I thought that, too. But then I came across this article:
https://liliputing.com/2016/05/nxpfreescales-64-bit-mx-8-chips-finally-coming.html
NXP is positioning the i.MX 8 chip family as a solution for connected vehicles and industrial or medical products. So donāt expect it to show up in a lot of consumer devices anytime soon.
That combined SIM/SD card slot is a bad habit of some phone manufacturers. What if the user wants to use two SIM cards and a SD card simultaneously? In my opinion, this should be abandoned in favour of separate SIM and SD card slots. Also, you canāt save much money there. You still need two SIM card readers and one SD card reader.
Please not! What if a user travels regularly between two countries and has to switch the 3G/4G SIM card at the border? Also, I donāt think that you can save much money with that.
Well, depends on when itās really coming and how much it will cost.
Iām not up-to-date on the etnaviv from, but Iād guess we could get the G7000(-lite) running with reasonable efforts (maybe less than 2ā¦3 MM). IPU and VPU donāt seem to differ much from the previous versions, so I guess ptx folks will catch up quite soon. Anything else should be mainlined pretty soon, too (IIRC fsl folks already submitted patches on lkml)
An interesting question is the power consumption and how fast it can switch between operation modes. I never used imx* on lo-power devices, so Iāve got no idea about it.
You can save space. In such small cases like smartphones, such slots make up a lot.
By the way, Iād like to add another requirement: very robust (IOW: industrial grade) connectors. I just have problems w/ my toshiba notebook where all USB connectors now have loose contact. Smartphones usually have only one - and if the usb slot its broken, you can the throw it away (I doubt anybody seriously replaces usb connectors in smartphones)
The logic is already in the SOC, so it doesnāt matter anyways. The only external parts are PHYs and slots. I doubt (havenāt checked) that combined slots are more expensive than two separate ones (actually, should be way cheaper). And Iād guess thereāre even combined PHYs, that even can detect the card type - maybe even 2 in one chip, or integrated in the socket.
I think you would be surprised to see what this community is up to 
Depending on whether itās still under guarantee, you could just order a new bottom module for the FP2 for ā¬20. Donāt get me wrong, I think using durable components is a very important feat, but equally you canāt avoid all defects. FP2 succeeded very well in having a repairable design imho, and Iād personally really like to see this continued in the FP3 and subsequent phones.
Ah, didnāt know that the usb connector is on an extra module. Thatās a great thing 
By the way: wh not just develop a new baseboard w/ some better SoC ?
That would include writing new drivers for all the components and the SoC is not really the weak point of the FP2. The week point rather seem to be Googleās dumb upgrade restrictions.
If they make a new main board with a different SoC, they can in the same go make a smaller screen and further improve the modular overall design.
In the end, I think it is also a question of customer āconfusionā: Which modification is enough to justify a version jump from (Fairphone) 2 to 3? Is it simply a different SoC, while the rest of the phone stays exactly the same? Or do other qualities of the phone also need to be modified to make the change recognizable for customers?
What I donāt like at all is the āalphabet confusionā Samsung creates (links to Wikipedia): Samsung Galaxy A, C, E, J, On, S, Y. In many cases they even have a 2016 and 2017 edition (or 2015/2016). One aspect that shouldnāt be forgotten in this regard is the aftermarket. The many versions and models make the smartphone landscape extremely fragmented and hard to maintain for community developers.
Just a side note: The iMX is originally by Freescale (US) and came to NXP when they bought Freescale. Some months ago Qualcomm decided to buy NXP and if no miracle happens the iMX will come from an US Firm again.
Most of it should be supported by mainline kernel, so itās mostly just writing a new DTS and config for the new base board. Pretty much daily business for us kernel hackers.
Whatās the big problem here ?
Take a recent android tree, put in new kernel (w/ board specific config) and new mesa.
Canāt be that big deal.
That would imply other changes, eg. new case.
IOW: a really new device.
Simple: When modules arenāt fully compatible anymore.
Iād call the variant w/ just new baseboard 2.1.
Simple: within each generation, the individual components are versioned separately.
Eg. if the cam module is upgraded, it will be version 2.
The only thing that might require some more thoughts is proper module probing
(if the modules dont already come w/ some generic identification mechanism, eg. a tiny i2c eprom, etc). But thatās all solvable, once the specs are on the table.
If everything so simply I guess we will see these changes in the FP3. 
So finally qualcomm become the new Intelā¦ 
I think, they would not bother if they can sell more of the chipsā¦ And as they use it for medical devices, long-time support will be granted (by law!)ā¦
Iāve been using LineageOS for a few weeks now and it works quite well. There are some issues but I had issues with the original OS as well. If you are willing to experiment a little bit and flash LineageOS than you can have Android 7 right now.
More info:
Yes, but not for Android Updates, that is for sure!
Now, as an Fairphone owner, but also being hungry for power and Android Updates, the FP 3 definitely needs a new processor. There is no guarantee, but it should be state of the art at release time. Throw in three or four GB ram, a better camera module and LTE on both slots ā done. The more I think about it, itās like the S models at Apple. The new slim case, the displayā¦ all great with the FP2 already, keep them ans the battery for compability . If they can keep some more modules compatible, the better.
Probably just design a new main module and a new camera module might already do the trick! All FP2 owner could just exchange the mainboard, if they want to, but they have not to! And if all modules of FP2 and FP3 are compatible with each other, you donāt get into trouble with spare parts as with FP1ā¦
That is possible, but manufacturers hardly use this.
See e.g. the Motorola Moto G5 Plus: two sim-cards on one side of the tray, the memory-card on the other side.
Motorola, did it right.
Dual SIM doesnāt help mutch if you canāt use them, when you need an micro SD.
FP3 would sounds great, but I think we need a The Fairphone or a Fairphone X, where we can change as much as possible, so about 85% should be modules, so every part could be updated.
Ben
Essentially Fairphone should base the next generation on the FP2 to minimize loss in manpower, parts and support. Should the FP3 be a completely new design I would even bet on failure of the company, as the support for FP2 will then most likely face the same obstacles as it did for FP1. Too few customers to enable further production of spare parts and batteries and no more software updates, as the number of working FP2 is shrinking.
The modular concept is a good one and any further steps should build on the same foundation, not just on the same idea.
Just my 2 cents.
