Update on the Solar Charged 18650 Blocks & the diyBMS - 12v Solar Shed

Ten days after setting up the Tracer A to charge my 18650 lithium Ion blocks I’ve learnt a few things about them and the diyBMS which I’ve built to keep them balanced. Oh and I answer some frequently asked questions about the diyBMS and how much it has cost me.
0:19 Tracer A Settings
1:54 diyBMS Software Update
3:48 diyBMS Hardware Issues
7:41 diyBMS FAQs
10:30 diyBMS Costs
Costs spreadsheet bit.ly/2OB9GPT
Items used in this video:
❗️These links are affiliated and I may earn a small commission if you purchase❗️
EPEver Tracer A: ebay.to/2JDJwsf
Wemos D1 Mini: ebay.to/2u12mUd
LM Buck Converter: ebay.to/2KlxMe6
Colin Hickey diyBMS GitHub: github.com/chi...
Stuart Pittaway diyBMS GitHub: github.com/stu...

[@QUADBOYification]

Great series of videos, you got everybody entertained. I see the fun in it as well. I would consider going a different route, buying 21 pieces of 3.7v Li-ion Aluminium cased 14Ah cells 300 gram each. Making a 3S7P 11.1v 6.8kg Pack with 98Ah Capacity = 1 KWh in Total. These cells cost $10 each and you would only need 1 beefy 3S7P 100A BMS. Building it all together in a tiny case or on a piece of wood on a wall is much simpler. An old UPS unit will suffice as an inverter and if needed a constant voltage adapter can even charge the BMS-protected battery pack if solar isn't available on rainy days. Having an AC load is the main purpose for this to be useful. You basically want it all to be mobile.

[@AubsUK]

I've added this to @Adam Welch 's "EPEver Tracer AN MPPT Solar Charge Controller Review - 12v Solar Shed" video at kzbin.info/www/bejne/q2W6dXWkpteekLc, but adding here for convenience as it relates to 1:48 of this video too. Just bought the Tracer1206AN, it comes with a user manual that is V2.4, the one on their website shows V1.0. I've asked them for an updated version. There's 9 battery choices (Sealed/Gel/Flood/LFP-4S/LFP-8S/LNCM-3S/LNCM-6S/LNCM-7S) and 1 for User. It has a table for each of the LFP and LNCM batteries, along with User, so it looks like newer revisions do properly support Lithium. For reference, the values for Li(NiCoMn)O2 7S are below, voltage in brackets taken from 1:48 of this video... Over voltage disconnect voltage 29.8 V (29.75) Charging limit voltage 29.4 V (29.25) Over voltage reconnect voltage 29.1 V (29.20) Equalize charging voltage 29.1 V (29.10) Boost charging voltage 29.1 V (28.70) Float charging voltage 28.4 V (28.00) Boost reconnect charging voltage 28.2 V (27.50) Low voltage reconnect voltage 24.5 V (23.00) Under voltage warning reconnect voltage 28.4 V (24.00) Under voltage warning voltage 24.5 V (23.00) Low voltage disconnect voltage 21.7 V (22.40) Discharging limit voltage 21.7 V (21.00)

[@AubsUK]

I got a copy of the PDF user manual that is V2.4 for the Tracer1206AN, I've uploaded it to: aubs.uk/solar/Tracer-AN-Manual-EN-V2.4.pdf - Download a copy if you want to keep it as it won't be there forever. Selecting the type of battery is on page 16, Lithium Battery voltage parameters are on page 20.

[@Royaltea_Citizen]

Thanks so much for sharing costs, links and doing all this work and taking time to.make the videos! Makes it so much easier for me to understand and get my own diy projects going.

[@AdamWelchUK]

You’re welcome. Thanks for taking the time to watch and comment! :-)

[@ahaveland]

Excellent video Adam, glad that Colin has finally implemented a balance threshold, as it is pointless balancing when the pack is half full. I'd seriously consider using this for my system, currently 14S50P or 7S100P, but the time and hassle of collecting the parts and ordering them is more daunting than the build! If someone could do all this and a kit with a markup on it, then I'm sure it would sell. Note to Colin/Stuart - next level balancing would be charge redistribution using a supercapacitor to slurp and pump from each cell via solid state relays and a multiplexed buck/boost converter. It's easy enough to do this manually with a couple of flying croc leads though tedious, but quite a challenge to isolate and address each cellpack for charge/discharge a bucket at a time and do it automatically.

