A little about the history of electronic loads and what’s going on in general
Lately, I haven’t been posting much, and someone even asked me if everything was okay. Unfortunately, I’ve had a lot on my plate, and I’d like to tell you about some of it today, since it turned out to be a timely topic for buyers and an interesting one for me.
I’ve actually prepared a lot of material recently. I’ve also nearly finished testing the new version of the CL24 load. There are some positives, but also some negatives. Either way, the reviews will be coming soon.
First, a quick announcement.
I get a lot of questions on all sorts of topics. Besides calling me, people write to me on the blog, YouTube, email, Instagram, Muska, Skype, Viber, Telegram, and even WhatsApp. Most often, they ask for advice or have various questions, but sometimes I’m either very busy and can’t reply right away, then simply forget, or I might miss your question, or something else might happen.
If your question is important to you and I haven’t replied within a few days, just send it again. That’s perfectly fine and even appreciated.
If you need a really quick reply, messaging apps are my priority—for example, Skype—because I might not check my email, or your question could get lost among a pile of mailing lists, spam, and so on.
All right, that’s enough introduction. Let’s get to the topic at hand.
A reminder: to keep up with new topics on the blog, I recommend following me on Instagram, where I’ll post notifications about all new topics and might also share something interesting - link to account
You can also subscribe to updates and new articles on the Telegram channel - https://t.me/KirichBlog
So, I’ll start with the first load model. These were based on the well-known DL24 board, which I’ve already covered before.

Then there were a couple of models based on the DL24M board. Many thanks to the person who gave me one of the boards.
These boards are discontinued, so for obvious reasons development of this product line has come to an end, although the boards were interesting in their own way.

Around the same time, there were two custom orders for loads built without enclosures. One was a 1000 W model based on the DL24 board; the other was more interesting: a 1500–1800 W model based on the DL24EW board, with battery charging support, fan control depending on the operating mode, and various other little features. As you can see in the photo, a submodule was made for the additional options and mounted on top of the original board. One feature is that the board can control a charger with a current of up to 40 A. I haven’t decided whether I’ll write about it, since this board was made as a one-off, and a more interesting version was developed afterward.

By the way, the customer later sent a video showing how they installed this load in their enclosure and even laser-engraved the front panel. I think it turned out really cool. This is also, in a way, my answer to the question of why I don’t include any cables with the loads. The power cable is a standard “computer” cable, and most people have one lying around at home. As for the power cables, everyone makes those to suit their needs, since the wire gauge, length, connectors, alligator clips, and so on can vary.

After that, as an experiment, I made two more versions based on the DL24 board. Here I’d like to give special thanks to my reader Dmitry for his invaluable help making the display bezels, and especially for managing to make them look neat even with poor-quality material!
The 500 W version just got a design change, while with the 1000 W version I tested installing this kit in an enclosure. My loads based on this board are a little cheaper, but I’ll probably make 2–3 more, since I still have a few boards, and then I won’t be making any more. The thing is, they’re functionally inferior and also more difficult to assemble. Installing the buttons alone is a pain... And the current only goes up to 30 A, with power up to 1000 W. The only advantage besides the price is the USB port.
The 500 W model.

And the 1000 W model.

And this is what I’m now making on an ongoing basis. In my opinion, it’s the best option: loads based on DL24MP boards.
Key differences:
1. Encoder control
2. Current up to 40 A.
3. Hardware CV mode. By the way, it’s also available in the 24EW models, with a separate op-amp and switch.
4. No USB; all connections are via Bluetooth, although I think USB could be added if needed, at least in theory.
5. External temperature sensor connector on the smaller model; available on request for the larger one.
Otherwise, it’s the same as the 24M version: 500–600 W for the smaller model and 1000–1200 W for the larger one. They’re essentially identical, except that the larger one has two power modules instead of one and a more powerful power supply. I actually like the look of these loads, too.

