New FP2 Motherboard

You should be able to invite people with their email address, regardless of them being a forum member.

1 Like

Or - @Leo_TheCrafter - you most likely will find them in this thread; though I have no idea how active they are in this forum:

And you will most likely find them on
http://ubports.com/

3 Likes

Thanks for you replies.

2 Likes

I would love to be at the moment when the question of warranty will be an issue: it would mean the project is finished and working :heart_eyes:

Warranty usually covers things like bad soldering (shortcuts, broken tracks, …) or design problems (power supply too small, missing current protection, …), which is the point of that work. Once everything works fine in labs (alpha test), beta test can be run on willing customers like us, then we can think about the future: it could as well be an open-source xGPL hardware design that people download and build with instruction. Like they say in the GPL: “THERE IS NO WARRANTY FOR THE PROGRAM, TO THE EXTENT PERMITTED BY APPLICABLE LAW. (…)

1 Like

So… you just opened the URL, saw the images and then jumped to conclusions… The post-truth era is real!!! :joy: :joy:

For those unaware, they speak of convergence as the ability of plugging your phone to a screen and having a computer (as cool as it sounds!)

Not possible for the FP2 due to hardware limitations.

Oh, man.

Maybe there’s an alternative. Still worth the price nonetheless.

2 Likes

For the precision:

  • “cable-convergence” is not possible on FP2 due to hardware limitations.
  • wireless “Miracast-convergence” is possible ! But it could give a bad user experience due to the ~1 second lag…
  • “DisplayLink adapters” is a working solution as well ! But from what I understood, limitation= no charging of your FP2 simultaneously…
4 Likes

@NeoTheThird is in the UBPorts team.

Indeed! They are very active in the Telegram supergroup :slight_smile:

2 Likes

Thanks for tagging me @oli.sax.

Congrats on the interesting project @Leo_TheCrafter. Looks interesting, but after skimming the thread i have a couple of questions and comments.

  1. Are you 100% settled on the idea of using x86?
    a) FP2 battery life is subpar as it is. Sure atom processors are a little more efficient than the one in your average desktop, but they’ll never compete with any RISC arch.
    b) What standard modules do you want to keep from the default device configuration? Have you thought about the proprietary binary blobs you need for that? Or will you reverse-engineer them to compile them for x86? :smiley:
    c) You could just as well address enough ram with a 64-bit ARM chip.
    d) The biggest performance-breaker on the FP2 is the poorly optimized Linux kernel. Sorry, but that thing is hacked and stitched together, it’s pretty ugly. I don’t want to insult anyone here. Fairphone is a small company with very limited resources, and what they achieved with that is pretty damn amazing. But if you compare the FP2 performance with that of other devices with similar specs (like the Oneplus One), the difference is depressing. That problem would of course go away if you had a mainline linux kernel, but you’ll circle back to the issue with the drivers.
    e) Do people really want windows? Even with a more powerful atom processor, a juggernaut like windows 10 would still be excruciatingly slow. It could probably never be more than a funny gadget, idk if that can justify the price. Linux is considerably better optimized and works on ridiculously cheap ARM hardware. For perspective, i’m typing this on my 100$ Pinebook, sporting a 1.2 ghz A53 and 2 gigs of ram. Try running Windows 10 on similar hardware, i dare you.
    f) Cooling. You referenced the GPD pocket, that one has fans, iirc. How big do you want to make your device?
  2. Since someone mentioned MHL: If size is not a big concern and you’re able to get a native video signal from somewhere else, you could also just slap a native micro-hdmi port on that thing. That’d circumvent the need for slimport, mhl or a usb-c standard with video-out that’d probably require licensing costs.

Sorry, got a little lost in my anti-x86 rant there, so some other points i wanted to make about other things may have slipped my mind. I’ll come back to this later, if you don’t mind.

Please understand my criticism, i don’t want to shit all over your project, just want to give you some things to consider. Even the most beautiful PCB design will be useless if can’t draw to the screen…

Best of luck!

