Preparing power supplies for a multichannel battery tester
To charge battery packs up to 19S, you need a voltage of almost 80 V. I had a DPS8005 converter lying around for regulating current and voltage, but for it to work properly, its input needs at least 84–86 V. Incidentally, this converter is what limits the maximum voltage of the battery packs being tested. If it were rated for 84 V, I could go up to 20S.

In my review of these power supplies, I praised them, and I’ll praise them again now: they’re genuinely great little units, and I even felt a bit bad putting them in a charger that doesn’t really care about ripple.
They output 50 V at up to 2.5 A and were clearly intended for PoE systems.

Only later did I think that I could probably have used a regular enclosed power supply with a 48 V output and doubled its output voltage by adding a couple of diodes and capacitors, along with a small modification to the circuit itself.

Especially since it has enough power, and more importantly, it’s a push-pull design based on a TL494 (KA7500), which makes it much easier to modify. What’s more, this probably would have been more interesting to my readers, but as they say, hindsight is 20/20. Maybe I’ll show a similar modification using some other power supply as an example.

The first thing to do before connecting the supplies in series is to make sure the negative output terminal isn’t connected to either the case or earth. Otherwise, you’ll get a short across the output of one of the power supplies when you connect their cases or earth terminals together.

The power supplies output 50 V, and I need somewhere around 85–88 V, so I need to lower their output voltage. There’s no adjustment, so I’ll have to open them up.
I remove 10 screws from each power supply to take off the cover! I’ve never seen this level of thoroughness, even from big-name brands.

The goal is to have both boards in front of me, with access to the feedback circuit components.

I already had the schematic, so this part is relatively straightforward. If you don’t have one, find the feedback circuit. It’s usually near the adjustable TL431 shunt regulator, if there is one, and consists of a voltage divider: one resistor is connected to the power supply output, the other to ground, and their junction goes to the TL431 input. There may be more than two resistors, and they may be connected in parallel or in series—for example, R22 and R26 here. This is done to fine-tune the output voltage when standard resistor values aren’t enough.

To change the output voltage, you need to change the resistor values in the divider. There are several ways to determine the resistor values and work out the whole process.
1. Simply replace the resistors with different ones. Desolder the originals, open a calculator, and enter the required parameters—the output voltage and the TL431 reference voltage (2.5 V). The calculator will suggest resistor values.

2. Add extra resistors. Adding one in parallel with the upper resistor will lower the power supply’s output voltage; adding one in parallel with the lower resistor will raise it. You can even build just the feedback circuit in a simulator: first enter the existing output voltage and resistor values, then measure the voltage at the divider output. After that, change the output voltage and add resistors in parallel until the divider output voltage matches what you measured in the first test.
In my experiment, the feedback voltage was 2.488 V with a 50 V output.
Then I set the power supply voltage to about 44 V and adjusted the resistor in parallel with the upper one. I had to disconnect the 150 kΩ compensation resistor (R26 in the power supply schematic), and then everything worked perfectly with an additional 430 kΩ resistor, which I had on hand.

Find the required resistors on the board, although I think people usually find them before this stage :)
By the way, you can choose the resistor values by experimenting, connecting resistors to the power supply, but you need to be extra careful not to burn it out.

SMD resistors would of course be more convenient, but I didn't have any. I did have some domestic leaded 0.062 W resistors, shown here next to common imported ones.

I desoldered R26, but didn't throw it away. Instead, I soldered it in so it wouldn't affect the circuit—just in case I need to put everything back the way it was.
I connected a new 430 kΩ resistor in parallel with R27.

I could have used a trimmer resistor, but that wasn't very convenient for me. Speaking of trimmers, it should only be installed in series with the lower resistor, and this is very important.
The thing is, trimmer resistors are less reliable and can lose contact between the wiper and the resistive element. This can increase the resistance or cause a complete open circuit.
So if the trimmer is on the upper side, the power supply output voltage will rise, or you could even lose control of it altogether.
If the resistor is on the lower side, then even if it goes completely open, the output will be 2.5 V (with a TL431), or the power supply will shut down.

Soldered it in, switched it on, and got 43.6 V at the output versus the calculated 43.7 V. Looks great to me.

