Playing with Play-Doh

[Zwerg01]

I know that you must get this question alot, but how long until the final products are ready to ship?

they have prototypes right now with few issues with the hardware, the case and some more open points with the software. I would estimate at least 6 month for production (in detail, bugs have to be fixed until sept/oct (aug is holiday season, then you have to scedule mass production, shipment of the finnished parts) and 1 month for assembly and shipment. Also to mention, there will be a first batch one and a first batch two. So only 500 pieces are produced and shiped to customers. If they have them, we have to wait some weeks and see what happens and if there are any bugs left. If so, they have to be removed and parts have to be changed and produced new. If not they can continue producing parts. If this sounds like chaos for you, you are not alone. So the first 500 preorderers will have the pyra around january. if everything goes right ( which was not the case for the last years) the next 500 will have the pyra around april/june 2018. If there are changes to be made, they have to wait for christmas 2018.

 

[oskda]

We do know the max sustained CPU speed in Pyra (with and without case-heatsink at normal room temperature)?

In a good quality and good design notebook or desktop computer (at normal room temperature) you can drive CPU to continuous max speed without problem. In a modern smartphone it is different: speed decreases; And I suppose in Pyra it will be the same (specially since it hasn't been designed specially for take heat out).

In a smartphone we don't bother too much, because a most of time we use them in a different way we use a desktop/notebook computer, but Pyra is a more hybrid device: it is a pocket computer, for me more a small notebook than a smartphone: and it has a computer SO (real Linux), not a smartphone SO.

I think it will be also interesting to know the pleasant (for human body) max sustained speed. Here, with pleasant, I refer to a case temperature which doesn't bother to our human body (for example our hands in contact with Pyra while we hold it and we type or play with it).

PD: I would love a Pyra more in focus with heat dissipation, for example I would put CPU card in display part and rear display cover in aluminum, connected with SOC/CPU, so it would act as a big heatsink. And in this way we would have lower part (with keyboard, gamead, battery) cool for our hands (and good for battery), being hot in the upper part (displat part) and taking away with big heatsink. Obviously this can be changed as we have other design yet done.

North or South America?

Or Central America. Don't forget the center.

 

[Binky]

PD: I would love a Pyra more in focus with heat dissipation, for example I would put CPU card in display part and rear display cover in aluminum, connected with SOC/CPU, so it would act as a big heatsink. And in this way we would have lower part (with keyboard, gamead, battery) cool for our hands (and good for battery), being hot in the upper part (displat part) and taking away with big heatsink. Obviously this can be changed as we have other design yet done.

I think that would make the lid both rather heavier, and thicker - both of which would make hinge design more difficult and the unit feel less balanced.

Plus, you'd need a much larger cable going through the hinge to connect everything in the base to the CPU.

You could argue that heat should be deliberately directed to the user's hands, as humans have excellent thermal regulation and fingers make good heat exchangers.

Anyway, most of the cunning schemes for heat dissipation I've heard involve a case redesign, so they may have to wait until our favourite dragon is curled up on a mountain of golden treasure and can afford to buy new case moulds.

 

[Grench]

I think that would make the lid both rather heavier, and thicker - both of which would make hinge design more difficult and the unit feel less balanced.

Plus, you'd need a much larger cable going through the hinge to connect everything in the base to the CPU.

You could argue that heat should be deliberately directed to the user's hands, as humans have excellent thermal regulation and fingers make good heat exchangers.

Anyway, most of the cunning schemes for heat dissipation I've heard involve a case redesign, so they may have to wait until our favourite dragon is curled up on a mountain of golden treasure and can afford to buy new case moulds.

The following is all hypothetical.

Actually - that would be a viable design alternative and could have several advantages. The weight difference would be negligible - amounting mostly to the aluminum lid heatsink - which would be ~ 100 grams. Compared to the battery in the lower portion, the lid would still be negligible.

The display backlight board becomes the SoC board - potentially a parts reduction (one less board), though it would be a larger multi-layer board - so it might be a wash.

Having the SoC connected to the lid via a heat pad would be a simple solution. If the outer lid surface were corrugated (ridges), it could even be a pretty effective radiator.

As for the connections lid to base - it would need a single USB 3.0 OTG set to cover:
- Power from the battery to the lid/SoC for the SoC & screen.
- Data to/from the base to/from the lid.

The key to this hypothetical device design working is to completely ignore every data transport option on the SoC other than USB 3.0 once it leaves the lid. eMMC could still live on the SoC board in the lid and be direct attached. Power button may need to live in the lid somewhere too.

A USB 3.0 hub (electronic equivalent in circuitry) in the base (getting power from the battery would then run all of the ports & slots & radios. Each SDXC slot becomes a USB-SDXC adapter (on the circuit board). Wifi is the equivalent of a USB WiFi dongle's circuitry. Keyboard becomes a circuit based USB HID device. Nubs become a circuit based USB game controller

3 card slots, keyboard, game controls, WiFi, Bluetooth, 4G, 3 USB hosts, keyboard & LED controller ~ 12 port USB 3 internal hub. 12 port USB 3 hubs exist - so their circuitry is possible to work out.

