be a common storage recommendation.

I ordered a 16w reptile warming mat that came with a thermostat, and gorilla-taped the sensor to the side of the battery.

Results so far: overnight temps here in the desert have been dropping into the 30s or 40s F. The mat holds the battery at set temperature and the temp probe reading is within a degree or two of the temp the BMS reports.

The battery has significant thermal inertia. At daybreak it was 59F and although ambient temps have risen 30F the battery temp has only increased 9F.

I'll keep an eye on it but I think 16w of heating will be sufficient for my mild-weather camping. If anyone has concerns about a warmer temp like 15C
vs 10C/7C please chime in.



UPDATE: I've now used the warming pad in ~freezing weather and it holds temp (within a couple 1/10ths) without issue. The pad turns on so infrequently the additional load gets lost in the noise of daily power use variability.
 
Last edited:
I'm not worried about the heaters so much as the temp sensor in the bms. I assume it's very close to correct if not bang on so I'm not that stressed to verify it, but I'd like to be sure.

I have three temp sensors on one of my batteries. I suppose you could count the Fluke 325 as a fourth. They all reported just what @time2roll did, within .5F of each other. And the interesting part of that is that two of them were taped onto the top of the center of the four cells and the third sensor was on the battery post.
 
The key to this is proper insulation.I just checked mine: it's -6C (21F) outside right now, and in the battery box (currently in an unheated tech building) sits at +10C (50F). All heated with a single 7W polyimide heating pad that's been running maybe 50% of the time (so taking what, 85Wh per day per battery).
I agree insulation is important, but there are practical limits to what we can do, and in the trade of using a little more battery I think good insulation makes more sense than great insulation. In my battery box build (I've detailed it here) I settled on 2" XPS rigid foam insulation on all sides. That already adds 4" in every dimension to the size of the box.

My box will be in the basement of our cabin (partially below grade) in Colorado at about 9,000 ft elevation. It gets very cold outside, but in that corner of the basement we think it will stay above 20°F. If it does, I think my current box and my current heat should be sufficient to work through the winter. If all the thermostats and heaters fail, I'm counting on my JBD BMS cutting off charging to prevent bad things from happening. The BMS is currently set to a low-temp cutoff of 4°C, but I may adjust that.
 
I agree insulation is important, but there are practical limits to what we can do, and in the trade of using a little more battery I think good insulation makes more sense than great insulation. In my battery box build (I've detailed it here) I settled on 2" XPS rigid foam insulation on all sides. That already adds 4" in every dimension to the size of the box.

My box will be in the basement of our cabin (partially below grade) in Colorado at about 9,000 ft elevation. It gets very cold outside, but in that corner of the basement we think it will stay above 20°F. If it does, I think my current box and my current heat should be sufficient to work through the winter. If all the thermostats and heaters fail, I'm counting on my JBD BMS cutting off charging to prevent bad things from happening. The BMS is currently set to a low-temp cutoff of 4°C, but I may adjust that.

I think you're going to be just fine. My batteries are inside my RV trailer that sits in a storage lot almost to Castle Rock. I checked last winter during one of the cold spells and the battery temps were looking good.

1" XPS on all the long sides of the batteries, no insulation on the short sides. I have the compression frame (3/4" plywood) on the short sides. The top is 2" XPS, but not well sealed. I have 1" XPS and then 3/4" plywood under the batteries as a shelf. It's not a four season trailer.

I don't have a data logger, so I can't tell you how often the pads had to come on.
 
I think you're going to be just fine. My batteries are inside my RV trailer that sits in a storage lot almost to Castle Rock. I checked last winter during one of the cold spells and the battery temps were looking good.

1" XPS on all the long sides of the batteries, no insulation on the short sides. I have the compression frame (3/4" plywood) on the short sides. The top is 2" XPS, but not well sealed. I have 1" XPS and then 3/4" plywood under the batteries as a shelf. It's not a four season trailer.

