DPH5005, an adjustable voltage converter with a universal input

This is already the third review of RuiDeng products that I’ve posted, but there are actually quite a few reviews of their products. So today I’ll try to add a little to the previous reviews, especially since this converter is interesting in its own right.
As always, I’ll take a look at it, run some tests, cover a little theory, and share my conclusions.


At the end of the review, you’ll find information about discounts at RD official store for 11.11

Since the converter under review is largely similar to another model, DPS5020, which I covered earlier, this review will be a little shorter, with the main focus on testing and the differences from the model I showed before. I’ll also probably mention another similar model, DPS8005, which is also similar, although noticeably smaller.


I ordered not only the converter but also a case for it. For some reason, I thought it would be the second case model, but things worked out so that I ended up with the one I’d already covered. Once again, it arrived packed in a large polystyrene box—full marks for that, as it’s pretty hard to damage.
The converter itself comes in a standard box listing all the models they make and the manufacturer’s contact details. It’s a little odd, though, to see ads for two somewhat outdated models on it: DPS5015 and DPH3205.


The converter comes in the usual expanded polyethylene tray, along with the included manual.


I ordered the full kit: the converter, a Bluetooth module, and a USB module. The store offers three package options:
1. Converter only - $36.94
2. Converter + USB converter - $39.94
3. Converter + USB + Bluetooth - $42.94

Of course, it’s up to you, but I’d go for the second or third option.


The manual is in Chinese and English, though most of it is easy enough to understand intuitively.


For me, the list of technical specifications was the most important part. The connection instructions may also be useful, although everything is clear here too from the labels on the circuit board.

Brief specifications:
Input voltage - 6–50 V
Output voltage - 0–50 V
Output current - 0–5.1 A
Output power - up to 250 W
Maximum input current - no more than 10 A.
Output ripple - up to 100 mV p-p
Cable length to the control module - 200 mm.
Voltage setting resolution - 0.01 V
Current setting resolution - 0.001 A
Voltage setting accuracy - +/- 0.5% + 2 least significant digits.
Current setting accuracy - +/- 0.5% + 3 least significant digits.

You may have noticed the odd specification: “Maximum input current - no more than 10 A.” This means the input current cannot exceed that value. Since the circuit board has a step-up converter, for example, with a 12 V input you can get 2 V at 5 A on the output, but 30 V at 4.5 A is out of reach.

The thing is, this model can both step the output voltage up and step it down relative to the input, which is why the title says it has a universal input.


The control and display module is exactly the same as the one used in the DPS5020.
Controls
1. Three buttons
Up — navigate the menu, change a parameter, or enable voltage adjustment mode. Press and hold to activate the first memory slot.
SET — enter/exit settings mode. Press and hold to display memory slot M0 and restore the settings to their default values.
Down — navigate the menu, change a parameter, or enable current adjustment mode. Press and hold to activate the second memory slot.

2. The screen displays the parameters and the control menu.
3. An encoder for adjusting parameters. The encoder is push-button operated: pressing it changes the adjustment step (on the main screen) or selects a menu item (in settings mode). Press and hold to lock/unlock the controls.
4. Output enable button


Inside, there’s nothing but the display, buttons, and encoder; the controller is on the main circuit board.


The connectors are labeled and plug into the corresponding connectors on the main circuit board.
The encoder knob has extra insulation on the inside. Earlier versions could malfunction when the metal knob was touched.


At first glance, the circuit board looks like the one in the DPS5020, but in fact it has far more differences than you might initially think. There are no particular complaints about the build quality; everything is fairly neat.


The input has a four-pin connector. A transistor is visible next to it, presumably for reverse-polarity protection.


On the other side, you can see a 20 A fuse, the current-sense shunt for the first converter, and its inductor.


Two MBRF20100CT diode assemblies are mounted on the heatsink, one for the first converter and the other for the second.
Even at this stage, you can spot something a little odd, which I'll explain later. I'm curious whether anyone will notice anything else unusual, and if so, at what point during the inspection :)


At the output there's a second inductor, a little smaller than the first, along with a fuse and a four-pin pluggable terminal block.


