HDMI dummy plug, or how to bring a PIPO X10 back to life

A little over a year ago, I bought PIPO's new mini PC, the PIPO X10. I liked almost everything about it: a large screen, a built-in battery, and an external antenna. But almost immediately, a problem became apparent: it wouldn't work properly. And it wasn't just me—the problem was widespread.
I played around with it for a while, then eventually put it on a shelf. One of my readers suggested a way to bring this computer back to life, so today I'll tell you how to revive the PIPO.


Before I get to the product review, here's some background on how this all came about.
As I said, I bought this little toy last fall and even wrote a review of it, where I mentioned the problem. As far as I know, it's the only review of this version.


I bought it so I could control things like a 3D printer, a laser engraver, and a mechanical engraver without being tethered to a computer—in other words, to almost completely do away with my main computer when working with these devices. I planned to power the devices themselves with a 24 V uninterruptible power supply, while the computer has a built-in 10000 mAh battery.
Besides the large screen, it also has four USB ports, one of them USB 3.0. So it was just right for the job.


I planned to control everything either directly from the computer or over Wi-Fi or a wired network. The Wi-Fi on this computer works pretty well, too.
I already have a review of the previous version, which I really liked, but I wanted the "new model."



The computer worked just fine for a while—maybe half a day, maybe a day; I don't remember anymore. But then an error started popping up, and the USB 2.0 ports would stop working. Interestingly, if only a flash drive was connected, everything was fine. But as soon as I connected anything else, it would drop out. USB 3.0 worked fine, but connecting a hub to it caused the same problem.


If I tried reconnecting the device, as Windows recommends, it would drop out again within a few seconds.


The Device Manager reported a device descriptor request failure.


A thread about this computer was started on the 4pda forum, where I learned that many users had the same problem, though not everyone. It affected computers running Windows only, as well as dual-boot Windows/Android models.


At first, the USB hub installed on the computer's circuit board was the prime suspect. But the X9 has the same hub and no such problems, and neither do other computers—fortunately, I've tested quite a few.


We eventually found that the problem only affected Windows. Everything worked fine in Android, so the operating system was where we needed to look.
We tried installing different Windows 10 builds and even Windows 8. Sometimes it would work for a while and we'd start to get our hopes up, but then the ports would drop out again.
After all that trial and error, we established that everything worked fine as long as the graphics card drivers weren't installed. By the way, after installing the operating system, it didn't install any drivers on its own, just the basic ones, and Device Manager looked pretty bleak. But one of the users shared a complete archive of drivers.


And fairly recently, one of my readers not only suggested a solution, but also showed that it actually worked. A huge thank-you to him.
The solution was right under our noses, and that may be why some users' computers worked fine.
Everything worked normally only when a monitor or TV was connected to the HDMI output. And the problem appeared right after installing the video drivers, but somehow nobody thought of that at the time.

As you can imagine, keeping a monitor connected just for this isn't practical, so he suggested using an HDMI dummy plug, or "monitor emulator."

So I ordered one for myself, too. At $5, it wasn't much to spend on an experiment.
It looks like a flash drive, but with an HDMI plug.


The metal case looks very neatly made.


It takes up very little space at the back and doesn't get in the way of the USB ports, which is great. Let's connect it and start testing.


A second monitor appears when connected.


The monitor is detected as a 28E850 digital TV.


When you select extended desktop mode, you can choose resolutions from 800x600 up to 4K, with a bunch of options in between.


But since I didn't need that, I chose duplicate mode (the default), which set the resolution to 1920x1280 (the PIPO's screen resolution).


As usual, the mouse “dropped out” during the process, so I had to use the touchpad, or switch it to USB 3.0.
Open Device Manager and simply disable/enable the USB Root Hub.
After that, everything works. You could just restart the computer, of course, but that takes longer :)


I got curious about what was inside this dummy plug and tried to open it, but no luck—it felt like the insides were potted in rubber. I had to look online for information, and found it on mining forums.
The thing is, miners often manage their rigs using Radmin/TeamViewer, and without a monitor connected, the software sets the resolution to 640x480 or 800x600. I usually had no problems with Radmin: I just logged in remotely and set the resolution I needed. TeamViewer, though, was trickier.



It turned out to be an extremely simple design, consisting of an EEPROM chip and a few components.


There's even a wiring diagram for the EEPROM, though it's a “stripped-down” version. People on the forum said this setup can cause problems, and that pull-up resistors may need to be added to the signal lines. I can't see how the lack of pull-up resistors could affect the HDMI port's reliability, but it's better to add them anyway. There's also a capacitor on the board, presumably on the power supply line.
The EEPROM stores information about the “monitor” and a list of supported resolutions.


What's more, it turned out that someone in Russia makes similar dummy plugs, too.


