GPD Win (x86 Computer / Palmtop)


Looks like GPD has pretty much settled on the following for the WIN 2 keyboard layout , port layout and shoulder button layout. I suspect small changes may still happen.


dxeRCxO.png

C2kowGa.png
I like that shoulder button layout better but still wish it had clickable sticks for L3/R3.
 
If you're going to leave off a key, ditch the 'Windows' caps lock key. It is a waste of time and space.
Fixed. Not to say you can't kill the Windows key, too. While your at it, would there be a great loss from throwing out the Xbox key, start and select, and L3/R3?

I haven't looked closely at any of the proposed layouts, and I don't know how taking out some of those would impact control functionality (in my mind it is simple to assign those functions to other keys), so I am probably missing important considerations. If the keys, which are already too small to hit without hitting nearby keys, are going to be made smaller, it seems they should first remove anything that isn't needed so they don't need to shrink it down so much. Having a keyboard is pointless if it can't be used without hitting all the other keys...that is also a reason to get the power button of the keyboard this time.
Your logic could be the same for 'Home', 'End', 'PgUp', and 'PgDn'. They can be left off too because there are arrow keys that will eventually get you there.
Good point! Look at all the extra space we are freeing up.


i understand keyboard threads about the pyra on the pyra forums.

i'm not sure i understand keyboard threads about the gpd win2 on the pyra forums. at least with the former there was a non-negligible chance that your design contributed to the final design...

but don't let me rain on your parades :)
Yeah, and that is why I am not looking too closely, or thinking too deeply, about the layout, plus I won't be buying one. But it's a keyboard thread, so, ya know...
 
Fixed. Not to say you can't kill the Windows key, too. While your at it, would there be a great loss from throwing out the Xbox key, start and select, and L3/R3?
Apropos: What's the deal with the L3-only redundancy on the keyboard?
 
No clue why the official one only went with L3. Maybe it was accidentally left, or it is used for something so important they need an extra one. It is their keyboard, so they get to make it however they want. If they really want it to be a usable part of the device (not something they added on just so they can say they have it, which ifs what it seems like) they should consider potential user concerns, and do some testing.
 
Looks like the WIN 2 will have an M.2 slot for swappable storage , Wade (Co owner of GPD) seems to have confirmed it. Mind you it's very early days and things can change

It's the slot to the right of the R2 Shoulder button


C2kowGa.png
 
Apparently GPD are looking at this stage , to put a Intel Core M3-7Y30 in the WIN 2. Subject to change given it's still early days.

Interesting because that chip is about twice as fast as the current 8750 they use in the GPD WIN
 
The current GPD Win already gets too hot when running demanding stuff, and the huge battery is drained fast as hell.... and the m3-7Y30 needs a bit more power when running at the same speed...

It's not clear how much power it uses compared to the z8750 on Intels pages... you can configure the TDP and max frequency, but it's hard to tell what speed difference it'll make.

According to passmark, it seems to has a typical TDP of 4,5W at 1,0GHz (whereas the z8750 has a TDP of 4W) - though TDP doesn't really tell us how much power it uses...
The passmark score of the m3 with 3906 is quite a bit higher than the z8750 with 1902, which sounds really interesting.

I wonder if that additional power usage / heat production can really be done, as the current GPD Win surely is at its limit now.

Well, the price for the m3 is higher than the retail price of a GPD Win, which means the GPD Win 2 would be at least twice as expensive as the GPD Win...

Really interesting, let's see where this leads to :)
 
They are planning to completely revamp the cooling system. That said, someone tested the M3 vs the current Z8750. Apparently at the same power usage, the M3 delivers a lot more performance - like almost twice the frame rate. It seems it's a lot more power efficient than the Z8750. You can see the video tests below by Phawx




 
A quick look shows you can get a M3 tablet in the UK for €495 delivered

Which is mind boggling looking at the cost of the CPU alone.


There has been a lot of discussion about what the M3-7y30 can be bought for on the Dingoonity forum. The list price is $281, but there is a tablet that has been selling for $290 - $360 delivered with the SOC . It has been that way from pretty much day one. So it seems the M3 can be sourced for a lot lot cheaper. Perhaps it's a regional pricing thing, or perhaps Intel sell it for far cheaper than the list price. I suppose we shall see in the end. By the time the GPD comes out , the 8y30 should be released by Intel making the 7y30 probably a little cheaper again.
 
