Adjustable Voltage Converter RD6012(W), New Version from Rui Deng

Last fall, a very interesting adjustable voltage converter, the RD6006, went on sale. It stood out from previous models with its large display, convenient controls, battery charging function, and WiFi connectivity. But discussions soon started: 60 V is certainly nice, but 6 A is a bit on the low side. The manufacturer replied, “Don’t worry, a more powerful version will be out in about six months.” In fact, it took longer, mostly because of the virus, but the manufacturer didn’t abandon development, and today I’m reviewing the version with an output current of up to 12 A.


Since I already reviewed the previous version, and the devices are very similar, this review will focus mostly on the differences between the old and new versions. I’ll copy over the most important details from the previous review, though.

I’ll start with the packaging. I ordered the kit with the converter and enclosure. At the moment, this kit costs $101.00. I’ve also already reviewed the power supply I bought from the same seller, which is specifically intended for this converter version.

As always, the seller took care with the packaging, but apparently underestimated what the postal services can do, since they managed to dent even a box like this. I’ll come back to that later.
The box with the converter was inside the enclosure, which provided extra protection.


No complaints about the converter module’s packaging: thick cardboard, a soft tray inside, everything neatly packed, and a brief list of specifications on the bottom.


The product page lists the full specifications and, conveniently, compares them with the 6006 version. I really like the manufacturer’s approach in general: there are detailed descriptions, some of them in Russian.


I ordered the full kit, which includes:
1. Converter
2. WiFi module
3. Temperature sensor
4. Pair of terminals
5. Fuse

The only difference from the basic version is that it doesn’t include the WiFi module. I’d recommend ordering it with the module, though, because it may be hard to buy one later.


1, 2. WiFi module based on the ESP 12F, with a voltage regulator for it already mounted on the board.
3. Temperature sensor with a long wire, about 1 m.
4. 20 A fuse. For some reason, there’s only one; I wouldn’t mind having two, just in case.
5. Pair of fork terminals.

6. Battery, not included! I added the photo to show what size you need.


To build a complete power supply, you’ll need the converter, enclosure, power supply, and a cable for connecting it to the mains.


Externally, the converter differs from the 6006 model only in the model number printed on it—in this case, 6012. Otherwise, they’re easy to mix up.


The dimensions, mounting method, and everything else are also identical to the 6006 version, since it uses the same enclosure, keypad, connectors, and so on.


Since the controls are exactly the same, I’ll simply repeat their description from the previous review.
A. Power button
B. Shift — additional control modes
C. MEM — access preset memory slots
D. Set output current (use Shift to set the overcurrent protection limit)
E. Set output voltage (use Shift to set the overvoltage protection limit)
F. microUSB port for connecting to a computer
G. Negative terminal
H. Positive terminal for charging batteries
I. Positive terminal for the main output
J. Turn on output voltage
K. Encoder: adjust current and voltage, select menu options, press to exit the menu
L. Up/down and left/right buttons: navigate the menu and select the adjustment step for current/voltage
M. Enter: apply the selected settings
N. Numeric keypad: directly enter current/voltage values, access memory slots and the menu, and lock the keypad.
O. Display


All the differences become apparent when you look at the back: the power board has clearly been redesigned.


They kept the input common-mode choke, but it is clearly rated for a higher current now. The four-pin terminal block for connecting to the power supply has two pins for each pole.


The auxiliary voltage converter section. I also noticed that small LM321 op-amps are used throughout the board, and found a resettable fuse, apparently in the power supply circuit for these converters.


A few notes on individual components
1. Input reverse-polarity protection transistor. At least, that seems likely from the way it is connected. I think it could be removed to reduce heating.
2. Two 330 µF, 80 V input capacitors.
3. GS8332 precision op-amp
4. Relay that disconnects the battery charging output. Since the current can now reach 12 A, they replaced the relay with a more powerful one.
5. The same PWM controller used in probably all of this company's converters: TL594
6. There are five capacitors at the output: three 220 µF and two 330 µF, all rated for 80 V. I couldn't trace how they are connected, but a continuity test suggests they are in parallel. Only one capacitor is connected after the shunt; the others are before it. The shunt is on the negative side.