[@AdamWelchUK]

Thanks Andy. All the module parts are available from places like RS and Farnell but they are more expensive - although cheaper in bulk if someone wants to start making kits or even having them produced. An hour searching through their website and the only other items you’d need to source would be the wemos bits and the pcbs. A few people are commenting about buying them pre made or in kit form. Perhaps our friends in China will build them? :-) Moving the energy is always more sensible and I’ve looked into this a little. A fair bit more complicated though than this. Cheers

[@dejayrezme8617]

Yeah sign me up too! You could probably look up sellers on ebay that do similar things and ask them if they'd make and sell this.

[@tg_privat]

I use REC active BMS. I don't have much experience with electronics so I was too afraid to use some generic BMS. I also didn't like the idea of wasting energy in resistors. When I heard, about active balancing idea I was amazed about it. Only the prices were just crazy. In the end I found out that a company 50km away from my home (Ljubljana-Slovenia) is making these things for big players. The price was manageable so I have one now. A bit of overkill for my 4s9p battery pack, but this 18650 pack is just for testing. I can control a lot of parameters and it can also trigger two relays. Charging and load. It is much more secure in my opinion this way. It also measures internal resistance of each pack and calculates SOC and SOH. Web interface is missing. The product is built for e-mobility after all. It is only 4S but enough for me. It can balance the cells with currents up to 2 amps. Top for LiFex. My project is somehow different while I'm trying to built a battery to work together with AC solar grid tied micro inverters (2 x 500W). So I could also charge it with AC charger (as solar dumper) and return the energy (slowly) over the night to my AC grid with grid tight inverter. Over winter I would go for lower and higher (night/day) tariffs. Hipoteticaly it could be better ROI than investing in solar. But for now I wouldn't "put money" on it. :-)

[@dejayrezme8617]

Thanks I didn't know about these!

[@tg_privat]

I'm not sure, but their activeBMS technology is used in last VW Golf-e

[@diydsolar]

Hi Adam. I have one thing missing. You have configured your MPPT with a lot of parameters. Let's we speak about boost or overoltage one. How these parameters work with the diyBMS parameters ? I mean, I we configure our MPPT with a voltage of 28.70v (4.1 per cell), and we configured our diyBMS with a maximum voltage of 4.15 per cell, probably we will never reach 4.15 per cell, but also, we can reach one cell to 4.15 and another one to 4.05, because MPPT will cut of at "28.7v total". The other way round. If we configure the MPPT to 29.4v, for example, to get 4.2v per cell, and diyBMS to 4.15v, the MPPT will never stop, we will get the cells balanced at 4.15v, by MPPT will continue putting 29.7v to the 7s cell pack, and all the cells we will "wasting" 0.05v each. How these parameters live with each other ?

[@nordio98]

Hi adam, I realy like your project, and I also have the same question from Fernando

[@MrKitesurfercharlie]

Thanks for the video Adam. I was assuming I was doing something wrong, but now I know to use a different resistor.

[@AdamWelchUK]

The original 510 might be ok - but as mentioned in another comment something around the 300 mark might be best. I’ll give these options a go soon.

[@CrankyCoder]

I have been working with Colin and using his firmware as well. The bypass status also gets passed now to emoncms and influxdb. So you can see all the bypass status in both systems.

[@AdamWelchUK]

Yes I’m aware of that - we talk pretty regularly. I’m expecting a new video from Colin pretty soon on the updates to the software and his grafana dashboard. I think our experiences have all been very positive - and it’s looking like a decent solution.

[@CrankyCoder]

absolutely! i have both emonscms and grafana running with it currently. Just so i can see what both have to offer. Been very happy with the project and really happy to see others using it to.

[@svvelanautica2374]

Hi, I like you project! and you great videos. maybe I missed something? What are the maximum charge and discharge currents handled by this system? I like to setup a system for my boat, but it needs 600Ah batteriebank, charging current up to 175A, discharge current, continuous 40A up to 150A max. Go on Oliver

[@RWBHere]

Thanks, Adam. Now all you need is a vented enclosure to keep it all clean and tidy. :-)

[@emceh]

Adam - excellent video - one question - how do you reprogram tracer?

[@TheFarleyGordon]

Very nice Adam And thanks for sharing.. I am about to the point to start my system and I believe these to be the next best thing to batrium. Would you agree?

[@AdamWelchUK]

I haven’t used the batrium system so I’m not fully qualified to comment. However if Batrium is your premium product and the Chinese BMS’s are you budget brand then the diyBMS is most definetly your mid market choice!

[@georgepene6097]

Hello interesting video, i do have some issues setting mine like you said. User settings and after I've input all the values when i press save its saying that the values are not correct?!.

