[Community] [OpenPhoenux Community] New cell phone design wanted?

james garftd at verizon.net
Thu Apr 14 23:26:35 CEST 2016


On 04/09/2016 10:50 AM, Chris wrote:
>
> Hello,
>
> james wrote:
>> I'd be willing to code on that *organized* project, if it existed.
>>
>> I want modern hardware and I'm willing to pay a lot more, as long
>> as I can put the software I want on the phone. In fact, I just prefer
>> to pay $1,000.00 usd and buy one and not have to 'venture' the phone.
>> So please let me know if such a phone exists.
>
> You could order one of those
> https://pyra-handheld.com/boards/pages/pyra/ to test your
> software developments for your use-case.

Hmmmm. Not bad for baseline hardware. Only, not to be rude, but
I hate systemd and therefor am done with debian. I'd rather a minimal 
embedded linux, alpine or gentoo be installed on the box. If they
offered embedded linux, I'd probably get one for for grins, maybe more.

Where is the spec on the crital hardware, including the data-sheets so I 
do not have to go looking?



> The next step, would be to evolve the pyra lessons (also community and
> software distributions) into a mobile phone form factor.

OK, I get agree with this approach, particularly if the device came with 
embedded gentoo, right from the vendor. There is a very active 
gentoo(arm) embedded channel, in case anyone wanted  check it out::

https://www.gentoo.org/get-involved/irc-channels/all-channels.html

Me, I'm a crusted up old hardware guy:: I hate irc as it appears as 
noise to this old fart....

>
>> I can work with one or 2 folks to refine it to a specification. Then
>> we float the idea and see what kind of response we get.
>> I take the financial risks, I'm going to get my seed money back on
>> the first 5,000.00 phones.
>
> Please be aware that there are a couple of folks here that have also not
> just refined some specification but gone through from projecting
> prototypes to selling series of devices already.

Sounds great, maybe those folks should drop me an email on what they 
have, what they are doing and where they are stuck.  My team build 
wireless cellular pay phones decades ago for the S. American market.
There are thousands of low volume products that embed cellular 
capabilities, but they are not advertise as they are for niche markets.


>  Want to seed?

What? go steady without the first date? Sounds too good. Seriously, 
there is tons of money floating around, but I'm very selective about
who and how I 'hang' on deals. Still count me in, at least as a 
peripheral player. Maybe more, depending on what I see an like. One 
thing that is huge for me. Later, I need armv8 multi-core with ddr3/4
ram for an extended feature set, that I'm not willing to talk about just 
yet. But building a baseline cellular dev/kit/dev/hack-box, sure I'm all 
in on that and most willing to follow and keep my issues to a minimum.


> If you really want a more powerful mainline linux phone,

YES


> and are in to fund a venture to accelerate this now,


Sorry, if I misled you about funding others, directly. NO.

> you may be able to fund current
> team members https://pyra-handheld.com/boards/pages/pyra-team/ right
> away,

A materials take-off list for 100% of the hardware you guys are 
proposing to build, would be most keen. As an old_fart Rf connected
industry hardass, I still have a few friends. I have many enemies and 
here in the US, some are well funded. So, no I do not look to get ' out 
in front of these efforts, for a wide variety of reasons. That said
I can help, from the peripheral. Now, like a stated before, a materials 
take-off list (100% of what is needed, down to surface mount components
and the corresponding spec sheets, from the vendors, is what would get 
me excited. THEN, I can 'shop the list' with folks I know and any number 
of companies could build product; that make no difference to me.
I do not know how to be more plain and direct. If you do not have (1) a 
matierals take off and the corresponding spec scheet(s) for those 
components, you do not have a robust specification, you have a box for 
sale.

OK, I'll buy the box, any box, as long as it has an armv8 core, SoC 
details, ram details and the details of the major components that the 
drivers interface with. Missing slots for cellular modem interace chips?
Not a problem, just clearly state that somewhere.


> or alternatively find and fund another established Phone and Case
> manufacturer,

Wow. I guess I'm not making myself clear. I select componts from 
manufacutres of the components and chips (uP,ram,etc) directly. I'm not 
wasting my time on some other build, unless it has an exisiting embedded 
linux stack without systemd.

That said, you have or anyone knows an established phone vendor (which 
opens the door for lots of 'huey') they can drop me an email and we can 
chat directly. I build dev kits for those kinds of folks, write the
bootstrap and  firmware. So I'd never go to those sorts of vendors
and ask to use their design. New components come out every day that 
changes how cell phones are designed. I'm looking to be on the bleeding 
edge by waltzing a design spec under the noses of manufactures and their
technical folks to get the latest in hardware. That's why it will cost $$$.