The product page had a nice picture of the converter taken apart, though it's very hard to tell what goes where from it.


But let's fix that oversight and take it completely apart.
The converter is fairly easy to disassemble (unlike putting it back together): first, use a powerful soldering iron to desolder the three output terminals, then release the clips and remove the board. It connects to the lower board via a multi-pin connector.


As far as I can tell, the control board is also used in the 6006 version and differs only in the firmware version.
I noticed three unused screw holes near the keypad, but the keypad is very soft, so nothing gets pressed through.


Microcontroller, flash memory, voltage regulators, another precision op-amp and, as far as I can tell, an RC04V FRAM memory chip.


A UART-to-USB converter and, unexpectedly, a seven-segment display driver right next to it. I assume it drives the keypad.


The underside of the power board is practically empty: a couple of transistors, a couple of capacitors, and that's it.


A pair of transistors is screwed to the heatsink underneath. Judging by the circuit, synchronous rectification is used. They are N-channel transistors rated for up to 120 A and 100 V, with an on-resistance of 6.8 mΩ.
The use of N-channel transistors suggests that they are probably driven here by a dedicated driver with bootstrap power.

On the other side of the board is the converter's power inductor. It looks too small to me. I didn't measure it specifically, but it looks the same as the one in the 6006 version, which is odd.


Probably the transistor driver and fan control section.
A small temperature sensor is visible on the left, near the heatsink, but it isn't pressed against it. I pressed it against the heatsink just in case, but you have to be very careful because it's mounted very low.


With the case removed, some components are easier to see: the input section, fuse, and temperature sensor connector.
It's interesting that the fuses are soldered in, but there's a fuse holder next to them, presumably for user replacement. I'd rather solder the fuse in, though—the holder could heat up at 10–12 A.


There's also a spot for a spare fuse near the output, along with a diode for output reverse-polarity protection (for example, when connecting a battery). It probably blows the fuse. Two 10 mΩ shunts connected in parallel are also visible. Since there’s room for a third shunt nearby, there may be a 6018 version based on this board.

There’s an LED near the connectors, and another one near the temperature sensor connector. Their purpose isn’t clear yet.


I put everything back together, cursing the designers who came up with those tight clips, installed the WiFi module—which neatly fills the empty space—and moved on to checking the features.


For the first power-up, I used the power supply I plan to use with it. Nothing burned out, and the converter started up normally, which is already encouraging.


On startup, it displays the same splash screen as the 6006, with only the model name changed. After that, the converter switches to normal operation.
During testing, I also powered the new model from the older version of the converter and caught a very curious glitch on the RD6006 after it had been idle: all the digits in the voltage input field went blank except the least significant one. Turning voltage setting mode off and back on restored everything, but this is the first time I’ve seen that in all the time I’ve been using it.
The third photo also shows the minimum voltage at which the converter starts up and tries to operate: 4.75 V. Its voltmeter reads much too high at that point, of course, but everything returns to normal with a 5 V input.


The display indicators are identical to those on the 6006, so I’ll simply repeat the description:
W — current date and time; works correctly only when a battery is installed.
X — sound notifications on/off
Y — keypad lock
Z — connection type: USB/WiFi/UART
AA, AB, AC — output voltage, current, and power
AM — input voltage
AL, AK — set output voltage and current
AJ, AI — voltage and current thresholds at which the output will be switched off automatically; trigger protection.
AD — active memory slot number
AE — CC/CV operating mode
AF — status: normal, overheating, fault, etc.
AG — battery charging mode; lights up red when active.
AH — amp-hours, watt-hours, temperature. The Ah/Wh counters reset when power is switched off and back on.


