SecondHandPower Cell tester build

Joined
Dec 2, 2017
Messages
79
Before taking on the DIY Powerwall,I have to build up my test processto characterize all the cells.

After a couple of weeks getting to know cell behaviors and setting up a basic test station to start practicing my test process, I have realized it is time to move up from my 4 cell Opus to something that will be able to keep up with my "pack whacking" skills.

Have you taken the 10 pack challenge for fun yet ? See the videos and go have some fun.
I can whack the packs and clean a stack of 60 cells in about 15 minutes. I use an 8 oz rubber mallet and it reallymakes the plastic fly,spreading the impact to a larger area of the case while minimizing collateral damage to the cells.

Consideringmy current setup, I canclean 60 cells in 15 minutes then load up my 4 banger Opus tester. Those 60 cells would take about 6.25 days to test if I were to feed it every 10 hours. The maths are( 60 cells / 4 slots * 10 hours ea). I can add some bulk charging to help speed this up, sothat is a baby step I plan to take when my batch of cell holders and TP4065's come in.

While I am waiting for these parts, I have been checking out different designs I could follow to build a more scalable tester and have decided to use MrConstatin's design as a pattern with some small changes to the layout to fit an easy to obtain computer case.

The reference builds:
** Edit**
I have been considering MrConstantin's and MrFercsa's builds, and removed thoselinks, as both are easily found on the forum here. This is an area where you need to do your own shopping and evaluation, and there are many excellent tester builds onthis site that I have reviewed and many more that I have missed. I will keep looking and comparing the parts cost as I go.

Basically I want to test with 32 cell holders and want to havescalability, temperature monitoring anddata logging plus the "standard" feature of charging and discharging. If it is possible for the micro controllerto connect out to the web or to my smart phonefor reporting, thats a bonus. Having a hook to share cell_serial_numberwith a cell handling system (tracking outside the tester) would be a bonus.
**End edit**

My starting point:
I chose a sturdy tower computer case (a 2007 era Dell Optiplex 745) because it has a nice latch that allows one side of the case to open like a big door. There is ample cooling already in placeand space inside for a bigger power supply, the charging board, discharging board, relay boards and Arduino control boards inside, and I did a quick layout with 32 18560's to see how it structure of the case would allow meto add the display and keypad as well.

image_pyjeqp.jpg
image_npzrzt.jpg


Also collected a few cables from the computer that Imay be able to use later in the build.
image_mzhmdy.jpg

image_wydjyr.jpg


Now, lets build something....
 
Reusing an old computer tower is a great idea! I have several around here. Hrmm......

Taking a lot of the "junk" out of the tower will give you plenty of room. You might be able to repurpose a power supply shell and use the connector as your means of connecting the batter to everything else (possibly putting the same plug on the other end as well so as to not accidentally connect to the mains outlet). Then you could have several towers sitting next to each other and that gives you the look of a computer warehouse or something, but the power of lithium! ;)

How many cells max do you think you can store in the tower? If you could have each tower be like 5KW or something, that'd be pretty neat.
 
Assuming your cells arent discharged completely , your time using your opus will most likely be less how much less is anyones guess.
assuming 2.8 volts at start of charge, a1000 ma charge rate, 2000-3400mah = 2-3.4 hours
2-3.4 hours x 3 charge/discharge/charge 6-10.2 hours a perfect world test cycle
if you precharge the cells on another device , do a discharge only test on the opus, do your recharge on the other device. cuts the turnaround to approx 2-3.4 hrs


like mike does in this video. wish my setup was as good.
MrConstantin's tester is nice. I am afraid I would mess up the construction of it. easier for to use off the shelf units.

later floyd
 
Korishan said:
...
How many cells max do you think you can store in the tower? If you could have each tower be like 5KW or something, that'd be pretty neat.
@Korishan, the outer dimensions of this tower is 6.5" w 16.5" d and 15.5" h. I have not built any packs so I am not sure about the fit if you were to use them as a new kind of "cell tower". There are some some oddcompartments welded in here and there, so all the space is not easily accessable on the interior.