I’ve already written about the power ratings: 600 W for the smaller model and 1200 W for the larger one are specified as short-term ratings. In fact, the cooling system has plenty of headroom. For example, I recently tested a power supply (review coming soon) and ran a heavy load at 1200 W for an extended period.
According to the sensor, the temperature was 80 degrees; I think it was actually closer to 90–95. The transistors are rated for 125 degrees in this operating mode.
I've seen people write several times that the transistor temperature in their loads is 50–60 degrees. Unfortunately, that's usually a measurement error.

Even in my case, where I press the temperature sensor directly against the transistor case and use sensors with a different resistance-versus-temperature curve, the readings are about 15 degrees too low. With the original sensor, the readings are way off.
For example, in the 24MP and 24EW series loads, the sensor is simply cooled by air from the heatsink, so it's hard to say what it's actually measuring.
By the way, you can see the shunts on the left. The originals are rated for 1 W, which is completely unacceptable to me, so I have to replace them everywhere with thicker ones rated for 3 W. I showed the difference once.

And now I've gradually come to what's happening now.
Along with the standard versions, I made a custom one. It differs in the type of connectors used and has a 24EW board, so it supports connecting a charger. The additional board was also modified slightly: the transistor was replaced, and the current can easily be increased to 20–30 A.
The load is in a large enclosure, but rated for 500–600 W. It's assembled so that you can install a second module later if you want, bringing the total to 1000–1200 W. In other words, everything is in the enclosure except the second module.
This was actually the first load where a problem came up. For some reason, its WiFi doesn't work. This was the first time I'd built a load with a customer's board, and the second time I'd used a 24EW board (the first was my own, tested one), so this was an unexpected snag. I didn't check it during assembly, partly because I didn't want to bind the board to my software. When the customer received it, they were disappointed. Resoldering didn't help, so we're figuring out what to do.
Update: it turned out to be an issue with the control module (not the only one). In the end, I had to add an operating mode switch and a USB WiFi adapter, and that fixed the problem.

Besides the upgrade option, the customer wanted to be able to apply a high load for short periods. I can say that the load can do that: up to 900 W for a few minutes without any issues. Beyond that, it will naturally overheat and the protection will trip. I didn't push it that far, since tests like these need to be supervised. I think it could handle 1000 W briefly, perhaps even more, but I haven't tested that.
Photos from testing above the rated power.

I got so interested in battery testers—that is, loads that can not only discharge a battery but also control a charger—that I decided to design my own charge switching board.
As a reminder, the original board gets very hot during operation, has a large voltage drop, and uses a flawed reverse-polarity protection circuit (which may not work correctly with CC/CV chargers). Here's what I ended up with:
1. The circuit is fully disconnected. Previously, current was interrupted only in the charger-to-battery direction, but could still flow in the opposite direction. This could short the charging terminals and cause a short circuit. It could also damage the charger's output if power was disconnected and then restored (with chargers such as the well-known DPS, DPH, RD, and similar models).
2. The voltage drop is about 50 mV at 10 A, and therefore 100 mV at 20 A. That means just 2 W of heat, so without additional cooling it can handle charging currents up to 30 A. With two modules, the current can easily be increased to 40 A.
3. Reverse-polarity protection. The original protection was completely unacceptable to me. I provide basic protection with diodes, so the load tolerates reverse polarity at its input, but things are more complicated with a charger. In the end, I added a protection circuit to the board that won't connect the battery to the charger unless they have the same, correct polarity.
Unfortunately, the protection works by inhibiting operation: by default, it allows operation and blocks it when the polarity is wrong. It would be better to do the opposite and allow operation only when the polarity is correct, but given how the board works, I considered this completely safe.

