68 V 800 W Power Supply from RuiDeng

A couple of months ago, one of my readers sent me a RuiDeng power supply to test. Well, it’s sold under that brand, although as far as I understand, it has nothing to do with the company itself. He sent it to me because he’d seen my review of the same power supply, bought from the same store, but it turns out quite a few changes have been made since then.

To stay up to date with new posts on the blog, I recommend following me on Instagram, where I’ll post notifications about all new articles and maybe write about something interesting now and then - link to my account

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

First, I’d like to thank the reader for sending me the power supply and giving me the chance to show you the current version. And I’d also like to apologize for the delay in getting this review out.

So, I previously reviewed a 800 W power supply bought from the store mentioned above. It’s a decent unit, but it turns out they’re now selling a completely different version, so we’ll test it and compare the two.

It came in the usual polystyrene box. That’s how the reader sent it to me, but Nova Poshta apparently didn’t think that was enough and packed it in an additional cardboard box. All in all, the packaging was top-notch :)


The power supply wasn’t just brand new; it was even sealed in plastic. The person who sent it to me bought two, used one, and sent me the other.


Of course, I was interested in comparing it with the previous model, especially since I still have that one lying around, which makes things convenient.
The first thing that catches your eye is the extra ventilation holes on the old model.

From here on, the new model is on the left or at the top.



The terminals are identical, but the trimmer resistors are different. More on that later.


The new unit is labeled RuiDeng. I only noticed later that, despite having the same power rating, the two have different output currents. That’s because the new one is rated for 68 V, while the old one is rated for 70 V.
The mains voltage selector on the new unit is also covered with transparent film, as on the power supply for the RK6006 and the 1500 W power supply.

Since the power supplies were clearly made by different companies, the case has changed too—more specifically, the fan grille. I have to say, it’s noticeably better now.


Naturally, we’ll open it up and take a look inside. That’s what this was all about, after all.


To an inexperienced user, the differences seem minor: the board is black instead of green, and the components are arranged a little differently. But in reality, these two power supplies have more differences than similarities—apart from, of course, their basic function.


Let’s get started with the comparison.
1, 2. The input mains filter is present in both versions, but it looks a bit neater on the new one. Also, it has two diode bridges without heatsinks, compared with one bridge with a heatsink in the old unit. It’s hard to say which is better; personally, I’d have preferred a more powerful bridge with a heatsink.
3, 4. The thermistor is more powerful, but they skimped on the relay: 10 A versus 16 A in the old unit.
5, 6. The input capacitors have one and a half times the capacitance, which is certainly a good thing.


The fans are roughly similar, although the new one seemed a little quieter to me. But otherwise, the changes are major: the auxiliary power supply is gone, and there’s noticeably more room inside.


1, 2. There’s still one transformer, which controls the transistors. By the way, you can clearly see here that the circuit design is very similar to the 1500 W power supply I showed recently.
3, 4. The main transformer is the same size, but a thermostat has been added to the output heatsink.
5, 6. The PWM controller is an NCP1252A, the same one I installed when repairing the LRS-350 power supply. The old unit had the older KA3525.

Yes, I think you’ve already figured out that the new power supply uses a half-bridge LLC topology, just like the LRS-350, while the old one used the more common half-bridge topology. The single-ended controller, the lack of a decoupling capacitor in the transformer primary, and the pair of diodes near the power transistors all point to this.
It’s hard to say right away which is better or worse, but personally, I think the “asymmetrical half-bridge” is better in this case.


The transistors are LSB65R099GF, quite interesting and, in my opinion, pretty good: N-channel, 650 V, 40 A, 99 mΩ.
There’s an op-amp near the transformer; I assume it’s part of the overcurrent protection circuit.


1, 2. The output diode assemblies are identical, MUR3060CT, but the new version has two of them.
3, 4. There are now three 680 µF output capacitors instead of four 470 µF ones, so the total capacitance is about the same. However, the new version uses 80 V capacitors, compared with 100 V in the old one. I’d still recommend replacing them with something from a more reputable brand. The output storage inductor in the new version is about half the size, but we’ll have to see how it performs thermally, since size isn’t everything.
5, 6. In the new version, current is monitored on the output side, using two shunts. The old version monitored current on the primary side using a current transformer.


There’s a powerful transistor near the fan connector. It appears to control the fan speed.


