Favorite Tablet / 2-in-1 Windows PC?

With the proliferation of Windows 10 tablets in the world, the idea of running PAC Display and SoftPAC on a single machine to read, process, and display data from industrial equipment seems within reach.

In practice, I’m finding the tablet to be a serious weak point in this system. So far I’ve tried an Asus Transformer T102HA, and a Chuwi Hi10Pro, but neither has been close to reliable enough to consider for use in a (clean) manufacturing environment. (The T102HA had power management problems, the Chuwi freezes and requires a reboot at least once every 24 hours). Both of those use an Atom Cherry Trail processor, so perhaps that architecture is to blame.

Does anyone have a favorite Windows 10 tablet that has proven to be reliable? I’ve stuck with the sub-$500 tier to experiment, and would hate to move out of it and run into the same problems. The application would justify spending money on a nice Surface, for example, if it is really better. I know that some members of this forum are great at finding cost-conscious alternatives.

I had looked at whether groov would be appropriate here, and concluded that the processing required to take the raw data would require a controller of some kind, so PAC Display seemed like the simpler approach. Any thoughts or suggestions are welcome.

(The data acquired from the industrial devices comes via TCP/IP, but not Modbus)

Also see this discussion: http://forums.opto22.com/t/touch-screen-industrial-pc

Many of our groov customers are moving away from Windows for just the reasons you mention.

What kind of “processing” do you need? (There are a couple of options for doing processing on a GROOV-AR1, some more highly published than others.)

In this specific application, the data is broadcast as an ascii stream, every 100 ms. The message length varies due to absence of leading zeros. The controller needs to read and parse that stream into 8 data elements, using delimiters. 4 of those elements represent floats, and both the range and geometric mean of those elements are calculated and stored. The range and mean are used to calculate a final set of values, which are presented on the display.

It looks possible to do all that in Node-Red, so for the cost of the groov box, and a “dumb” display, I think I’d get much more robust hardware. Development skill and time is hard to come by, though. I’ve got a strategy and PAC Display screen already completed, so it would be nice to use those with good enough hardware instead of starting over.

Ah. That makes sense.

But just curious, a few more questions for you:

  1. How big does your screen need to be?
  2. What’s your timeline?
  3. How much I/O (if any) do you have?


The screen does not need to be big… 6" - 10" is fine. Timeline is asap, at this point. I’ll probably go with a regular desktop and monitor for now, but it would be nice to make the system more compact. We may have several of these to install eventually, so continuing to experiment with tablets in search of a good one might still be worthwhile. The next step here might be to just wipe the hard drive and see if I can get a better version of windows to install.

No I/O… the only information needed comes through TCP/IP. At one point I was even looking at a Raspberry Pi with their 7" screen, but the idea of introducing yet another architecture and programming system into our manufacturing environment is not attractive, despite the savings in hardware cost.