Be cautious ordering generic blue wrap LFP cells

[Alloy]

Good to know, I've done it for large battery lugs (blow torch, fill cup with solder, insert 2 gauge wire, hammer crimp quickly) think I saw that somewhere on youtube and it made sense given that my hammer crimps weren't always precise. I did it for the small uninsulated balance leads on the portable packs I built just to get the tiny bit of wire poking out of the tunnel on to the surface of the ring terminal flat. I'll just up my crimper game and get something ratcheting. If I did more large battery cables I'd throw the hammer crimper in the trash and get a hydraulic one.

I charged the pack up to 3.45Vpc last night and then started a discharge @ 20 amps before going to bed. It's still going, 170ah consumed. Under this constant load and with 170ah taken, the cells are fluctuating between 10-15mV difference. If I flip off the load they go back to 1-3mV. Cells all reading 3.15v with load and at this low SOC. Would assume the seller claims of capacity are accurate but I don't want to push it and have to watch the cell monitor like a hawk. Also would never draw 170ah in service so i'll conclude this discharge momentarily. Pretty thrilled right now. The BlueSea terminal mount bus bars will be here Saturday, the last thing I need to complete assembly in to its ammo can home.

Only thing I might add if I can find something of quality, are adhesive backed 12v heat strips. Use load output on charge controller to operate a 12v temperature controller that powers the heat strips below 45*F. That'll keep it warm and capable of accepting charge this winter and I can't assume heating a near sealed ammo can would require a ton of power even with ambient temps well below freezing.

While that'd respect the battery by operating in configured boundaries of the load output, I think it'd also make the heater operate at midnight when it doesn't need to. Maybe I can add something between load output and the thermostat, like a light sensor. If bed of truck is pitch black, don't bother heating it. That might cause some delay in charge kicking on in the morning (sense light, work for an hour to get the box up to temp when it could be charging had it been warming earlier) but worth a test. Any ideas?

I used to solder anything over 8ga. Later found the solder had melted out of some. If I was doing rework the copper would be brittle close to the solder lug and the wire would be green even if it was cut back 4"-6". I learn later that the insulation on the wire is water permeable. Everything I do now is tinned wire / crimped / and coverd with adhesive lined heat shrink.

If you used heat pads leave an air gap between the pads and the battery. Heat will damage Lithium just as much as cold does.

 

[ZWilson07]

Final update on this before I create a build thread as I'm throwing it in to an ammo can. This will include all the stuff learned from this thread including getting a custom pack in to operation and will be the steps I follow to a T on my next build. Yeah, there will definitely be another as I scope out rural property for a SHTF/cabin deal.

I'm a dipsh!t, balance leads or ring terminal fuvkfest caused issues, a lot of these posts wouldn't exist had I known that was my culprit.

Pack charged up in above post and I got it to end balancing at 1mV. So, it's time to discharge and see if I'm still observing a 200mV difference on one cell, indicating it was borked.

Now hitting it with 270w load and i'm floating between 5mV and 15mV imbalance. This is almost 4x the average load I'll have on the pack. Everything is staying tight. If I switch off load it jumps back to 1-2mV difference instantly.

I'll see after many hours and 100+ Ah taken how balanced things are under that sustained load, and when cutting load.

I'll charge up to 3.45Vpc after this, discharge again until I'm content with what I've pulled or a cell starts indicating it's gonna go weird on the low end. That'll be my configured pack size on the Victron BMV.

Thanks for sticking with me.

Dont beat yourself up too bad. I for one learned a lot from this thread and I never wouldve came across it without all the struggles.

 

[hour]

If you used heat pads leave an air gap between the pads and the battery. Heat will damage Lithium just as much as cold does.

Been thinking about this for hours. I suppose heat strips would be best suited on the bottom walls or absolute bottom of the ammo can, and if on the absolute bottom the battery could be sitting on top, elevated-with an air gap.

Also no clue on the wattage to use for the heat strips. Luthj mentioned a 90w setup but I'm guessing low and slow would be better? A ton of heat output would find its way to the top of the can and potentially tell the BMV temperature sensor that it's good to go, when the top third of the can is 70*F and the rest of it, and the cells, are 20*F.

Wondering if this could be done directly from the load output of the SmartSolar utilizing the street light function, but inverted (can't figure out in the app nor literature what will happen, and it apparently needs to learn so I wouldn't get any quick results running tests indoors)

What would simplify this greatly is:
Sunrise setting from streetlight function reads greater PV > min detection (11.6v in manual) and satisfies condition #1 of operating load output
BUT only if battery voltage is in operational range defined in the Load Output setting (streetlight already respects this so I read)

So night has ended, battery is in an acceptable voltage -> power up the load output. Load output runs to a temperature controller inside the can which turns on. If it's too cold, flips to operate the heat strips.

