RuiDeng RD6006P hybrid adjustable converter

At last, the new product from RDtech that I’ve genuinely been waiting for for a long time has arrived—and I even posted its announcement.
Is it really that interesting, and does it live up to expectations? I’m curious to find out.

I’ve already reviewed all three previous models: 6006, 6012, and 6018. Aside from minor tweaks, though, the only major difference between them is the maximum output current and, consequently, the power.
People have repeatedly brought up the idea that it would be nice to increase the voltage resolution, and especially the current resolution. At the time, though, the discussion was about more powerful models. There were also suggestions and complaints about why the manufacturer still wasn’t using separate DACs/ADCs. Not long ago, a model finally appeared that added not only those, but also the long-awaited hybrid mode.
Anyway, let’s take things in order and, as always, start with the packaging.

As usual, I have no complaints about the packaging: a large foam box containing my entire order.
As before, the converter comes in its own cardboard box. The only difference from previous versions is the color, and there are a few changes to the specifications listed on the bottom.


As usual, the package includes a temperature probe, a pair of terminals, a spare 10 A fuse, and, in my case, a WiFi adapter.


On the outside, it’s the same enclosure many people have known since the RD6006. During the review, though, I noticed that it seems a little lighter in color, but the difference is barely perceptible.


These are the module specifications from the product page. The translation has some errors, but I think it’s still understandable.
The key differences are:
Current and voltage measurement resolution is one order of magnitude higher.
For some reason, the current and voltage setting accuracy is also worse, which doesn’t make much sense.
Ripple is noticeably lower.


The only real external difference from the RD6006 is the letter P at the end of the model name.


The internals, however, are substantially different. There are now two heatsinks, and clearly quite a few more components.


The input power circuitry and the basic PWM controller are unchanged, but there’s no synchronous rectifier here. If I remember correctly, the RD6006 had one.


1. Auxiliary voltage converters that power the board’s electronics.
2. Input common-mode choke, reverse-polarity protection transistor, and a high-power resistor. This section looks much like the one in the regular 6006.
3. The main DC-DC converter’s power stage uses a standard diode assembly and a P-channel MOSFET.
4. A connector for an external temperature probe is tucked between the heatsinks.


There are noticeably more components on the other half of the board. In fact, this is where the model’s main “special feature” is concentrated.


1. There’s a 470 µF capacitor after the DC-DC converter, followed by 220 + 47 µF after the linear regulator.
2. The shunt and the transistors for the parallel current regulator are also located here. They’re responsible for discharging the output capacitors, or at least that’s my assumption.


To remove the board from the enclosure, you first have to desolder three terminals. It was the same on all previous versions, though.



The microcontroller and its associated circuitry.


UART-to-USB section


The underside of the power board has only the TIP35C transistor for the linear regulator. On the left, you can see a spot for a heatsink beneath the auxiliary converters, but the heatsink itself is missing. By the way, the transistor heatsink for the linear regulator isn’t screwed down; it’s stuck to the board with double-sided tape. A questionable solution.


The section responsible for measuring and adjusting current/voltage, as well as controlling the linear regulator.


The most interesting part—let’s take a closer look.
You can make out:
three optocouplers
122U31 - digital isolator
TM7707 - 24-bit ADC
DAC8562S - 16-bit DAC
Stm8s003f3 - microcontroller
GS8592 - operational amplifiers.

The overall design is very similar to the one I showed in my review of the RPM-416 recorder. There’s a main microcontroller and an auxiliary one, connected through galvanic isolation. The DAC/ADC and analog circuitry are connected to the auxiliary controller. The main one controls almost everything, while the auxiliary one handles measurement and adjustment and “talks” to the main controller.


1. The input section from another angle.
2. Power DC-DC converter inductor and relay for the battery charging section
3. 15 mΩ shunt. Interestingly, and very importantly, unlike all previous versions of the converter, this one has the shunt on the output line rather than on the ground. So, if you want, you can power two converters with a common ground from a single power supply.
4. Galvanically isolated converter that powers the measurement/control section.


And here I’d like to say a huge thank-you to forum user UniSoft for providing the device schematics. He does warn that there may be errors, but I’ve looked them over and haven’t spotted anything serious so far.

