A little about the new 120 A, 2.7 kW electronic load and Mibox enclosures

At the end of last year, I wrote about a 2400 W electronic load. At the time, it was essentially a pilot unit ordered by one of my customers.
After the New Year, I ordered several enclosures designed specifically for this version and decided to take it beyond the prototype stage—even if not into full production, at least to the next level.

A reminder: to keep up with new posts on the blog, I recommend following me on Instagram. I’ll post notifications about new articles there and might also write about something interesting - link to the account

You can also subscribe to updates and new articles on the Telegram channel - https://t.me/KirichBlog


I’ll say up front that I won’t go into much detail about the load, since almost everything is as described in this article.
But there are a few differences:
1. The power output is higher thanks to the two intake fans. The fans weren’t added to increase power; without them, I would simply have had to make an enclosure with more air intake holes.
2. The enclosure is now metal, because making it from two standard plastic enclosures is simply inconvenient.
3. I used higher-current terminals. With the previous ones, I could only specify a maximum of 100 A, even though the power stage can handle 120 A.
4. A few minor changes: for example, the front fans are controlled by a controller with automatic speed adjustment from 50% to 100%.

But I’ll start not with the load, but with the enclosure—or rather, with Mibox, the company that makes them. I’d like to thank them for being so accommodating: they’ve made custom-panel enclosures for me at least twice now, with the holes laser-cut. There’s nothing unusual about that in itself, but try finding someone who’ll do it for you without CAD drawings, working from a simple sketch—and at a reasonable price, too. I tried....

The enclosures are mostly pretty good, though not without their flaws. Today I’ll talk a little about both sides of the story.

As before, I agreed on the service, lead time, and price, sent over the drawings, paid, and about four weeks later received a box with my enclosures (they quote up to one month for custom orders).
I used this enclosure as the starting point.



Yes, there are three enclosures in this little box :)
Everything is packed neatly and carefully. I wouldn’t call it absolutely perfect, but it’s quite well done and secure.


There are lots of panels inside. At a quick glance, everything is there. They’re coated with powder paint, and look neat and nice.


The rear and front panels were the most important ones for me. Though, as I’ll mention later, I made a small mistake in the rear-panel drawing: I didn’t account for the fact that they’re narrower than the front panels. My mistake, but it didn’t affect the end result, even if some parts were a tight fit.


I was also worried about the controller’s dimensions, since there’s no way to shift anything around here.


But everything worked out perfectly; the controller fit like it belonged there. I only ordered the main openings to be cut; I didn’t order the mounting holes and drilled those myself. By the way, I’ve already run into cases where two fans of the same size but from different manufacturers have slightly different hole spacing.


Unfortunately, I walked right into a trap I hadn’t even thought about. Because of it, I first had to get out the angle grinder, spend half a day cutting, and then spend more time cleaning the balcony, since I work in an apartment :(


It all comes down to the panel layout. The enclosure is tall enough, but the top and bottom covers have flanges at the front, about 10 mm wide, and they get in the way of both the fans and the controller.
The drawing shows a side cross-section; I think it’s clear what I ended up having to cut off.

I’ll say right away that metal enclosures are noticeably harder to work with, because you have to drill lots of holes and tap them. Even a basic task you don’t think about when assembling a plastic enclosure—installing four rubber feet—takes more effort.

Unfortunately, this operation left a few small scratches on the covers. They’re not deep, but still, they were enough to upset me :(



Unfortunately, those weren’t the only problems I ran into. The previous one was mostly due to my own inattention, but these are entirely the manufacturer’s responsibility. I noticed them during final assembly.
First, the side panels. When I assembled the first enclosure, I found that there were five panels instead of four. Was one a spare?
Second, the rear panels had a problem with the threads. They were either clogged with paint or tapped only with the first tap in the set. In any case, if you try to screw in the supplied screw, its threads just get stripped.
Third, even during the initial assembly I noticed that the rear panels here don't have a hole on the side, even though it clearly belongs there. I thought about writing to the manufacturer, but it turned out to be more interesting. When I was installing the side panels, I noticed they had a hole—and suddenly... the rear panel had one too. I thought I must have imagined it, but no: it turned out that one of the three rear panels had holes on both sides, while the other two had just one, on the right. So no, I hadn't imagined it; the first time I'd simply picked up a panel with only one hole and was looking at the side without it.

Yeah, in my opinion they still need to work a bit more on quality control. The first enclosure I ordered didn't have any of these problems.
The problem is that the manufacturer makes the holes by partially punching the metal, which gives the threads more depth. For obvious reasons, all I can do is drill a hole and cut the threads myself. The metal is 1 mm thick, so it's best not to overtighten anything, as there's a risk of stripping the threads.



