Zketech EBC-A20 software frozen

[filippomasoni]

Hello,
I recently purchased the EBC-A20 to charge/discharge and capacity test some Lifepo4 batteries. As I've seen some video of it recommend by the Off-grid garage.
I installed the software provided on the website and run the first cycle correctly: charge, discharge, and charge back up to 50% of a 280Ah cell. The PC software worked for the entire cycle of about 60h and I saved data and curve.
On the second test, the software froze while still charging, I couldn't control the unit with software and the buttons on the unit didn't work. It still displayed charged Ah on the display and it was still going, but the PC software and curve stopped after 2h30m. I had to shut off the unit with the button on the back. I tried again but couldn't connect, I had to reinstall the TTL drive to make it work again. I started charging again and managed to finish this time since I only had a few Ah left. I then started a new cycle of discharge and after only 2h50m the software stopped responding again. The timer was still going and the unit still discharging according to the display, but the graph and data on PC software was stuck. I then turned it off again and stopped everything.

I am now concerned the unit could over-discharge or overcharge my battery as the software can't be trusted like this and I can't leave it do it's thing as it was supposed to.

Do you have any recommendations?
I tried contacting the manufacturer but it's not responding, I bought it through AliExpress, but I really don't want to go though sending it back and get another one. I would just like to easily test and cycle all of my 8 cells.

 

[Gizmo740]

What version of software are you running? Current version is V1.8.5 and can be downloaded from forum...search file section beginner corner.

Also check settings in control panel for power saving options (Microjunk software). The only thing allowed to sleep should be your display. Depending on version of windoze they added usb and other power saving snoozer settings.

The fact it ran ~60hrs once and you had no magic smoke from box leads to some other change/problem.

The A20 is designed to be a stand alone tester. You can program settings and step processes using built in display and buttons. Software is adapted to work with and provides data logging/plotting but not absolutely necessary for device to function properly. Not a bad idea to verify calibration of any new piece of equipment with a DVM and Amp meter.

 

[filippomasoni]

What version of software are you running? Current version is V1.8.5 and can be downloaded from forum...search file section beginner corner.

Also check settings in control panel for power saving options (Microjunk software). The only thing allowed to sleep should be your display. Depending on version of windoze they added usb and other power saving snoozer settings.

The fact it ran ~60hrs once and you had no magic smoke from box leads to some other change/problem.

The A20 is designed to be a stand alone tester. You can program settings and step processes using built in display and buttons. Software is adapted to work with and provides data logging/plotting but not absolutely necessary for device to function properly. Not a bad idea to verify calibration of any new piece of equipment with a DVM and Amp meter.

I was running the latest 1.8.5 version downloaded from the Zketech website installed on Windows 10 updated to the latest version. Of course, I had performance mode on and set everything to not turn off apart from the screen.
I checked voltage and amp draw with a meter before starting the long test and they were spot on, also checked various times during the first 60h cycle and they were always correct.

I also tried updating the firmware of the unit following the instructions of the manual, but got an error and couldn't do that, not sure if that's supposed to work.

After my post, I uninstalled everything from the PC, cleared temp files, restarted, and installed everything back again following the order of the manual (EB software first, then TTL driver). This time I installed the software files from the forum (version numbers are the same as of the official website, but size is slightly different). I first had some trouble connecting it, it would say connected and - PC - on the unit but showed No device on the parameters mode. Turning on and off a couple of times I managed to connect. I started the discharge, and it actually finished in about 9h. I planned it so I could be there when it was close to the low voltage cutoff I selected of 2.9V and it properly worked.
After saving data and curve I then started charging back up to 40% to store the batteries for a few months, it's charging at 5A (maximum for the A20) 3.32V with 2.5A cutoff. I found with the first cycle this to push about 100Ah into the battery which is 35% and about what it shipped with.
It's still going now after 12h, fingers crossed it was the different installation file.

