PIPO X10, or variations on the theme of a desktop tablet

This computer is actually the third in the series. The first was the X8 with a 7-inch display, followed by the X9 with a 9-inch display.

It all started when I eventually had to sell my previous X9, something I later deeply regretted.
Then, around the beginning of the year, I saw the announcement of the new X10. Saying I wanted to buy it would be an understatement, but after looking for information online, I found out that it was only an announcement for now, and nobody knew when it would actually be available.

Then, about a month ago, I saw it for sale and ordered it right away. With points, it came to $186, down from the original price of $194. I was a bit slow on the uptake here: I could have just used an 8% off coupon and bought it even cheaper, but what's done is done.
The computer is currently selling for more; the new price is shown in the title. There's one difference, though: mine has just one OS—Windows—while the one on sale now is dual-boot, with Windows and Android.
Fortunately, I wanted a Windows-only computer anyway, as I mentioned in my previous review.

Anyway, I ordered and paid for it, then had to wait for it to arrive at the warehouse and be shipped.
It was delivered by Nova Poshta—quickly, with tracking, and right to my door.

What arrived was a pretty substantial box. I expected it wouldn't be very small, but it turned out to be noticeably bigger than the X9's box.



The package lists the basic specifications.



Here they are in full:

Operating system: Windows 10
Processor: Intel Cherry Trail X5-Z8300
Graphics: Intel® HD Graphics (8th generation)
Memory: 4GB RAM + 64GB ROM
Wireless: WiFi IEEE 802.11 b/g/n
Ethernet — 100Mbit
Bluetooth: BT 4.0
Audio output — 3.5mm audio jack
Display: 10.8 inches (1920x1280) + HDMI output
External interfaces: 1x USB 3.0 + 3x USB 2.0
Memory card support: 1 SD slot, max. 128GB
Battery: 10000mAh
Dimensions: 220 x 150 x 60 mm (excluding antenna)
Weight: 680g

These days, specs like these won't impress anyone. The processor is the entry-level model in the Cherry Trail lineup, but it does have 4 GB of RAM and 64 GB of flash storage. The key point is that this is essentially more of a desktop computer, albeit one with a battery, and it also has a built-in 10.8-inch display.




The package includes:
1. PIPO X10 computer
2. Power adapter
3. Manual

Yes, that's a very modest bundle.



The manual is fairly detailed, but as usual, not very useful.
The power adapter wasn't a European plug, and no adapter was included. The one I received is rated for 5 V, 2 A and gets quite hot while the computer is running and charging the battery at the same time. But while I was preparing this review, I noticed on the product page that they now include 5 V, 3 A power adapters.

I didn't take the power adapter apart, since they're selling a different one now anyway, and I plan to replace it. But I should point out that the X7, X8, and X9 computers came with 12 V, 2 A power adapters. In my opinion, switching to a 5 V adapter is a step backward; they should have kept the old one.



There's nothing else to look at, so let's move on to the computer itself.
The computer is indeed quite large, although the design itself is unchanged from the previous models.



There were two films on the screen: the main one and a protective shipping film.



Below the screen is the standard Windows button, although the word “button” is a bit of a stretch here, since it’s actually a touch sensor.



As before, the right side has a pair of USB ports, volume buttons, the power button, and one of the speakers. There’s also a small LED that’s supposed to shine through a tiny hole, but in practice it lights up the USB port more.
Speaking of USB, one of the ports is USB 3.0, but the volume buttons are still “backwards,” just like last time.



All the other ports are on the back, just like on the X9.
Non-removable Wi-Fi antenna
Memory card slot
Pair of USB 2.0 ports
HDMI port
Ethernet port.
5 V power input.



The only thing on the left side is the second speaker, just as on the previous versions of this computer.



The computer’s case is plastic, but the bottom cover is metal, with quite a lot of ventilation holes.



Power it on, and Windows displays its standard prompt to enter all the localization details.



From here it’s easier—we can take screenshots now :)
The familiar Windows 10 desktop that many of us know by now.



Windows 10 Home is preinstalled.



Since this computer doesn’t have Android, 44 GB of the 64 GB is available to the user, which is generally enough for comfortable use.



There’s also a hidden recovery partition on the drive, but unlike before, it doesn’t contain a system image—just the bare minimum. You can tell from its size: only 480 MB.



There was also a folder on the drive with various useful and not-so-useful tests. The only one I ended up using was the touchscreen test.