In the end, this effectively became a new model. The second custom order had almost the same configuration as the previous one: a large enclosure, 500–600 W, upgrade capability, a 24EW board, and charge control support. But this one already has the new version of the control board.
Yes, on loads based on the DL24EW, I don't use the original UART-to-USB adapter. Instead, I use a version with galvanic isolation and a USB-C connector, which I think is better.
The model now has the following specifications:
Load-mode voltage up to 100 V.
Charging voltage up to 55 V
Power: 500 (600) or 1000 (1200) W
Charge current switching up to 30 A (500 W version) and up to 40 A (1000 W version)
USB Type-C with galvanic isolation and WiFi
Adjustable charge cutoff current, enabling proper CC/CV charging. After my recommendations, the Atorch programmers updated the board firmware.
Hardware CV mode, i.e. You can test not only power supplies, but chargers too.
And thanks again to Dmitry for the neat display panels! Why doesn't Atorch make external displays for all their models...

But there have also been some changes for the worse. It looks like they removed the four-wire connection and replaced it with a three-wire one; I'm still figuring that out. This is actually the third version of the DL24EW board. The first was covered in the review; the second got a firmware update adding a cutoff current for charging; and the third has a redesigned circuit board, with the connection changed and a small DC-DC converter replacing the linear regulator.
The first and second versions are on the left, the third on the right. On the third, they added a 2 mΩ resistor between the A- and V- lines... and changed the regulator circuit.

While I was at it, I built an external battery for a UPS. It even ended up with a power bank function, and I also tested eight more EVE 33140 15 Ah cells.
The idea was to be able to use it instead of the UPS's original battery, since some models don't like the battery voltage going above 14 V and go into protection. It also needed to charge from an external charger via Type-C. I had to reconfigure the power bank board so it wouldn't output more than 14 V, and while I was at it, add a high-current, low-dropout voltage regulator. The result is 13.80 V. The regulator also conducts in the reverse direction, which lets me draw up to 100 W from the Type-C output.

Naturally, I measured and balanced the cells, then assembled them into a pack matched by capacity. The connections are soldered with tinned wire. I wanted to use nickel strip, but the plan was to draw up to 40 A continuously and 60 A briefly, so I decided soldering would be more reliable.

The result is a very compact device with a capacity of 30 Ah, or about 385 Wh, charging via Type-C (up to 5 A), an output of up to 20 V at 5 A, 60 A protection, an active balancer (because it operates at a reduced voltage), a longer service life (also thanks to the reduced voltage), and a build made with care.

And, tangentially related to electronic loads, I finally got my little drill press. I was getting fed up with drilling heatsinks using a cordless drill: it's difficult, and the quality isn't what I'd like. It was stuck in customs for three weeks... But it runs so quietly! At 500–700 rpm, it's about three times quieter than a cordless drill.

That's about all my news for now. I've shuffled my workload around a bit, and I hope to write a few reviews based on material I've been preparing along the way.
What's next? Next on the list is building a battery tester with a built-in charger. I've already bought a couple of DC-DC converters, and I'm also planning a 1200 W load with a current of up to 100 A.
People often ask me about the warranty. Friends, I'm not a stickler for formalities, and I also believe my word is worth more than a line in the manual. If needed, I'll help you after a year, two years, or even five. And as you've seen, the actual specs are clearly better than the stated ones—not like with Chinese products, where everything is right on the limit...
A special thanks to the patient people, because recently I ran into... damn, I don't even know what word to use here—suggest one yourselves. Anyway, one Friday evening, an “anonymous” person contacted me with a question about a UPS and external battery setup. I replied, they asked another question, I replied, and then they sent another email with a question, which I didn't see until quite a bit later. I didn't reply right away because things got hectic at my day job, and it took up both days of the weekend. Then on Monday morning I received this email, hidden under a spoiler because it contains profanity.
The question is, how are you supposed to feel about people like this—and whether they're even people... Just a reminder: I have my own work and my own life (at least for now), and if I haven't replied, there's a reason or other priorities. Besides, our military personnel periodically ask me for help, and they are my top priority.
I’ve actually prepared a lot of material recently. I’ve also nearly finished testing the new version of the CL24 load. There are some positives, but also some negatives. Either way, the reviews will be coming soon.
First, a quick announcement.
I get a lot of questions on all sorts of topics. Besides calling me, people write to me on the blog, YouTube, email, Instagram, Muska, Skype, Viber, Telegram, and even WhatsApp. Most often, they ask for advice or have various questions, but sometimes I’m either very busy and can’t reply right away, then simply forget, or I might miss your question, or something else might happen.
If your question is important to you and I haven’t replied within a few days, just send it again. That’s perfectly fine and even appreciated.
If you need a really quick reply, messaging apps are my priority—for example, Skype—because I might not check my email, or your question could get lost among a pile of mailing lists, spam, and so on.
All right, that’s enough introduction. Let’s get to the topic at hand.
A reminder: to keep up with new topics on the blog, I recommend following me on Instagram, where I’ll post notifications about all new topics and might also share something interesting - link to account
You can also subscribe to updates and new articles on the Telegram channel - https://t.me/KirichBlog
So, I’ll start with the first load model. These were based on the well-known DL24 board, which I’ve already covered before.