However, I am going to use the open space where the motherboard was... that area is 5"x10"x11" and should be enough room to put all the tester boards inside and use cell holders outside.


floydR said:
Assuming your cells arent discharged completely , your time using your opus will most likely be less how much less is anyones guess.
assuming 2.8 volts at start of charge, a1000 ma charge rate, 2000-3400mah = 2-3.4 hours
2-3.4 hours x 3 charge/discharge/charge 6-10.2 hours a perfect world test cycle
if you precharge the cells on another device , do a discharge only test on the opus, do your recharge on the other device. cuts the turnaround to approx 2-3.4 hrs


like mike does in this video. wish my setup was as good.
MrConstantin's tester is nice. I am afraid I would mess up the construction of it. easier for to use off the shelf units.

later floyd

@floydR,
The power supply in the tower is rated for 22A at3.3v . I have seen some folks using the 3.3v leg of the PC power supply to do bulk charging (I would assume they match the number of cells to the number of amps to get 1000ma per cell). I was thinking to try it once I get more than 20 cell holders. I would also put the Opus on the PC 5V supply because it is available.

I started going through all MrC's videos and notes and schematics to make an updated list forall the parts... having ot shop in three different carts online and it got late on me. Will show the list once I get it all together. Hopefully I wont mess it up too bad !
 
Just a quick update on the project. I had ordered some charger and discharger modules a couple of weeks ago when I first started thinking about scaling up, and in the same timeI decided to go aslightlydifferent route with my build, at least on the discharging side.

Now Ihave a pile (20)ofthese HW-586capacity testermodules(ZB2L3 design?)to sell or trade if anyone is interested. PM or e-mailme if interested in a fair trade at cost ofwhat I paid... you canget thesequicker from me than you can from China!


image_ptiodu.jpg


I like these modules because they are programmable, and that makes them flexible for whatever parameters you want to use for testing.I will probably use a couple ofthese to do bulk discharge for4-pack or 8-pack stringsof cellsto drain them toa specific voltagefor longer term storage since I am not sure how long it will take me to collect enough cells for my 80P packs (a long time, at therate I can test them on one Opus !)

FYI...BU-808 says that research indicates thecapacity of LI-Ion cells is better maintained at a voltage of 3.92v because storing at higher voltages (and in higher heat) can actually decrease capacity.

I am interested in some TP4056 chargers with protection if you have some to trade. I need about a dozenmore of those.
 
The cell holders came in today.... I decided to go with ahorizontal layoutbecause that is the most natural way to write the mAh on the cells.There are 4 banks of 8 cells, and all the cells in abank will beindividually controlled by aFCSA slave module. Inside the case, there will be 4 slave modules,the master module and a 40A power supply. Parts are on order...

image_eiuvvd.jpgimage_drfcqa.jpg
 
Looking good!

I have the reverse problem on my FCDS build, all my cell holders never showed up but have the rest of the parts!

Cheers!
 
neurocis said:
Looking good!
....
Thanks!!

I added some plastic covering for the cell holder screws. I Plan to mount the cell modules and slave modules to these plastic rails somehow. I hope the spacing available works out with the i2c connectors on the slaves.... I have 8-3/4 inches for the length of the slave modules and connectors .


image_drbskf.jpg
 
Heres a couple of shots showing underneath the cell holders.


image_oyzntl.jpg

image_wfasyy.jpg


When i get all the parts (in the next month or so)...I will be wiring up each of the (32)cell holders to work independantly with oneTP4056 moduleand a#FCDS discharge module for each port.

I drilled for the thermistors after these shots, and [temporarily]wired up a row of 8 cells to a couple of TP4056 chargersso I can now charge 8 cells at once.


image_pqxaiv.jpgimage_someas.jpg

The brown board below the cell holders has (6)zb2l3 modules that Iam checking out. I found that two of these do not discharge to the same voltage as others, even when set the same target voltage.These will be used to drain cells to a target voltage if i can get 6 that work within 2-3 mv of each other

The Opus is along for the ride on the brown board as well, everything is running off of the computer power supply.

I have to look inside the case tosee the reflection of status LEDs forthe TP modules, as shown in the photo to the right. The stuff on the brown board is not part of my tester build.... just sharing one power supply temporarily.

See Fercsa's#FCDS build page for details on how the control boards for this 32 port build will all come together.
 