15 Likes

You are not shitting all over the project, Neo. Your reply is worth a celebration.

That being said:

OH, YESSS!
:tada: :firecracker: :confetti_ball::man_dancing::dancer:


Where are you based @Leo_TheCrafter?

2 Likes

Please be aware that I can only answer some of your questions.
a) The Intel Atom processor has a SDP of only 2W, but it can vary quite a bit, beacuse the Intle CPU has got a factory programmed voltage vs clock table thats indvidual to each CPU.
c) Most ARM (based) Processors (like snapdragon, Mediatek or others) use RAM thats layered on top of the processor. It’s still possible tu have something like 8gb RAM, but you’re limited to only one memory chanel.
e) Win10 runs on the GPD Pocked quite nicely. But still most people would run Android or Desktop grade Linux.
f) I don’t think I will need fans, beacuse most snapdragons (Including the one that is already build in) also have a SDP of 2W. And because of the higher efficiency the ATOM processor will reach the 2W not as often as the Snapdragon 801, which still uses 28nm processing.

3 Likes

My understanding of that project is that it’s an “open-hardware” phone so the idea is to get some new chips (LTE, WiFi, SIM/radio, …), so drivers would have to be written from their specs. That would of course be open source :innocent:.

1 Like

Thanks Matthieu for your respons. Thats the concept of the phone and even the CPU has datasheet availeble. There is a 300 pages datasheet for hardware and then a 8000 pages datasheet for software development.

4 Likes

Ok, but what hardware components do you want to keep from the original FP2? I wouldn’t be so sure there’s x86 blobs for most of them, let alone source code…

2 Likes

I plan to reuse the bottom module, the top module, the screen, and the camera. The signals (or most of them) are processed in hardware (eg. the signal processor for the cameras). And for the software development of the CPU, well there exist a 8000+ pages datsheet. you can take a look at it if you have the time.

3 Likes

Naaa…:frowning_face: don’t consider this, we are talking about more performance aren’t we?..
Single channel in 2018/19?

For my digital experience in the last few decades I strongly would prefer to have better performance over more ram. So, of course have dual-channel.

I consider 4GB ram for such a device as totally sufficient. I somehow cannot understand the reasons for some users wanting to have more than 4GB ram in such a device. What are they doing with it? Reaching out for mobile virtualization?? :thinking:
I mean having a computer replacement is one thing, but staying realistic is another. We are talking about a (mid-range) smartphone, not the latest top-notch on the market. Please stay a bit more down to earth. FP is not yet far enough to approach Apple, Samsung or other big boys heights. Three iterations are simply too less for this goal.

I think many users wanting to have 8GB in their handset probably don’t even keep a standard notebook with more than 4GB physical ram. So what’s the deal?
And still this is a Linux based OS not Windows. Yes, Windows may be run on this hardware, but if that’s the goal I am afraid there need to be two hardware options depending on the OS the user prefers to run.

I prefer to stick to (any) kind of Linux based OS, it’s unlikely that I want to pay for a 8GB ram unit knowing that it may take years to come needing that much ram, hence having to pay a noticeable higher amount for more (just in case/nice to have) ram I would not like and probably would decide against a purchase of such an oversized unit.

Actually I am following the launch of the (new) AMD Ryzen mobile APUs e.g.
5 2500/2600 & 7 2700/2800.
There are new notebooks out there (Acer/Lenovo) with a single channel ram design but keeping the stronger APU 7 line next to the the weaker APU 5 line with dual channel design which after all outperforms the stronger APU due to this imo design flaw just for the sake of low-costs. I´m helping to save costs and won’t purchase one such unit. :neutral_face:

Which one you would go for if having to take a decision and also in price comparison? (this one should be easy…).

There are FP2 users out there complaining about a lack of performance. Some complain about regularly running low on ram e.g. while browsing.
Both could be helped with 4GB ram as dual-channel setup.