Then there were a couple of models based on the DL24M board. Many thanks to the person who gave me one of the boards.
These boards are discontinued, so for obvious reasons development of this product line has come to an end, although the boards were interesting in their own way.

Around the same time, there were two custom orders for loads built without enclosures. One was a 1000 W model based on the DL24 board; the other was more interesting: a 1500–1800 W model based on the DL24EW board, with battery charging support, fan control depending on the operating mode, and various other little features. As you can see in the photo, a submodule was made for the additional options and mounted on top of the original board. One feature is that the board can control a charger with a current of up to 40 A. I haven’t decided whether I’ll write about it, since this board was made as a one-off, and a more interesting version was developed afterward.

By the way, the customer later sent a video showing how they installed this load in their enclosure and even laser-engraved the front panel. I think it turned out really cool. This is also, in a way, my answer to the question of why I don’t include any cables with the loads. The power cable is a standard “computer” cable, and most people have one lying around at home. As for the power cables, everyone makes those to suit their needs, since the wire gauge, length, connectors, alligator clips, and so on can vary.

After that, as an experiment, I made two more versions based on the DL24 board. Here I’d like to give special thanks to my reader Dmitry for his invaluable help making the display bezels, and especially for managing to make them look neat even with poor-quality material!
The 500 W version just got a design change, while with the 1000 W version I tested installing this kit in an enclosure. My loads based on this board are a little cheaper, but I’ll probably make 2–3 more, since I still have a few boards, and then I won’t be making any more. The thing is, they’re functionally inferior and also more difficult to assemble. Installing the buttons alone is a pain... And the current only goes up to 30 A, with power up to 1000 W. The only advantage besides the price is the USB port.
The 500 W model.

And the 1000 W model.

And this is what I’m now making on an ongoing basis. In my opinion, it’s the best option: loads based on DL24MP boards.
Key differences:
1. Encoder control
2. Current up to 40 A.
3. Hardware CV mode. By the way, it’s also available in the 24EW models, with a separate op-amp and switch.
4. No USB; all connections are via Bluetooth, although I think USB could be added if needed, at least in theory.
5. External temperature sensor connector on the smaller model; available on request for the larger one.
Otherwise, it’s the same as the 24M version: 500–600 W for the smaller model and 1000–1200 W for the larger one. They’re essentially identical, except that the larger one has two power modules instead of one and a more powerful power supply. I actually like the look of these loads, too.

I’ve already written about the power ratings: 600 W for the smaller model and 1200 W for the larger one are specified as short-term ratings. In fact, the cooling system has plenty of headroom. For example, I recently tested a power supply (review coming soon) and ran a heavy load at 1200 W for an extended period.
According to the sensor, the temperature was 80 degrees; I think it was actually closer to 90–95. The transistors are rated for 125 degrees in this operating mode.
I've seen people write several times that the transistor temperature in their loads is 50–60 degrees. Unfortunately, that's usually a measurement error.

Even in my case, where I press the temperature sensor directly against the transistor case and use sensors with a different resistance-versus-temperature curve, the readings are about 15 degrees too low. With the original sensor, the readings are way off.
For example, in the 24MP and 24EW series loads, the sensor is simply cooled by air from the heatsink, so it's hard to say what it's actually measuring.
By the way, you can see the shunts on the left. The originals are rated for 1 W, which is completely unacceptable to me, so I have to replace them everywhere with thicker ones rated for 3 W. I showed the difference once.