The catch is that the USB OTG would have to be set into permanent host mode and the device could never be a USB target.

I could see it being theoretically feasible. It is a huge departure from everything else though - so it would be risky too.

All hypothetical. But - I could see it as being theoretically feasible if someone really wanted to build it that way.

 

[oskda]

I think that would make the lid both rather heavier, and thicker - both of which would make hinge design more difficult and the unit feel less balanced.

I was writing a PD (I didn't expected discussion about this topic as Pyra design is in other way and we can go back to origin).

Actually display case part has a board -display board-, so installing CPU board there wouldn't increase thick (nothing or only a small little). Even you save one board: now with actual design with we have 3 boards (base, CPU and display), but the way I propose we would have 2 boards (as display board and CPU board would combine in one).

On other side I think base case is too thick: moving CPU board to display case would made base board less thick, and easier on our hands.

Plus, you'd need a much larger cable going through the hinge to connect everything in the base to the CPU.

Grench explained this much better I can do.

Actually we must pass display data (usually the bigger data transfer in a personal computer) from lower to upper case, by a custom complex and delicate cable. If we move CPU/SOC to upper case, we save that enormous amount of data.

On other side, how do you connect keyboard/gamepads/etc on a modern computer? By USB. So we end up no with more cables, but less cable: one USB/power between up/down: and that cable is cheap/standard and I think more strong. Less pins and more robust cable is possible in this way.

You could argue that heat should be deliberately directed to the user's hands, as humans have excellent thermal regulation and fingers make good heat exchangers.

Of course you are joking

My fingers and hands say they don't like too hot items

PD: I must repeat: I assume actual Pyra design, as now we can't change it.

 

[levi]

My laptop seems to use the touchpad as a radiator when it gets overly hot, which was a surprise when I first noticed it and a little uncomfortable at times, but not actually painful.

 

[gardrek]

My laptop seems to use the touchpad as a radiator when it gets overly hot, which was a surprise when I first noticed it and a little uncomfortable at times, but not actually painful.

Mine does as well, tho it actually reaches a temperature that I would call more than "uncomfortable." Then again, I have sensitive skin.

 

[DrHAX]

As a former tig welder and a person who works with 195c temp dies for extruding PVC and acrylic, I think a little heat isn't going to bother me.

Granted I doubt the general populous has thick skin and hold onto hot objects without burning as well.

 

[Kippykip]

My GPD-Win gets pretty bloody warm at times but it doesn't bother me, in fact in most cases it's actually was nicer to hold since the weather here is freezing.
If the Pyra gets hot, I don't really care

 

[mecksg1]

im good with it being warm.

 

[Hồng Thất Công]

The following is all hypothetical.

Actually - that would be a viable design alternative and could have several advantages. The weight difference would be negligible - amounting mostly to the aluminum lid heatsink - which would be ~ 100 grams. Compared to the battery in the lower portion, the lid would still be negligible.

The display backlight board becomes the SoC board - potentially a parts reduction (one less board), though it would be a larger multi-layer board - so it might be a wash.

Having the SoC connected to the lid via a heat pad would be a simple solution. If the outer lid surface were corrugated (ridges), it could even be a pretty effective radiator.

As for the connections lid to base - it would need a single USB 3.0 OTG set to cover:
- Power from the battery to the lid/SoC for the SoC & screen.
- Data to/from the base to/from the lid.

The key to this hypothetical device design working is to completely ignore every data transport option on the SoC other than USB 3.0 once it leaves the lid. eMMC could still live on the SoC board in the lid and be direct attached. Power button may need to live in the lid somewhere too.

A USB 3.0 hub (electronic equivalent in circuitry) in the base (getting power from the battery would then run all of the ports & slots & radios. Each SDXC slot becomes a USB-SDXC adapter (on the circuit board). Wifi is the equivalent of a USB WiFi dongle's circuitry. Keyboard becomes a circuit based USB HID device. Nubs become a circuit based USB game controller

3 card slots, keyboard, game controls, WiFi, Bluetooth, 4G, 3 USB hosts, keyboard & LED controller ~ 12 port USB 3 internal hub. 12 port USB 3 hubs exist - so their circuitry is possible to work out.

The catch is that the USB OTG would have to be set into permanent host mode and the device could never be a USB target.

I could see it being theoretically feasible. It is a huge departure from everything else though - so it would be risky too.

All hypothetical. But - I could see it as being theoretically feasible if someone really wanted to build it that way.

Speaking about hypothetically speaking, I would add a second SoC in the lid so Pyra can have dual SoC design so I can run PS2 emulation on it

Better yet, if the screen can be detached then I can use just the screen as a tablet ....

