Adjustable DC-DC Converter DPS8005 + USB + Bluetooth

I've been asked about reviews of converters like these many times, both in the comments and in private messages. I've seen plenty of reviews of similar devices, but somehow never got around to trying one myself. Partly because I already have quite a few different adjustable power supplies, and partly because I have a general idea of what these converters are and how they work. But as it happened, I needed exactly this kind of converter for a project, so I decided to order one for a review.
This won't be a very long review. It'll mostly be about my personal opinion of the converter itself, its design and controls, and my overall impressions compared with models from ZXY.


First, a little about why I needed this particular converter in the first place.
I've mentioned before that a friend of mine works with various batteries, including repairing battery packs. Also, in my review of the electronic load kit, I wrote that I was planning to build a test and balancing station for large battery packs. But an electronic load can only discharge batteries, and they also need to be charged somehow. That's what this converter is for.

Now, why did I choose this particular model?
Current: 1–2 A would have been enough for me, and as far as I know, the 5 A model is the smallest in this product line, so there wasn't much choice.
The voltage is a different story. Most DPS models are rated for 50 V, while ZXY models go a bit higher, up to 62 V. I did come across a description of a 120 V version, which would have been just right for me, but it's no longer made (and was probably never mass-produced).
Reviews of the ZXY6005S and two reviews of the ZXY6020S: one, two.

While we're at it, I'll answer one question: why are such high-voltage converters needed? I'll skip over the fact that there are models rated for 400 V and higher, but those aren't available as kits, so they don't suit me. I need to build a device for testing battery packs from 1S to... at least 15S, preferably more. But even 15S is already 63 V, and I've already found that even the ZXY6005 doesn't quite cut it.
This photo has already appeared in one of my reviews. It shows a 2x15S2P battery pack that I needed to balance.


All in all, the best candidate turned out to be the DPS8005, which has an output voltage range up to 80 V and can supply up to 5.1 A. That makes it possible to charge batteries up to 19S, or 79.8 V. Of course, 20S would have looked nicer, but my friend and I decided this would be enough.
I should point out right away that this review won't feature a finished product, and neither will the next one, since quite a few separate components are still needed. I'll tell you about them as I go, but each review will bring us closer to building what we originally had in mind.

I've often seen kits like this arrive in foam boxes, but this one came in much simpler packaging: a plastic case and a tray made of expanded polyethylene.


So, I ordered the kit in the "full configuration", which includes:
1. DPS8005 converter
2. USB-RS232 TTL module with cable.
3. Bluetooth module.
4. Instructions.

The store offered a version without any extras, as well as one with a converter + USB, but I didn't know what I would need, so I ordered everything—especially since I had no spending limit.


The instructions are fairly clear. You only need a basic knowledge of English to get the gist; after that, it's easier to figure things out by experimenting.
The ZXY instructions were only available as a PDF, and I had to track them down, so that's a point in DPS's favor. On the other hand, the ZXY documentation included information about the control protocol, so you could write your own software.


Each add-on board comes in its own bag.


The add-on boards are used to connect the converter to a computer. If you don't need that functionality, you can skip them, but I'd recommend buying at least the USB board just in case—it may come in handy.


Basically, the interface boards are just RS232-to-USB or Bluetooth converters. The USB board provides galvanic isolation using optocouplers.
In my power supplies based on ZXY boards, I used a dedicated ADuM chip to isolate the interface. But since the data rate is very low, optocouplers are perfectly adequate.
You can also use any RS232-to-USB converter chip, or a Bluetooth module if you have one on hand, since the converter's connector has the standard four pins: power, ground, RX, and TX.



If you want to use both interfaces at once, I once wired them up like this: the first connector (2-3-4-5) was used to connect to the ZXY module, the second (1-2-3-4) to a Bluetooth adapter, and the ADuM provided galvanic isolation for RS232-to-USB.
The circuit is from my ZXY6020 review, but the principle is the same here. Keep in mind that devices like these must not be connected without galvanic isolation.