And now I've gradually come to what's happening now.
Along with the standard versions, I made a custom one. It differs in the type of connectors used and has a 24EW board, so it supports connecting a charger. The additional board was also modified slightly: the transistor was replaced, and the current can easily be increased to 20–30 A.
The load is in a large enclosure, but rated for 500–600 W. It's assembled so that you can install a second module later if you want, bringing the total to 1000–1200 W. In other words, everything is in the enclosure except the second module.
This was actually the first load where a problem came up. For some reason, its WiFi doesn't work. This was the first time I'd built a load with a customer's board, and the second time I'd used a 24EW board (the first was my own, tested one), so this was an unexpected snag. I didn't check it during assembly, partly because I didn't want to bind the board to my software. When the customer received it, they were disappointed. Resoldering didn't help, so we're figuring out what to do.
Update: it turned out to be an issue with the control module (not the only one). In the end, I had to add an operating mode switch and a USB WiFi adapter, and that fixed the problem.

Besides the upgrade option, the customer wanted to be able to apply a high load for short periods. I can say that the load can do that: up to 900 W for a few minutes without any issues. Beyond that, it will naturally overheat and the protection will trip. I didn't push it that far, since tests like these need to be supervised. I think it could handle 1000 W briefly, perhaps even more, but I haven't tested that.
Photos from testing above the rated power.

I got so interested in battery testers—that is, loads that can not only discharge a battery but also control a charger—that I decided to design my own charge switching board.
As a reminder, the original board gets very hot during operation, has a large voltage drop, and uses a flawed reverse-polarity protection circuit (which may not work correctly with CC/CV chargers). Here's what I ended up with:
1. The circuit is fully disconnected. Previously, current was interrupted only in the charger-to-battery direction, but could still flow in the opposite direction. This could short the charging terminals and cause a short circuit. It could also damage the charger's output if power was disconnected and then restored (with chargers such as the well-known DPS, DPH, RD, and similar models).
2. The voltage drop is about 50 mV at 10 A, and therefore 100 mV at 20 A. That means just 2 W of heat, so without additional cooling it can handle charging currents up to 30 A. With two modules, the current can easily be increased to 40 A.
3. Reverse-polarity protection. The original protection was completely unacceptable to me. I provide basic protection with diodes, so the load tolerates reverse polarity at its input, but things are more complicated with a charger. In the end, I added a protection circuit to the board that won't connect the battery to the charger unless they have the same, correct polarity.
Unfortunately, the protection works by inhibiting operation: by default, it allows operation and blocks it when the polarity is wrong. It would be better to do the opposite and allow operation only when the polarity is correct, but given how the board works, I considered this completely safe.

In the end, this effectively became a new model. The second custom order had almost the same configuration as the previous one: a large enclosure, 500–600 W, upgrade capability, a 24EW board, and charge control support. But this one already has the new version of the control board.
Yes, on loads based on the DL24EW, I don't use the original UART-to-USB adapter. Instead, I use a version with galvanic isolation and a USB-C connector, which I think is better.
The model now has the following specifications:
Load-mode voltage up to 100 V.
Charging voltage up to 55 V
Power: 500 (600) or 1000 (1200) W
Charge current switching up to 30 A (500 W version) and up to 40 A (1000 W version)
USB Type-C with galvanic isolation and WiFi
Adjustable charge cutoff current, enabling proper CC/CV charging. After my recommendations, the Atorch programmers updated the board firmware.
Hardware CV mode, i.e. You can test not only power supplies, but chargers too.
And thanks again to Dmitry for the neat display panels! Why doesn't Atorch make external displays for all their models...