If the load opens at (11.6v in manual) to signify sunrise, that's still a little bit from the panel voltage rising enough to initiate charging, +5v over battery voltage. Possibly time to get things mostly warmed up. Run 60w or less through temperature controller, in turn safe for the limits of load output.

Have a feeling they didn't design it to work like that.

 

[Alloy]

Been thinking about this for hours. I suppose heat strips would be best suited on the bottom walls or absolute bottom of the ammo can, and if on the absolute bottom the battery could be sitting on top, elevated-with an air gap.

Also no clue on the wattage to use for the heat strips. Luthj mentioned a 90w setup but I'm guessing low and slow would be better? A ton of heat output would find its way to the top of the can and potentially tell the BMV temperature sensor that it's good to go, when the top third of the can is 70*F and the rest of it, and the cells, are 20*F.

Wondering if this could be done directly from the load output of the SmartSolar utilizing the street light function, but inverted (can't figure out in the app nor literature what will happen, and it apparently needs to learn so I wouldn't get any quick results running tests indoors)

What would simplify this greatly is:
Sunrise setting from streetlight function reads greater PV > 0 and satisfies condition #1 of operating load output
BUT only if battery voltage is in operational range defined in the Load Output setting (streetlight already respects this so I read)

So night has ended, battery is in an acceptable voltage -> power up the load output. Load output runs to a temperature controller inside the can which turns on. If it's too cold, flips to operate the heat strips.

If the load opens at anything over 0v, that's still a ways off from the panel voltage rising enough to initiate charging, +5v over battery voltage to initiate, or whatever the # is. Possibly time to get things mostly warmed up. Run 60w or less through temperature controller, in turn safe for the limits of load output.

Have a feeling they didn't design it to work like that.

Yeah I went through all the senarios. There is no liquid in Lithium so low and slow is the best option.

Glycol is the best way to heat the cells. This way each cell could have 40F-59F liquid surrounding it.

We'll head out when it is 20F which meant starting the heat (BUT NOT THE SOLAR CHARGING) 2 days beforehand. This and the possibilty of forgetting to disconnect the solar charge on after each winter trip is why I went with FLA.

 

[hour]

Yeah I went through all the senarios. There is no liquid in Lithium so low and slow is the best option.

Glycol is the best way to heat the cells. This way each cell could have 40F-59F liquid surrounding it.

We'll head out when it is 20F which meant starting the heat (BUT NOT THE SOLAR CHARGING) 2 days beforehand. This and the possibilty of forgetting to disconnect the solar charge on after each winter trip is why I went with FLA.

2 days before hand? I'd consider trying to power the temperature controller constantly and routed through a LVD (or load output terminals configured), would be interesting to know how much power it takes to maintain the ammo can and what could be gained from adding an insulated blanket to it. But a foot of snow on the panels for a few days would mean the can was unnecessarily heating itself when nothing was going to be charged anyway. Really need a way to detect PV voltage inline, evaluate, and switch a relay fed from load out.

 

[deeznutz206]

Wow such a long thread, lol. So the packs you got were around the 190AH rated capacity? Would you recommend these then?

 

[hour]

Wow such a long thread, lol. So the packs you got were around the 190AH rated capacity? Would you recommend these then?

I guess my only gripe left was in first post, had to screw around getting a stud to thread in straight. And the slight curvature to the cells may have been reduced but is still present. Shamus said his cells straightened out. Plus his were the same cost as mine and by CALB. I don't feel like I got burned though. And I can only say that there was 175ah in the battery and more, just dunno how much

This screen shot looks a little screwy but you can see my discharge ending today (and last discharge) netted 175ah. It only says "Number of full discharges: 1" because I reduced the capacity from a presumed 190ah to 175ah before initiating recharge. I still had 5% left with it set to 190ah.

And I think the min battery 12.12V in screenshot is from the 150ah discharge I did the other day and maxing out my inverter. I think they were still all at 3.1X at the end of 175. I logged that several posts up. Anyway, FWIW. Oh yea, time since last full charge 20 hours ago but I'm -1Ah currently, so today's re-charge is about done.

 

[Alloy]

2 days before hand? I'd consider trying to power the temperature controller constantly and routed through a LVD (or load output terminals configured), would be interesting to know how much power it takes to maintain the ammo can and what could be gained from adding an insulated blanket to it. But a foot of snow on the panels for a few days would mean the can was unnecessarily heating itself when nothing was going to be charged anyway. Really need a way to detect PV voltage inline, evaluate, and switch a relay fed from load out.