The converter module itself is very compact. The cutout dimensions are 76x40 mm, and the external dimensions are 79x43x42 mm (excluding the encoder knob).


The front panel has:
1. Three buttons
Up — navigate the menu, change a parameter, and 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 defaults.
Down — navigate the menu, change a parameter, and enable current adjustment mode. Press and hold to activate the second memory slot.

2. A display showing the parameters and the control menu.
3. An encoder for adjusting parameters. It's a push-button encoder: pressing it changes the adjustment step size (on the main screen) or activates a menu item (in settings mode). Press and hold to lock/unlock the controls.
4. An output enable button

I didn't like the control logic, but more on that later.


Basically, the connections are just two wires for the input and two for the output. But keep in mind that input ground and output ground are not the same, since a current-sense shunt is connected between them on the board. In other words, you can't connect the converter using just three wires. Well, you can, but it won't work properly and may fail.
The terminal block is the right type: removable and convenient, but impractical. The thing is, you usually connect a device just once, during installation, while the extra contacts are there to cause trouble all the time.


The converter is very densely packed. The power section is on the rear board, along with the capacitors, power transistor, output diode, inductor, and shunt.


1. The input capacitors, from a manufacturer I don't recognize, are rated at 330 µF and 100 V. They're connected in parallel, giving a total capacitance of 660 µF. There's almost no voltage headroom, since the board is designed for an input voltage of up to 90 V.
2. The connector for the computer interface is located between the input capacitors and the case. It's not in the most convenient spot, but since you only connect it once during installation, it's fine.
3. The power semiconductors are mounted on a heatsink:
The HY18P10 transistor is rated at 80 A, 100 V, and 28 mΩ. As you can tell from the part number, it's a P-channel transistor, as on the ZXY boards. That's not the best choice, since P-channel transistors usually have worse characteristics than N-channel ones, but they're easier to drive because they don't need a dedicated high-side driver.
The VF40100C diode assembly is rated at 100 V and 40 A (20+20 A). No complaints here.
4. In my opinion, the heatsink is a bit small compared with the one on the ZXY6005. It's also worth noting that the ZXY6005 uses an IRF9540N, rated for only up to 23 A and with an on-resistance of 117 mΩ, compared with 28 mΩ for the unit under review. The ZXY also uses an ordinary fast-recovery diode, while this one has a Schottky diode.
5. The inductor, on the other hand, is very compact—about half the size of the one in the ZXY6005, which is rated at 300 W versus 400 W for the unit under review. Inductor size is directly related to output power, but also depends on frequency. The inductor is wound with fairly thick wire, so no concerns there. I'll check its temperature during testing.
6. Since this device is a switching voltage converter, it also has a fairly large output capacitor—in this case, 150 µF, 100 V. An output capacitor is not ideal, of course, since it can cause a current surge when a load is connected and the power supply switches to CC mode, especially when there's a large difference between the set voltage and the actual voltage at the load.
Unfortunately, this capacitor is necessary here, and designers usually face a trade-off: a larger capacitance improves the converter's performance, while a smaller one reduces the current surge when switching to CC mode. In my view, 150 µF is a reasonable value. The ZXY6005 has two 1000 µF capacitors. The main thing is that this doesn't affect the converter's stability at maximum load current.


I don't think I'm too far off if I assume the circuit is similar to the ZXY6005 series: the microcontroller sets the reference voltage, which is then compared with the output voltage and current using op-amps. The op-amp output signal is fed to a standard PWM controller, which in turn drives the power transistor through a basic high-side driver.
ZXY6006 circuit


But the ZXY is more solidly built.



Bend back the four clips one by one and lift the boards at the same time to remove the whole assembly from the case. You'll also need to remove the encoder knob first.


I'd like to say that the device is easy to take apart, but although the module comes out of the case easily, that doesn't really help, since the rest is effectively non-serviceable.
What we end up with is a module containing the display, encoder, and power stage.


