NVVV S-800-70 Power Supply for the RD6012 Converter

Some time ago, a new version of the converter called RD6012 went on sale. Since it also needs a power supply, I ordered a suitable one from the same store—in this case, an 800 W unit with a 70 V output.
This review won't be very long; why will become clear as we go, but I'll still run some tests.


I ordered the power supply and converter at the same time, and they were even shipped together, though in separate parcels. As it turned out, the converter was delayed a little in transit, so I'll start with the power supply. Also, while the converter was sent to me for review, I bought the power supply because Ruideng simply sells these units; they're made by a completely different company.

As always, the packaging gets top marks: a large polystyrene box, with the power supply firmly secured inside in a separate polystyrene insert.


The power supply is sealed in plastic, and a small bag containing a resistor was included.


In short, if the power supply emits a high-pitched noise when there's no load, connect this resistor in parallel with the output to load it. The resistor is 2 kΩ, so it will dissipate 2.45 W.
If the power supply doesn't make any noise, or you're happy with how it works, it's better not to install the resistor—there's no point in dissipating extra heat.


The cover is almost mirror-finished. The power supply measures 240 × 124 × 65 mm.


The terminal block has a cover. To its right are the power indicator LED and a trimmer for adjusting the output voltage.
There are ventilation holes for the fan on the opposite end.

According to the label, the power supply has the following specifications:
Input voltage: 100–120 or 200–240 V; 200–240 V is selected by default.
Output voltage: 70 V
Output current: 11.4 A
Output power: 800 W


Initially, the output voltage was about 70.1 V, but after just 10–20 seconds it dropped to 69.92 V and stayed there.
The adjustment range is 57.35–73.44 V. The adjustment is fairly smooth, and its resolution is better than what's usually needed for power supplies like this.


Now I can explain why there's no point in writing a long review. I've already tested a similar power supply, and something tells me the converter manufacturer chose this one for the new model for a reason. If you're interested, here's the review of the 600 W model; you'll find a lot of similarities with the one I'm reviewing here.


1. The input has a common-mode choke and Y and X capacitors, but no varistor. The fuse is also a letdown—they could have used a full-size one.
2. The bridge rectifier is mounted on a heatsink.
3. Here's an important difference from the previous power supply: after startup, a relay shorts out the thermistor.
4. The input capacitors have a capacitance of 1000 µF.


One thing stood out right away: the isolation capacitor is installed so that it's pressed against a high-power resistor on one side and a small capacitor on the other. It looks rather untidy; at the very least, I would have raised it a little.


As in the previous model, the current-limiting circuit uses a current transformer. The trimmer to its left sets the protection trip current.
There's a fan connector between the trimmer and the current transformer. A resistor (gray) is installed to its left; its value determines the voltage supplied to the fan during normal operation.


The PWM controller is the same 3525, but from a different manufacturer. An LM358 op-amp is installed nearby as well.


1. Unlike the lower-power model, the thermistor isn't right up against the heatsink; it's set a little farther away and essentially measures the air temperature inside the enclosure, which I don't think is quite right.
2. The output has a large energy-storage inductor and four 470 µF capacitors, from the same manufacturer as the input capacitors.
3. The inverter transistors are mounted on a fairly large heatsink, which is itself screwed to the enclosure.
4. The output diode assembly is an MUR3060CT, also mounted on a heatsink. Unlike the transistor heatsink, this one is fully isolated from the enclosure.
5, 6. The transformer is a decent size, but it’s wound with ordinary wire. It would be nice to see litz wire used by now.


For a closer inspection, I had to remove the board from the case. By the way, it’s held in place by just three of the five screws; I recommend adding at least one more in the center of the board.


I’ll say this up front: the fan moves plenty of air, but it’s quite noisy, especially at the second speed—which, fortunately, is rarely used. That’s why I mentioned the resistor near its connector earlier; I think the fan speed could be reduced considerably without any downside.


A separate circuit based on a VIPer22A handles the auxiliary power supply. The inverter transistors are IRFPS37N50A, and the two transistors look noticeably different from each other.


The input capacitor values are close to the stated figures. Since they’re connected in series, the total input capacitance is 477 µF, which is rather low for an 800 W power supply.


The soldering is reasonably neat—in fact, quite clean.


There are safety slots, and the solder mask isn’t continuous either: there’s a strip without it separating the primary and secondary sides..


I found just one issue, with one of the transformer leads: the solder pad wasn’t completely covered with solder.


I started testing with a pair of electronic loads rated for a maximum input voltage of 200 V. With no load, the power supply draws about 6 W.


