ED P2 successor notes - and a suggestion.

[Grench]

In this thread/post ED states that he potentially has interest in creating a Pandora successor.

http://boards.openpandora.org/index.php/topic/14012-trying-to-find-a-fair-solution/?p=265585

If it happens:

It will have another name.

It will be independent of Craig.

There will be no orders accepted until devices are ready to ship.

The funny thing is, if ED came out tomorrow and said he had it all figured out, gave solid specs and was taking pre-orders for under $1000 each for his version of the P2, I'd likely be willing to pre-order it.

So - cheers to ED for considering it!

ED, unless you've already got it sorted out...  Maybe consider creating an invitation only 'advisory council' of whomever you consider to have the best ideas on here.  Then put them under an NDA, and get some collaboration going?

The ideas here in the P2 forum range all over the extremes - which they should.  Maybe a small working group could assist in distilling those down to what is high performing yet practical, efficient to produce, meets functionality needs and aesthetic needs, while still coming in at an 'affordable' price point.

I'm looking forward to seeing what the next iteration brings.

 

[Foxgod]

Same here, as much as i love my Pandora, i also love to see the next one, i will follow these forums like a sniper wolf hunting a snake when the P2 days are on us.

I rather spend that money on a P2, then on any apple device.

 

[Timstertoo]

If it's a portable clamshell with a keyboard and gaming controls I'm in. I'll leave the designing up to the rest of you.   

Although, not being held back by any knowledge on these matters an option for a mechanical keyboard would be stellar. Otherwise just have a good look at the Psion devices for other ideas for great portable keyboards. 

 

[_wb_]

As far as I understand, ED wants to make a Pandora successor that is not radically different in design/concept from the current Pandora, but just upgrades all the parts (screen, SoC, etc) and addresses all the issues with the current Pandora.

The most important thing to decide on is the SoC. The problem with SoCs seems to be that vendors don't want to sell in low quantities, because they're afraid it will cost them too much effort in support. So the question is not: "what SoC is the best?", the question is "what SoC can we get?". If anyone has good contacts with SoC companies (e.g. Qualcomm or Samsung), please tell ED.

The "advisory council" thing is not a bad idea, but I would prefer to just keep discussing things on this forum. To make the discussion more "official" and focused, I think the best approach would be for ED to start threads on all subjects he wants input on, perhaps with polls when the options are known.

 

[kasp]

I would say make the device as sleek and sexy as possible.

That will give it a larger market.

 

[_wb_]

I would say make the device as sleek and sexy as possible.

That will give it a larger market.

Assuming we want the same or better functionality, it's not going to be much sleeker (in the sense of smaller/slimmer) than it is now - the current design is pretty compact already, so not that much margin for shrinking stuff. Also for reasons of ergonomics, I wouldn't make the unit (at least its base) too slim.

But of course it can be sexy as hell

I don't think the looks matter that much, the functionality and how it feels in your hands are way more important. And I think most of the Pandora successor target audience also prioritizes functionality (which includes grip) over beauty. I'm not saying it has to look ugly, but beauty should be a secondary design goal that shouldn't reduce functionality in any way.

 

[Klumpen]

Another type of keyboard would make it more slim (and it would speed up writing).

The physical keyboard of the P1 is the only design flaw for me,

but back then, all the good flat tiny keyboards were not out there, I think.

Well, I know that some people with big thumbs like the current design though.

 

[Mr_Loon]

@ Grench : What do you think OP GmbH would gain by having a select group under NDA, compared to being open and trying to engage as many Pandorians as possible.


I'd say go for the latter option and cherry pick the best ideas.


To have a select group working behind closed doors and to present the final design as a fait accompli would be a mistake imho.

 

[ghostpatch]

As for the specs versus price, we should go for a compromise to not screw any group completly over. Let's prioritise on parts availablility, coast efficincy and quality.. I'd also look at the drivers beforehand - are they open source or at least good quality?

A lot depends on software optimisation, software, support etc. (needles to say that won't be a problem - too much awesome people)

; D

one thing, plz keep the awesome Burr Brown DAC or select an even better one! I have a lot of fun listening to music on my Pandora ; )

Also I'm not scared of preordering, once the design is finalised and we need funds for parts to get going!

 

[hipnotoad]

I for one am satisfied with the current pandora's overall design. Updating the technology and addressing it's current flaws would be enough to have me sold.

 

[WizardStan]

The problem with SoCs seems to be that vendors don't want to sell in low quantities

Yes and no. There are vendors that have overstock and have no problem selling low quantities, but then you don't know when you might be able to get more. ED could source several hundred, maybe a few thousand, and then everything dries up for 6 months. So it isn't that you *can't* get SoCs in low quantities, it's that you can't get them with any sort of regularity, which ultimately amounts to the same thing but is still a worthwhile distinction to avoid "here's a company that has 1000 parts in stock!" overly hopeful posts.

 

[EvilDragon]

The only company I know selling SoCs without any contract and NDA is Freescale. We could easily buy a quad-core Cortex A9 1,2GHz.


There's no company selling a big.LITTLE without any NDA and contract, and .most won't spend ressources unless you want at least 100,000 of them.


And even if you could get them with a workaround, you still need the design guide and drivers for Linux...