The next step is connecting the power supplies in series.
I'm often asked: can power supplies be connected in series?
Yes, if they're the same type.
A less common question is whether they can be connected in parallel.
Usually not, although these particular power supplies can be, since they have a CC mode and will limit their current themselves at high power.
What are the drawbacks of connecting them in series?
1. For safety, it's best to reduce the total power a little. For example, with a pair of 125 W power supplies, don't draw more than 230–240 W, so there's some headroom and the power supplies don't go into protection.
2. If the power supplies are identical, their different switching frequencies can cause beat frequencies. For example, one power supply runs at 100 kHz and the other at 101 kHz. At high load power, this results in 1 kHz ripple, which is harder to filter than 100 kHz ripple. Whether it's a problem depends on the situation.
3. Protection diodes need to be installed.
4. It's best to connect the load only after both power supplies have started up.
5. If you use step-down converters after the power supplies, one power supply may go into protection due to overload when the other is switched off. This won't happen with a resistive load, where the current drops in proportion to the voltage.

Protection diodes are needed so that if one power supply shuts down, the current flows around it, as shown in the second diagram, rather than through the transformer, diode, inductor, and power supply circuitry. In the first case, current flows through the transformer winding, which can cause problems when restarting it.

I have a review of a powerful adjustable power supply where I also connected power supplies in series.

The protection diode must be rated for the power supply's operating current, and in some cases it also needs to be mounted on a heatsink, as I did in that review.

In my case, the current doesn't exceed 2.5–3 A, so I used some standard 1N5408 diodes. Something with a lower forward voltage drop would have been better, of course, but these will do. The diodes can be low-frequency types.

There wasn't much room to install them, so I put them where the film capacitors had been, connected in parallel with the power supply output. In this case, those capacitors aren't really needed: the ripple is already low, the power supplies aren't heavily loaded, and there's another converter downstream.

Connected everything, switched it on, and got 87 V at the output, or 43.5 V from each power supply. Great.

When I connected the converter, it turned out that the maximum it could output was 79.91 V. That's not a measurement error—the multimeter shows the same thing.
But let's do the math: 79.9119 = 4.206 V per battery, with a little headroom.

The final step is testing it under load. For the test, I connected a 150 W load, drawing 1.85 A, and left it running for a while.

About 40 minutes later, I checked the temperature.
It turned out that the converter's heatsink got the hottest, at around 72 degrees, while the power supplies themselves were just warm, at 55 degrees.

That concludes this stage. Next, there will most likely be some additional work and assembly, which I'll cover separately. I hope this information was useful :)
последовательное или комбинированное соединение нескольких блоков питания, схема последовательного подключения импульсных блоков питания, последоваьельное соединение источников питания, 4, последовательное включение импульсных блоков питания, https://www.kirich.blog/obzory/bloki-pitaniya/807-podgotovka-blokov-pitaniya-dlya-mnogokanalnogo-testera-batarey, самый простой способ переделки компьютерного блока питания в регулируемый, последовательное подключение импульсных блоков питания, обучение блок питание, последовательное соединение блоов питания me, соединение импульсных блоков последовательно защитные диоды, как развязать последовательные блоки питания диодами?, электроника мк 59 схема, как соединить два блока питания в двухполярное питание из двух блоков питания с общим нулем, тестер блоков питания, kirich.blog/obzory/bloki pitaniya/807 podgotovka blokov pitaniya dlya mnogokanalnogo testera batarey.html, подключяение тестера от блока питания, подготовка блоков питания, тестер стабилитронов от блока питания, через диоды нужно соединить. более подробно прочтите здесь kirich.blog/obzory/bloki pitaniya/807 podgotovka blokov pitaniya dlya mnogokanalnogo testera batarey.html, тестер блоков питания с резисторами, можно два импульсных бп соединить последовательно, последовательное соединение импульсных источников питания, электроника мк 59 схема блока питания, блок питания на lm317 и кт827 схема
Questions and answers
How did the author change the output voltage of the 50 V power supplies?
He modified the feedback circuit, disconnecting the 150 kΩ R26 resistor and adding a 430 kΩ resistor in parallel with R27. Each supply then produced 43.5 V.
What voltage did two power supplies produce in series?
The combined output was 87 V, or 43.5 V per supply. The converter itself could deliver a maximum of 79.91 V.
How did the power supplies perform under load?
With a 150 W load drawing 1.85 A, after about 40 minutes the converter reached 72 °C and the power supplies 55 °C. The author notes that series-connected supplies need protection diodes and some power headroom.
The figures from this review are in the measurement database: Power supplies
Discounted products
You may be interested in
Discounted products
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!
Comments3
Only 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.