The control menu is also almost identical to that of the previous model. Some items have been renamed, but their positions and functions are unchanged; I described them in more detail earlier.
1. Main settings: sound, connection parameters, brightness, and time.
2, 3. Main screen display modes: standard and graph plotting.
4. Preset current/voltage mode settings
5. Firmware version (1.31), board temperature near the heatsink, and serial number (which, by the way, is lower than the 6006’s).
6. There’s a small error in the description and specifications: they list 60 V, 12 A, and therefore 720 W, but the converter can actually output 61 V at 12.1 A, or 738 W. So, allowing for the board’s efficiency, an 800 W power supply leaves almost no headroom.


There’s not much else to inspect, so I’ll move on to the tests.

Test 1: checking the accuracy of the input voltage measurement. I’m not entirely convinced this measurement is necessary, but the converter passed the test in my view. The largest difference was at the minimum of 10 V, where it was +0.03 V.


Test 2: checking the accuracy of the output voltage setting and the converter’s own voltage measurement.
I tested at 0.5, 1, 2, 3.3, 5, 10, 15, 20, 30, 35, 40, 45, 50, 55, and 61 V. The results fully satisfied me in every mode. There’s a small measurement error, but I don’t consider it significant. The biggest difference was at the maximum output voltage. It was set to 61 V; the actual output was 60.96 V, while the built-in voltmeter showed 60.98 V.


Test three: current setting and measurement accuracy.
The test was performed at the following settings: 10, 20, 50, 100, and 500 mA, and 1, 3, 5, 7.5, 10, and 12 A.
Almost all the settings were fine, except for 10 mA: there was no output current at all—in fact, it was no different from when the output was disabled. Instead of 20 mA, I got 13 mA. Everything was fine from there up to 7.5 A, but at 10 and 12 A the difference was 40 mA: at 10 A it was lower, and at 12 A higher.
Ignoring currents below 20–30 mA, it was generally fine.


Test four: how well the output voltage is maintained across a load-current range of 50 mA–12 A, at 3.3, 12, 24, 36, 42, and 50 V. Testing a wider range was very inconvenient because of the electronic load's limitations; it was already running at 1.5 times its rated load, dissipating 600 W instead of its 400 W maximum.

Up to and including 24 V, there was a slight positive dependence. At 36 V there was no difference, and at higher voltages there was a slight negative dependence.


Test five: checking the minimum required input-to-output voltage difference at 10 A, with input voltages of 36 and 57 V.
I got a difference of about 3 V with a 36 V input and 4.7 V with a 57 V input. At a 12 A load, the difference will be a little greater, but in general, when powering it from a switching power supply, I'd recommend setting a difference of at least 7–8 V.


Test six: measuring output ripple amplitude in different operating modes.
Output set to 30 V, input at 68 V.

1. No load
2. At 4 A
3. At 8 A
4. At 12 A.

Well, what can I say? On the one hand, it's disappointing: ignoring the spikes, ripple amplitude at 12 A is 200 mV. On the other hand, the manufacturer specifies up to 250 mV at 6 A, so I suppose it meets the spec.


Here the load current is 12 A in every case, but the output voltage is 5, 12, 24, and 50 V.
Basically, it's the same thing shown above, just from a different angle.


Although the ripple is within the specified limits, I was quite surprised: the RD6006 performed noticeably better. Even though the test was carried out under the same conditions, I decided to take a control measurement with an oscilloscope running on battery power, at these settings:
1. 30 V, 8 A
2. 30 V, 12 A
3. 12 V, 12 A
4. 50 V, 12 A

And got essentially the same results as above. I tried both the new 800 W power supply and powering it from the RD6006. I didn't try a linear power supply, but that would probably have been overkill.