What are you going to be used to power these on a regular basis when the pack is put together? My setup is using a 3p 10s configuration and all the cells are tested out from 1800mA to 2000mA and have installed aDiymore 3S 20A 12.6V Li-ion Lithium Battery 18650 Chargers PCB BMS Protection Board for every 30 cells. Any suggestionson what to use to keep these charged up?
 
larry1024 said:
What are you going to be used to power these on a regular basis when the pack is put together?
...
Any suggestionson what to use to keep these charged up?

Hi @larry1024,

With respect to what to use to charge your packs...
Most of the powerwall systems on this forum use solar panels and a suitable charge controller with BMS to charge and keep their packs balanced. I have seen a couple of people on the forum using hydro to generate the charge. Wind could be an option if that is a natural resource that works for you.

I am considering solar, but I have not gotten any of that type of hardware yet. I will be building some 18650 packs for smaller projects to use many of the lower capacity cells i am harvesting, and to learn more about charging and balancing packs.
 
The 32 portparts kit from Fercsa.com is in !


image_gdqkig.jpg

image_nwbiqi.jpg


To get the buildgoing in the right direction, Fercsa included one pre-assembled discharge module (DM)... Ineed to add a header to aTP4056 charger board so it will snap onto the DM, and the DM will snap into the Slave Module (SM).

There will be 8 ports per slave module, and the 4 slaves will be connectedbyi2c bus jumpers to the Master Module (MM).


image_mmboqh.jpgimage_jmnkzg.jpg


I have been collecting parts for the master module from various sources over the space of the last month or so... The power supply also came in today.

The only bummer is my new combo soldering station forfailsto power on. The heat gun (for SMD) works OK,but I get nothing in the temp control lights when I power on the soldering iron : l

The station is a Youyue 8685... There is a note saying "3) Plug the outlet power. Don't forget ground connection."


image_nhvssl.jpg

image_apqmzc.jpg



There is an unused connector near the power cord, butno photo referencein the manual and I cannot seem to find anything resembling aground wire part besides the traditional 3 wire power cord going to my properly grounded wall outlet.<hmmmmm>
 
That's for the anti-static antenna :p

I have no idea either. Kinda strange. Unless they re-used a case from another device that used that port and this setup doesn't.

Are we gonna be privy to a video of the assembly? And/Or review of it functioning?
 
Korishan said:
That's for the anti-static antenna :p

I have no idea either. Kinda strange. Unless they re-used a case from another device that used that port and this setup doesn't.

Are we gonna be privy to a video of the assembly? And/Or review of it functioning?

It's hard to say from the pictures. But itlooks like its for a bullet style connector or a banana plug. My guessprobably for an anti static wrist strap.
 
Confirmed... I opened it up and verified the bananna plug jack conects to ground. Also found a loose wire on the power switch so the iron will work but still not 100% since the temp readout is blank. I will have to open it again, but lets see what the seller says.
image_hbluyx.jpgimage_xiywlo.jpg


NOTE: To be perfectly clear, the soldering iron is from a seller on Ali Express... nothing to do with the Fercsa parts, everything is good with the parts from Fercsa, I just need to get in the groove with this new iron
 
Some progress on the FCDS master module.I have aFTDI USB programmer on orderto load the firmware, so that will take a few days


image_szalye.jpg

image_rebkir.jpg


The soldering iron is working good. Its worth the investment since there's a lot of soldering in this project.
 
I made it through the firmware download and update processes and my#FCDS master module is online!


image_uiixho.jpg

image_cvocjz.jpg


I had missed one wire ( a 3v3 line for the processor) but once that was fixed, it booted up OK and presented the status you see on the tiny display.

I am working through network configuration so I can connect to it using a console window before Iconnectit to the slave module and a discharge module.
 
Here is a photo showing the #FCDS master module cabled to the (4) slave modules. These components are upside down because they will be mounted underneath the cell holders.


image_ctvrir.jpgimage_dxbnxm.jpg



The photo on the left also shows one discharge module that will be connected to a modified TP4056 module. I have 31 of these to build, but will start by verifying thefirst one I buildworks to make sure I dont duplicate any errors across the build.

The photo on the right is showing the charting capability on one of the cells Fercsa tested... the channel is available from a link on his web page.
 
Back
Top