ZKEtech EBC-X 0510 Eight-Channel Battery Tester

I think many of my readers remember that I often test batteries, and that I've had my eye on a dedicated battery tester for quite a while. Well, friends, today we're looking at the EBC-X electronic load and battery tester, which I finally bought :)


First, I want to thank the person who helped me place the order, since I have some trouble ordering through Mist Express. The tester came to $271 (the device itself plus the commission), plus $34 for shipping, for a total of $305. I ordered it here; it's also available on AliExpress, but at a higher price.

I placed the order in early May, and by mid-June I'd picked up the box from the collection point. The whole process took about 35 days, even though the parcel came by train, which takes longer but costs less. The parcel weighed almost 7 kg. I bought the tester from ZKEtech's official store.


The cables made up a pretty large part of the weight. I didn't weigh them, but they definitely weigh quite a bit.


Besides the tester, the box included:
1. A power cord. It's decent, but unfortunately has the wrong plug. Luckily, it can be replaced with a standard computer power cord, which most of us have lying around.
2. A cable with a converter for connecting to a computer; it has a four-pin connector.
3. A manual and a flyer from the shipping company.
4. They also included a warranty card. The manual explains everything in detail, but as you can imagine, it's all in Chinese.


There was also a large bundle of cables for connecting batteries—eight cables, to be exact. The standard version comes with clip leads, but you can request alligator clips instead. Neither option worked for me, since I was going to modify them for use with holders anyway, so I left them as they were.
The connection is four-wire. The cables are neatly made: 16 AWG (1.3 mm²) wire for the power leads and 20 AWG (0.52 mm²) for the sense leads. The cables are very flexible, as they always are.


A while ago, I ordered six holders specifically for this tester. I reviewed them before; they were ordered from the 100MHz store—link.
I ordered six because this particular model was out of stock when I placed the order, and I figured six would be enough anyway.


I ordered the holders unassembled and without wires, which was cheaper; the wires came with the tester.
They're not difficult to assemble, just tedious.


The result is a holder with a cable for connecting it to the tester. I chose this particular model because it's the most convenient to work with, and it's rated for currents up to 10 A. Long-term use of my previous pair has shown that their quality is excellent. By the way, I could resolder the connectors on my previous loads to make them compatible, giving me eight holders in total. Right now, the connectors on my old holders have the exact opposite polarity.


And now I've finally made it to the subject of this review. It definitely doesn't look like a household appliance: a large case, with handles on the front for carrying it (which turned out to be useful).
The case measures 350 × 285 × 135 mm, excluding protruding parts and feet, and 355 × 320 × 145 mm overall.



This series has three models, differing in charge/discharge current, power, and current-setting accuracy.
Mine is the 0510 model, with a charge current of up to 5 A and a discharge current of up to 10 A. I could have bought the 0510A, of course, but I don't think it's still being made. Besides, I don't see the point of such high charge currents for capacity tests, and it's also more expensive and heavier.


The front panel is very simple: eight connectors and the same number of LEDs.


There are ventilation slots on the left and a 120 mm fan on the right.
On the back are the power switch, the input connector, and the computer connection port. There's also a label with the model name and manufacture date: March 30, 2020. The top has a warning label about high temperatures.
It also has a warranty seal, but unfortunately, it doesn't look like this company plans to open any service centers here.


The bottom has a bunch of screws securing various modules, some of which are insulated from the case.


To get inside, you need to remove the top cover. There's nothing unusual inside: a power module, power supplies, and a control board.


1. The main power supply is the same as the one in the EBC-A40L: 5 V, 60 A.
2. The auxiliary 12 V power supply appears to be made by Sanmim.
3. There are also two heatsinks inside, like the ones in both the A40L and the A10H.
4. The fan is seriously powerful: 120 mm, 12 V, 0.45 A. It blows hard and is only slightly quieter than a vacuum cleaner. It runs whenever a program is running under load; there’s no speed control.

Overall, it’s put together reasonably well, but the fan’s power cable was dangerously close to the blades. It would probably have been sucked in and torn off.


Unexpectedly, the power board has... fuses. This is my fifth device from ZKEtech, but it’s the first one where I’ve seen fuses.


The transistors are arranged in pairs: 4 × 2 on one side and the same number on the other. You can also see the reverse-polarity protection diodes.


The manufacturer didn’t bother choosing anything new and went with its usual IRFP250M transistors. The thermal switch is visible on the right; there’s one on only one heatsink, although there’s room for a second one diagonally opposite. Installing it would mean drilling the heatsink. In previous models, it switched the fan to full speed. Here, it just prevents the fan from shutting off when the software turns it off.
The fan starts as soon as the program begins running, but the software turns it off after about 15–20 seconds.


Almost the entire front panel is taken up by a backplane board packed with modules. The modules aren’t just plugged into connectors; they’re also secured with solder at a special tab on the circuit board.


The design makes sense and is logical, but I was a little disappointed because I’d planned to add local control and had bought a tablet for that. Mounting it on the front panel will be tricky, though, since I’d have to take the whole thing apart, desolder the connectors, and move them elsewhere. So for now, that’s a problem...


The boards aren’t much different from those in the devices I’ve shown before: the same controller, the same supporting components, and very similar traces.


The difference is that this one has a ninth board, with its own microcontroller and opto-isolation. This board “talks” to the computer and controls the other eight controllers.


As far as I can tell, all the boards share a common bus, so their installation order matters. That’s why each board has a corresponding label.


