SecondHandPower Cell tester build

Announcement - Help us fight the BOTS! Please report all spam including stuff in your inbox!
Joined
Dec 2, 2017
Messages
79
Time for an update on building the #FCDS Discharge Modules (DM's). It took me a couple of weeks to get around to starting this phase of the project, but once I fixeda dedicated space for my soldering station and a place to put all the tiny parts, it was ON.

I spent a lot of time looking at the schematic on the web, and studying the pre-assembled example provided by Fercsa with my parts kit order.... here's how they look after 3 day's worth of project time spentsoldering componets and clipping leads.


image_dgjpsv.jpgimage_lndvfg.jpg

I am using egg cartons to organize the parts. Once I learned the suggested order to install the components, I decided to put a single component on each one of the 30 PCB's to minimize the chance of putting the wrong component in the wrong hole.

I will have 32 of the DM's when done... these 30 plus the pre-assembled example plus the first one I tried to assemble. The photo on the right shows me staging5 boards to add a yellow capacitor.


image_qbgohe.jpg

image_upqskk.jpg

I found amagnifying visor that works good to help me see what I am doing... in particular getting the soldering iron in the right place and applying the correct amountmount of solder. The photo on the right shows a redneck soldering draft fanI made from acomputer power supply cover. It really helps pull the fumes away from my face when I am in close.

An alligator clip on a flexible rod is helpful to hold the part in a specific direction while soldering.


image_fbrneg.jpg

... a little more than half way there now.
 
Joined
Dec 2, 2017
Messages
79
To connect the TP 4056 to the DM, there is some surgery required. Here is how I lifted a leg on the TP module so the DE could be connected to the discharge module. I used a strait pin and the soldering iron. There are two wires to solder on each TP module to its respective DM.

image_jcfjlc.jpgimage_vvscto.jpg


here is a photo of the first string of 8 modules as I start thinking about how to hang these underneath the cell holders (on the other side of this cover)


image_jjkoti.jpg

image_pyibay.jpg
 
Joined
Dec 2, 2017
Messages
79
Been doing a lot of business related travel, so my progress has been slow due to being away and having very little spare time for projects.


image_zrlvfh.jpg

image_sutpsx.jpg


Got the 5v power supply cabled up and mounted in the case. There isone string of modules assembled and hanging underneath the cell holders and thenew FCDS Master module powered on...

There are some issues I am sorting out to operate the new Master Module and get the wireless set up... Those issues have eat up a day or two here and there when I have time to work on it.

So far, the issues I have encountered were are all self inflicted.
 
Joined
Dec 2, 2017
Messages
79
Over the past week and a half,I took the opportunity to re-work the bank of discharger modules using new boards with the known good variety of TP chips. Since I had to pull everything apart I decided to put it all back together without being in a rush. This time, I invested extra time in making connectors for all connections on the discharge/charge modules so I can pull them out individually or swap around cell connectors for troubleshooting.


image_iounig.jpg

image_nrhgnc.jpg

I replaced all the TP modules and made the connections to the discharge module to fully automate the discharge/charge cycle and updated the firmware as suggested by Fercsa.
Note: I did not use the pin configuration shown in the picture on the right. That was too tall and was the first attempt for power connections that was used for individual poiwer wires to each board as shown in the bottom of the next photo below.


image_tdhsbr.jpg

image_tyszpj.jpg

I added a couple of 90 degree pins to each of the TP modules ,one on each side of the board so that the power rails could be run next to each other. I decided to use 14ga house wire, and this stiff solid wire works really good tomake all +5v and GND connections all at once.This works much better than using smaller wimpy wires that have a bunch of unsightly bends. The small wires did not want to stay attached and would have been a reliability pain. Thephoto on the right shows thelower profile, 90 degree pin arrangement for the cell connections.

This made a shorter "header stack" and a way to mount the SM where the center of gravity is... around the heat sink area, at the bottom of the header stack. Thisalshelps distribute the weight across two rows of pins instead of just one row.(remember, I am hanging these upside down underneath the cell holders).




image_uewrqp.jpg

image_lqlzfc.jpg

Then I added a 4 pin female connector to each set of cell holder wires in the tester top. +B and Gnd for cell power and 2 wires for eachthermistor.


image_yaqgdd.jpgimage_ypexzg.jpg

Next, I found a way to fasten the populated SM/DM assembly across the top of theflat connectors for the cells. Used more 14ga wire with the insulation on, just bent the ends into a loop and soldered them, then screwed it to a large paint stirring stick. Added a 10A fuse block for the input to the (8 cell) power rail...

I used the wireless telnet session from the my "service processor" Rpi andran calibration and set up the test parameters and loaded up a bank of 8 cells...
It all works !

I have one thermistor connection to fix, but otherwize I now have an 8 port FCDS tester up and running....


image_htwdnv.jpg

image_lszqhl.jpg

I will be trying out the Grafana script for entertainment and working on a script to store barcodes so I can accociate each cell slot to a device under test. That gives meaning to the endless stream of log data that the FCDS can collect.

Fun stuff to doas I take a break from soldering for a couple of weeks.
 
Joined
Dec 2, 2017
Messages
79
Thanks Brett.
Still need to add some barcode stickers, and the electronics for 3 more banks of 8right?