Test seven: transient response.
1, 2. Applying and removing 5 V at the converter output with no load.
3, 4. Response to connecting/disconnecting a load. The output was set to 12 V, with a 30 Ω resistor permanently connected and a 3.3 Ω resistor connected in parallel with it (or disconnected). This gives a current step from 0.4 to 4 A and from 4 to 0.4 A.
5. Transition from CV mode to CC mode: voltage 15 V, load current 4.5 A, current limit set to 3 A.
6. Same load and settings, but in the opposite direction, from CC to CV. The stepped transition is odd. It didn't happen every time, but it did in most cases; the time from load disconnection to the linear part of the waveform could vary by a factor of 2–3.

The response to connecting and disconnecting a load is much better than on the 6006, most likely because of the higher output capacitance.


“Soldering iron test.” The TS100 soldering iron is connected as the load here. As I recall, the RD6006 failed this test because it had noticeable overshoots. The RD6012 performs much better: upward overshoots are practically nonexistent, and downward ones are much smaller than on the 6006. The increased output capacitance is making a difference again.


Test eight: measuring the converter's efficiency.
The input voltage was 70 V. A multimeter was connected in series with the power supply, so as the load increased, the voltage at the converter input dropped slightly due to voltage drop in the wires. That's why I raised the voltage to 70 V, even though I initially set it to 68 V.
The test was run at a load current of 12 A and voltages of 5, 12, 24, 36, 48, and 60 V.
Output voltage — efficiency — power loss
5 V — 80.6% — 14.47 W
12 V — 90.7% — 14.7 W
24 V — 94.8% — 15.57 W
36 V — 96.8% — 14.36 W
48 V — 97.7% — 13.4 W
60 V — 98.4% — 11.6 W

In my opinion, the efficiency is very good. Of course, there is some measurement error. In this case, I simply multiplied the current shown by the multimeter by the input voltage measured by the converter to get the input power; the converter itself displayed the output power.
But even if there are errors, I don't think they're very large, and they can be ignored.


I also ran the converter for a while at 11.5 A and 50 V. Of course, that's not the maximum, just 575 W, but it's what I could manage. After it warmed up, I measured the component temperatures.
1. Overall view. You can see that the heat is concentrated in the lower-right corner and on the left.
2. The auxiliary power converter on the right gets warm, but not critically so; it doesn't even reach 60 degrees.
3. An inductor is installed on the left. Its temperature is 50–52 degrees; I think the fan airflow helps.

By the way, about the fan: I didn't check the temperature-based activation, but I did check the current-based one. It turns on at 8 A, regardless of the output voltage. The fan is very quiet, and against the noise of the primary power supply, you can hardly hear it at all.



As I wrote at the very beginning, I ordered a case along with the converter. There are two versions of the case, for the 6006 and the 6012. The difference is that the 6012 version doesn't include a fan or its controller, so it costs less.


1, 2. And this is the other thing I wanted to write about: the case arrived damaged. How the postal service managed that when the parcel was shipped in a sturdy polystyrene box is a mystery... But a little hands-on work fixed everything.
3. Inside the case was a box containing the converter and a set of mounting hardware for the case.
4. The accessories are packed in a bunch of large and small bags.


I couldn't resist and immediately tried fitting the power supply. On one side, everything fits beautifully: even the fan holes not only line up, but the cutouts have the same shape and size.
On the other hand, there's a gap of about 10 mm between the power supply and the case. You can't move it any closer because the power supply's own casing gets in the way.


The kit included not only mounting hardware, but also wires, a power connector, and a switch. I'll say in advance that I used everything they included; nothing was left over.


1. Credit to the manufacturer for reinforcing all the threaded holes with welded nuts. ZKEtech, for example, doesn't bother with this, so the threads eventually strip.
2. There are standoffs at the back of the case for the fan controller board, but not only is it not included, it probably wouldn't fit with this power supply anyway.
3, 4. You can start assembly with anything, but I usually install the switch, power socket, and feet first. The socket takes two M3 nuts and two countersunk screws; the feet need four screws with captive washers.


The manufacturer has basically brought to life the idea I wanted to implement with the RD6006: the power supply fan also serves as an exhaust fan. Originally, the power supply fan would blow hot air into the case, only for the case fan to exhaust it afterward—a lot of duplicated work.