This is the device's biggest drawback: it's practically impossible to repair. The boards are soldered together in a sandwich, and taking the whole assembly apart is almost impossible to do without damaging it. The boards are connected by two 8-pin headers.
Although I'm familiar with the circuitry of converters like this, I really wanted to trace at least part of the power stage to answer two questions people have asked me:
1. How to tap the feedback signal to enable a three-wire connection
2. How to route the CC/CV signal to an external LED.

Unfortunately, the only option here is to break it apart. Even a simple replacement of the power transistor or capacitor could turn into a major undertaking. Overall, I'd give the design two out of ten.
But for those who asked me the questions above, don't be discouraged just yet: I have another 5020 module here. It's built on two boards using the same circuit design, and I'll try to investigate it more thoroughly.


The board also has a microcontroller and other components, including a TPH8R8 MOSFET. I don't know what it does, but it's clearly a power component, and there are reports of it failing in higher-power versions. Unfortunately, I don't know which circuit it's in.


1. The converter is powered from the input voltage, so the board has a small DC-DC converter marked 7472, though I can't identify the exact model.
2. The board also has an additional SS5200 protection diode.


Unfortunately, there's nothing more to examine, so I'll move on to a few tests and explain how to operate this converter.
I haven't received the power supplies intended for this device yet, so I connected it to the ZXY6005 I'm used to. It's served me faithfully for about 4 years, and its specifications are very similar to those of the DPS8005 under review. I'll be comparing the two from time to time, especially in terms of operation.


A brief translation of the control and indicator labels.



Now for the controls and features.
1. When powered on, the splash screen and firmware version appear—in this case, V1.5.
2. The display then switches to the main screen. I can say that a lot of information is displayed on a small screen, so people with farsightedness may have trouble reading the small digits.
Also, it was only while writing this review that I noticed something I miss here, but which the ZXY has: an amp-hour counter. Yes, despite its simple display, the ZXY also shows the capacity in Ah that has passed through it. A small thing, but an annoying omission.
3, 4. At first glance, the control menu has a lot of items, but they’re actually very straightforward:
U-SET - sets the voltage that will be applied at startup
I-SET - sets the current that will be applied when switched on.
S-OVP - maximum output voltage, above which the output will be shut off automatically
S-OCP - maximum output current, above which the output will be shut off
S-OPP - maximum output power, above which the output will be shut off
B-LED - backlight brightness
M-PRE - save settings to the corresponding memory slot. All the settings listed above are saved.
S-INT - automatically turn on the output when power is applied
5, 6. Earlier versions didn’t have an auto-start option. Now they do, and it works without any problems; you just need to set the mode to on.
7, 8. Brightness can be adjusted across six levels, from 0 to 5. There’s no automatic dimming, but you can assign a different level to each memory slot.
9. Select a memory slot for saving settings. I honestly tried to figure this out, but the mode is so unintuitive that, in my opinion, it gets in the way more than it helps. That said, this is just my subjective opinion.

A few words about the controls, display, and features.
My ZXY-based power supply has almost all these functions, but they’re much more conveniently implemented than on the DPS.
First, navigating the menu: to adjust a parameter, press SET, use the buttons to select the parameter you want, press the encoder, turn it to select the value, then press SET to return to the main menu.
This part - press SET, use the buttons to select the parameter you want, press the encoder - annoyed me the most. I kept instinctively pressing SET again to enter parameter-setting mode, and of course that kicked me back to the main screen. Nonsense.
Now, about the display. The ZXY shows just 5 parameters plus one extra (time since startup or reset):
Voltage
Current
Power
Capacity
CCCV

This one shows 11 -
Set voltage
Set current
Output voltage
Output current
Output power
Input voltage
Lock
Fault
CCCV
Memory slot
Output status

So how is it that the ZXY displays fewer parameters and still measures capacity?
On the ZXY, the set current/voltage is displayed while you’re adjusting it. If you stop turning the encoder, the display switches to the measured value after a second. And if you’re changing the voltage, the current indicator continues to show the measured value. Convenient? Very.
The power display is the same on both converters.
Input voltage display - why is it even here? You can’t set the output voltage higher than the input voltage minus 1 V anyway; the software protection simply won’t let you.
The lock indicator is probably useful, I won’t argue with that. Same goes for showing the memory slot.
Both converters have a CCCV indicator, but on the ZXY it’s duplicated by an LED, which is much more convenient than a label on the screen. The ZXY doesn’t show output status on the screen, but when the output is off, the display uses a completely different layout. It also has a separate third LED for the output.
The ZXY has no indicator for overcurrent/overvoltage/overpower protection trips, nor does it actually have those protections.