2 days.......yeah. There's no way to measure internal temp and I didn't want to take any chance on damaging 3x200Ah Victron

For heat my thought was a 1/2 gal tank with a 12V silicone heating pad and a 12V circ pump

bayite BYT-7A006 DC 12V Solar Hot Water Heater Circulation Pump Low Noise 3M Discharge Head 2.1GPM - - Amazon.com

bayite BYT-7A006 DC 12V Solar Hot Water Heater Circulation Pump Low Noise 3M Discharge Head 2.1GPM - - Amazon.com

www.amazon.com

 

[hour]

This

Damn, what a pain in the ass. I can understand why you went with FLA given that potential setup... but yeah, the inability to measure internal temp...

I've discovered a handful of threads and githubs across the web for reading data from the ve direct port on the smartsolar (5v) and the BMV (3.3v). My objective is to read PV voltage so I'll be connecting to the smartsolar using an arduino nano to start, and then an ESP chip later.

Nano will read PV voltage over ve direct port and determine if it's capable of (or on its way to) being able to initiate charging. I might be able to get the temperature of the BMV from the SmartSolar, since the SmartSolar is receiving that information wirelessly from the BMV. It might just give me its own temperature reading which wouldn't be useful. In any event, I'll add some additional temperature probes to the arduino.

If PV looking good and temperature low, run some heat strips off the load output (skip all streetlight stuff, operate solely on battery voltage being within defined range). Still try to figure out something to ensure BMV temp sensor doesn't start reading optimistically, and speaking of optimism - make that an optional mode on the arduino where it'll anticipate charging the following day and start warming well before sunrise.

 

[hour]

For what it's worth...

a $5 Arduino Nano clone and some C-work and I'm reading values from SmartSolar MPPT over VE. Direct. Also got it working on ESP8266 units but having serial communications issues intermittently so I can't even deploy to one anymore. Will deal with tomorrow but can read anything/everything. This works on BMV units too, but you have to buy a 5v->3.3v step down if you want to use a 5v arduino with them. TL;DR - BMV operates on 3.3v, SmartSolar on 5v. A 5v arduino could damage the BMV. An ESP32/8266 would work for both.

---

Anyway, here's where the ammo can gets smart. I'll be using a Particle Photon (3.3v) and it'll be in constant operation, not going to bother with the load output on the MPPT. I'll be installing an SMA to u.FL jack through the can for an external antenna so the photon can speak outside of the box.

Relay will be controlled by photon and operate heat strips. Photon will measure temperature in a couple of places and based on those readings, battery voltage, and panel voltage, operate the heat. I might incorporate a real time clock module in to this if I find I'm missing a lot of morning light only starting to warming up the box when the sun is coming up, or that I should start slowly warming up a few hours before sunrise and if low panel voltage by mid morning cut it off (foot of snow on roof)

Since I have an LTE router in the truck that's to be operated by the house battery I can push all the MPPT and BMV data to the cloud and do some data logging way cheaper than the off the shelf solution. If I swap out the Photon for an ESP32 I can do all of the above + communicate with it over bluetooth as well as wifi. Then I could build a cross platform app to view all of the stuff I want to see on one screen that would give me the same data whether I was physically near the truck (bluetooth) or fall back to wifi with reduced refresh intervals. And graphs that don't just run for the duration that the app is open (VictronConnect). If I didn't have a router I'd probably get one of the cellular dev boards. Oh yeah... push notification alerts and ability to control system from anywhere.

Anyway, that's my idea for temperature related stuff and data collection.

Then there's this other thing I've been thinking...

I ordered two bluetooth BMS units the other night and they should be able to talk to the brain via wire, or even wireless if I use a bluetooth board. VE Direct is also two way (though less to be found on sending info) so I think I might be able to instruct the MPPT based on readings of individual cells. Reduce charge current, disable charger completely. Also the ability to do what a BMV traditionally does and communicate with BatteryProtects. Selectively shutting down loads and taking control versus the traditional BMS behavior of on or off.

Just adds some additional layers of safety and should communications fail those devices can just operate on their own settings.

I hate having to trust a chinese BMS, whichever of the two I use in my project will be a last line of defense but mostly used for reporting readings to something I can actually program. If there were an assembled board available to simply talk to an arduino a person could add a wholeee lot of jazz to an existing setup for $20. They couldn't cost that much given bluetooth models for $40 with all the other BMS capabilities..

 

[hour]

So, the above sprawl worked out.

I received my Dalygreen non-smart BMS yesterday after like 40 days waiting. Thankfully I didn't notice until this morning when I was going to work because the two bluetooth BMS units I had ordered arrived this afternoon after like 8 days waiting. Both from China.

