Sousim Electronic Load Controller, or Another Device for Using Up “Excess” Electricity

Last year I reviewed two electronic loads made by Sousim. I successfully put one of them to use, and today it’s the turn of the second one, which originally could only be controlled using a computer.


I should probably start with some background.
I originally planned to use the electronic load with a computer, but that turned out to be very inconvenient because the bundled software was so buggy that I couldn’t get it to work properly. At the time, I also mentioned that a standalone controller was available for this electronic load, and even posted a photo of it.


In the end, I got excited about the idea of building a standalone electronic load with this controller. As it turned out, the process wasn’t as simple as it might seem at first.
For starters, they aren’t sold anywhere, but I found a seller on Taobao who sells products from www.xldn.net,, the company that made this controller, and had posted information about it on its website. It says there that the load can handle currents up to 50 A, but when I placed my order, I didn’t know that version existed, so I ordered the 20 A model.
Since I didn’t know how to buy it, I asked the manager at yoybuy.com for help, and I’d like to thank her separately for that.
In a nutshell, the process went like this:
I message the manager, she messages the seller, and the seller contacts the company that makes the controller.
The manufacturer makes the controller, the seller buys it from the manufacturer, receives and checks it, adds it to their store, and I buy it from them the usual way through the agent.

That’s the short version, though. There were complications: the manager wasn’t a technical person, we had to communicate in broken English, and technical terms had to be translated from English into Chinese. All in all, quite an adventure :) The whole thing took about two months, and in the end I ordered and received the long-awaited little board.
It was very expensive—about $28 with shipping within China, plus the agent’s fees. But by then there was no turning back, especially after I’d gone to all the trouble of setting the whole process in motion.

There were two items in the order. One was very small; I’ll probably review it at some point, especially since it looks like an interesting piece of gear. The package was quite large, with a total weight of 220 grams.
The controller was wrapped in so much film that it would have been very hard to damage.


The kit is as simple and straightforward as it gets:
1. Controller
2. Two sets of wires and connectors
3. Instructions.


Since the controller is fairly universal, the instructions aren’t specifically for this model. In any case, all I really needed was the wire color mapping, and I could have managed without the instructions altogether.
The two-wire cable is for connecting the RS485 interface.
The five wires are for the power supply and measurement circuits: red + black = power, yellow = voltage measurement, blue + white = current measurement. The last pair of wires was soldered together, so the hint was pretty clear.
At first, I couldn’t figure out why it needed to measure voltage and current when the same data could be read from the electronic load. As it turned out, that ended up saving me later, in a way.


The display, with a resolution of 128 × 64 pixels and a CoG (Chip on Glass) controller, takes up almost the entire front.


On the right are three control buttons and a two-color LED. Further to the right are contact pads where you can solder on a connector to connect the controls and indicators separately. I was hoping I could connect an encoder, but no—just hardcore, just three buttons.


Only later did I notice that a decorative bezel was also available. On the other hand, there’s no guarantee everything would have fit in my enclosure with it.


The display is mounted on a separate board, which in turn connects to the main board.


But when I turned the board over, I actually froze for a moment... About $30 for a handful of components?
Still, there were no alternatives. I wouldn’t have been able to reverse-engineer the control protocol, let alone write my own program for the microcontroller. In short, if you can’t use your head, use your hands money.


It’s only this empty and simple in my case. In the normal version, the board looks much more substantial. For example, here’s a photo of the same board, but configured to monitor a 32-cell battery. In this case, the manufacturer simply used this board as the basis for a simpler design.


And here’s one of the display layouts shown on the controller.


When I placed the order, the seller asked which input voltage range I needed. There were two options: 35–200 V and 6–35 V. Since my supply voltage is 12 V, I chose the second option.
The nice surprise was that the regulator is galvanically isolated, which immediately solves the problem of ground loops and generally makes things easier. Remember, the RS485 interface on the load board is also galvanically isolated from the power section.
I found another button on the side. Experiments showed that it resets the board to its factory settings.


There are very few components.
1. RS485 transceiver
2. Operational amplifier.
3. EEPROM

I also found pads for connecting a pair of temperature sensors, but from what I can tell, they won't work with this feature set.