Btw I now capacity-tested 2 LF280K EVE cells from 3.5V to 2.9V at 274Ah and 275Ah and I'm pretty pleased with that since I'm not using the full range, as I want to test the actual range I'll be using in my setup.

Side note, I'm not using the provided power supply as that was faulty for the off-grid garage guy, it shoot 26V and damaged the unit, also some other people on the forum say to stay clear of it. So I'm using a laptop charger that I've been using for many years without issues, it's a 20V 4.5A and tested it delivers 20.5V with no load and 19.5V with 4A load, while the original it's actually 19.5V without load and 18.2V with 4A load.
I don't think this means anything for the software but at least I can trust this power supply.

 

[Gizmo740]

Great you are up and running again! Firmware fail may have been issue and fixed when reloaded software. I've been leary of fw upgrades on most non-mainstream purchases from China. If it works ok when I get it...don't upgrade/fix it. Seen lots of magic smoke over the years. Sometimes "good" ideas get implemented into packet and use feed back from end user to test. Own risk type deal.

You may consider running full factory spec check on at least one cell so you have an actual full to empty plot. 2.9-3.5 are fine but you may see both can be adjusted up a few mV to optimize your setup. I have EvE 304's and tested my sweet spot to 3.58-3.15 which yields over rated cap usable power. Look for beginning of knees.

 

[filippomasoni]

I've only considered updating the firmware after having some issues, hoping they would fix with that. Otherwise I wouldn't have bothered. But the update didn't work anyway...
It's working now, I'm currently charging the third cell. But every time I start it, I have issue connecting it. When I press connect it shows PC on the unit but nothing in the software. So I have do try at least 4 or 5 times. Closing and reopening the software, turning off and on again and replugging the USB. Once connected though it seems to work ?

You're charging at 3.85?! Isn't that too high? I always thought 3.65 was the upper limit.
Looking at the curves they are already pretty steep at both limits, so I know there's not much more than that. Specs sheet from the supplies stated about 288-290ah for the cells, so I'm not loosing more than 5% of capacity and I'm staying inside safer limit for a long life.

 

[Gizmo740]

? dyslexia kicked...been few late nights! 3.58 to 3.15. Corrected post. I may adjust limits with more testing but finding mV matter with this chemistry. 3.65 to 2.5 being accepted hard limits to prevent damage. Seems to be more than expected (based on published curve data) no load V drop after charge. It is consistent across the few cells I have fully tested but warrant further testing. Part of my issue is re-training my brain from LA. Old dog...new trick!

 

[filippomasoni]

Got it, I might run one cell 3.65 to 2.5V to actually check the full capacity and look at the curve.
The unit is now working, but the software is still buggy, sometimes the graph and data output freeze and stop recording. I assume the reading on the display to be accurate since that is still going.

 

[Gizmo740]

Probably already tried this but throw it out...clean usb contacts on data cable from tester and check continuity from end to end. Also may want to try another input port to computer. Memory serves mine was flakey on one of my laptop ports. Switched ports and been rock solid for weeks now.

 

[Herb]

Glad I found this thread, well today I fired mine up and the fan started making what seems to be a worn bearing noise, the unit is pretty new, also the maximum discharge amperage won't go above 6.48A out of 20A I thought it was supposed to start the test at 20A, not happy with this unit, I contacted the seller and hopefully they can explain if this is by design or if is defective. I'm using the power supply that came with it by the way. Comments please? I still got 7 more batteries to test!

 

[m0onspell]

Same. Discharge seems to work fine, but in a charge mode the software is frozen two times now.

The unit is now working, but the software is still buggy, sometimes the graph and data output freeze and stop recording. I assume the reading on the display to be accurate since that is still going.

So were you actually able to solve the problem?

 

[filippomasoni]

Same. Discharge seems to work fine, but in a charge mode the software is frozen two times now.

So were you actually able to solve the problem?