Reading data over ble like a champ right now. Guess this thing tracks amps in/out too but I haven't calibrated any of that. Speaking of calibration, you can adjust each cell for complete accuracy. I found it to be a perfect match to my iCharger X6 and the quality variant of RC cell meter I own OOB.. I confirmed it was balancing the cells when set to balance at 3.40v, and did so to .001

"What does any of this mean?"

Well, I paid $46.20 for a 60-amp BMS with bluetooth that I can communicate with from a $6 ESP32. The BMS alone and the goofy chinese app (and a third party one) are adequate.. but pairing to another device for complete control - A+

So, from that ESP32 I can either serve data to my phone/pc/other device over bluetooth / wifi, or I can push that data to the cloud given wifi connection or using a cellular microcontroller. You can see the values being printed out in pics.

Operate relay on PV in based on cell voltage. Operate relay / BatteryProtect based on cell voltage. Turn on heat/fans. Anything else you could want to do.

Best buy of 2019. If I could go back in time I'd probably omit the BMV from my system. The VE. SmartNetwork voltage sharing from BMV to SmartSolar doesn't work for ******** anyway which was 50% the reason of buying it, and the temperature control ability (the other 50%) would be better done with a microcontroller and a few probes around the box. Hell, I already get two temperature readings from the BMS.

 

[hour]

 

[Rando]

This looks like fun!

Which bluetooth BMS did you end up going with?

 

[hour]

This looks like fun!

Which bluetooth BMS did you end up going with?

The 'lifepo4 60a with UART' version 4S - 5S (the text is clickable, i dont know if links turn blue on latest forum update)

and the 'lifepo4 30a with bluetooth module' version 3S or 4S

both came with bluetooth modules

from lithiumbatterypcb.com

Note that $58 gets you a beefy ass 100 amp version with bluetooth module

I've done some more testing and these things are the bees knees. The Dalygreen non-smart BMS I got was $45 or something for 80 amps and is a nice unit, but I'd much rather pay $13 more for a giant 100 amp like the link above and have all the smart stuff.

On to the smart stuff... not everyone would want to bring in another microcontroller and the crap I'm doing. How about if you just want a BMS that you can set custom low voltage/high voltage alarms and cut off on, see power consumption, or shut it off in a temperature range? And set custom balancing ranges? That's still a whole lot of usefulness and the xiaoxiang app from app store (confirmed to work fine on iOS from like 60+ feet away) delivers. App probably works better on android even. In short, anyone buying a BMS for a custom build should give these things a whole lot of consideration.

 

[hour]

Bee's knees. Got it all hooked up. Have some massive issues reading data from the MPPT reliably which was no good for the purpose of performing actions based on values.

Basically when parsing something like panel voltage, 'VPV 220933" I'd drop characters. So I had good luck reading data with an arduino nano and built that to be a slave to the particle photon. Only problem is then I had problems reading data from THAT. It's like the process of analyzing BMS data is overwhelming to then accomplish some sort of data reading with SoftSerial.

Anyway, my main goal of talking to the MPPT was to find out its panel voltage, so that I could leverage that for calculating when to run heat. I ended up sending a 0 if < 13v, a 1 if ~13v, or a 2 if 19v+ (can't charge, might be able to charge soon, could charge now). Light enough to parse from serial stream and with a high enough frequency of success to use for heating. Since I'm pushing all the data to the web and broadcasting it over my local area network, I figured I could start building the app.

Note this app has about 15 minutes in to it, and was just to set up parsing for the data. Pulls every 4 seconds on LAN.

Best thing I decided to keep track of is the highest cell voltage (CMAX). On my old pack I'd stand watching solar get close to entering float, and stare at an RC cell meter to see what the highest cell peaked at before the charging got cut. Now that logs automatically and I can reset it, though I might autoreset it every night so I can see that information after every day of charging. As you can see, the charger entered float @ 13.810 volts with a high cell of 3.457. So I'll bump the MPPT down a tenth.

I'll spruce this app up over the course of tonight and get push notifications working and a settings page for temp on/off, alarms, whatever. Relay already connected and coded, installed, just waiting on heat strips. I got bike handlebar heaters for a test, they're thin junk. Some better silicone ones are arriving weds, 48w total. I don't have any room to place them on the sides/beneath the battery - or not without complete disassembly, so I'm going to stick them on a 5x5" aluminum plate that'll be screwed in to the wood face plate that clamps the batteries together.

Will test, insulate outside of can maybe. Next 4 nights are supposed to be below freezing so I should get to test it thurs



edit, fixed up enough to use for next week or two testing.