> and goldelico.com engineering, to work together and
> combine their expertise and access to component sourcing, case design,
> production, free software development and the corresponding custom
> hardware development.


OK they look cool. But, in the past, my teams have built things for 
vendors like that. I know the guys that design the new Rf chips:: BUDS!

So if they want my help, all they have to do is ask. But they look 
competent and self sufficient, so I'm a bit confuse as to what exactly 
you are proposing?

I need multicore armv8, at least 4-gig of ddr3/4 ram and access to the 
key components. Debian is a no go for me. That is what Linaro is 
offering. I hate systemd and it has no place in embedded anything.


> A powerfull, unrestricted mainline linux mobile phone device certainly
> has a market.
> http://plasma-mobile.org
> https://wiki.debian.org/Mobile

I'd take lxqt(5) raw over plasma. Lumina, from the bsd folks looks even 
more suited to minimalistic work.


>> I do not read the list, so somebody drop me an email, when you guys
>> coalesce into some form of a converged specification.
>
> If it is not just investment interest, maybe you are the person that can
> bring together a larger phone manufacturer, that is able to source
> proper hardware at bulk prices for a side-project, with the already
> existing developments from the open source software and hardware
> development players? Connecting the last piece of the puzzle?

Now I think you get it. Please focus and get one or serveral options 
together that

(1) have a 100% material take-off list. This is the list of 100% 
electronic components needed to build the proposed phone.
(2) the companion 'spec-sheets' (or data file) from the original 
manufacture about the components.


THEN I can do majic by shopping where the silicon is 
made/designed/packaged and argue (converserse) with folks I know about 
component substitions and buildouts to test the proposed designs. Yes, 
there is actually quite a bit of magic what I do, but that has happend 
for 30 years and cellular<everything> is right up my ally. In fact, I 
share this tidbit with you. It's really about shopping the (IoT) aka 
California slut style, to get folks with money on board. I already have 
a pre-eminent goups of lawyers (yes butt_hole lawyers) cause that is 
going to be a key component to get this thing build in the US for my 
home market.

Actually, there is absolutely why dozens of similar designs cannot find 
there way to 'pick and place' machines for production runs in dozens of 
different countries.

OK? We cool?




>
> In any way, Nikolaus, does the project software at goldelico allow to
> provide a wiki page for gathering a table of components with say
> desired/available/selected/supported/... status?
>
> A project wiki seems like a good idea.
>
> Cheers,
> Chris

(50) atta-boys for Chris. I knew you guys already had the core team.
Now get me what I ask for. That pathway, getting the (2) things I asked 
for, will get you organized and you will not need me. But, I shall shop 
the design(s) just to ensure that somebody somewhere builds several 
thousand prototypes, so we can being to ferret out hardware stacks
on actual hardware.

You want me to code? Excellent, NO SYSTEMD, is non-negotiable for me. In 
fact, if I may be bold, ditch debian and get a real developers OS,
Gentoo or Alpine. I love Gentoo!

>
> PS:
> The basic prerequisite for to run arbitrary free software stacks and
> distributions are hardware drivers. A fully functional device requires
> that the mainline linux kernel has driver modules for all hardware
> components. (Yes, they may still depend on some closed firmwares.)
>
> As Dr. Schaller wrote, just making hardware and expecting that the open
> software community develops and maintains drivers does not work. If
> hardware drivers are missing, the open source drivers needs to get
> written in order to sell hardware that is "linux mainline" ready
> (usable with common linux distributions). With the golden-delicious
> engineering company he primarily selected components for the GTA04
> board which already had existing drivers, and payed kernel developers
> to write or improve drivers where necessary. Concluding from his
> emails, goldelico ordered the parts and had the board assembled by a
> production plant specialized on custom electronics devices.
>
> Another thing was implementing proper bootloader and system
> installation methods.

I know a coder that runs circles around Russell King (arm embedded 
hall-of-fame) The bitch, as she is called by her friends, has over 300 
product designs and has fixed many problem in arm compilers for vendors 
over the decades. I like what the good doctor espouses above. I need 
armv8-multicore and ddr3/4. Don't worry about the circuit board layout,
I've got a pal that does designs over 100GHz for breakfast, so we'll
'get er done'.

Specs--> material take off --> component spec sheets. Get that together 
and throw up a wiki page or 2.


OK?


James




More information about the Community mailing list