Build our own device
 

I was sat in the pub a few days ago and generally having a moan at the state of the mobile phone industry with someone I know who sources stuff and gets things manufactured in China/Taiwan in a totally non-phone related industry. The challenge I had was "Why don't you get someone in China to make the phone you want?"

For me that's something like an N9 (or Sony Z compact) with modern mid to high end internals and a removable battery but I realise for others it's an N900 replacement. Let's not get hung up on that now.

Suitably lubricated we fired off some emails.

A few emails later and we've offers to supply some fairly generic Android based handsets and preliminary discussions about custom work. That wasn't difficult at all but just a start.

So, really, how hard would it be to spec something for an "easy" community based device ?

Presumably something Qualcomm based rather than MediaTek so that a Sailfish port was easier.

The minimum order numbers don't look massive to me. Certainly doable with a kickstarter campaign.

So, punk ethic to the fore, why aren't we doing this ourselves ?

Re: Build our own device
 

Well, have you heard about the Neo900. I think to make your own device isn't possible. At least if it should be affordable and up to date. The Neo900 is still not ready. When it is ready (if ever) it is not outdated it's an oldtimer.

Re: Build our own device
 

Isn't easier to port SFOS on Sony Z5 compact instead?

Re: Build our own device
 

Having a phone made is easy, there are plenty of willing Chinese factories, well, not factories, but I guess you can call them resellers. Hovever, some of them can deliver whatever you want.
Even in lower quantities (some 1000 devices).
Copyrights, well, not sure lol.
You should also be aware that what you might get is not what you asked for (at least OS vise)
Drivers, sure, for what they deliver.
Source code? No way.

Re: Build our own device
 

Or:
1) we list all interesting devices with the possibility to port SFOS on it
2) open a vote thread
3) the most voted phone is our new sfos device
4) ask jolla what will it cost to get a license for with Alien Dalvik Support
5) hope that jolla Support in the future this phone, too

Re: Build our own device
 

If it was this easy, we'd already have a Neo900, I guess.

Edit: For clarification:

It's like nieldk already said: We can totally get a device made by these guys, the problem is, that this device won't be the device, most of us are looking forward to. We won't get any sources or special-optimized OS with it.
As far as I know, the (generic) hardware gets bought in bulk and gets thrown into a big magical machine, where in the end, smartphones fall out...no seriously, we would probably get yet another Chinese android phone/clone, running android. :/
Guess why there are so many Chinese smartphone brands currently floating the market...

Re: Build our own device
 

We don't even have to build the device. Let's do like Jolla an create some images, a Indiegogo campaign. then we realize that you need millions to build the device...then we simply buying a few neo900 for the money we got and use it in coding competitions.

Re: Build our own device
 

My thoughts went along similar lines. We can't even have a decent port to devices that exist. What makes you believe that making yet another one would make a difference?

This is BTW kind of my reservation to Neo900 too (other than the device itself has turned out to be a vapourware, that is). Even if it had seen the light of the day it would still be a useless piece of silicon, metal and plastic without an OS. Yet I see even less signs of that ever happening than finishing the hardware.

Re: Build our own device
 

Isn't Mediatek the one that actually releases sources (eventually)?

Re: Build our own device
 

I have and OpenMoko and such like. No offence to the people doing that but it strikes me as an almost impossible task with the outcome being a slightly better N900 that is still years behind what is expected today... and then there's the price.

Probably but they've stopped making those now. It's replaced by the X Compact which incidentally looks like a slightly more oblong N9 - I'd love a port to that but would we get all the hardware working? The camera alone seems like a challenge to get the OIS and laser focusing supported. Maybe better to start with something with a less fancy camera.

Plus the idea was to spec base hardware and let people with 3D printers mess about with form.

neildk - I'm sure you have to be careful with specifying what you want delivered hence this topic. If we can spec it all out, including deliverables such as drivers then we might have a chance at a community device with less binary blobs that cause issues later on.

cvp - the problem with a poll is that there's one winner and like we've had in the UK recently, it's not always the informed, knowledgable people that win. It would be a waste of time if device XYZ wins but it's impossible to port to. The idea was we spec core hardware (eg. a Qualcomm CPU 6xx, Adreno GPU xxx, modem xxx, sensors etc) and then the form factor such as keyboards, screens, battery sizes is something we can fiddle with after device no. 1 is done.