Power board - link


Control and display board - link


And here’s the 6006P next to a standard 6006 board for comparison.


For testing, I used a 600 W power supply from this review. It was installed in a case with an RD6006 board. In other words, everything was completely ordinary and stock—no low-noise power supplies.


The control menu has been noticeably revamped. They’ve added an output power limit option, as on the 6018 model, as well as the features I’d asked them about repeatedly: choosing the temperature units and the charge termination current.
There are a few quirks, too. You can adjust the current at which the charger considers charging complete, but only in 0.0001 A increments. There’s no way to change the adjustment step, so it takes ages to dial in. I got as far as 216.9 mA before I simply got tired of it.
You can also set the shutdown temperature anywhere from 0 to +100 degrees, which is very convenient.
The connection interface and time settings are all as before.


Russian has been added, but something went wrong here too—or, more precisely, something seems to have shifted, and some menu item names ended up misaligned.


1, 2. Several options are now available for displaying parameters on screen.
3, 4. Two “digital” display options: either just the left side or both sides use seven-segment digits.
5, 6. You can also display all the available information at once, but I think it ends up looking pretty cluttered.
7. As before, there’s also a graph view.
8, 9. The memory cell save menu and information about the module and firmware.


I decided to skip checking the input voltage measurement accuracy. It’s traditionally fine; I just gave it a quick check against another adjustable power supply.

First test: output voltage setting and measurement accuracy. Interestingly, despite the 6006P having a finer display resolution, its stated accuracy is 0.5% + 4 digits, compared with 0.3% + 3 digits for the standard 6006. However, the 6006P’s four least significant digits represent 0.004 V, while the 6006’s three digits represent 0.03 V, so the 6006P should be more accurate in this respect.

So, during the test I set the output voltage to 0.1, 0.2, 0.5, 1.0, 2.0, 5.0, 10, 20, 30, 40, 50, and 61 V. The actual setting accuracy turned out to be quite high. The largest error occurred at the maximum output voltage and was about 0.15%, against a stated 0.3%.

There was one “catch”: during measurement, the penultimate digit kept “wobbling.” For example, it might show 5.006 V, then 5.016 V. For the result, I used an approximate average, or the value closer to the worst result, depending on which reading appeared most of the time. The fluctuation was present both with and without a load.



Test two: current setting and measurement accuracy.
The stated accuracy here is 1% + 6 counts. The regular 6006 was rated at 0.5% and 3 counts, but the least significant digit behaves just as it did with voltage.

The test was performed at currents of 0.1, 0.2, 0.5, 1.0, 2.0, 4.0, 10, 20, 50, 100, 200, 500 mA, and 1, 2, and 6.1 A.

At very low currents (0.1 and 0.2 mA), the power supply did set the output close to the requested value, but its meter displayed nothing at all. That is still within the stated accuracy, though (those 6 least significant counts).
At higher settings, the current is set much more accurately, but below 10 mA the built-in meter is noticeably off, although again, it falls within the stated accuracy.
Across the rest of the range, accuracy stays roughly within the stated limits, but given the separate DAC/ADC, I expected somewhat better results.


Test three: how well the output voltage holds up under load. This looks good, though the 6006 also did fine here.



Of course, I could have recalibrated the converter, but I don't think the results would have been much better. And if you don't have a sufficiently accurate instrument, it's best not to try. Just in case, here's a video from the manufacturer showing the process.


Test four: the minimum input-to-output voltage difference at which the converter can still regulate.
1, 2. Input voltage was about 35.6 V; at around 6 A, the difference was 2.5 V.
3, 4. Input voltage was 55.6 V; at the same current, the difference was 3.6 V.

In both cases, the difference increases slightly as the unit warms up.

In my opinion, that's not much, especially considering this is a hybrid converter. For it to work properly across its full range, you need a power supply of at least 66 V; the converter's maximum input voltage is specified as 70 V.



By the way, about the linear regulator stage: I was curious how much voltage it drops. It was awkward to reach the transistor, so I measured at one of the power inductor's terminals and before the shunt, which means there's a small measurement error.
Below are the results at output voltages of 2, 5, 30, and 50 V, and currents of 1 and 6 A, respectively.
At an output voltage of 2 V and a current of 6 A, the drop is about 1.5 V; at 50 V, it's about 1.4 V. That's actually pretty good. Here, more is better, but the greater the drop, the more heat the transistor has to dissipate. In my opinion, 1.5 V is a reasonable compromise.