And of course there's a second shunt, since the board has two converters. This is the main shunt, used to measure the converter's output current.


There are no fewer than six main capacitors. I'll try to explain what each pair does:
1. The first pair (from left to right) is at the board input: 2 × 330 µF
2. The second pair is between the first and second converters: 2 × 470 µF
3. The third pair is at the output of the second converter, i.e. the board output: 2 × 330 µF.

All the capacitors are rated for 63 V.


Auxiliary voltage converters, regulators, PWM controllers, and a microcontroller.


It also turned out that there are more components under the heatsink. To get to them, I had to do more than just unscrew it—I also had to peel it off.


The heatsink was glued to a pair of transistors with something that looked like silicone sealant, but was gray. Most likely, it was a type of thermal paste.


The heatsink and fan are exactly the same as those on the DPS5020. It runs fairly quietly and has temperature control, so it only switches on under a decent load.


Since I won't be able to attach the heatsink to this silicone again, I'm scraping it off.


On the left is the first converter's transistor, AOD2810, an N-channel MOSFET rated for 80 V and 46 A, with an on-resistance of 8.5 mΩ. There's room on the board for a second transistor in parallel with the first, so there may be another version of the converter, or you could add one yourself to increase the maximum input current and thereby lower the input-voltage range needed to get full output power.
On the right is the second converter's transistor, AOD409, a P-channel MOSFET rated for 60 V and 26 A, with an on-resistance of 40 mΩ. In my opinion, this transistor is right on the edge of its ratings, and it would have made much more sense to use one with a higher voltage rating here, say 80 V.


Now for the answer to the question about paying attention. If you didn't notice in the previous photos, here's the explanation: the board has diode assemblies in fully insulated packages, and they're mounted to the heatsink with insulating pads. The heatsink itself is completely isolated from everything.
That's all well and good, of course, but it reduces heat transfer. Since I removed the adhesive, I ended up mounting the diode assemblies with thermal paste, and applied some to the converters' switching transistors as well. I recommend doing the same if you've bought one of these converters.


A fairly large part of the board is taken up by the auxiliary converters and regulators. They power everything: the microcontroller, op-amps, display, and other PWM controllers.


The other components on the circuit board:
1. Microcontroller, the same as in the DPS5020
2. TL594, the PWM controller that controls the second converter.
3. Dual rail-to-rail SGM8582 op-amp, presumably used to measure current and voltage.
4. Looks like a small op-amp. Since it's installed next to the current-sense shunt, it's probably there to amplify the shunt signal.
5, 6. Input and output fuses. There's also an LED near the output fuse, connected in parallel with it. It's a little strange that the board has fuses F1 and F3, but no F2.


The underside is clean and tidy.


Among other things, the board also has an extra little board with another chip on it. I don't know where the idea came from that this board powers the controller or improves and stabilizes its operation, but I can say for certain that it doesn't.


This chip is nothing more than a standard PWM controller — LTC1871EMS


It operates at frequencies from 50 kHz to 1 MHz and can be used to build both step-up and universal SEPIC converters. In this case, though, it isn't working as a SEPIC converter, as you might expect, but as a standard step-up converter.
What's more, this chip can measure the current through the power switch without a shunt. I was about to criticize RuiDeng for not using this feature, but then I realized why they didn't: they needed to monitor the current with the microcontroller.


So the first (step-up) converter is controlled by the LTC1871EMS, while the second one, just like in the DPS5020, uses a TL594. In effect, you could say this is a DPS5020, but rated for 5 A instead of 20 A, with a step-up converter on the input.

A simplified schematic of the power stage looks something like this.
You might ask: why not just regulate the step-up converter?
You can, but there's one problem: in this configuration, the output voltage can never be lower than the input voltage minus the voltage drop across diode D1. Of course, you could replace it with a transistor and make a synchronous rectifier, but that brings its own complications.
You could also use a SEPIC converter, but that has its own issues with the coupling capacitor, which is prone to failure.