Re: Build our own device
 

If so, I've a list of flippin hundreds of handsets I can have customised costing about $50 each. :D

Re: Build our own device
 

They're speccing an entire phone platform from scratch and not using libhybris so the choices they have are quite narrow. I was thinking more along the lines of finding an existing phone model and running with that and modifying it to match.

Which is exactly what I want to use and abuse. It's exactly what Jolla did and what Intex did. You start with a known working Android phone where all/most of the R&D has been done. That's how you get a cheap, rapidly developed, modern phone.

But, what needs to be carefully done is working out what the base "donor" device spec is so that ports aren't going to see red boxes in the Mer port table.

Re: Build our own device
 

The blobs is the issue.
You won't get the sources of those.
This is what is blocking this kind of development.
Qualcomm et al will deliver the binary blobs for what your need is (currently, updates will cost), given payment, but never the sources.

Re: Build our own device
 

Theres lots of available devices if you look on Alibaba.

But they will be running android, and of course the kernel sources wont match the blobs they give you and then theres all the porting work.

Whilst its not that hard to imagine something like the turing phone but without all the weird stuff, i honestly think we're much better off with the Xperia XC and using the rather nice information that sony provides to get a good port together.

Re: Build our own device
 

By the way, what's wrong with the fairphone? Why do we need another device? For me official support and alien dalvik would be enough.

Re: Build our own device
 

It's extremely expensive (and also too big).

Re: Build our own device
 

And you think building your own device is less expensive? At least there is a very good reason for this high price.

I'm with you, in respect of the size. But I think many others not. I don't think you will find a device that fits the needs of all community members. We want a smaller devices others a large screen. I want a hardware keyboard others not. Some people think the camera is really important others not.

So I know it sounds disenchanted. But building the perfect community device is an illusionary dream.

Re: Build our own device
 

The FP2 is a highly customised device, taking an OEM offering with zero modifications would definitely be cheaper.

I agree that getting everyone to agree on specs is nigh on impossible, but not everyone has to agree for it to be a worthwhile venture. The point of this would be to provide something that the industry at large isn't already providing, so if you want a >=5" screen then you already have lots to choose from, some of which already have working Sailfish Ports.

The hardware keyboard is a sticking point for a lot of people, and I really wonder whether you can get an OEM device with such a thing without it costing Neo900 level prices.

Re: Build our own device
 

https://www.youtube.com/watch?v=8eaiNsFhtI8

Would be interesting to see a project like this done with a raspberry pi zero and a nice 3d printed case.

Re: Build our own device
 

Don't expect devices like a N900 or N950 in the near future.

Re: Build our own device
 

Sadly, Raspberry Pis are quite power-hungry. You'd need a really large battery to actually make a smartphone you could use for a full day. Notice that the guy in the video doesn't keep it powered up all the time.

There's a reason the GTA04, neo900 etc have not had great successes. Building a cell phone just isn't something a few friends can do in their spare time.

Re: Build our own device
 

I'd like to think that part of the problem is that the usual operating systems that are installed on Raspberry Pis lack the power management that something like Maemo has. This is a software problem and is fixable.

Also, the Raspberry Pi Zero which was suggested, draws significantly less power than the traditional model B series. I'd personally like to see a phone built from the Pi Compute Module (draws a bit more power than the Zero but not close to model B). It's easier to embed than the Zero and can potentially be upgraded. A new Compute Module based on the Pi 3 is due in a few months.

There's a lot going for the Raspberry Pi and it's community is huge compared to ours. I think a Pi phone has potential to be successful if done right. I bet the Android Pi community would love to have a Pi phone.

Re: Build our own device
 

Don't forget that the price includes continued open support.

Fairphone is doing most of the things that people on this forum claim they want a manufacturer to do (open access to source for the hardware as far as they can etc.).

I guess it shows how much all the talk about "open hardware is the most important thing" is worth, when people start complaining that it costs more than the competition ;)

Re: Build our own device
 

Suspend power consumption is dominated by ram refresh.
There are hardware differences in ram, lpddr consumes 3 times less than ddr.
On an atmel a5 for example suspend to ram with lpddr is 9-15 ma @ 3.3V, with normal ram around 33-40 ma.

The Pi Zero is currently at 100 ma (AFAIK). Would be amazing if someone could fix this to 10ma, but i wont hold my breath on this issue. And the newer versions, powerwise, are going in the wrong direction.

So if you want to build something in the pi range with power management you might consider something like this
http://www.micro-technic.com/Webshop...A5D3X-MT0.html
or
http://www.emacinc.com/products/syst...dule/SoM-A5D36

lpddr2 seems to be a rare commodity, somewhere i read the DIPFORTy2 "All Programmable Linux" is "delayed", because even 64MB lpddr2 seems hard to get.