Standard information from CPU-Z.



The motherboard model doesn't mean anything to me, but the BIOS manufacturer is very familiar, since it appears in practically every mini PC I've come across.
As promised, 4 GB of DDR3 memory.



The first test I ran was a thermal test. When I bought it, I didn't yet know that this computer had active cooling, since even small computers with this processor all use passive cooling, let alone tablets.
The fan isn't very loud, but if you're sitting nearby, you'll definitely hear it.
On the plus side, there's no overheating, at least in this test, as you can tell from the fact that performance doesn't drop during the test.



Although 3D performance doesn't matter for my needs, I included a 3DMark test anyway.
What can I say? The results are quite predictable and a little higher than those of the mini PC with the same processor.



In Cinebench, however, the result is slightly lower: 5.75 in the 3D test, compared with the modified Wintel.
This is partly because Windows 10 is constantly going online to look for updates, using 20–40% of the CPU in the process. To be honest, it was annoying.



A couple more quick benchmark tests.
7-Zip



WinRAR



In my reviews, I've often mentioned that I like this manufacturer's computers because of their fairly fast flash memory. They didn't disappoint this time either, although the previous models performed better.



For comparison, here are the flash storage speeds of other mini PCs.

Pipo X9
Pipo X7
Pipo X7S
MeeGoPad T02
Pocket P1
Vensmile W10
Teclast X98 Pro
MeeGoPad T03
Wintel Pro CX-W8




The AS SSD Benchmark results are a little more modest. The test was run on a 5 GB file to rule out any influence from the operating system cache.



To test the built-in card reader, I tried two memory cards: a fast 64 GB card and a basic, inexpensive 32 GB one. The result with the first card was more or less predictable, but the second one was a surprise.



I decided to run the second test after seeing the read speed of this no-name memory card.



Since the computer has a USB 3.0 port, it would have been unfair not to test that too.
First, a 64 GB memory card connected through a USB 3.0 card reader.



And now an external 500 GB hard drive. The drive can actually deliver over 100 MB/s in sequential reads, but the result in this test was a bit worse.
By the way, as my experiments showed, the hard drive absolutely refused to spin up, just clicking as it tried to start. In the modifications section, I'll show what I had to do to fix this problem.
There was a second problem, too, and I don't even know exactly what caused it. The drive absolutely refused to work when connected directly; it only worked through a hub. Before you ask: no, the hub didn't have an external power supply. Most likely, it was a connector compatibility issue, and the hub's cable was a better fit.



If you think I decided to stop here, you're mistaken.
After confirming that the computer handled the standard tests without any problems, I decided to push it harder.
Let's fire up the OCCT test utility and get started.
First, let's load the processor—more specifically, its math unit—without putting any load on the GPU.
But even here, I couldn't get the processor above 75 degrees.



Linpack heats it up even less: 73 degrees. Both tests ran for the standard hour, with 54 minutes of that devoted to heating up.



As they say, this is no fun. Let's fire up the "furry donut."



I ran this test for a little over half an hour and stopped it simply because the temperature had stopped rising. The peak temperature reached 81 degrees; the maximum was 80 degrees.



Temperature data from HWinfo.



Since the temperatures seem fine so far, I moved on to the included touchscreen test, which checks the maximum number of simultaneous touch points. It was supposed to support 10 points, but for some reason I could only get 9, though I can't imagine why anyone would need more than 5 anyway.



With the main tests done, I'll move on to the Wi-Fi test. I was quite impressed with its performance last time, mainly thanks to the external antenna.
I ran the scan with the computer sitting on my desk and it picked up 38 access points.



Strangely enough, though, the connection quality turned out to be worse than on its predecessor.
In order:
1. 12 m from the access point, very poor conditions, reinforced-concrete walls.
2. The same distance, but with fewer obstacles.
3. Moved closer to the AP, to about 6-7 m away; no line of sight, several interior walls.
4. In the room next to the router; line of sight is more theoretical than real.
5. Clear line of sight, about 1 m away.
6. Direct wired connection.



And how could I not take this beauty apart? Of course, I had to open it up.
The computer is fairly easy to disassemble: peel off the four rubber feet, remove the four screws, and yank ("remove" doesn't quite cover it) off the bottom cover.



Interestingly, the ventilation openings are covered with a fine mesh, apparently to keep out insects or dust and prevent the fan from getting clogged.