I ran into this when assembling the first enclosure, too; the situation here is exactly the same. Due to a not-quite-right calculation of the stiffness and configuration, there's a small design flaw.
The bottom cover is attached to the rear panel at four points: two on the sides and two underneath. But since it essentially "hangs" from the rear panel, two points clearly aren't enough. Even with my relatively light load, it sags when the enclosure is standing on its feet, so I had to add another mounting point in the middle.

A scale drawing to make it clearer.




I decided to build two loads at once, since I still hadn't settled on the specs for the third one. I assembled the front panels, then the power section and part of the enclosure, took a look, and realized something was off.


What was "off" was that the enclosure really needed labels and markings. A decorative panel wasn't an option here, and sticking on paper just didn't seem right, so I got out the laser engraver and used it to make the labels. The panels kept trying to slide around from the vibration while being engraved, since they were long for the engraver, so I had to use whatever I had on hand to hold them in place :).




Personally, I think this looks a bit better :) I engraved the rear panels for all the enclosures and the front panels for two of them.



Next came testing, calibration, and more testing. Since the enclosure is metal, I also had to use the ground contact on the mains connector and connect it to the enclosure.


Even after the initial, fairly rough calibration, the result is pretty good.


For further calibration I needed a higher current, so I used a 230 Ah battery, for which currents like this are well within its normal operating range.
I set it to 100 A and initially got 100.1, but as it warmed up, the reading dropped slightly to 99.7–99.8 A, which is perfectly normal.


Next I hit the maximum current my clamp meters are rated for. Although their stated limit is 100 A, they can measure up to 108–109 A, but I needed more.
I pulled an old clamp meter out of the drawer and was genuinely surprised: it showed a noticeably higher current.
Instead of 100 A, it showed 103.6; instead of 120 A, it showed 124. So it seemed like the readings were proportionally high. And yes, of course I know you're supposed to zero them and so on. In fact, I even checked another way: I clamped the meter on with no current, zeroed it, and then switched on the load. Same result.

I lowered the current to 30 A and compared it with my main clamp meter; the results were exactly the same.


I hooked my main clamp meter up to the power supply—nope, everything was fine, the measurement error was within the normal range. So the old clamp meter is just plain lying... What a shame.



In the end, this is what I came up with, with the following specifications:
Continuous power: 2400 W
Short-term (relatively speaking): 2700 W.
Current: up to 120 A
Voltage: up to 100 V
Connection: 4-wire; for a 2-wire connection, you don't need to install jumpers, since the controller has two measurement points.
PC connection: USB-C connector
Fan speed is temperature-controlled, but the main fans run continuously at low speed.
Thermal protection for all four modules.

I was a little disappointed by the fan grilles. Black ones would have looked better. Oddly enough, the grilles themselves aren't the problem—black ones are readily available (and I already have plenty of chrome ones). The problem is that I couldn't find black mounting screws.



I said the maximum power was relative. The thing is, the modules actually installed can handle more. For example, one user runs a load with a single module at 650–660 W. There are four modules here, so that works out to at least 650 × 4 = 2600 W—and this setup also has booster fans.




You can also fit rack ears, but I'm not including them in the kit yet. The problem is that I misjudged the position of the mains switch, and it ended up too close to the side panel. Short screws don't get in the way, and neither do the long screws used to attach the rack ears. But if you screw in a long screw without the ears, it will break the switch or at least push it out. I'll fix this in the next revision.



The rear is marked accordingly, and has vents for the hot air to escape.


So far, I've built two of these loads. I'm not sure what to do with the third: build another one like this, or make it a 160 A model, maybe even 200 A. The thing is, as current goes up, the voltage we can work at with that current goes down.
Roughly speaking, the current version can load a 20 V power supply at 120 A without any problems. If I make a 200 A load, then at 2400 W that would be just 12 V. In other words, it could test a 4S LiFePO4 battery at 200 A, but the current would still have to be reduced as the voltage increases.



As you can imagine, I didn't build these for myself. One load has already gone off to test drone batteries, since it can test 6S packs at currents up to 110 A. The only extra thing the owner asked for was an XT60 connector, to make testing smaller batteries more convenient. And this is where the problem with metal enclosures showed up: the XT60 male connector is installed from the inside, so the cutout has to be much neater than for the female connector, which covers the cutout. But I think it turned out pretty well.




That's all I wanted to say today. In a few days, I plan to show you another model, this time for the “consumer” segment, with a charge switching function. But I won't get ahead of myself.
As for the loads, get in touch and ask away. If you need something, though, it's best to describe the task right away, so I can find a suitable ready-made option or suggest the best custom solution.