In the end, I can say that with its 2-line, 16-character screen, the ZXY is more convenient to use than the unit under review with its trendy TFT screen. I know that sounds strange, but it’s true. And I forgot to mention above that the DPS screen updates twice per second, while the ZXY updates at least 10 times per second. Sometimes the high refresh rate is annoying, but it makes operation much more responsive. For example, I can turn the encoder quickly and see the steps 4-5-6-7-8-9 on the ZXY, while the DPS jumps through 3-7-12-15-19-24.


1. To set the voltage, press the "up" button, then turn the encoder to change the value and press it to change the increment.
2. The same goes for current, but press the "down" button. In both cases, the increment cycles from smaller to larger. The ZXY does it the other way around, but that’s probably just a matter of habit.
3. As I mentioned above, the input voltage is displayed, but there’s no indication that it has dropped below the set input voltage. So why display it? For example, the ZXY doesn’t show this voltage on screen, but it does indicate when the power supply has stopped regulating.
4. Pressing and holding the up or down button activates the first or second memory slot.
5. Pressing and holding the SET button activates memory slot zero.
6. In the last mode, turning the encoder changes the slot number from M0 to M9.


Above, I mentioned that the settings include the S-OVP, S-OCP, and S-OPP options. These parameters set the maximum output voltage, current, and power, respectively.


1, 2. In the example above, I set the power limit to 2 W. At 2.03 W, the output is still active, but if I increase it further, an OPP warning appears and the output shuts off.
3, 4. The same applies to current and voltage.

It all looks nice and interesting, except for one thing—or actually, in this case, two things.
1. The fact that the converter has not only CCCV mode but also a trigger-based shutdown function is genuinely useful. But why does the output shut off if, for example, I set the limit to 3 A and then use the encoder to set the output current above 3 A, even with no load—in other words, with no current flowing at all? What’s the logic? I can set the output to 5 A, but the protection should only kick in once the preset 3 A limit is reached.
2. But that’s a minor issue. The real problem came when I got to the power limit function. I once asked whether DPS units had this function and was told they did, but it works quite differently from the way it does on the ZXY.
On the DPS, the output shuts off when the set value is exceeded, acting as a protection feature.
On the ZXY, the controller adjusts the current and voltage so that the output power never exceeds the set value, without shutting off the output. For example, with a 50 W limit, 10 V, and 5 A, as soon as I try to set the voltage to 20 V, the output current is automatically limited to 2.5 A.

This function is needed to protect the power supply. Here’s a specific example: in this case, I need up to 80 V at up to 2 A, or 160 W. I’m planning to buy a pair of 48 V, 100 W power supplies and connect them in series, giving me 96 V at 200 W. With the ZXY, I can simply set OPP to 200 W, and the converter will never deliver more than 200 W at the output, while all its functions continue to work normally. Here, though, I’d constantly have to press the On button because the converter would keep shutting off the output.




A few tests.

The converter has three measurement circuits: two main ones and one auxiliary. The main circuits measure output current and voltage; the auxiliary one measures input voltage.
I started by checking the auxiliary measurement, and in my opinion its accuracy is excellent—five stars. By the way, the converter starts up at an input voltage of 9 V.


For output voltage setting and measurement accuracy, I set the voltage to 0.1, 1, 5, 10, 15, 20... 55, and 61.38 V. I can say the accuracy is quite good: excellent from 0 to 40 V, and not quite as good in the upper half of the range, but still decent. The bigger issue is the setting accuracy, not the measurement accuracy, which is fine.