Overall, it looks no different from the previous version.



Interestingly, even though it has active cooling, the designers left in a thermal pad that transfers some of the heat to the bottom cover.



The pad is about the size of a matchbox and roughly 6 mm thick. I'd even say I don't see much point in it, especially at that thickness.



The touchscreen controller is from the same manufacturer as the one in its predecessor, but it's a different model.



There's practically nothing on the back of the board, apart from a small picture of a girl. What's interesting, though, is that some of the markings are printed in mirror image.



At first glance, even the additional board with the connectors looks the same, but in fact it has also undergone some significant changes.



On the right side of the case (when viewed from inside), I found an interesting compartment. It was apparently designed for something, but the manufacturer later decided not to use it. Perhaps it was meant to hold a power supply, or... batteries :)



And here's the main board. To get it out of the case without too much trouble, I recommend disconnecting the antenna connector.



You can immediately see that it has not just a large heatsink, but one equipped with a fan. The previous models had nothing more than a strip of sheet metal.



For a closer look, I'll remove the heatsink. Fortunately, you don't even need any tools for this: just squeeze the plastic clips and pop them loose.



Components installed on the motherboard.
1. Processor—or, more accurately, SoC—X5-8300
2. Samsung RAM: 4 GB, made up of two 2 GB/16 Gb chips.
3. Toshiba eMMC flash memory. This is probably why its transfer speed is lower than in previous PIPO computers, which used Samsung memory.
4. AXP288C power management controller. This controller is used in almost all devices of this type; I've come across only a couple of exceptions.



1. RTL8723 Wi-Fi and Bluetooth module from Realtek. The X9 had the same one.
2. Davicom Semiconductor Ethernet controller, 10/100 Mbps USB-Ethernet DM9621.
3. Audio is handled by the ES8316. This is a change from the previous version, which had an ALC5640.
4. As before, the BIOS firmware is stored in a Winbond 25Q64 with a capacity of 64 Mbit (8 MB).



1. Probably the most classic USB 2.0 hub, the GL850G. Yes, the USB 2.0 ports are created by “cloning,” so their speed will drop when they're used simultaneously.
2. Accelerometer. As before, I still have no idea why a stationary device needs an accelerometer.
3–6. Various DC-DC power converters scattered across the board. This seems a little odd, given that there's also a multi-channel converter. Apparently, the designers needed more than it could provide.



There are also unpopulated footprints for additional connectors on both sides of the board.



As I mentioned above, the daughterboard looks similar to the one in the X9, but in fact there are some significant differences.
1. The board connects via a flat ribbon cable, which we'll “reinforce” a little later.
2. The board has a USB 3.0 connector, and since it needs more connections to the “motherboard,” a ribbon cable is used.
3. The audio amplifiers are also located here.

Even so, the designers left the volume buttons swapped. This time I honestly couldn't be bothered to rework them as I did last time, especially since it's more difficult to do here.



As was clear from the specifications, the computer has a built-in battery with a fairly respectable capacity of 10 Ah.
Let's skip over the fact that I only found out about it after receiving the device (the battery wasn't listed when I bought it), and that I don't really understand what it's for. Still, I consider it a plus rather than a minus, especially for how I use it.
But I really can't understand why they put the battery where they did, so that replacing it requires peeling off the screen, when there's a more convenient spot for it inside the case.
The only explanation I can think of is that the battery has short wires and this was more convenient for the manufacturer.
The battery consists of two prismatic lithium cells connected in parallel.



To check the battery capacity, I connected it to a battery tester. I also forgot to mention that, besides the battery, the computer has a coin cell to power the clock, so it can certainly be used without the battery.



I charged it at 1 A up to 4.2 V, which is roughly how it charges inside the computer.
The discharge test was run at 1.5 A down to 3 V. I wanted to test it at a higher current, but the battery wires are very flimsy, and I was using only two of the four wires (or five, to be precise—the fifth is for the temperature sensor).
Since the test took a very long time, the software behaved a little strangely at one point: it stopped increasing the Ah counter in the upper window, although the count in the lower window continued correctly.
The result was a battery capacity of 10.32 Ah, or about 41 Wh, which is not only pretty good, but also higher than the manufacturer's rating.



And of course, I couldn't resist making a few modifications. First, I removed the heatsink and fan.



The fan is rated for 5 V and draws 180 mA, which is quite a lot—about 1 W.