Last edited:
Still going to have to hold it in your hands.
The whole form-factor isn't catered to high loads.
That being said, looks like they went with another fan-based design.

5.5 key wide spacebar is going to be a challenge.
 
Last edited:
That being said, looks like they went with another fan-based design.

I still maintain the only way to cool something like this properly is to put the SoC behind the display and use an aluminum lid as a massive heatsink. Moving the SoC and supporting chips to the top half, in addition to removing the need for a fan/heatspreader in the lower half, would leave you a lot more room down below for a bigger battery.
 
I still maintain the only way to cool something like this properly is to put the SoC behind the display and use an aluminum lid as a massive heatsink. Moving the SoC and supporting chips to the top half, in addition to removing the need for a fan/heatspreader in the lower half, would leave you a lot more room down below for a bigger battery.

That would make the whole device pretty unbalanced to hold, unless you put some weight into the basement as well.
 
That would make the whole device pretty unbalanced to hold, unless you put some weight into the basement as well.
Something like a large battery perhaps?

Moving the SoC and supporting chips to the top half
That idea sounds strangely familiar ;)

The GPD win 2 sounds like it could be rather interesting, I might well approach them with the ideas in the above linked post, which I still believe would make a superb system. Looking at the above linked post it appears that the associated images have disappeared, guess this was during forum migration?
 
Last edited:
It's basically how tablets with connected keyboard work, especially in designs that have a large secondary battery in the bottom half which charges/powers to screen/tablet half while connected. If the hinge were made permanent you wouldn't need any battery at all in the top half.

The only other problem I can think of is that on a small device you expect all of the ports to be in the base part. You'd eliminate needing to carry LCD and camera data through the hinge, but have to wire all the ports through there instead. I guess you could eliminate the 5V and ground lines from USB ports making them just the data pairs (for USB2 - extra superspeed data lines for USB3), and run them directly off the battery. Similarly with HDMI, though presumably there more of the pins carry data and fewer carry power in comparison. And eSata commonly only has 5V and ground power and a couple of data lanes.
 
That would make the whole device pretty unbalanced to hold, unless you put some weight into the basement as well.
This would be more of a problem than it sounds. If you look closely at current phablet devices their weight is concentrated at the part where you hold them and even then they put a strain on your wrists. I had to get a smaller phone because my wrists started to hurt using an 8" tablet as heavy as the Pandora.
 
I would expect the base to contain the power management components. You would have power going up, and HDMI, audio, and USB (a single channel going to a hub on the base PCB?) going down. I assume the KB & game controls would all be USB. You would need a handful of other lines for things like LEDs, power switch, etc.
 
If you wanted to pursue this CPU/SoC in the display idea...

Put a USB 3 hub in the base. The only lines from the base to/from the SoC/display section is then power (from the battery in the base) and the USB 3 back to the base.
Run an internal USB 3 - HDMI adapter off of it for the external display.
Connect the Keyboard & controls via USB.
The base would need to be semi-intelligent to handle battery charging independently of the SoC.
Radios & antennas could be in the lid OR USB based. I.e. an 802.11n USB adapter equivalent on the board in the base. Same could be done with Bluetooth.
Card slots in the base effectively become USB card readers.
eSATA and/or wired Ethernet would both be possible via on-circuit-board USB adapter equivalents.

The display backlight board becomes display backlight on one side and the SoC board on the other, effectively making the whole unit a 2 board device. Replacing/upgrading the SoC would be done by replacing this whole board.

Having an aluminum display back piece would also make a good heat sink & radiator.

I can see how this could work. An obvious hurdle would be having a curled up cable through the hinge that can handle the power requirements of the SoC. It is also a pretty radical departure from everything out there today.

A shortcut to this might be to take a commodity smartphone with a USB 3 OTG connector on it for the top half. Something like a Galaxy Note 3 would work for proof of concept (removable back & battery, USB 3.0 OTG). I wonder if a USB hub and related components could be built as a replacement for the SoC module in the Pyra keyboard board.
 
Back
Top