So the 6–50 V input voltage first goes to the first converter (T1, L1, D1), where it's boosted to 54.5 V. Then there's a regulated step-down converter (T2, L2, D2) with an output voltage of 0–50 V.
The efficiency tradeoff is offset by the simplicity of this design—or, more precisely, by the fact that it's essentially just two independent converters.
By the way, although the manufacturer specifies an input voltage starting at 6 V, I wouldn't recommend using less than 10–12 V. At that input voltage, you'll get about 100 W of output power. To get the rated 250 W, you'll need more than 26–27 V.

Voltage across the intermediate filter capacitors.


Connecting the board is very simple: two ribbon cables to the display and control module, and the USB or Bluetooth module plugs into its own connector.


Speaking of USB and Bluetooth, they're exactly the same as on the DPS8005, DPS5020, and similar models. The USB module has galvanic isolation, while the Bluetooth module doesn't need it for obvious reasons.


Just in case, here's a photo showing how to connect it.


The input and output use four-pin pluggable terminal blocks. In my opinion, they could have used better ones: these have a simplified clamping mechanism rather than the more reliable elevator-style type.



I planned to power this converter with a 24 V, 300 W power supply, which I expected to have plenty of headroom.


The only problem was that to get the full 250 W, the input voltage needs to be at least 26–27 V, preferably higher.
To raise the voltage a little, I found the feedback circuit on the board, which needs a slight adjustment.


The section on the left is the feedback circuit; the component designators and values are as shown in the original schematic.
To increase the output voltage, you can either increase the value of R13 or decrease the values of R14 and/or R15. One option is to replace R15 with a 470 Ω trimmer potentiometer.
The feedback circuit's job is to maintain 2.5 V at the indicated point, so the divider must be set up to provide that voltage at the desired output voltage.

It's easier to decrease a resistor's value than to increase it: you can simply solder another resistor in parallel with the one already installed, which will raise the output voltage. In the schematic, this resistor is marked Rx; you can connect it in parallel not with the entire chain shown, but with either R14 or R15.


I estimated how much the voltage would change if R15 were shorted completely. It didn't look like it would be too much, so I simply shorted it with a piece of wire. In the end, I got a little more than I needed, about 27.3 V.
I could have left it like that, but the greater the difference between the input voltage and the adjusted voltage, the harder the power supply has to work. So instead of a jumper, I soldered in a 24 Ω resistor, which gave me the 27 V I wanted.


Connected it, powered it up, and it works.


Firmware version 1.6, just like the DPS5020 from my previous review. Basically, everything here is exactly the same as on the DPS5020; the only difference is the maximum output current: 5.1 A here, rather than 20 A on the DPS5020.


For a full description of the controls, see my review of the DPS8005 board. Here, I’ll include an excerpt with an image.


Since it has the same firmware version as the DPS8005 and DPS5020, the same issues apply.




And of course, the tests.
First, a check of the input voltage measurement accuracy.
1, 2. The board starts up at around 4 V, but almost nothing is visible on the display.
3. At 5 V it works fairly reliably, but the input voltage needs to be at least 6 V for normal operation.
4–8. Everything is basically fine from here on. The readings are slightly high: at 50 V, the voltmeter reads 0.06 V high. I don't think that's critical, since this parameter isn't specified for the device.


The output voltage setting accuracy is much more critical, though.
Here I increased the converter's output voltage from 0.5 V up to 50 V. The output voltage is slightly low across the entire range. It's the voltage set by the converter that's low; the built-in voltmeter measures it quite accurately.
The photos show output settings of 0.5, 1.0, 2.0, 3.3, 5.0, 7.5, 10.0, 12.0, 15, 20, 30, 35, 40, 45, and 50 V.