I had to take all the modules out to work on the one in the middle (hardest one to get to)sotook the opportunity to make another round of changes to the buss bar connections so I can remove a bank of 8 modules when needed.


image_htmsko.jpg


Also, here is a photo of aGrafana dashboard that is based on the work that @Fercsa posted on his build page . I am hoping this approach will help me keep the tester busy... Look, its time to putnew cells in slots 0, 2 and 7. The new cells will automatically start the charge/discharge cycle ! Just plug 'em and let 'em run, then write down the mAh.


image_azslur.jpgimage_wmyqpr.jpg


Table on the right shows the status after changing out cells in slots 0, 2 and 7. These are automatically drained then fully charged and then discharged to get the capacity.
 
Joined
Dec 2, 2017
Messages
79
Just a quick note since it has been a year since I joined the forum.

This update covers the "stopping point" where I got to before the weather changed and I got too busy with outdoor projects.

[image_czfbcx.jpg


My tester build is 3/4 complete... I have 3 banks of 8 cells connected and I have 20 of the 24 ports testing reliably. There were a couple with the TP4056 enable pin pulled off and a couple with wild thermistor readings, so I run the system with about 20 cells under test and am really happy with it so far.


image_hfziqz.jpg


I damaged the tiny OLED display when I accidentally plugged in a home made i2c connector with reverse polarity, so that made me to use the RPI as a system console. I have a new display and with colder weather on the way, plan to spend some time working on the final bank of 8 cells so I can try to test up to 32 cells at once.

image_ulkfgk.jpg


Having a 20-32 cell tester is a great starting point for the new year, so I am optimistic that I will be able to test a bunch of cells for a battery bank in 2019.
 

Cherry67

Member
Joined
May 13, 2018
Messages
493
For liion with Graphite Anode i have newer data saying There ist another aging decrease under 3.6 Volt And in lower Temperaturen Like below 20 degrees C.
 
Joined
Dec 2, 2017
Messages
79
Cherry67 said:
For liion with Graphite Anode i have newer data saying There ist another aging decrease under 3.6 Volt And in lower Temperaturen Like below 20 degrees C.

So is this to say that cells age faster (losing capacity) when stored with less than 3.6v or 20 deg c for extended time ?

This FCDS tester can make an automatic record for the starting voltage for untested cells. The trick is to get a cell serial number physically on the cell and into the database.

I have a barcode label printer and a scanner that I am planning to use for creating cell serial number and tester slot labels so I can associate cell test data to the tester slot that generated it.

How do you "check in/check out" a cell from your tester and data logging process ?
 
Joined
Dec 2, 2017
Messages
79
I have been exercisingthe tester over the past week, with some good fortune on my rpi support processor. I used @Daromer's rpi image as a starting point because I liked the way the Grafana charts looked... so I requested help from @Fercsa to get the FCDS tester to write to Influx and he provided a python script and directions to pipe the tester tic data over to Influx, and provided an example Grafana dashboard for one slave module.

Here is a screen shot of an updatedGrafana dashboard to keep an eye on the testing process with up to 4 slave modules:

image_sdxmff.jpg


The rpi is an amazing machine, and I am really amazed at how the rpi isrunning a python script to shovel data from the FCDS tester over wifi, storing the tic data intoa database, and also runningtheGUI charting application for presentation of real-time and historical cell test data. I think this is too much for such a small (but capable) device... look at the CPU on the bottom right of the task bar... over 90% utilizationtrying to update the charts on the screen.

To get some more scalability, I have been working on a Windows laptop to host my CHARGE database... (Cell Handling And Report Generation Engine).

So I am finding it is not so easy to use Windows, but I will continue to work on the laptop. Hopefully most of this development work can be used on the laptop as well if I can get the connector tools that run on Windows. For now, on the laptopI have a small test database in a spreadsheet for capacitytest results (no tic data) and building discipline to track cells by a serial number. Alsomaking progress on barcodes and how that will help tie a specific cell to the test data recorded by a particular tester slot.

I have many cells that were tested with anOpus and charged by TP modules, sometimes discharged with a HW-568 , and notice the FCDS results areabout 5-10% lower mAh than the Opus or HW discharger, so I am cutting over to use the FCDS,and will be runningmy "keeper cells"on it to be consistent... sad to loose 100 mAh on most of my cells, but hope to pick up some efficiency once I learn how to get things flowing smooth.
 
Joined
Dec 2, 2017
Messages
79
Update on the Grafana dashboard.


image_wrhdfj.jpg


On the left side there is a table that shows the status of all the tester slots. Rather than use a table for each group of 8 cells, this version has one table to show the status of 24 cells. There is a scroll bar to see the additional rows.

The columns can be sorted to group the data according to interest, for example: Temperature or Status, etc.

The temperature entries are color coded automatically (green/yellow/red) by the limits defined in the properties section for alerting .

On the right is a graph showing the results of the DEPLETE / CHARGE / DISCHARGE process for the cell of interest .

This panel refresh is quick for the table where Influxdb can easily locate the most recent entries for status, and it takes about 5-10 seconds during the refresh to pull the data and when re-draw the chart. During this time the Rpi CPU utilization increases to 30-35%.

I use this dashboard with automatic refresh turned off to conserve resources on the Rpi and just click the refresh button when I am around to check on the test progress.

This dashboard JSON fileis posted over on Fercsa's tester thread at:https://secondlifestorage.com/t-FCD...station-MM-giveaway-inside?pid=43560#pid43560
 

Bubba

Member
Joined
May 9, 2018
Messages
300
Hey Gary,

Could the FCDS modules be connected to the RPI i2c instead of going through the ESP32?
 
Top