I don't have a data logger, so I can't tell you how often the pads had to come on.
Yeah, I've got my eye on you and your results (could you sense me looking over your shoulder? ;) ). I even changed from my Centigrade thermostat to the Fahrenheit 24V model of what you used based on your posts here.
I don't have a data logger, so I can't tell you how often the pads had to come on.
If you really want to log it, I may be able to help. It takes a single voltage logger (unless you wanna log all the temps like I do). I currently have a spare voltage logger until next winter when it goes back to the cabin. If you want to give it a try, send me a PM. I think we must live within 15 miles of each other, so I bet we could make the hand-off.
 
I should just buy one for myself. I need one that can get ambient temp and the battery temp as well as when voltage is being applied to the heating pads.
 
I can't recommend the datalogger I bought strongly enough. The thing is great because I can leave it in my camper outside then go in the house and Bluetooth to my desktop computer. Exports to excel with very minor settings tweaks and a quick import.

Won't inherently do voltage on to the pad but my solution to that is just to stick a probe next to said pad.

Alternatively I've been looking at whether it's possible to convert the output from voltage into something the unit can read safely. Then whatever temperature that winds up displaying as I can just consider the "on" state.

Seems to me a shunt resistor might be possible here to get the expected mv range needed to interface with the temp logger but I've not gotten much further than that.

I'd think anything between 0 and 50mv @ 12v would work here, with the value chosen to read relatively low on the temp chart so the scale doesn't get blown out of proportion with a 1500 degree reading or some such nonsense.

Just don't accidentally short it out and put full voltage into it lol

 
Last edited:
I can't recommend the datalogger I bought strongly enough. The thing is great because I can leave it in my camper outside then go in the house and Bluetooth to my desktop computer. Exports to excel with very minor settings tweaks and a quick import.

Won't inherently do voltage on to the pad but my solution to that is just to stick a probe next to said pad.

Alternatively I've been looking at whether it's possible to convert the output from voltage into something the unit can read safely. Then whatever temperature that winds up displaying as I can just consider the "on" state.

Seems to me a shunt resistor might be possible here to get the expected mv range needed to interface with the temp logger but I've not gotten much further than that.

I'd think anything between 0 and 50mv @ 12v would work here, with the value chosen to read relatively low on the temp chart so the scale doesn't get blown out of proportion with a 1500 degree reading or some such nonsense.

Just don't accidentally short it out and put full voltage into it lol

I only glanced at the logger you bought. Lots of functionality, but a bit pricey.

It looks like it uses thermocouples, vs. NTC probes like mine all do (and like most BMS's use). I believe that means that the meter is reading a voltage (created by the thermocouple) instead of a resistance (like from an NTC). Correct me if I'm wrong.

If that is true, then if you want to use one channel to measure a voltage in your battery box you just need a couple of resistors for a voltage divider. The shunt would be useful if you want to measure current, which may be useful but is not what you said you want to do. Otherwise, the shunt is not the right answer.
 
just throwing these out there

adafruit qt py samd21 : 7.50 usd
mcp9808 thermometer : 4.95 usd

these are what i use, although it’s important to note that does not necessarily mean other people should also use them. however, these two inexpensive components have enabled real time very precise temperature logging for my experiments.

a note on the thermometer mcp9808…
typical accuracy of ±0.25°C over the sensor's -40°C to +125°C range and precision of +0.0625°C. They work great with any microcontroller using standard I2C

to get the data, it is connected to a computer with USB cable and writes to a text file on the computer over serial ascii data.

if anyone wants help with getting the right code for their setup, happy to help with that.

< 20 dollar for multi channel temperature logging -40C to +125C with 0.06C resolution.
 
That is interesting, although for my purposes I don't anticipate this being something I will deploy with the battery box. This is just for my testing this winter, then I will probably repurpose the sensors / loggers for other things, like making sure my garage refrigerator keeps my beer at an ideal temperature. ;)
a note on the thermometer mcp9808…
It is kinda funny that in one sentence they say the accuracy is +/- 0.25°C and the precision of 0.0625°C, so they are pretty much telling you they are giving you more digits than are meaningful, unless someone wants precision without accuracy. :oops:
 
That is interesting, although for my purposes I don't anticipate this being something I will deploy with the battery box. This is just for my testing this winter, then I will probably repurpose the sensors / loggers for other things, like making sure my garage refrigerator keeps my beer at an ideal temperature. ;)