But it would all work well if only the manufacturer had thought to isolate the airflow somehow so it wouldn't flow back into the case. The modification is extremely simple. I used some self-adhesive weatherstripping left over from renovating my apartment and stuck it on in a U shape, so the strips overlap when pressed together.
The result is excellent: the power supply fan easily pulls air through the entire enclosure, drawing it in from the front and cooling the converter board along the way.


During assembly, you may be surprised that they only included two nuts. It’s actually quite simple: the power supply is secured with screws in the holes used to mount the board inside the PSU, but from the other side.


The wires connect without soldering: one end has a female connector, and the other has a spade terminal. Note that the connectors are different sizes: the one for the power socket is smaller, while the one for the switch is larger.
One complaint: they didn’t include any kind of insulating sleeve to keep the wires from lying loose inside the enclosure. After all, they don’t have double insulation.


1. The converter’s power input wires are paired. I recommend stripping them a little more than they are from the factory; otherwise, they may not clamp securely in the connector.
2, 3. Everything comes together quite neatly. The converter’s input terminals are even laid out so the wires don’t cross over each other.
4. A temperature sensor came with the converter. It’s intended for external use, but I slipped the thermistor into the PSU’s air outlet, so now I can monitor its temperature.


I ended up with the spade terminals and fuse that came with the converter left over.
The finished unit looks fine, but the slightly bowed top cover and the enclosure’s considerable depth detract from the overall look—and that depth isn’t always convenient. It would also be nice to have the power switch on the front panel, or to add a standby mode, but that would require another small power supply.


Test nine: battery charging.
Last time, I complained that charging didn’t stop at the set current or at 1/10 of the charging current, but only once the current dropped to 10 mA. I hoped the manufacturer had fixed this, but it turns out they hadn’t.
We connect a battery and set the charging current to 12 A, but charging stops at 100 mA, not 1.2 A. Of course, that’s much better; at least you can charge battery packs with a balancer. Still, it would be better to have the option to adjust this threshold.


A few words about the software: you can download version 1.0.0.7 here.

The software hasn’t changed much. It’s the same interface, letting you fully control the power supply over either USB or Wi-Fi. In the latter case, you can configure access through your home network.
When you launch it, you’ll be asked which series of device you plan to control.


Once you select and connect the device, the software switches to control mode, and the power supply’s keypad is locked.


An advanced mode lets you run a user-defined program. You can specify the sequence of voltage/current changes and how long each step lasts. You can also set up a sweep: specify the current, starting and ending voltages, and the duration of each step.


The software can also control several devices at once over RS485. In that case, you’ll also need to set each device’s network address.


And finally, a face-to-face comparison—even though there’s not much to compare when they’re practically twins :)


And, of course, the verdict.
The converter is very similar to the previous version, but it also has some notable differences: twice the power and output current, higher efficiency, and the charging cutoff current has been raised from 10 to 100 mA, bringing it closer to a normal charge termination current, along with a few other minor changes. The converter is very well calibrated and maintains good voltage regulation accuracy despite large load changes. I really like the extra features in the PC software, which turn the converter into a truly programmable power supply. Wi-Fi no longer feels out of place, though I’d suggest the manufacturer include an Ethernet module as well.

I’d be ready to praise it: overall, it works well and in some respects even outperforms the RD6006, especially in how it responds to load changes. But, as always, there’s a catch: the ripple amplitude is higher than I’d like, even though it remains within the manufacturer’s specifications. The 6006 had 95 mV; the unit under review is closer to 250 mV.
I should also mention the coarser current adjustment and measurement: 10 mA versus 1 mA on the 6006. But that depends heavily on your requirements; personally, I’m fine with this level of adjustment.

In short, it works, and there are improvements, mostly for the better, but there are a few “quirks.” I tried to test it as thoroughly as possible in this review, but if you have any more questions, I’ll do my best to answer them.



That’s all for today. I hope you found it useful.