Test five: measuring output voltage ripple.

The test was performed at currents of 0, 1.25, 2.5, 3.75, 5, and 6 A, with an output voltage of 30 V, or half the maximum. The probe was connected directly to the output terminals, with no additional capacitors.

The specified peak-to-peak ripple is 20 mV. In reality, I measured about 4–5 mV of the main ripple and 15–20 mV in the form of spikes, which is clearly better than specified. As I mentioned, the converter was powered by a very ordinary power supply that doesn't have particularly clean output to begin with. Also, both the load and the oscilloscope have their own power supplies, which could have introduced a small measurement error as well.


The current was 6 A throughout, while the output voltage was varied: 3.3, 5.0, 12, 25, 40, and 60 V. The highest peak-to-peak ripple occurred at an output of 12 V, but even with the spikes included, it stayed within the specified 20 mV.


Test six: response to transients.

As before, the tests were performed in several operating modes:
1, 2. Applying/removing 5 V at the output.
3, 4. Response to connecting/disconnecting a load. The output was set to 12 V, with a 30 Ω resistor connected continuously and a 3.3 Ω resistor connected in parallel with it (or disconnected). So this tests the response to a current step from 0.4 to 4 A and from 4 to 0.4 A.
5, 6. Switching from CV mode to CC mode and back. Voltage was 14 V, load current was 4.5 A, and the current limit was set to 3 A.

The results look very good. A very small overshoot appears only when voltage is applied with no load, and it's about 0.2 V. Otherwise, the traces are noticeably better than those of all the models in the “regular” series.


Test seven: using a TS-100 soldering iron as the load. All the previous models showed noticeable fluctuations when the heater switched on and off.
I was expecting better here, in principle, but I didn't think it would be this much better. The converter actually manages to handle all the current surges, so the voltage stays stable.


Test seven: evaluating the converter's efficiency.
Input voltage was around 60–61 V, load current 6 A, and output voltage 5, 12, 24, 36, 48, and 56 V. The converter's own no-load current draw at a 60 V input fluctuates between 20 and 30 mA.

Output voltage — efficiency — power loss
5 V — 60.0% — 20.1 W
12 V — 78.3% — 20.0 W
24 V — 87.7% — 20.27 W
36 V — 91.6% — 19.84 W
48 V — 93.8% — 18.9 W
56 V — 94.6% — 19.1 W

As expected, efficiency is lower. You can feel the losses in the linear regulator; that's the unavoidable price of a cleaner output and a fast response to load changes.


To check the temperatures, I ran the converter for an hour at around 350–360 W, with it lying on the table heatsinks-down, which made cooling worse.
In the end, temperatures were quite reasonable, except for the largest resistor on the board, which heated up to almost 120 degrees. The same resistor also got noticeably hot in the regular 6006, by the way.
1. Overall view
2. Input section; the highest temperature is at the power resistor
3. Output terminals, shunt, and measurement section


During testing, I often connected and disconnected power to the module, and connected and disconnected the load, sometimes with a little sparking, and ran into a couple of "glitches."
The first significant one was caused by frequently switching the input power—for example, when power is applied, immediately removed, and then applied again. It showed up as an output voltage measurement error.
That is, the output voltage was set correctly, but the converter's voltmeter showed something nonsensical. Simply cycling the power fixed it.

The second oddity was that at some point, sparking in the load circuit changed the screen interface from the one you see in the photo to the full display, which shows all the parameters. But I can't confirm that it was 100% spontaneous, although I don't remember changing anything during testing.

Both "glitches" happened once each.



Along with the converter, I ordered a shortened S12D case for this review. I'd already been thinking about buying one, but since I had the opportunity, why not take it?
The thing is, the two case options I already had are convenient for housing power supplies, but awkward to use because their overall depth, including the power cable, is at least 40 cm—which is really a lot.

The smaller version is considerably shorter: it's 110 mm deep on its own, or about 170 mm including the wires, more than twice as short.