It is kinda funny that in one sentence they say the accuracy is +/- 0.25°C and the precision of 0.0625°C, so they are pretty much telling you they are giving you more digits than are meaningful, unless someone wants precision without accuracy. :oops:
the accuracy offset is static per device from my understanding. so two devices, out of the box, might report 0.25C difference at the same actual real temperature. after a one time offset, they ought to read the same “forever”

to narrow it more, comparing it over the entire range with a more accurate thermometer would be necessary, but 0.25C is good enough for me maybe. makes me wonder what my personal temperature difference discrimination threshold is in different conditions.. ?

haven’t intensely tested this yet.

all the ones in my experience have reported very close temperatures during periods of low heat flux, when they are mounted to the same copper sheet.

your concern motivates me to better characterize the thermometers i already have!
 
Pretty sure it's the total error, including offset, gain, non-linearity, etc...

But in any case 0.25 °C accuracy is very good so nothing to worry here ;) actually after looking the specs it's actually +/- 0.5 °C accuracy and only over the -20 to 100 °C range, it gets worse at the extremes (probably something like +/- 1 °C if I had to guess), the 0.25 °C spec is the typical value, it's not guaranteed.


That is interesting, although for my purposes I don't anticipate this being something I will deploy with the battery box. This is just for my testing this winter, then I will probably repurpose the sensors / loggers for other things, like making sure my garage refrigerator keeps my beer at an ideal temperature. ;)

It is kinda funny that in one sentence they say the accuracy is +/- 0.25°C and the precision of 0.0625°C, so they are pretty much telling you they are giving you more digits than are meaningful, unless someone wants precision without accuracy. :oops:

That's pretty much always the case, resolution (sometimes called precision in hope people will mistake it for accuracy...) is always better than the absolute accuracy, and it's a good thing since usually you can ignore the rounding error since it's small compared to the other error sources. Also, high resolution with lower accuracy is still useful in some cases, like when you want to be able to see a small change but don't really care about the exact value.

But sometimes it gets ridiculous and is just an attempt to sell things (first example that comes to mind is some cheap DMM), they announce something like 1 mV resolution but +/- 1 V accuracy... ?
 
I only glanced at the logger you bought. Lots of functionality, but a bit pricey.

It looks like it uses thermocouples, vs. NTC probes like mine all do (and like most BMS's use). I believe that means that the meter is reading a voltage (created by the thermocouple) instead of a resistance (like from an NTC). Correct me if I'm wrong.

If that is true, then if you want to use one channel to measure a voltage in your battery box you just need a couple of resistors for a voltage divider. The shunt would be useful if you want to measure current, which may be useful but is not what you said you want to do. Otherwise, the shunt is not the right answer.
A shunt resistor creates a voltage difference proportional to current though. Which then can just be read in mV.

I mean maybe I'm wrong as I've never made or tested one but that's what I recall from class.
 
A shunt resistor creates a voltage difference proportional to current though. Which then can just be read in mV.

I mean maybe I'm wrong as I've never made or tested one but that's what I recall from class.
Yes and the mV meter is calibrated in amps. Basic analog ammeter. No clue on how to get a digital circuit to log the information.
 
A shunt resistor creates a voltage difference proportional to current though. Which then can just be read in mV.

I mean maybe I'm wrong as I've never made or tested one but that's what I recall from class.
A shunt (very low resistance) is designed to measure current by creating a tiny voltage drop in the shunt, as per ohm's law. A 50mV / 100A shunt is spec'd to create a 50mV drop when the shunt sees 100A. An outside app can interpret the voltage drop as a direct indication of the amount of current through the shunt. So a shunt by itself doesn't measure any voltage at all. That's why I suggested the voltage divider, since it seems the meter you have is looking for a voltage (from a thermocouple), rather than a current.
 
No. The shunt does not "measure" voltage. It produces a voltage difference that a meter can read.

The shunt does not output current.
It's just part of the measurement circuit.

A divider would of course work fine but a single resistor would be much simpler
 
Back
Top