Let’s put it back together and move on to the tests.
As for the indicator, there’s just one LED, and it’s single-color. It stays lit while the program for the corresponding channel is running, regardless of whether it’s charging, discharging, or paused.


Before the tests, I’ll tell you about the software, especially since it differs from the software for the other devices.
The manufacturer offers four versions of the software: two for each language (Chinese/English), and separate versions for the EBC-X and all other devices.

Once the software is launched and the device connected, the screen shows information not only for all eight channels, but also for channels that aren’t connected. The software supports at least six eight-channel devices—that’s 48 channels!


Right-clicking a channel brings up a control menu where you can open its settings or start that channel individually, start a group, or start all channels on the device at once.


Unlike the previous models, this load can’t be controlled directly in different operating modes; it only runs programs. That made the voltage/current measurements rather inconvenient: each time, I had to set the mode, start the program, take the measurement, stop it, and go back into the settings. There are also no preset charging modes or CP mode, but that’s all down to what this device is designed for.


Double-clicking a channel displays its operation, including a graph and status. Double-clicking any area again returns the display to its previous mode.


Click the List view button to see detailed information about the current channel status.


There is some kind of search by tested capacity, though I haven't quite figured out all the details of how the program works yet.
Also, while the channels are running, a colored indicator appears in their windows: red indicates charging, and green indicates discharging.


The fourth button at the bottom, Data report, displays the results in a condensed format.


There is a function for analyzing test data, but again, I haven't quite figured out how it works. Basically, the software automatically saves all the results, so you can open the data later. For some reason, though, one function uses the CSV format and the other uses DAT. I wouldn't say I need this—the results from 100 cycles are usually a massive wall of data—but I think it's worth looking into anyway.


First, I tested voltage measurement accuracy, which determines how accurately the cutoff thresholds work during charging and discharging.
In general, the load channels must not be connected in parallel, and batteries must not be tested this way either: all connected batteries must be completely independent. But that doesn't matter when testing voltage measurement accuracy, so I checked all four channels at once.


I checked the voltage at four key values: 2.5, 3.0, 3.65, and 4.2 V. The voltage was supplied by an adjustable power supply and monitored with a multimeter.


In the end, I got results for all eight channels, for a total of 32 measurements.
You can see that the voltage measured by the device differs from the voltage I applied. Specifically:
1. At 2.5 V, the difference is in the last digit.
2. At 3 V, the largest difference is on channel 1: 3 mV. The datasheet allows for 0.2% ± 0.003 V, or a 12 mV error.
3. At 3.65 V, the error is 3 mV.
4. At 4.20 V, the error is 3 mV.

So, with a stated maximum error of 12 mV, I actually got no more than 3 mV, which is excellent in my book.


The device can actually measure voltages above 4.5 V, but it won't operate and will report an error, even though its measurements are more accurate. I also tried applying higher voltages, around 6–7 V, and the result was the same.


The current measurements are a little worse. For this test, I supplied voltage from an RD6012 power supply, with the tester in discharge mode.
I tested at currents of 100, 200, and 500 mA, and 1, 2, 5, 7.5, and 10 A. The datasheet allows for an error of 0.2% ± 5 mA, or up to 25 mA overall.
Up to 5 A, things were generally fine, but above that the error exceeded the stated limit.


Worse still, at currents close to the maximum, the readings drift as the unit heats up—by about 15–20 mA from the initial current during the first minute.


Above are the results for channel 1. I also checked channel 4. I could have tested all of them, but it would have been very inconvenient.
The results here are a little better, though not by much; accuracy is higher only at low currents.


The results aren't great, of course, but ZKE loads have one big advantage: they can be calibrated, and the process is fairly simple. The only requirement is a sufficiently accurate, verified multimeter.
For example, just out of curiosity, I calibrated the upper half of the current measurement range (calibration is usually done at two points) and got some very good results.
Since the readings depend on temperature, let the load run in the relevant mode for at least a minute before each calibration step.


The same error is specified for charging mode, but the maximum current here is 5 A, so the allowable error is 15 mA. The device exceeded the tolerance only at maximum current. You can try calibrating it yourself here too, and temperature stability should be much better, since noticeable drift occurred only at currents of 8–10 A.


Channel 4 produced roughly comparable results.


The following auxiliary devices and components are already ready for the upgrade (in addition to the holders):
1. An uninterruptible power supply, which will probably replace both power supplies
2. LiFePO4 batteries
3. A tablet for local and remote control (Radmin, TeamViewer)
4. A converter with an output current of up to 20 A.

I have some doubts about the converter, since 20 is a bit on the low side—that's only 2.5 A per channel. So I ordered some others from Taobao. Their output current is a little lower, but they're branded, and I got more than one :)
I also need to buy some kind of USB hub, preferably one with a built-in network adapter, and maybe a fan controller, since the fan is pretty noisy and its speed doesn't depend on the temperature at all.



Based on my limited experience using it, I can say that I like the device. But I can also say that it's genuinely specialized and clearly designed for long-duration tests. Of course, you can use it as a regular battery tester, but its functionality is limited in that role, and it's not as convenient to use.
Until I modify it, I doubt I'll use it, since testers like this need at least a UPS for the tester itself and the control computer; otherwise, it's easy to lose all the test results.

It's a shame the manufacturer didn't implement channel aggregation—for example, 1×80 A, 2×40 A, 4×20 A, or any combination. That would have made it an absolutely fantastic device.

Anyway, if anyone else is planning to run similar tests, I recommend it. That's all from me for today; I hope this was useful.