Since I saw a fairly high temperature in the “hairy donut” test, I decided to make a small modification. That said, this test is more of an extreme case; real-world applications won't generate that much heat.
I replaced the original rubber pad, about 0.7 mm thick, with a 1 mm-thick piece of rubber from this review. I used a leftover piece from my experiments rather than a new one.



I also noticed that the fan sits very low. Usually, a fan is positioned so that the airflow runs along the fins, with the fan itself raised 10–40 mm above the heatsink. The reason is that the closer the fan is to the surface, the lower the airflow becomes. If you bring it almost right up against the surface, it will hum and spin, but only cool the small area directly beneath it.
What's strange, though, is that there's room for a fan inside the computer case, yet the manufacturer chose to mill down the heatsink and install the fan incorrectly, rather than move it 15–20 mm and put it where there was more space. One option would be to cut away part of the plastic case.
But the fan wouldn't fit there either, so in the end the manufacturer placed it almost flush against the heatsink, effectively “choking” the airflow.



For my modification, I decided to raise it at least a little, so I put a couple of pieces of wire insulation underneath it.
I couldn't raise it too much, because without modifying the case, the fan would start to “suffocate” from above.

This modification was really more of an experiment. It would be better simply to replace the heatsink with a larger one and get rid of the fan altogether.



The fan was assembled and put back in place, and I moved on to the next, more important modification.
During the computer tests, a few significant problems came to light.
1. If a mouse or keyboard was connected to the computer, it would "drop out" after about 10 minutes of use. At first, the computer had worked fine for several hours; the problems started later. We'll come back to this issue.
2. When an external hard drive was connected to the side ports, it couldn't spin up. The voltage sagged badly at startup.
3. There was a slight interference with any sound, regardless of the volume setting. Connecting the hard drive made the interference much worse.

I didn't solve the first problem, but I did fix the other two. The manufacturer had used a cable that was too thin between the main board and the auxiliary board, causing a significant voltage drop. On top of that, the audio amplifiers were on the auxiliary board, so any voltage sag caused audio interference.

The fix is very simple: use thicker wires for ground and the positive supply. The audio is affected only by the ground wire, but for USB it's best to upgrade both.
After finding that the power to all the USB ports was simply connected in parallel, I connected the corresponding contacts on the boards, and found more convenient pads on the boards for the ground wire.



After the modification, I ran a few of the tests again.
First, the math test: the temperature dropped by 8 degrees. Last time it was 75, now it's 67—a pretty noticeable difference. The screenshot also shows the processor running in Turbo mode, with its frequency boosted by 11%.





In the "furry donut" test, the processor frequency dropped during the test, and the same thing happened after the modification..
Only later did I remember that I'd forgotten to "tune" the BIOS. I raised the minimum temperatures from 50–60 degrees to 80.



But as it turned out, this changed nothing: the frequency still dropped by 66%.



The temperature did drop by 10 degrees, though, so I suspect the frequency reduction has nothing to do with overheating.



After a small modification, it's clear that overheating the processor is now impossible.



While I was at it, I tested the battery capacity during video playback. I put on my "test" movie and timed it.

The test started at 3:40



About 8 hours later, the computer automatically went into sleep mode.



So far, I've been talking about the good things about this computer, but unfortunately, I have to move on to its drawbacks. More precisely, there's just one drawback, but it's a significant one :(

While testing it, I ran into a pretty serious snag, courtesy of the manufacturer and the well-known company Microsoft.

The snag was this: after spending the whole evening playing around with the computer and running various tests, I turned it on the next day and saw this message.

Actually, it wasn't quite like that. The computer was sitting nearby; I used it for a bit, then got distracted by work on my main computer. Then I heard notification sounds. At first, I didn't even understand what was going on. The computer said there was a problem with the last connected device. The devices connected were a wireless keyboard with a touchpad, a mouse, and a couple of flash drives.

After a series of experiments, disconnecting different devices and reconnecting them in various combinations, I figured out that the mouse and keyboard were disconnecting.



The Device Manager showed a similar message, but after disconnecting and reconnecting the device, it changed to the one shown in the screenshot.