Well, that would be Pyra 2

 

[Swordfish II]

I was thinking about this. Playdough contains a large quantity of salt making it both conductive and corrosive when moist: is that really the best material to be doing this with? Maybe some other modeling material, like plasticine would be better?

The Pyra isn't on, and it isn't a permanent thing. There shouldn't be issues

 

[Bokkertoff]

they have prototypes right now with few issues with the hardware, the case and some more open points with the software. I would estimate at least 6 month for production (in detail, bugs have to be fixed until sept/oct (aug is holiday season, then you have to scedule mass production, shipment of the finnished parts) and 1 month for assembly and shipment. Also to mention, there will be a first batch one and a first batch two. So only 500 pieces are produced and shiped to customers. If they have them, we have to wait some weeks and see what happens and if there are any bugs left. If so, they have to be removed and parts have to be changed and produced new. If not they can continue producing parts. If this sounds like chaos for you, you are not alone. So the first 500 preorderers will have the pyra around january. if everything goes right ( which was not the case for the last years) the next 500 will have the pyra around april/june 2018. If there are changes to be made, they have to wait for christmas 2018.

This is demoralizing ...
The SoC will be 5 years old. Hope they arent end of life by then ...

 

[WizardStan]

The Pyra isn't on, and it isn't a permanent thing. There shouldn't be issues

It's sticky. Something transfers. I didn't mean to suggest it would be a major issue, just that it could be a minor issue and that this being such a common practice you'd think something that was guaranteed not to be an issue at all would be used. But others pointed out, rightly so, that this is one unit and a prototype at that so manufacturers probably don't care at all about whether it causes any damage, whether minor or major.

 

[levi]

The SoC will be 5 years old. Hope they arent end of life by then ...

ED already has thousands of them in stock at the fab plant, IIRC, and there are bound to me more hanging around in sales channels for a good while yet (although at what price I couldn't guess). I'd imagine parts will still be available after ED's designed a new CPU board and stopped buying TI's OMAP parts.

 

[Grench]

It's sticky. Something transfers. I didn't mean to suggest it would be a major issue, just that it could be a minor issue and that this being such a common practice you'd think something that was guaranteed not to be an issue at all would be used. But others pointed out, rightly so, that this is one unit and a prototype at that so manufacturers probably don't care at all about whether it causes any damage, whether minor or major.

Likely a physical prototype, but not likely a working one. The last board population endeavor was not all perfect after all. ED would have a few populated boards that this would be a good use for.

 

[WizardStan]

Also an excellent point.

 

[DrHAX]

ED already has thousands of them in stock at the fab plant, IIRC, and there are bound to me more hanging around in sales channels for a good while yet (although at what price I couldn't guess). I'd imagine parts will still be available after ED's designed a new CPU board and stopped buying TI's OMAP parts.

I think this is the point of having the modular CPU boards, also I think developing a new replacement 2.0 board for the pyra will be easier once they're out in the wild.

Especially considering Linux' versatility, I know its not 100% easy as including the new drivers and kernels in the Pyra os for the new board, but it should be pretty close. Heck I used to swap a single hard drive of ubuntu back in 2009 between 4 different machines of varing hardware without flaw.

I feel the only thing that could mess up the CPU board is an archeticture change from ARM to x86, but doing so would also ruin our community's software repository, let alone the Pyra OS.

I am not well versed with this field of development, but that is my uneducated blue collar pleb opinion.

 

[Letalis Sonus]

Especially considering Linux' versatility, I know its not 100% easy as including the new drivers and kernels in the Pyra os for the new board, but it should be pretty close. Heck I used to swap a single hard drive of ubuntu back in 2009 between 4 different machines of varing hardware without flaw.

That doesn't work on ARM, you always need an entirely new kernel specially build for your target platform even before you need to worry about drivers. Work has been done to change that, but it isn't of any use in this case AFAIK.

 

[levi]

I don't see why that would be the case to be honest. Not having the fallback of VGA display on everything would be a bit of a pain when trying to bring a kernel up, but you can still do that over some sort of serial link before the heat death of the universe I'd have thought. While it's true that you can run a kernel built for i386 and VGA on the latest and greatest, and use it to build a more optimised kernel on x86, the corresponding arm version from when the i386 was a thing would be the old armv2, which put its status registers in the top 8 bits of PC (leaving 26 bits as an actual program counter), and had a PC-instruction offset of 3 words. So you'd need something a bit newer than that, but if you can get something build for say armv5 or later in 32 bit PC mode I think you'd be dandy for getting something useful out of serial over USB. If you were trying to run it on something armv8 you might need to watch out for the swp instruction I guess.

It's certainly more of a minefield of compatibility than x86, because ARM can't afford to put all of the compatibility transistors in their specs, but to say they're completely incompatible for relatively complex programs is overegging it I'd say.