The current readings showed the opposite: the board ran slightly high. Again, it was mostly the output current itself that was high, not the reading on the built-in ammeter, which was quite accurate.
The photos show current settings of 1, 10, 50, 100, and 500 mA, then 1, 2, 3, 4, 5, and 5.1 A. At low currents (up to 50 mA), the output current is significantly higher than the set value.


The fourth test checks how accurately the output voltage holds under load. The left side shows the voltage with no load, and the right side under a 5 A load, which is almost the maximum.
I was surprised that the board raises the voltage slightly under load at low voltage settings, but lowers it at higher settings.


Also, for some reason, with the output set to around 12 V under load, the voltage fluctuated slightly between 11.88 and 12.02 V. I'm not sure what caused this behavior.


The fifth test measures the output ripple. I connected directly to the output terminals and powered the board from a 24 V, 300 W power supply with its output voltage raised to 27 V.


Input voltage: 27 V; output voltage: 25 V; load current: 0, 1, 2, 3, 4, and 5 A.
At a load current of 5 A, the ripple reached the specified maximum of 100 mV.


And now for a slightly different test: the input voltage is still 27 V, but the load current is 5 A throughout, while the output voltage changes:
3.3, 5.0, 12.0, and 50 V. The results are noticeably better here. It looks like in the previous test I happened to hit the worst-case condition, with the output voltage at half its maximum.
The maximum ripple in this test was about 75 mV at 12 V.


A rough efficiency estimate at different output voltages. The input voltage is shown on the converter display on the V-IN line, and the current is shown on the multimeter. All the photos in the review can be enlarged by clicking on them. With no load, the board draws very little current: just 22 mA at 27 V.
So, with an input of around 27 V and a load current of around 5 A, the efficiency came out to:
2. 5 V — 78%
3. 12 V — 88%
4. 24 V — 92.5%
5. 35 V — 94%
6. 50 V — 94.7%

The efficiency was pretty much as expected. It dropped noticeably at low voltages, but was already at a perfectly respectable level for this kind of device at 12–24 V, especially considering the double conversion.


The sixth test checks how hot the converter gets.
The board first ran for about 20 minutes at 50% of maximum output power, then at 100%—or rather, 98%, since the current was 5 A instead of the maximum 5.1 A. My load operates in CC mode, so I can't raise the current any higher. The input was still 27 V, the minimum needed to deliver 250 W at the output.


Thermal images taken after 20 minutes of testing at 50%, followed by almost 20 minutes at 100% current. In my opinion, the results are excellent. The diode assemblies got the hottest, but even they reached only 56°C.


This is a thermal image of the power supply after 40 minutes of testing, with a small, quiet 60 mm fan placed near the output inductor, blowing air toward the power transformer. You can see that the temperature dropped significantly, reaching a maximum of 62°C. The thermal camera picked up that 62°C reading partly from the output resistor.
The output resistor itself, which wasn't cooled, reached 120°C. I recommend replacing it with a 1 kΩ resistor; the original is 560 Ω.



Now we can move on to installing the board in the case. The process is described in detail in this review, so I’ll keep it brief here.
Originally, the idea was to get the DPS5005 model with the same output specs but in a more compact version, build a compact 250 W power supply, and fit the whole assembly into its original case. For some reason, though, I wasn’t given one, so we’ll be installing the DPH5005, which is noticeably larger and takes up almost all the space inside the case.

Link to the case, price $23.37 (option 2).


The case is specifically designed for DPH and DPS series converters. There’s a second version of the case too; I wanted to try it, but ended up getting the same one as last time :)


The kit includes everything needed for assembly: a fan, power board, wires, screws, nuts, terminals, a switch, and so on.


First, install the terminals, fan, and USB converter board, then stick on the feet. The feet are usually attached at the very end, but it’s better to do it now, since the screw heads protrude from the bottom of the case and could scratch the table.