[@johnarmstrong3782]

Super project! A couple of layman’s questions. Where will you draw power from? The battery bank or the charge controller? Is there any advantage in drawing power from the load output on the controller? (Always wondered about this for lead acid as well.)

[@clintoncoker6]

Only an advantage if the charge controller gives you additional functionality that you don't otherwise have, e.g. low voltage disconnect, or energy metering. Otherwise, no.

[@clintoncoker6]

Depending on the current rating of your controller vs the capacity / current rating of your batteries, the load output of your controller may be very limited e.g. 20 or 30 amps, but your batteries may be ok to draw 100A from, otherwise (particularly the case with lithium).

[@xanataph]

Just four days to the Day of Imbolc here is NZ! Once upon a time this was all about when the sheep began to lactate (that's what the name means, literally "Ewes Milk", in Gaelic I think.) However, I reckon it's an important day in these modern times when it comes to off-grid! It's the day which ends that super-poor part of the solar power year! :)

[@camolas71]

Hi, I have a small 12v inverter so I can run it at 15.5v max. I made a 4s20p battery pack and I need some advice how to set the parameters on the triron 2210n my PV panel is 320w 24v mono. Can you please hellp? Thanks in advanced

[@timrauscher3965]

Good context Adam. Thanks for that information. The only question I have is will these BMSs Handel a battery pack in the range of 100p 7s?

[@AdamWelchUK]

Each module has been designed to have a maximum bypass current of two amps. As long as your 100p packs have been reasonably well sorted that should be fine - I guess in theory you could pop two modules in one pack if you wanted. They should both read the same voltage and that shouldn’t effect the average balance algorithm in theory.

[@timrauscher3965]

Adam Welch I might have to give that a try. Thanks for your response. Tim

[@MrFreeway99]

I'm wanting to try this project for myself but I can't figure out how to order the pcbs

[@rezaashk1000]

please make a video with your pure sine wave inverter (EPever) under the maximum load and see what happen to the waveform of inverter on oscilloscope

[@theophileiradukunda5769]

Hi, i have a simple question if I have following specification 1. Solar panel: 20W 18V 1.1A max 2. LED Lamp: DC 12V input:2W ,200lmX1.1W,100lmX2 3. Battery cell:LiFePO4 battery up to 2000 cycle life 4. System capacity: 6.4V 40WH 5. System input : DC 18V 1.5 A max 6. System output: DC 5V 2A max/DC 12V 2A max . can you guide me on how I can calculate the battery capacity

[@nosafetyswitch9378]

About R12, the green LED has a voltage drop of around 1.8v and you dont want more than 10mA flowing through it, I d say 5 mA is enough if it is SMD. R12=(3.3-1.8)/0.005=300 Ohms. So something around there should do the job.

[@AdamWelchUK]

Cheers for the maths. I’ll be replacing them soon.

[@nosafetyswitch9378]

no worries! Glad I could help with my electronics 101 knowledge :P

[@johndoe-rd3hq]

Nice. What are you powering ? An ice machine? Air conditioner? Microwave? And with all this lithium ion in one spot wouldn't this be a BOMB in your shed in case of massive electronic failure?

[@brettmcconnell8040]

be great if you update speedsheet with cap voltages to make it easier for people to find parts

[@WildLifeBackyardCamera]

Hi Adam, you got a 7 cell with 20 18650 each (7S20P - 40Ah) system, we can assume that charging a single cell at 4.2V with 20 A is ok too. On your system, in average, how much Amp receive each and every 18650 ? How much Amp in total you got out of your Tracer ? Thanks a lot :)

[@AdamWelchUK]

I’m treating these old cells very carefully and only put a maximum of about 4 amps into them day by day. Yes I could charge these packs of 20 cells at 20 amp - but I haven’t the roof space for that amount of solar sadly. Charging a single 18650 cell at 20 amps might be a bit aggressive but some are able to be charged at quite high levels.

[@WildLifeBackyardCamera]

@@AdamWelchUK thx adam for ur reply, 4Amp per cell / 20÷4 ~ 200 mA per 18650 / 7*4 ~ 28 Amp from tracer. It sound like it must take you a while to charge, right, smth like 3.5H with 200W solar panel at ~ 50% discharge battery ? I'm trying to do do the same system as you just with 3S10P ,smartsolar mppt and stuart DIY BMS :) EDIT: realised u already mention the time you need in one of your video, sorry :)

[@AdamWelchUK]

As I said I’m very gentle with them. Even when I discharge them it’s usually at low levels. Good luck with yours.