I ran two tests in parallel, measuring ripple and efficiency, but ran into a problem: the two loads shown above have a combined power rating of up to 450 W, and I needed more. A third load could have brought the total to 700–750 W, if not 800 W, but its maximum voltage is only 50 V. I had to set the power supply to its minimum output voltage, though in this configuration the electronic load was operating outside its normal range.


At an output current above 11.5 A and a voltage of 57.13 V, the power supply began lowering its output voltage, apparently trying to switch to constant-current mode. The maximum output power at this voltage was 657 W, with about 750 W drawn at the input, giving an estimated efficiency of 87%—quite good.


The efficiency graph may be a little inaccurate because measurements were taken at different output voltages, the steps weren’t tied precisely to power levels, and the current setting wasn’t very accurate either. Still, it gives a general idea. It shows the output power range from about 140 to 657 W, divided roughly into five sections.


I measured ripple over the same power range. Here are the results:
1. Almost clean with no load.
2. At 140 W, about 150 mV peak-to-peak.
3, 4. Almost the same across the 280–420 W range: 450 mV.
5. At the maximum load I could manage—almost 660 W—ripple barely increased, to 500 mV.
6. Interestingly, in all operating modes the ripple amplitude would periodically drop by about 1.5–2 times. Here’s another result at 660 W.

At first glance, the ripple may seem very high, but that’s not the case. The oscilloscope trace shows that most of it consists of very short spikes, which are easily removed with a simple LC filter. The main ripple component—the part that’s difficult to suppress—is only 20–40 mV across the 140–660 W range. That’s very good for a power supply with a 70 V output.


For the thermal test, I used all three loads, for a combined power of about 630 W. Input power was 721 W, so about 90 W was dissipated in the power supply.


The test lasted an hour, though active-cooled power supplies don’t really need to run that long. After about 20 minutes, the fan began periodically switching to full speed and continued doing so until the end of the test: roughly one minute at full speed, then three minutes at medium speed.

1. At the start of the test, the power supply was already slightly warm from the efficiency and ripple measurements.
2, 3. After half an hour, nothing had changed significantly, but I found a clear hotspot above 200 °C, while the other components were at 60–85 °C.


After another half hour, the component temperatures were unchanged, which means the system had reached thermal equilibrium. Any further change would depend only on the ambient temperature.
1. Overall, the transformer temperature is still the same 60 degrees.
2. The snubber resistor is around 80 degrees. In my previous review, I complained about how hot this section got.
3. The output inductor was still at 85–86 degrees, just as it was half an hour ago, but there was still a component nearby running at a very high temperature: 225–230 degrees.


It turned out that the resistors in the RC snubber networks connected in parallel with the output diodes get extremely hot. Their temperature ranged from 180 to 230 degrees, depending on the fan speed and whether the cover was on.
The photo below shows these resistors. You can even see that they have changed color slightly in the middle. I recommend replacing them with higher-power resistors of the same value (68 Ω).


To sum up, I liked this power supply more than the previous 600 W model. In fact, the two are similar only in their circuit design: the circuit board layout has changed significantly, a relay has been added to short out the thermistor, the snubber values in the high-voltage circuit have been changed, the transformer and input capacitor capacity have been increased, and the inverter uses transistors rated for 1.5 times the current. The overall build quality isn't excellent, but it's pretty good.
The unit is also more stable in operation. I didn't notice the voltage drifting as it heats up; it only drops slightly after startup, so it's best to set it 20–30 sec after switching on.
The ripple level is also comparable to that of the lower-power model. At the very beginning, I mentioned that it came with an “anti-whine” resistor, but I didn't install it. If there was any whine with no load, it was very faint, though this can vary from unit to unit—some people may be less lucky.

There are a few shortcomings, too. I wasn't happy with how hot the output diode snubber resistors get; they definitely need replacing. I also wasn't impressed with the input fuse or the rather loud fan. If you're opening it up, you can reduce the fan speed while you're at it.
The output current is limited to 11.5 A, compared with the specified maximum of 11.4 A. You can adjust it, but I'd leave it as is—or increase it only slightly. If the limit is exceeded, the power supply switches to constant-current mode.

By the way, in theory, you could try making a similar power supply adjustable, since it has an auxiliary power supply and the PWM controller's stability doesn't depend on the output voltage. What's more, power supplies like this can even be connected in parallel.

In short, it's certainly no Mean Well, but overall it's a decent power supply, though it does need a little work.

That's all for today. I hope you found this useful.