I ordered the case from this link. The case itself is inexpensive at $11.40, but shipping is paid and expensive: $19.72 :(

It's wrapped in film, undamaged, with a bag of accessories taped inside.


The case also comes with:
1. Four rubber feet with metal washers inserted, plus four screws.
2. Wires, terminals, and a switch
3. The terminals are reasonably good; I've used ones like these and had no complaints
4. The switch, however, is unbranded, and I do have a complaint. Switching DC above 30 V with this kind of switch is unsafe for the switch. I once had one stick on the very first switch-on at 50–60 V.


There's space at the back of the case for a 50 mm fan, but the fan and its control board aren't included.
There are plenty of air intake holes, some on the bottom and some on the sides.


Assembly is as simple as it gets: screw on the four feet, install the terminals and switch, connect everything, and hook it up to the converter.
The only tricky part is that the wires have fork terminals on the ends, so connecting them to the converter's original terminal block is both awkward and unsafe, since you could catch the resistor next to it.


But I don't like it when a converter is separate from its power supply, and I think this is a dead-end idea, although it does make sense in some situations.
As you can see, there is still some free space, and in theory it could fit a power supply with a board up to 165×80×60 mm, which is quite substantial. Though that's without accounting for the fan and switch.


I have a 300 W Chinese power supply at home; I already wrote about it.


It turned out that this power supply not only fit inside, but there was also room left for a switch and a fan.
And a fan is definitely needed here, since it will have to cool a fairly hot power supply.

One drawback is that the power supply is rated for 24 V. There seems to be a 36 V version, but that's still not enough. I think I'll try modifying it for 72 V. That will mean either rewinding the transformer or reconnecting its secondary winding, and in any case replacing almost everything on the secondary side, including the output inductor, which overheated during operation anyway. The output diodes will probably run cooler, though, and they can also be attached to an aluminum plate, which can then be screwed to the bottom of the case.
I think that after the modification and adding cooling, it should be possible to get more than 300 W out of it.


So there's a very real chance of getting a 360 W adjustable power supply in this case. And if you don't need that much power, you can use a less powerful PSU and set a power limit in the settings, giving you either 61 V or 6.1 A.
I'm already thinking about how to do everything neatly, and yes, I'll also have to cut a hole for the power connector.



For comparison, I put the two case options side by side. It's a shame there's no in-between option: one is really compact, while the other is just a bit too long.


A couple of words about the software. It's been updated to version 1.0.0.9, but I didn't notice any differences from the previous version. I assume they simply added a new converter model to the list of supported devices.


The graph window still doesn't scale properly. I hope they eventually fix that.


So, let me try to sum up everything I've tested above.
Current and voltage setting/measurement accuracy. It falls within the stated error limits, but somehow I expected a little better. Still, it's more than good enough for an adjustable converter, especially one in the hobbyist class.
Ripple and response to load changes. In my opinion, both are very good, and definitely better than on all the models in the previous lineup. Personally, I'd be willing to pay extra for that alone.
Efficiency and voltage headroom. They're certainly worse, but that's to be expected, although I was expecting a poorer result. The linear regulator drops 1.5 V, which is pretty good. The only thing I might have done differently is use a MOSFET in this stage, increasing the adjustment headroom until the DC-DC converter reaches the appropriate operating mode.

A lot has changed in the firmware too, including new display themes. But the most significant change for me is the option to set the charge cutoff current. The developer did get a little carried away here, though, setting the adjustment step to 0.1 mA, when a 10 mA step would have been more than enough.

Would I buy this converter to replace the previous model? Probably yes. The only reason I'd say no is that I'd like a version rated for 10–12 A, just to have some extra headroom.

Now, about the case. I turned it this way and that, and realized that yes, with a bit of work, it's entirely possible to fit a 300–400 W PSU inside. But it will obviously need active cooling, since the converter itself dissipates up to 20 W as heat.
At first, while I was tinkering with it, the small case seemed somehow... stubby, like a chopped-off piece. But it's been sitting on my desk for almost a week now, and I've gotten so used to it that it actually looks convenient. The large case, on the other hand, has started to seem enormous... Unfortunately, the small case isn't well suited to more powerful models, although if you really want to, you can fit a more powerful PSU inside.

That's all for now. I hope this was useful, and as always, I'm happy to answer any questions.