So yes, a next-gen SoC is hard to get, an A9 would be easy to get.


But I want to try to get a big.LITTLE one - so before I didn't try everything here, I will not revert to an A9.

 

[Mr_Loon]

The only company I know selling SoCs without any contract and NDA is Freescale. We could easily buy a quad-core Cortex A9 1,2GHz.


There's no company selling a big.LITTLE without any NDA and contract, and .most won't spend ressources unless you want at least 100,000 of them.


And even if you could get them with a workaround, you still need the design guide and drivers for Linux...


So yes, a next-gen SoC is hard to get, an A9 would be easy to get.


But I want to try to get a big.LITTLE one - so before I didn't try everything here, I will not revert to an A9.

Forgive my ignorance but what do you mean by Contract & NDA?


I was going to suggest using Samsung’s Exynos 5410 like the Odroid XU after reading in the article that 'You could buy this sort of device to use as a standard Android or Linux computer'. Guess they must have ordered more than 100,000

Basing P2 on a SoC that is outdated by at least one generation upon its release would make it a very hard sell as a premium priced device IMHO.

 

[EvilDragon]

Samsung Germany wants to support me, but Samsung Korea don't. They won't sell their SoC to us, since they think it would need too many of their ressources.

You can't just buy these chips off the shelf.

Contract: We'd need to sign a contract to be able to buy the chip..
NDA: Non disclosure agreement. Would be needed to get the techsheet of the SoC. Without that, you could't even design the device...

 

[vcoleiro1]

Samsung Germany wants to support me, but Samsung Korea don't. They won't sell their SoC to us, since they think it would need too many of their ressources. You can't just buy these chips off the shelf. Contract: We'd need to sign a contract to be able to buy the chip.. NDA: Non disclosure agreement. Would be needed to get the techsheet of the SoC. Without that, you could't even design the device...

That's probably a good thing, Exynos have had there fair share of over heating issues and suspected thermal throttling.  Personally I wouldn't even touch them for a P2.

 

[Mr_Loon]

Samsung Germany wants to support me, but Samsung Korea don't. They won't sell their SoC to us, since they think it would need too many of their ressources.


You can't just buy these chips off the shelf.


Contract: We'd need to sign a contract to be able to buy the chip..


NDA: Non disclosure agreement. Would be needed to get the techsheet of the SoC. Without that, you could't even design the device...

Would it make sense to approach Odroid and see if you could use a slightly modified version of the Odroid XU as a 'motherboard' for P2?

By slightly modified I mean exactly as per the off the shelf version but with bare headers where things like USB & HDMI are?

 

[vcoleiro1]

^ The Odroid XU is slightly too big for a Pandora being 94 x 70 mm  - forgetting the port heights. The Exynos 5410 has serious issues anyway - you wouldn't touch it.

If worst comes to worst, and ED cant get a SOC, (I'm sure he'll get one though), he could always try getting PC's sticks PCBs(which will fit) and working out a way to hook up a touch screen to them, perhaps one of the chinese manufacturers would make a custom one.  Actually, now that I mention it, I know they do that, I have seen somewhere where they talked about customisation options.  Might be the way to go IF ED cant get a SOC.  

Certainly be the cheapest option considering you can get a RK3188 Quad core PC Stick with BT4, Wifi + External antenna, 2gb ram, 8gb internal storage, HDMI out, OTG + USB, and micro SD + cables all for $62 inc delivery at retail.  Imagine what the wholesale cost would be. If there was a way to engineer the P2 to just plug a PC stick into it, I'd say that would be the best option, certainly future proof.

 

[ghostpatch]

That's probably a good thing, Exynos have had there fair share of over heating issues and suspected thermal throttling. Personally I wouldn't even touch them for a P2.

Samsungs Exynos 5 Octa (I belief it was the 5410) had a serious silicon bug which led to performance issues. But Samsung has two new big.LITTLE models in the pipeline (bug fixed of course)

The heat caurse is probably the Cortex A15 - its powerfull but also takes a lot of power

The A7 is more interesting to me because its smaller, faster and more power efficient than the A8 used in the OMAP! ; )

By the way the A7 is better than the A15 when you just look at power consumption versus processing power

Personally I boubt the need of 8 cores. 4 cores (2x A7 and 2x A15) would be enough. You simply can't parallelise everything

 

[ShleeDragon]

Hopefully in 12 more months... there are more A15 cores on the market. and ED will be able to source a supplier. 

Good to see ED is looking at setting the bar at a SOC with at least 2x2 big.LITTLE and some kind of recent GPU.. I'd love to see GLES3

 

[rohezal]

Personally I boubt the need of 8 cores. 4 cores (2x A7 and 2x A15) would be enough. You simply can't parallelise everything

I am very interessted in the open cl code. When you can write your own rasterizer for the gpu of a console (lets say Dreamcast) you can use almost the whole processing power of the gpu. This means more cpu time, means more consoles to emulate. The rasterizing / shading is the biggest part on emulation, so it would be great to let the gpu do it, since it is best in doing this things.

Maybe do vertex transformation in the gpu (not so hard, just one matrix multiplication per vertex) and the rest in gpu. Then the cpu can handle the code (ai, collision detection, sound) and the gpu most of the graphics.