[@Marcos_Ledinich]

Hi Adam, i will configure my EPever tracer for a 3s set up in my houser using your posted parameters (for each cell), do you think its ok if i go with that? i want to go that way as i already have a very expensive 12v inverter for my house

[@oldtimeengineer26]

What ever happened to the interference you were having? Also how many modules can the software handle? Reason I ask is I have 7 shelves with 7S modules each can that be setup? Good video

[@AdamWelchUK]

Colin has been having interference issues which look like the inverter and or wiring is causing some i2c communication issues. I’ve not seen the same issue, but I’ve not got an inverter connected as yet. Each controller can do 14 modules and due to the fact that the wemos is isolated from the bank it is monitoring I guess you might be able to use one controller for two different 7s packs. As I showed in the video though, the wemos bit is the cheapest. If it were me, I’d be tempted to either consolidate my packs down (repack them or run them in parallel) or have a wemos controller on each pack. Cheers OTE

[@Chickey]

Yup that was me and I've solved it finally. Video shall be coming soon :-)

[@dama054]

hi Adam nothing to do with this video, have you come across any type of relay board that can switch using frequency as a trigger so if mains frequency goes above 50.2 HZs it would turn on a relay, or could you come up with a pcb that could do this Thanks David

[@tengelgeer]

A standard green led has a drop of 2V, only the extreme high power (phosphor based) ones are 3V. But 2k2 might be a bit high. Although a modern led should be visible just fine with that 0,5mA. And yeah, 22Ohm might be fine now the pack itself (aka the capacity) is balanced. But if it ages and capacity of modules changes you probably need more balancing. And remember, having a lower value balancing resistor is not, I repeat NOT, more wasteful ;) To balance you need to get rid of a fixed amount of energy. With a lower value resistor that simply happens quicker aka the time the resistor is turned on just becomes less.

[@simplyflipflops]

Adam, Can't you run the Wemos from the 3.3v source powering the ATtiny (ISP1 pins 2 and 6) and skip the buck converter? There's even a easy to access plug ready to go.

[@AdamWelchUK]

In theory you could power the wemos from one of the modules, and as you say from one of the ISP headers. It would remain isolated from the rest of the packs, so there shouldn’t be a short. The trouble with that method is that that single module (and therefore that block) would be powering more kit than the rest. It would be powering one side of all the ADUM1250ARZs (which admittedly won’t be much current) but also the wemos - and they can pull about 80mA when transmitting. That would probably lead to an imbalance in the pack you’d end up fighting all the time. Besides, that might be approaching the current limit of the REG710. Good idea, but I’m not sure it would work in practice very well.

[@simplyflipflops]

Wouldn't the buck converter be pulling more current all the time from each pack at less effiency anyway? Why pass the buck (pardon the pun) when you wouldn't have to take the loss?

[@AdamWelchUK]

My theory is that although possibly not quite as efficient, powering the wemos via a buck converter powered by all seven packs will ensure they are all carrying the same burden and help with balancing.

[@simplyflipflops]

Adam Welch doesn't hurt to try now does it? :) Could be a fun test to try and report. Theory and practice are two different things.

[@simplyflipflops]

Perhaps I misspoke. I thought each battery module had its own wemos. You're talking about one wemos for the whole system. I get where you're going here. But maybe that's why you balance charge anyway. :)

[@electronic7979]

Excellent

[@tl1024]

What restricts/prevents this system from managing more than 14 cells? I could really use this setup with 16x 3.3v cells.

[@AdamWelchUK]

Looking at the code again more closely Stuart has allowed for up to 24 modules - however I’m not aware of anyone testing it with that many.

[@tl1024]

Adam Welch, Thanks! I really appreciate your answer. PS, your channel is about to be huge. Loving it. Thanks!

[@MadMods4u]

great info... thanks

[@MikesDIYTeslaPowerwall]

Should be able to go way more than 14s, the limit on i2c is over 100 from memory

[@AdamWelchUK]

I guess it may well support more than 14, but to my knowledge that hasn’t been tested. Didn’t want to suggest something and be wrong. Cheers Mike.

[@clintoncoker6]

According to my knowledge *if it wasn't for the BMS* what's the point in using the boost function at all? I have an EPEver LS3024 powering some big lead acids, but my system is very lightly cycled, so I basically don't use boost (boost reconnect is very infrequently reached). It's even less useful for lithium, I would have thought. Even with your BMS, why couldn't you set the parameters to not initiate boost charging (set reconnect very low, or make float and boost voltages the same) and set the float voltage to something in between your BMS's max cell voltage and the balance voltage, eg 4.13v/cell? As I see it, saying "float voltages" are the enemy is a bit incorrect, really it's just sustained overvoltage that's bad. So I guess it depends on your usage as well (I realise it's minimal atm though). Or are you not happy with the cells being held at 4.13 (for example) during the day? But they can't br dropping far below 4.10v at all atm. Just my 2c.