Next, solder the wires. As last time, I have three comments:
1. The kit didn’t include a connector for the fan, so I simply soldered in two component leads and used them as connector pins. It’s a pretty crude workaround, of course, but I didn’t have any connectors like that left at home.
2. The holes in the circuit board for the wires to the switch are small in diameter, so 0.75 mm wire is about the largest you can use—maybe 1 mm if you really try.
3. There wasn’t much wire in the kit, just barely enough. I ended up with about 5 cm of red wire left, and that was it. I used my own wire for the switch. You can use the supplied wire, of course, but it’ll be stretched very tight.

The supplied wire has a cross-sectional area of 2.5 mm² and is very good quality.


Install the board and connect the power wires. As last time, I recommend removing the stock heatsink and fan, fitting a larger heatsink, and connecting the fan to the board’s original socket. It’ll be noticeably quieter.


Install the control and display module, stuff its wires wherever they’ll fit route them neatly, and connect the USB converter while you’re at it.


You can also connect the Bluetooth adapter, but the metal case will shield it for the most part, significantly reducing its range.


Close it up with the cover, and that’s it. This time the assembly went more smoothly, by the way: the screw holes lined up without any trouble, and the gap between the cover and front panel was smaller too.

Connect the power supply and test it.


There’s not much point in describing the software in detail: it’s convenient and intuitive. The main thing is to test it without a load at first, so you don’t accidentally burn it out.
You can download the software, the USB converter driver, the manual, and the protocol description at this link.




The Android software worked without any problems either, though naturally it connects via Bluetooth.
Install the APK, then use your phone’s Bluetooth settings to connect to the RuiDeng adapter that appears. The password is 1234.
Then launch the software, select the RuiDeng adapter from the list, and that’s it. In my opinion, though, the Windows software is more responsive.


Another option is to use the Excellway® EF01 case, described in detail in this review.
With this option, you can easily fit both the power supply and the converter board inside. Bluetooth will also work better, and there’ll still be a little room left for a separate converter with a fixed output voltage.


A few words about the DHP5005’s features.
Its power isn’t particularly high, of course, but it covers most of the needs of radio amateurs. Its main advantage is that you can use different power supplies—for example, 36–48 V, or a 24 V supply modified to 27 V.
By the way, regarding the power supply: the manufacturer recommends feeding 28 V into the input to get the full output power. My tests showed that the board can already deliver up to 250 W at 26 V, although the input current is a little over 10 A. At 27 V, everything works beautifully. I could have raised the voltage to 28 V, of course, but I didn't want to push the power supply too hard. A 10% variation is generally considered acceptable, which gives us 26.4 V; 15% (27.6 V) is probably already the limit. Branded power supplies can usually handle even more, but with an unbranded one, you risk burning it out.


What can I say about the converter itself?
The calibration is very good. I have a few minor concerns about voltage stability under load at an output voltage of 12 V, but I plan to look into that separately; there may be something unusual about my test setup.
In the worst case, ripple was the stated 100 mV p-p; in other operating modes, it was about one and a half to two times lower.
The board barely gets warm. That's why, when installing it in an enclosure, I recommend not using the board's fan and relying on an enclosure fan instead—though you should use a larger heatsink, of course.
By the way, about the heatsink: in the review, I mentioned that it's better to remove it, scrape off the sealant, and replace it with thermal paste. Given how little heat it produces, you don't really have to do this, but I still think paste is better.

Otherwise, it's an interesting converter. But if you have a 48 V power supply, I'd be more inclined to recommend adjusting it to 54–55 V and using it with DPS-series boards. The model reviewed here is better suited to those who are fine with 250 W of power and only have a 24–36 V power supply on hand—or one below 24 V, though in that case the output power will be lower.

There's one odd thing about the 10 A input current limit. I think it would have been quite possible to add another transistor in parallel with the boost converter's switching transistor, especially since there's room on the board, and get, say, 15 A instead of 10 A. Then at 12–14 V (for example, from a battery), you'd get 150–180 W instead of 100–120 W, which would be much more useful. But the manufacturer made a different choice, unfortunately.

That's all from me. I hope the review was useful, and as always, I'm happy to answer your questions.