Along the way, I found out there are a couple of other options: use a VGA dummy plug (where possible) or an HDMI-to-VGA converter. The latter has both advantages and drawbacks. The advantage is that you can connect a VGA monitor; the drawback is that people report it can sometimes damage the graphics card.
I used a similar converter for a long time and it worked fine, but knowing how hot it gets and how much power it draws, I can see why people are concerned. Besides, I doubt the interface designers expected anyone to connect such a power-hungry load to the port.


Sometimes there's another problem, though: the converter also needs its own “dummy plug” to make it think a monitor is connected. Fortunately, it's a very simple design consisting of three resistors.


In the end, the computer works, though not without a few quirks. After a while, the mouse disconnected again, this time after the computer went into sleep mode, so I had to disable sleep. I also disabled the option to turn off power to the USB controllers.



I was going to wrap up the review here, but I couldn't resist and decided to investigate further—maybe it would be useful.
First, I tried reading the EEPROM chip, but couldn't get it to work, despite trying several different ways. I used the wiring diagram above to connect it to the programmer.
Eventually I gave up and decided to forget about it, but I still really wanted to figure it out, so I tried opening this “flash drive” and finally managed to crack it open.
It turned out that if you press hard on the rubber part on the side opposite the connector, it opens up. But the inside is filled with something rubber-like.


I've written more than once that I try to avoid destructive methods when testing devices. So I carefully gnawed out cut away a small piece of rubber, although removing it all would have been much quicker.
At first, though, I started cutting on the wrong side of the board, going by the photo I showed above.


Once I had “access to the patient,” everything became much easier.
I was interested in the connection pads marked SCL, SDA, and GND. On the connector pinout diagram, I highlighted the pins actually needed for the emulator to work.


But I realized my mistake when I saw the full schematic. The thing is, the wiring diagram I found on a miners' forum had several errors.
1. The SCL and SDA pins were swapped. They weren't labeled in the diagram, and I was too lazy to open the EEPROM chip's datasheet.
Naturally, the chip couldn't be read.


But while the previous mistake was partly my fault, there was no way I could have anticipated the next one. The GND pin isn't connected to the connector pins themselves; it's connected to the case! So even with the data lines connected correctly, I would never have been able to read the chip. The only way to figure that out was to open up the emulator.
Anyway, I connected ground, swapped the SCL and SDA pins, and it worked. Hooray, my first small victory.


The chip read successfully, and the information it returned included the name of the monitor being emulated.


But since read errors can happen, I checked it by reading the chip again and comparing the results. The check passed with flying colors, so we can consider the dump valid.


I was about to leave it at that, but then I thought: why not test it? I also sketched out a schematic of the emulator I was reviewing, in case it comes in handy.
The capacitor can be anywhere from 10 to 100 nF, and the resistors from about 2.2 to 6.2 kΩ. The unused pins are simply left floating. That's not good practice, but it's exactly how the emulator is built.


I didn't have a 24C02, but I did have a few 26LC16B chips.


I programmed the chip with the firmware I'd downloaded earlier. In the programmer, I selected the chip as a 24C02, though that doesn't matter.
Then I put together this contraption with “sticks and string.”


Power it up, and it works just like the original :)


Now for the savings.
The chip costs about 10 cents. There are cheaper ones, but they come in smaller packages. You could also salvage one from an old circuit board.
The easiest option for the connector is to use a piece of HDMI cable. It can be the cheapest kind, costing about $1, and you can make two emulators from it.
Three resistors and a capacitor—let's say another 5 cents.
You can skip the socket.
That comes to 10+50+5=65 cents per emulator.


Of course, you'll say the programmer costs money too.
It does, but there are three options: find a radio amateur you know and drop by, build a programmer yourself, or buy the simplest CH341A. Luckily, it costs about $2.58 including shipping
Even if you buy a programmer, it still works out cheaper than buying a ready-made emulator :)


As a bonus, here's the wiring for a DVI-to-VGA adapter for anyone without a VGA output but with DVI and the extra pins for analog output (above and below the wide pin)—DVI-A, DVI-I.


With this setup, you could try using three 75 Ω resistors. No guarantee it'll work, but it takes almost no effort to try.



I can already anticipate the questions, so I'll try to answer them.
Initially, I tried the following to solve the problem:
1. Disabling power saving
2. Doing the same through the registry.
3. Reinstalling Windows with different versions and builds
4. Using powered external hubs.
5. Looking for alternative drivers
6. Changing BIOS settings
7. Contacting the manufacturer and sellers.

None of these attempts worked. The result was always the same: the USB ports would drop out.

I tested the computer for several days. The ports disconnected once, during another transition into sleep mode. I disabled them, and everything was fine after that.
This solution is obviously a hack, but if it works, it's better than nothing. Still, I'd recommend trying an external monitor or an HDMI-to-VGA converter first (if you can borrow one). If everything works fine, then buy one of these dummy plugs.
I chose the seller simply by price; I don't think there are any real differences.

Almost forgot the most important thing: the EEPROM firmware itself, included in the archive along with the wiring diagram.
hdmi_plug.zip [ 0 b ] ( downloads: 387 )


That's all. I hope you found it useful.