No, I got a partial refund from the seller and stopped using it since it was unreliable. I guess with this kind of more complex pieces of Chinese hardware quality control is an issues and it's all about luck.

 

[m0onspell]

Upd: I disabled windows power setting for turning off USB outputs and at first it seemed that it helped as it was running for a whole day without any issues. But then it froze again.
It seems to me that the reason might be in the supplied USB cable. It is very flimsy and feels loose in the socket. I also noticed that when I touched it very slighlty it immediately disconnected. I don't know, apparently it is just a regular USB cable, will try using some other one.
But still, their software is so raw, simply unfinished. Like "Continue" button which just clears out all previous measurements instead of continuing the test. It messed up a whole set of tests for me which I was running for days, cause a part of the measurements was lost.

 

[filippomasoni]

Upd: I disabled windows power setting for turning off USB outputs and at first it seemed that it helped as it was running for a whole day without any issues. But then it froze again.
It seems to me that the reason might be in the supplied USB cable. It is very flimsy and feels loose in the socket. I also noticed that when I touched it very slighlty it immediately disconnected. I don't know, apparently it is just a regular USB cable, will try using some other one.
But still, their software is so raw, simply unfinished. Like "Continue" button which just clears out all previous measurements instead of continuing the test. It messed up a whole set of tests for me which I was running for days, cause a part of the measurements was lost.

I've always had the PC with all of the power settings turned off and checked them as well, but still bad problems.
Now that you mention the cable that could be the issue. I haven't though of that. Let me know how it goes if you try a different one.

I agree, the software is rubbish, but I would have accepted it if it worked as supposed to, considering is a cheap Chinese product.

 

[jv16bar]

I had a similar problem with my EBC-A20. After connecting to my laptop and running the software i randomly lost connection to the tester. Sometimes it worked hours and then disconnected, but there were cases when i lost connection after a couple of minutes. I suspected that the serial to usb converter cable was the reason and i ordered an isolated TTL to USB converter from duppa.net (i found this site on Offgridgarage yt channel).
This is the product:

Isolated USB to TTL UART RS232 – DuPPa

www.duppa.net

I soldered the 4 pins to the tester and since than i have no issues, the software is running days without connection being lost.

 

[filippomasoni]

I had a similar problem with my EBC-A20. After connecting to my laptop and running the software i randomly lost connection to the tester. Sometimes it worked hours and then disconnected, but there were cases when i lost connection after a couple of minutes. I suspected that the serial to usb converter cable was the reason and i ordered an isolated TTL to USB converter from duppa.net (i found this site on Offgridgarage yt channel).
This is the product:

Isolated USB to TTL UART RS232 – DuPPa

www.duppa.net

I soldered the 4 pins to the tester and since than i have no issues, the software is running days without connection being lost.

I think I have one of those I bought for future use on the JK BMS, I'll give it a try. Can you tell us the wiring order you used? Still using the mini-USB provided from the tester side?

 

[jv16bar]

I soldered directly to the PCB, i don't use the mini USB port.
RED = +5V
White = GND
Yellow = TX on tester and RX on converter
Green = RX on tester and TX on converter

 

[obd.tech]

Hi,
i've just got an EBC-A20 without the Blue USB Cable included.
Will a PL2303TA cable be compatible with this tester? (I have 3)
I've tried linking to your pcb connections,as shown above, but no comms or PC showing in the tester window for me?
Running the software on a Win7 Laptop, using comport 9, recognised in device manager & driver working.
Anyone here that may be able to help please?

Update .....
I took the PL2303TA USB interface cover off and inspected with a microscope.
I found that the 2 sided pcb had insufficient soldering (chinese el cheapo $2) and the GND link 'through hole' not connected to 1 side. A nice Pb resolder/rework required.
Back together & connected as above & ..... 'It WORKS!!'

Red = +5v
Black = GND
White = RX on PL2303TA
Green = TX on PL2303TA
Pin header photo attached