Current setting and measurement accuracy at 2, 10, 50, 100, 200, and 500 mA, and 1, 2, 3, 4, 5, and 5.1 A
Overall, the results are much the same as for voltage. Of course, this isn’t a measuring instrument, but it performs quite respectably.


Next, I tested it under load. Since it’s not convenient for me to provide the required 90 V at the input right now, I decided to use a 60 V power supply as the source. The remaining conclusions will be based on these readings.


To evaluate efficiency, I tested the converter at 5 A, with the output voltage ranging from 5 to 50 V—that is, from 25 to 250 W—and I’m pleased with the results. At an output power of 250 W, losses in the converter were about 6.5–7 W, giving an efficiency of around 97%, which is very good.
It’s also worth noting that the converter was operating at almost its maximum output current, with an input-to-output voltage difference of just 8 V at the minimum. In practice, increasing the output power will only increase losses in the power inductor, since the converter’s transistor and diode were operating almost the same way they would with 90 V input and 80 V output.


But I wouldn’t say the temperatures were particularly low. After about 20–25 minutes of operation at 200–250 W, I got these readings:
Transistor — 92 °C
Diode assembly — 88 °C
Inductor — 65 °C.

I was most concerned about the inductor temperature, but it turned out there’s still a pretty decent margin. So I think the converter can definitely handle 350 W. I’d still recommend using at least some active cooling, though, since I tested it as a bare board on the bench, not inside an enclosure.

But there were some problems too. When I turned the board over and measured the component temperatures, the thermal camera showed that one of the transistors was over 120 °C.


This is the transistor. As far as I can tell from the board layout, it’s connected to the gate of the power transistor. Given that the transistor was operating at its maximum, perhaps this was intentional, but personally I don't like it, and I'd recommend screwing at least a small plate onto it, especially since the transistor package has a hole.

At first I suspected the shunt, which would have made more sense to me, but it turned out to be the transistor after all.


Now for the “weak spot” of many converters: ripple measurement.
First, here's the waveform as-is, with the probe connected directly to the converter output, without any filters or anything like that. The tests were run at load currents of 2.5 and 5 A, with output power of 125 and 250 W, respectively.

I'd say that's pretty good—very good, actually.


People have asked whether the performance can be improved further. The question was mostly about more powerful models, such as the 5020, which I also have on hand right now, but I tested this one too.
For this, I grabbed the first inductor I could find on a motherboard—fortunately, they have a winding with a decent cross-section—and three capacitors:
0.68 µF
2.2 µF
68 µF

The inductor's inductance is just 2.7 µH, and its DC resistance is 2.4 mΩ.


And now the results.
1, 2: 0.68 µF capacitor; the load current here and in all the other cases is 2.5 and 5 A
3, 4: 0.68 µF + 2.2 µF
5, 6: 0.68 µF + 68 µF.

In my opinion, the 0.68 µF + 2.2 µF option is quite sufficient for this converter, especially since increasing the output capacitance is highly undesirable.


But don't get too excited: the power supply plays a huge role, since in some situations you can get a result like the one shown here. The presence and capacitance of the interwinding capacitor have a particularly strong effect—I often mention it in my reviews—as does the overall build quality of the power supply.
Scary, huh? :)


That's it for the tests for now. I'll say a little about the software.
Installation is standard: run the file, click Next, Next, Next, OK.


At the end, it asks you to reboot, but I ignored it. Besides, my computer was busy and had no time for my problems with various unnecessary programs.


Just out of curiosity, I first tried connecting via Bluetooth.


I selected the appropriate COM port in the program, but alas, “something went wrong.”


But when I connected via USB, it worked right away, especially since the USB-COM driver installed on my system almost immediately after I installed Windows.
I can give this program a solid 5. Of course, I'd have liked to give it a 5+, but 5 is a fair rating. The program works reliably and correctly, and what's more, it's very easy to use.


What's more, besides its main functions of adjusting settings and plotting graphs, the program lets you run various adjustment sequences—for example, gradually increasing the current or voltage to a set value in adjustable increments.