Most components in the new version are SMD, while the old version uses through-hole parts. The new version also has a multiturn trimmer for setting the voltage, and there’s clearly a spot next to it for a second trimmer. I assume there’s a version with adjustable current.


Now we can move on to the tests.
First, I checked the power consumption and output voltage adjustment range.
The adjustment range is excellent: 52.5–74 V. But once again, I noticed that with no load, power consumption is highest at the rated output voltage and lowest at the ends of the adjustment range—exactly like the 1500 W power supply. The only difference is that this one doesn’t “buzz” and is completely silent in general.


For the rest of the tests, I used one of the electronic loads I build. That’s partly why this review was delayed: sometimes I don’t have the time, and sometimes another load has gone off to its new owner :) By the way, I gave it a little “facelift” and replaced the power connectors with larger ones.


I started by measuring efficiency. The test covered 1–13 A in 1 A increments. Yes, 13 A is clearly an overload, but the power supply handled it just fine.
Peak efficiency was 90% at half the current, then it started to drop slightly. In my opinion, that’s a pretty good result for this kind of power supply.


I had to improvise for the load test. I didn’t want to open the case while it was running, so I put together this setup, which lets me remove the top “cover” at any time, take a thermal image, and close it again.


I also checked how accurately the voltage was regulated as the load and temperature changed. However, since the feedback components are near the edge of the board, on the air intake side, the voltage shouldn’t depend on the temperature inside the power supply itself.
As for the voltage, here’s what I found: with no load, it’s about 0.3 V higher than under load, and it barely depends on the load current—whether it’s 1 A or 12 A.

1. Test begins, current 6 A.
2. End of the first test stage, current 6 A.
3. Test begins, current 12 A.
4. End of the second test stage, no load.

I seem to have lost the photo from the end of the second test stage at 12 A, but the voltage was the same as at the start of that stage.


1. Thermal image after the first test stage at 50% load.
2, 3. Thermal images after 20 minutes at 100% load.

The results are excellent. The hottest parts were the load resistor on the power supply output and the transformer winding, at around 95°C. The core itself only reached 74–75°C, while the output inductor was at a similar temperature, around 80–85°C.
These are good results, especially considering that the room temperature was around 30°C and the power supply was running at 100% load, or even slightly above.

I should also mention that the fan was relatively quiet, and the exhaust air was just warm, definitely not hot. There were no unusual smells either, such as overheated varnish.


You’ve probably noticed that I measured the voltage with a clamp meter rather than my usual UT181A. Here’s why I had to do that.
I ran into a strange multimeter glitch.
1. Power supply with no load.
2. 8 A load, everything’s fine.
3. But as soon as I increased the load to 9–10 A, the multimeter really started going crazy. I tried restarting it, but that made no difference. The device started showing 680 instead of 68 volts, jumping between ranges, indicating an overload, and absolutely refusing to work properly.
4. Even with the load turned off, it still showed nonsense. More precisely, it seemed to be displaying the correct value, but the range indicator had gone haywire.


The clamp meter and the load's voltmeter, on the other hand, worked perfectly, allowing for their measurement error, of course.


Naturally, my first thought was about the output ripple and what kind of hell was going on there if it was making the multimeter lose its mind.
Imagine my surprise when the oscilloscope showed that the output was actually looking very good.
Below are the waveforms with no load and at currents of 4, 8, and 12 A. Yes, 2 V of ripple may seem like a lot, but bear in mind that this is 2 V at an output of 68 V, which is a good result for an 800 W power supply.


The low-frequency performance was also perfectly normal, thanks to the high input capacitance. After all, 750 µF at 800 W is genuinely pretty good.


Let me brag a little: the other day I shipped a customer a pair of custom electronic loads. They're basically the standard 1000–1200 W model, but with different connectors and a power button (the customer wanted those and supplied both the connectors and the buttons). While I was at it, I also added the option to power them from a 5–6S battery, such as one from a cordless drill. The idea is simple: tests can take a long time, and the power might get cut off, so in that situation the load switches over to battery power. It can't charge the battery itself, but that's better than interrupting the test.



Conclusions.
I liked the new version of the power supply, and in my opinion it's noticeably better than the old one, both in terms of circuit design and components, as well as performance. I've covered and described all the differences in the review, so I have nothing to add.

In short, I really liked this power supply, so I can definitely recommend buying it.

That's all for today. I hope you found this useful, and as always, I'd be glad to hear your comments.