Re: Build our own device
 

Does the Pi have multiple RAM chips? If so, I wonder if some sort of hybrid suspend would be possible, where all but one of them are swapped off to flash memory and powered down.

Re: Build our own device
 

Apparently the Pi 3 uses LPDDR2:

http://uk.rs-online.com/web/p/proces...-kits/8968660/

Of course it now has other niceties like a faster CPU, wireless LAN and Bluetooth which consume extra power.

Re: Build our own device
 

Yes, the Pis are optimized for low cost, not low power. They allow a whole new class of projects to be done. But you can't just tweek the software and make them low power too.

Re: Build our own device
 

http://www.ingenic.com/en/?newton/id/13.html is one of the more interesting kits I've encountered over time.

Re: Build our own device
 

It's a great idea and I'd love for all manufacturers to go in the same direction so prices would be lower, but for now, if I can't afford it, I can't support it. If we want to talk about 'our' device, it should be something which doesn't cost a fortune, otherwise whose device is it really?

Re: Build our own device
 

Hmm mix that idea with a provider (GSM) service, that profits from bitcoin, etherum or similar mining. Giving free talk/data service world-wide ?

Re: Build our own device
 

I'd would love such an initiative, but there is the already mentioned question of the blobs, but also a not mentioned one: how to get an agreement on the phone external design?

some would want a hw keyboard.
some would want a 5" screen
some would want a 4" screen
etc etc

Re: Build our own device
 

It was stated before but there are already countless devices out there with big screens so the only thing they need is a Sailfish port, and we don't need yet another 5+ inch slab to add to the pile. Almost none have a smaller screen, and none have a hardware keyboard. One device with a 4- inch screen and a hardware keyboard wouldn't be a bad idea. Put some Apple or Google marketing people on it and it'll be hailed as 'Innovative Small Smartphone With Real Keyboard Ready To Take The World By Storm'. It's certainly an untapped market.

Re: Build our own device
 

Thread discussing power consumption of the rpi zero

Rephone might be better as a starting point. Some interesting projects are discussed on their forum.

Re: Build our own device
 

I think nthn is right that all the slab hardware is out there, what isn't out there is a small device (Xperia X Compact is the best you'll get and even that isn't N9 small) and keyboards (much beloved by this community, although personally I can take them or leave them) have all but dried up too.

So we're going to ask a Chinese ODM to make us a Cyanogen-powered Lauta, right? Then feverishly try to get the hardware adaptation done before the thing becomes obsolete, which probably (unless one of us is a hotshot with acres of free time) means paying someone to do it full-time (if it's even possible with the level of non-cooperation you'd get from your average ODM), then paying Jolla to license SFOS?

Sign me up.
I think.
Oh and make it an OLED screen so we can get our LPM back please!

Re: Build our own device
 

I vote Xperia mini pro :)

Re: Build our own device
 

There are no Sailfish OS MediaTek ports yet, but it's possible given that Ubuntu Touch works on MediaTek devices and there is going to be official SFOS device from Oysters with MediaTek CPU. The port is doable as long as they can give kernel sources (Android device tree would be nice too, but is not a must since most of MTK SoCs have device trees leaked already).

The question is what kind of customizations can they offer and what will be the price/minimal order. Hardware keyboard is going to be hard, probably, since there are no such devices manufactured currently.

Re: Build our own device
 

I also agree there are plenty 4.5" and bigger smartphone on the market.

But, let's not talk in term of screen diagonal, when we say we want a small screen, I'm pretty sure we mean we want a 60 mm maximum wide phone. With such a width, we can expect a 4.3" 720p screen.

If we want to make things a bit more realistic we've to choose a hardware which is the closest possible of something already existing on the market. So, personaly, I don't vote for a HW keyboard.

And we've to find how to get a realistic external design.

Re: Build our own device
 

If the port issues with the keyboard layout and some other issues can be solved, I expect there will be a sudden surge in second hand Xperia Mini Pros sold - even though it has an old processor and not a lot of RAM. For anyone interested, there are a lot of Xperia Mini Pros being sold on the Polish and Portuguese versions of OLX.

Re: Build our own device
 

nthn, if xperia is ok for you then fix keyboard layout by yourself. Sailfish use xkb so it's "easy".

Re: Build our own device
 

Maybe we need to convince sony to make a new Xperia Mini Pro :P