Considering that all is working in form of more or less large memory shifts, it’s even more vital to have better performance (bandwidth) the more memory is available. I think performance is always the center key to focus on as it directly takes influence in the user experience.

3 Likes

Hi @Leo_TheCrafter ,
I have been reading these posts with interest and have a few questions. Some might seem simplistic, but would like to understand them if you don’t mind:

  • What is the primary purpose of designing this motherboard ?
  • What would be any secondary objectives anyone would benefit from ?
  • How many people would need to buy your card in order to manufacture & distribute them ?
  • Any rough idea on what the cost would be (cost + cogs + your margin) ?
  • You asked for features. USB-C is what I instinctively would respond in order to connect it to obvious peripherals, but I guess that wouldn’t work as you would need to cram-in the connector somewhere… ?
  • What is the current status of your project?
  • Why on earth would you want to have a phone running any M$ products? On traditional hardware they are a nightmare already, let alone on reduced specs as in a phone.
  • I noticed a few ‘good luck’ comments on the UB-ports website. Any evolution in that? It would really be key to ensure to have a good relationship with some SW folks in order to have you MB light-up any OS.

My 2cts. really.

6 Likes
  1. I want to update the FP2 so that people who want a Fairphone don’t have to live with a lack of performance.
  2. This would mean that there will probably be software updates forever, because there will always be a newer Linux version as long a the specs support it.
  3. Roughly about 300-400€. I don’t want to talk too much about pricing, but my price margin won’t be high because I use high quality components (Not only the obvious one like RAM,CPU, but also high qulity SMD components)
  4. I can make the Board USB-C compatible, but there won’t be any bottom module to support it.
  5. I have already finished most of the board design, but the are still some minor things missing, like BIOS. Currently im drawing the scematics, which will (most parts) be open to everyone.
  6. Most people will probably run Linux.
  7. I think some may have misunderstood my project, because it’s not compleatly new. There is a Pocket-Laptop called GPD Pocket, which can perform and run Linux aswell as Windows very nicely.
5 Likes
  1. sound just ok.
  2. 5-10 years would be fine :smile: ; well still it’s a phone, but software should not be the limit as we actually have with Android 8+
    The good thing with standard Linux is that there is hardly a piece of hardware out there which it does don’t support. Even my main machine is 10+ years old but runs smooth with Linux Mint. After all it has plenty of generic drivers.
    Still there would be enough space for tinkerers to have a play with your board bending the kernel to their needs.

Sure, anything else would speak against longevity, which this is all about. The price range sounds decent, slightly higher than the actual core module, but also more advanced.

  1. Actually I don’t know why many people are so eager to get their fingers on USB-C. I think often it’s due to the lack of more in-depth knowledge about the technique. The actual FP2 core module’s SoC could do quick charge, but the board design does not utilize this feature. So if it was only for quicker charging utilizing the QC feature for USB would do as well.
    But yes, USB-C is more versatile and in this forum we could read about users wanting to use the phone as replacement for a computer. Since there would be a standard Linux running such functionality could be made available.

Bottom module…I don’t know if you have more technical details on the actual FP core module and bottom module.
Maybe there could be done some preparation on your board? A final USB design for the actual module should not be too much work I think and a partial USB-C design in correspondence of a new module (2nd revision by FP supporting USB-C) so the main parts could be designed in the module itself but anyway connected to your board. Just an idea of mine that could work.

2 Likes

It is important to differentiate between USB-C as a physical port and the protocols that run only over USB-C. The Intel processor has got 2-3 USB-C/USB 3.1 to be correct ports.

3 Likes

I have to tell you that I will be using a 3G module instead of a 4G module. Th egood news is that 3G is only marginal slower than 4G and consumes less power. The reason for all this is that I just realised that my 4G module has no fallback 3G or 2G. The Downloadrate of the 4G module is 10mbps and the 3G module has got 7.5mbps. I want to hear if this is a issue or if you like that move, because some would prefer reliability over speed and 7.5mbps isn’t exactly slow, is it ?

2 Likes