The manufacturer's page claimed:
it uses all imported 18-bit AD sampling voltage and current
This means it uses an 18-bit ADC, so I removed the converter board to get a look at the microcontroller. I assume it's a basic STM, but I don't know whether they make any with an 18-bit ADC.


There's not much else to inspect, so I'll move on to how it works.
1, 2, 3, 4. The top button lets you select one of four operating modes: CC, CV, CW, and CR. It also switches to parameter adjustment mode, where the lower buttons increase or decrease the selected value.
5. The middle button turns the backlight on and off; holding it down resets the Ah counter.
6. The bottom button turns the load itself on and off.


Pressing the top button repeatedly takes the controller into the additional parameter and settings menus.
1. Display the current input voltage.
2. Adjust operation with an external current-measuring shunt.
3, 4. Set the minimum voltage and current thresholds at which the load shuts off.


1, 2. Set the time. I haven't tested this, but as far as I understand, it sets a limit on the operating time.
3. Compensate for the voltage drop in the negative lead to the terminal. The adjustment is very fine and lets you account for the voltage drop under load. Set it with a current of 1 A; after that, it adjusts the displayed voltage proportionally to the load current. In effect, this provides a “virtual” four-wire connection.
4. Device address in the system.

The language is a separate issue. I asked several times for English, but ended up with Chinese. Fortunately, there aren't many menus.


The screen looks roughly like this during operation:
The top line shows the time spent operating as a load, the voltage limit, and the current limit (basically just reminders).
Below that are the operating mode and the set parameter value.
At the very bottom are four values: measured voltage, measured current, calculated power, and capacity in Ah.
Voltage and current measurements are quite accurate; I'll come back to that later.


The manufacturer's page shows roughly the same thing.


After playing around with it, I moved on to assembly. I used the controller, electronic load, a 12 V 0.5 A power supply, an enclosure, and a few other bits and pieces.


Electronic load video review


I used an enclosure that had been sitting in a cupboard for a long time. I'd even forgotten its name; if I remember correctly, it was something from the KM series. The enclosure measures 178 × 140 × 69 mm.
The main advantage of this enclosure is that it has no posts joining the top and bottom halves, so you can make better use of the front panel.


The load module needed a few modifications: I moved the buzzer and RS485 connector to the opposite sides of their boards and shortened the standoffs. This was necessary to reduce the overall width of the assembly.


While we were talking, the seller said that proper operation required a separate current-measuring shunt with a standard 75 mV drop. I tried to explain that the load board already has a shunt and doesn't need another one. In the end, though, it turned out to be easier to agree. I did refuse to buy a shunt from him, because:
1. I wasn't sure it was necessary
2. It costs less here than on Taobao.

But I had a faint hope of extending the load range to 50 A, so I wrote to them asking whether there would be a 50 A shunt :)


The test proved me right about there being no need for an additional shunt, and also showed that the controller can be reconfigured to work with the load's original shunt. I had to select a 75 mV, 14.3 A shunt in the settings.
I was worried that with this setup I wouldn't be able to set the current above that value, but it worked out: I set it to 20.1 A without any problems. I was a little disappointed, though, since the load is still limited to a maximum current of 20.1 A, unfortunately :(
Just a photo taken while checking different modes.


After that, assembly followed a fairly standard process. First, I cut a piece of plastic to cover the heatsink.


This was necessary to make sure the air was blown out of the enclosure.


I used three motherboard standoffs to mount the board. I screwed two of them into existing holes and drilled a separate hole for the middle one. The standoffs have an outer thread of about 4 mm and an inner M3 thread.


I secured the board inside the enclosure.


I did find a black 50 mm fan guard at home, but I wasn't happy with the result, so I simply modified the 60 mm version a little. I also installed the power connector. Later, it occurred to me that I could put a connector for connecting to a computer on the back, too.


The front panel caused no problems at all, except that I had to cut a fairly large window for the display. The controller board was almost exactly the right height for the enclosure.
The controller board is held in place by four plastic standoffs, which I glued to the front panel.