Here's what I found:
1. It's not just the mouse that disconnects; the entire class of input devices is affected.
2. A flash drive continues to work normally in the adjacent port (on the same hub).
3. The problem only occurs when connecting to USB 2.0, regardless of which ports are used.
4. The problem also occurs with USB 3.0 if you connect everything the same way through a hub. In other words, the problem affects this class of devices when they're connected through a hub. Since the USB 2.0 ports are wired up this way, that's where the problem showed up.
5. Increasing the cross-sectional area of the power wires doesn't help.
6. The problem is entirely software-related. The devices work fine in Android on Android devices.
7. Windows 8 isn't affected by the problem, but then you run into driver issues.
8. According to information on 4pda, the problem appears after installing the integrated graphics driver.
9. This problem is fairly common on other computers running Windows 10, too.

The countdown to the problem starting isn't based on how long the computer has been running; it accumulates the total time the device is in use. For example, I can connect the mouse after half an hour and use it for 10 minutes, or use it for 5 minutes after half an hour, then another 5 minutes half an hour later.



I don't remember which version of the OS was originally installed on this computer, but it updated itself to version 1607 while I was using it. That didn't change anything, though.



Remembering that the computer had worked fine at first, I ran a full system rollback. The result was almost the same: everything worked normally for a while (about an hour), then the problem started all over again.

After that, I went a step further and decided to reinstall Windows completely.
I took photos along the way so that, if it worked, I could write up the process for those who had already bought one or were planning to. But since luck wasn't on my side, I won't be publishing those photos.
After reinstalling, I saw a screen with very tiny text and icons. It felt like the resolution had doubled.



There were lots of unknown devices in Device Manager, but in the end only one remained.
You can download the drivers here.



During my experiments, I tested various builds of Windows 8 and Windows 10. I found that the problem occurs in Windows 10 when video drivers are installed.
Windows 8 works fine, but most drivers aren't available for it, which is probably why everything works properly with it.

If you install Windows 10 without installing the graphics card drivers, the ports work normally, but of course you can forget about gaming and watching demanding video files with hardware acceleration.

I originally planned to use this computer to control my "zoo" of various devices: a 3D printer, laser engraver, milling machine, and regular printer. In principle, nothing prevents me from using it for that. Since the computer has a touchscreen, it doesn't really need a mouse (that's how I happened to notice the glitch), so I could make do without one. The problem is that it also affects USB-RS232 converters, which is completely unacceptable for obvious reasons.

This problem isn't limited to the computer I bought from GearBest. In the thread I started, buyers who purchased the computer elsewhere ran into the same issue. That thread also links to a mention of the problem on the Microsoft website, and searching online turns up the same advice: disable power saving for USB ports, and so on. I think we all understand that these methods have been checked repeatedly.

But that's not all. Just a couple of days ago, I was assembling a computer with a Gigabyte motherboard, MSI graphics card, and Core i3 6xxx processor—all brand new. So when, after installing Windows 7 from a flash drive, the computer recognized nothing but the keyboard and mouse, I wasn't surprised anymore.

That's probably all for now. Let's sum things up.

Pros
Large screen
External antenna
10,000 mAh battery.
Two speakers.
Excellent cooling

Cons
Active cooling system
Poorly positioned battery.
Problem with USB 2.0 ports in Windows.

My opinion: This computer really does have a lot going for it, and overall it's better than its predecessor, the X9, and especially the X8. My first laptop had a 10-inch screen with 256 colors and a resolution of 640x480. Now a 10.8-inch screen with a resolution of 1920x1280 doesn't surprise anyone. It's noticeably more comfortable to use.

I waited a long time for this version of the computer, but the manufacturer and Microsoft have really thrown a wrench in the works with this USB port problem. I'll probably find a solution eventually, or at least I hope so, but it's already cost me a lot of time. I can certainly use the computer as it is, without the graphics card drivers, but I'd like it to work properly.

I'm sure people will ask again why anyone needs a computer like this when there are tablets. I'll repeat what I've already said.

1. Large screen.
2. 4 USB ports, one of them USB 3.0.
3. External antenna
4. Wired connection to a local network, though only 100 Mbps.
5. No cooling problems.
6. Full-size HDMI output
7. Better sound than a tablet.

The next question will probably be why anyone needs a computer like this at all.
1. When connected to a projector, it gives you full control—for example, during presentations or in a home theater. What's more, this computer handles 4K video quite easily.
2. As a work computer for controlling all kinds of CNC equipment, small machine tools, printers, and so on. In this case, it's even more convenient to use than a laptop, and it takes up less space.
3. As a POS terminal
4. Or even just for "receiving" IP radio, with extra features.