Overall, ZXY got thoroughly beaten on the software front, since its software is just plain awful.


But this converter also has an Android app, so I decided to test that too.
First, I simply connected the converter to my phone via Bluetooth, installed the app, and launched it.


Then I spent a long time trying to figure out what the app wanted from me. It turned out it was complaining that the input voltage was lower than the value I wanted to set. All right, I raised the voltage, and it let me set the output 1 V below the input, which is fine.


Then things got weird. I'll skip over the strange current graph with a resistive load..., but the zero output voltage in the first screenshot, with output power at nearly 180 W, did make me a little uneasy.
Then I played around a bit more, looked at the pretty graphs, turned the virtual adjustment knob, and tried changing the converter's screen brightness from the app. Well, it looks nice and is even convenient in some ways.


And then I saw a genuinely interesting glitch. I don't even know what was glitching. Anyway, it started when, during another attempt to adjust the output voltage, a 500 V spike appeared on the graph. Of course, I understand that's technically impossible, especially with a resistive load. But here's a screenshot, here's the graph—see for yourself.

It gets even better: it turns out the software can freeze. It keeps regulating the voltage/current, adjusting the screen brightness, and even drawing the graph, but the lower-left window of the program—the one showing the input and output parameters—gets stuck.
The screenshot clearly shows that the output is enabled, the input is 62 V, and the output is 17 V, even though I set it to 7 V, not 17...

Then the program “unfreezes” and updates the readings, but not for long.


Even when it freezes, it still controls the converter correctly. The photo shows the set voltage, the voltage displayed by the converter, and the reading shown in the software.


To avoid ending the review on a downbeat note, here's a photo of what else I tested the converter with, and what it will have to work with going forward. More precisely, one of the things, since the battery packs shown are very basic—just 10S2P.


The video review will be coming a little later, but probably not for at least a week.


Now for the conclusions, starting with a comparison to the ZXY6005.
As far as I can tell, the circuitry in both devices is roughly the same: the same DC-DC step-down converter and the same P-channel MOSFET. In this case, though, it has been improved somewhat by using components with better specifications, resulting in higher efficiency.
When it comes to controls and ergonomics, the ZXY is clearly better. For me, it's still the benchmark among Chinese DC-DC converters when it comes to simplicity and ease of use. For example, the ZXY has just 5 buttons and an encoder, but in practice you only use 3 buttons and the encoder, and they're easy to use—unlike the unit under review, with its 4 buttons and push-button encoder.
The same goes for the screen. It certainly looks nice and displays a lot of information, but in practice it's overloaded, with a lot of unnecessary data. The meters also update at just 2 readings per second, compared with about 10 on the ZXY, and there's no amp-hour counter.
And, of course, the power limiting is implemented somewhat poorly, although the developers deserve credit for the voltage and current trigger protection feature.
In short, the ZXY is simpler in terms of controls and ergonomics, but much more convenient to use. It's a tool for getting work done, not for showing off.

Now, as for the converter itself.
Its specifications are quite good, both in terms of voltage and current measurement and output ripple. Unlike the ZXY, it has a significantly lower output capacitance of just 120 µF. It uses a good switching transistor and diode assembly, which helps it achieve high efficiency of around 97–98%. It also has a wide output voltage range. But there are a few flaws:
1. One of the driver transistors gets very hot (presumably).
2. The device is completely unrepairable—a case where compactness has done more harm than good.

Software.
It's definitely a cut above the ZXY's software, and I really liked the PC software in general: feature-rich and convenient. Five stars.
But the Android software is a completely different story so far. It has a strange tendency to freeze, which makes it very difficult to work with. I'll try testing it with other smartphones, though.

Overall, I liked the converter's hardware, and the software is pretty good too, but its ergonomics without a PC are a complete mess.

That's all from me. As usual, I'm looking forward to your questions. In the meantime, I'll keep testing the DPS5020. If you have any ideas for tests or modifications, feel free to leave them in the comments too.