First test fit. As last time, I soldered the power supply to the power connector, though in this case it was mostly for convenience, since there was actually enough room.


When planning the front panel, I was worried I might accidentally hit the power switch, so I used a toggle switch instead of a push button and moved it well over to the right. In practice, though, I found I'd overdone it a bit.
I installed the terminals on the left, since there wasn't room for them at the bottom. The extra terminal holes face left and slightly upward to make connections easier.


There are also considerably fewer wires this time, mainly because everything is controlled through just a couple of wires on the RS485 interface.


After the main assembly was done, I calibrated it at two points: 1 and 11 A. I chose the second value for two reasons:
My multimeter can still handle 11 A without any trouble.
Because the software is awkward to use, it was easier for me to switch between 11 and 1 A than between 10 and 1 A. I wrote about this in the corresponding review.

After that, I connected the wires from the controller board and checked the result. Everything seemed great. Almost great, anyway: it turned out the display has virtually no viewing angle, and when viewed from above (or below, for that matter), the image becomes inverted.


That's pretty much it for the internals. There are so few wires that the whole setup looks very simple.


As last time, the switch switches the 12 V supply, while the power supply runs continuously. I also had to trim the connector that links the power section to the control board a little; otherwise, the enclosure wouldn't close. I expected that, though.


And that's it. To my undemanding eye, it turned out pretty well, although it's still a long way from perfect, and there are a few things I'd do differently now.


For example, I should have painted the heatsink black and installed a connector for connecting to a computer.
I might also add more air intake holes.


The enclosures of the new and previous loads are roughly the same width and depth, but differ in height. I actually thought about using this enclosure for the 300 W version last time, but I was a little short on space and had to go with the Z2A.


I could have ended my story on this rather pleasant note, if it weren't for one significant flaw that I just couldn't overcome.
The thing is, no matter how precisely I calibrated the load, after a while the calibration had an annoying habit of drifting. And it could drift either up or down. For example, above I showed the calibration results. A day or two later, I turned on the load and saw 0.87 and 9.43 A instead of 1 and 10 A. I spent the whole evening calibrating it again, taking extra care and using slightly different reference points (0.5 and 5.5 A at 24 V). The next day, I got 970 mA instead of 1 A and 10.02 A instead of 10 A. But the next time I started it, the readings were completely different: 1.02 and 10.5 A, respectively.

It was also noticed that in some situations the fault didn't just shift the reading; it also made the current setting respond instantly. Normally, setting the current to 10 A looks roughly like this: it jumps to 9.5 A right away, and the rest "catches up" over about half a minute.
I even considered leaving it like that and just changing the shunt's nominal value a little, but the change wasn't linear: at 1 A, for example, the error was 3%, while at 10 A it was already 5%.

But eventually I figured out the problem. I started thinking along these lines: the voltage calibration had gone smoothly, and that reading hadn't "drifted" at all over several months, while the problem was only with current calibration. It turned out that one of the wires in the ribbon cable connecting the power board to the "brains" had a poor connection in the connector. More precisely, the control line for one of the power transistors was making intermittent contact. It wasn't disconnecting completely, though; instead, the load current was changing.
So I recrimped the cable, shortening it to about a third of its original length in the process. That fixed everything: nothing "drifts" anymore, and the current itself is more stable too. Before that, the last digit would fluctuate from time to time, apparently due to vibration from the fan. Later, as an extra measure, I added capacitors to the power input and the 5 V line that powers the interface converter.
By the way, it turned out there are two 5 V lines: one powers the microcontroller, and the other powers the analog section, which also has a -5 V converter.


Video review


As for the controller, I have only two complaints: the screen has a narrow viewing angle, and it doesn't display the heatsink temperatures, although the latter is much less important.
As for the device overall, I'm happy with it. Sure, there are ready-made alternatives, but they don't have what I need, namely:
Constant-power mode, which is very rare.
Voltage-limiting mode, which is even rarer and is essential for testing chargers. Without it, it's very difficult to test a charger properly.
Constant-resistance emulation mode, which is usually found only in branded equipment.

That's all. I hope the review was useful. As always, I'll be glad to answer any questions or read your comments.