But there have also been some changes for the worse. It looks like they removed the four-wire connection and replaced it with a three-wire one; I'm still figuring that out. This is actually the third version of the DL24EW board. The first was covered in the review; the second got a firmware update adding a cutoff current for charging; and the third has a redesigned circuit board, with the connection changed and a small DC-DC converter replacing the linear regulator.
The first and second versions are on the left, the third on the right. On the third, they added a 2 mΩ resistor between the A- and V- lines... and changed the regulator circuit.

While I was at it, I built an external battery for a UPS. It even ended up with a power bank function, and I also tested eight more EVE 33140 15 Ah cells.
The idea was to be able to use it instead of the UPS's original battery, since some models don't like the battery voltage going above 14 V and go into protection. It also needed to charge from an external charger via Type-C. I had to reconfigure the power bank board so it wouldn't output more than 14 V, and while I was at it, add a high-current, low-dropout voltage regulator. The result is 13.80 V. The regulator also conducts in the reverse direction, which lets me draw up to 100 W from the Type-C output.

Naturally, I measured and balanced the cells, then assembled them into a pack matched by capacity. The connections are soldered with tinned wire. I wanted to use nickel strip, but the plan was to draw up to 40 A continuously and 60 A briefly, so I decided soldering would be more reliable.

The result is a very compact device with a capacity of 30 Ah, or about 385 Wh, charging via Type-C (up to 5 A), an output of up to 20 V at 5 A, 60 A protection, an active balancer (because it operates at a reduced voltage), a longer service life (also thanks to the reduced voltage), and a build made with care.

And, tangentially related to electronic loads, I finally got my little drill press. I was getting fed up with drilling heatsinks using a cordless drill: it's difficult, and the quality isn't what I'd like. It was stuck in customs for three weeks... But it runs so quietly! At 500–700 rpm, it's about three times quieter than a cordless drill.

That's about all my news for now. I've shuffled my workload around a bit, and I hope to write a few reviews based on material I've been preparing along the way.
What's next? Next on the list is building a battery tester with a built-in charger. I've already bought a couple of DC-DC converters, and I'm also planning a 1200 W load with a current of up to 100 A.
People often ask me about the warranty. Friends, I'm not a stickler for formalities, and I also believe my word is worth more than a line in the manual. If needed, I'll help you after a year, two years, or even five. And as you've seen, the actual specs are clearly better than the stated ones—not like with Chinese products, where everything is right on the limit...
A special thanks to the patient people, because recently I ran into... damn, I don't even know what word to use here—suggest one yourselves. Anyway, one Friday evening, an “anonymous” person contacted me with a question about a UPS and external battery setup. I replied, they asked another question, I replied, and then they sent another email with a question, which I didn't see until quite a bit later. I didn't reply right away because things got hectic at my day job, and it took up both days of the weekend. Then on Monday morning I received this email, hidden under a spoiler because it contains profanity.
The question is, how are you supposed to feel about people like this—and whether they're even people... Just a reminder: I have my own work and my own life (at least for now), and if I haven't replied, there's a reason or other priorities. Besides, our military personnel periodically ask me for help, and they are my top priority.
This page was found when searching for:
электронная нагрузка своими руками
электронная нагрузка своими руками
Questions and answers
Which electronic loads is the author currently building regularly?
They are models based on DL24MP boards: 500–600 W for the small version and 1000–1200 W for the large one. Both support up to 40 A and use an encoder for control.
Why might an electronic load show an inaccurate transistor temperature?
The author says that even a sensor pressed directly against a transistor case reads about 15 °C low. The stock sensor may give even less reliable readings.
What electronic projects does the author plan next?
He plans to build a battery tester with an integrated charger and a 1200 W load rated for up to 100 A.
Discounted products
You may be interested in
Discounted products
Support the project
All the content on the site and the advice are free, but we’d be grateful for your support of the project and the channel!
Comments26
Discuss on the forumHave a question or some experience to share? Discuss the article with the author and readers: the forum is a convenient place to share schematics and photos.
Go to the forumOnly registered readers can comment. It takes a minute to sign up, and you can log in with the same account you use on the forum.
Log inRegister





Comments are published in their original language.