[@AdamWelchUK]

Thank you for the comment. The boost charge is going to be unavoidable, as it’s can’t be removed from the charge controller, but as you say with light usage, it may take a number of days before the boost reconnect voltage is reached. In the mean time, any use during the day is taken from the cells rather than the solar panels, which is why I’m worried about wastage. If you look at any lithium charging profile of any change commercial charger you’ll see that the charge terminates usually at 5-10% of the initial charge current. Its generally accepted that trickle or float charging lithium isn’t optimum and causes plating within the cell. Time will tell I guess, but my experimentation with my smaller 7s4p pack I ran for a year worked well. The individual cell capacities dropped very little after 365 cycles. Cheers.

[@clintoncoker6]

With my LS controller you can set the boost charging time in increments of minutes from zero to 180, so I can effectively switch boost off completely. I would have thought a Tracer has that ability too (could be wrong). I was thinking that if the controller float voltage was only just above the balancing voltage, the trickle current at float wouldn't actually be putting current into the cells - it would essentially only be powering the BMS bypass resistors etc. If you were still worried about plating you could just drop all those setpoints down i.e. BMS balance at 4.05/cell and controller float at 4.06 or 4.07 (obviously you'd have to check for voltage accuracy of the controller and BMS when you're running them so close). Maybe it wouldn't work but I'm not sure why not. I was thinking it was one solution to increase the utility of your panels.

[@AdamWelchUK]

Yes I think you can set it to 0 minutes. I’ll test this out and report back once I can. I’m still a bit concerned about relying on the float parameter but as you say I could burn that off. Interesting idea, thanks.

[@clintoncoker6]

Hmm you might be right about the float charging. In any case, you wouldn't want to float above 80% soc.

[@clintoncoker6]

The other option is to just increase the boost reconnect voltage so it recharges the pack more frequently.

[@cirelefebure5485]

Thank you sir.

[@alexandreesquenet3736]

Great video.

[@Reg-th6zq]

Adam I have the list but what is the pcb type number?

[@Reg-th6zq]

Others seem to have bought these pcb and are building, so where can they be bought?

[@jmanbrosef689]

Where can I find the 3d design for the wemos din rail mount? Thanks.

[@ollieb9875]

I think the links are in the description of the last video or so.. around there kzbin.info/www/bejne/gWPOlaeQltCLj9U

[@buttondeity3933]

Really need those andro bms made to purchase. Batrium is way too expensive

[@AdamWelchUK]

The batrium solution isn’t too bad when it comes to price - but I guess it is relative to how much you spent on your batteries, or how much time you spent putting the bank together. My biggest issue with the batrium is that the software is windows - which means you need a reasonably high power machine running to monitor it.

[@buttondeity3933]

yes that is a great point. I would rather the andro type system you are using. Are you able to monitor remotely? say from work or a friends house?

[@AdamWelchUK]

You could open up the port for the diyBMS on your router and make the webpage public, but as there is no authentication that would allow anyone to change the parameters. Stuart originally coded it to upload information the emonCMS dashboard and Colin Hickey has expanded the to Grafana as well. Both of these systems allow you to remotely monitor the pack and allow for authentication. Using either of them you can watch your pack from the internet. Another positive over batrium I think?

[@DIYTechRepairs]

Adam Welch i run My batrium mainly from Linux via the software i built. You only need windows for setting it Up ;)

[@AdamWelchUK]

DIY Tech & Repairs Useful to know. Thanks for that.

[@saadaldossri5348]

Thanks a lot.

[@pirateman1966]

So this is what a 1KWh setup looks like!

[@rontech34]

What is the point of having them batteries for?

[@AdamWelchUK]

They’ll be powering stuff soon enough. Granted at the moment they’re not doing much.

[@trickyriky1]

That's cheaper than expected. :D I would still like to get your opinion of my DIY solar monitor, i have a live dashboard with by the second reporting that i would like to show you, but don't want to make the link public, feel free to contact me in private.

[@AdamWelchUK]

Cheers tricky. Sorry I’d forgotten about that. I’m often using the mobile platforms and it’s harder to do the messaging thing. I’ll hunt you out shortly. :-)

[@chuxxsss]

Stay safe. Join the diy Charger Discharger group on FB if you would like I am one of the Administrators.