Update on Librem 5 hardware?


Certain? No. This is a phone that isn’t released yet.

It was commented in this forum though by someone, based on the published schematics: 3G/4G modems - Alternatives?

Yes, I knew that. What I didn’t know, without researching, is exactly which set of GPS services that GPS device supports. Looks like we have answered that issue then.


Within this link that confirms Quectel EM06-E certification is to be seen standardized modem card size (30x42mm) and its M.2 “B key” edge connector (if Librem 5 main-board will support this one as optional).


My bad. That will make make the European, Chinese and Japanese nationalists happy! :wink:

I should have clarified that I was talking about the USB 2.0 connection on the Gemalto modem, not the M.2 slot. There is no reason for Purism to limit the M.2 slot to USB 2.0, so future upgrades to modems with USB 3.0 and G5 will probably be possible. One difficulty is that you will also need a different antenna for 5G (especially if using mmWave), but I see that Qualcomm has figured out how to make tiny 5G antennas, so it will probably be possible to include both the 5G chip and antenna on the M.2 plugin card, but to get good 5G reception, you need to include 5G antennas on all four edges of the phone, because your hand can block reception if you only have antennas on one or two edges of the phone.

The more that I learn about 5G, the more that I think that it is a really bad idea. It will be very expensive and incredibly wasteful to setup millions of microstations with only a 250 meter range, and we have no idea what will be the health effects for humans living next to all those microstations. The way that 5G will be implemented in the US with mmWave sounds like a nightmare to me.

The one cool thing is that the large size of the PLS8 chip package means that Purism will have to use a standard-sized 3042 M.2 card in the Librem 5. Because the PLS8 is 2mm thick, plus another 1mm for its circuit board, this means that the Librem 5 will probably have a thick case (I’m guessing 12 mm or thicker), but it also means that we can buy a standard replacement cellular modem from any supplier as long as there are Linux drivers for it.

This is frankly awesome. I guarantee that every review is going to criticize the Librem 5 for being too thick, but I am going to love having a phone that isn’t designed for planned obsolescence.


Yes, the mmWave part of 5G is terrible, but that’s just part of it. Don’t throw out the entire spec just because of one of the frequency bands.

You only need a new antenna if you want to use the higher frequencies (along with their pathetic range, which will be a wonderful boon for those who track your location based on cell towers). They’re going to re-use the existing frequencies for 5G signals once 2-4G get retired in the future, and an antenna which works for one frequency now will work just as well for that frequency until the end of time (assuming that it’s not falling apart).

The newer signal encoding methods which it uses have a higher spectral efficiency than what we have now, translating directly into a higher data throughput (so: either faster data transfer, or more concurrent users at the same transfer rate). That part of the spec is good and actually useful.


@Caliga They likely won’t connect the Gemalto PLS8 GPS antennas which should render it neutered. Even if it does power up the GPS function it shouldn’t be able to pick up anything as without an antenna the signal gain will be below the noise threshold. Looking at the datasheet I don’t see any references for an on-board antenna so leaving it unconnected should be sufficient.

Could you confirm this @nicole.faerber?


Yes, this is the idea. For some reason pretty much all current cellular modem modules have a GPS/GNSS function included, which we do not want to use for obvious reasons (starting with the modem having access to precise location data and the mystery code of the firmware being able to (mis-)use it). Luckily most modem have a separate antenna input for GPS/GNSS which we will deliberately leave unconnected.

Instead we implement a TESEO LIV3 dedicated GNSS chip which is pretty new and supports all current constellations and can track up to three constellations simultaneously. Which three can be determined by software at runtime. That should give pretty good results (with Galileo soon offering down to 20cm accuracy this will be pretty cool even :slight_smile:



Hi Nicole,
since there is still all the discussion about add-in cards; could we get any information on what is actually used?
Gemalto only ships the PLS8 as a LGA surface mount package, to be integrated directly on the PCB.
Add-in cards will have to be sourced from a third party vendor.

Which will it be, and if it’s indeed a card, what vendors are in the race?


Did you check out the NaviSoc chip (from Poland) ? It is quite small 9x9mm and most important is a dual frequency GPS. It works with all constellations but has Galileo core. I would very much like to have a dual frequency GPS because it could give up to 10 cm accuracy (or so they say). I mainly plan to use my Librem 5 for mobile applications (not desktop) and GPS is one of the most important. Another important hardware is of course the LTE modem. Hopefully we will get local LTE base stations (Nokia already has design such) connected to our fiber network. The operators do not have so good coverage especially on higher frequencies.


I think they picked a pretty good one


I am no expert on GPS chips but I noticed that LIV3 has not dual frequency. On the other hand dual frequency is still new and it could be a good idea to wait until the chips have been tested.


The July update just came out, and it’s entirely about software. Will there be an official hardware update any time soon?


The silence on hardware is deafening.


I suspect that we are seeing a build up to an announce. Notice that this months update was pretty big and it’s all about finalizing software so we are coming to the end of new features and it’s all bug fixes. So that means that the software platform is basically done and the usefulness of the DEV board is kind of complete. We should be hearing hardware news within the next month is my guess (guess being the operative word).


If it was nothing but good news they’d be screaming it from the rooftops.

I don’t think they have everything they need yet. If they do it’s likely to be pretty rough and they’re waiting for the last possible moment to tell us.

I’ve bought an old nexus 5 that I’m going to use with postmarketos OS :slight_smile:


i’d be interested in seeing a video of the hardware sensors app (temperature, frequency, for the cpu, igpu etc maybe integrated as a plugin in the system monitor app)


I bet they are not gonna tell us anything till the price goes up. Because any more information would indicate that they have finalized the hardware and give an “advantage” over the “risk” early adopters took. Last raise was with the shipping of the dev kit if i remember correctly. So they habe taken this points which give more certainty of success in the past. And i think it’s kinda fair to the people whose shouldered the risk of early commitment.


To be honest, I was expecting way more openness and community involvement in the process of the development of the final specification… They shared a lot of informations about the software but, so far, not much about the hardware. That’s quite disappointing in my opinion. I am still waiting to know if they decided for an alternative modem (as Nicole stated a while ago) or if they stick with the current one, that unfortunately put myself out of the game (I currently live in a country that has different frequencies for LTE). In that case I have to ask for a refund, but that’s not what I want!


I agree.
Final hardware is ready, they will show it when they will announce the general availability and final price in August.


My impressions is that the expectations of many people are still to high. This isn’t easy and there are so many things to compromise on. And the absolute top thing is to get this done while keeping the project alive. As many other have tried and fail to enter the mobile world.
So from my perspective even such harsh things as leaving out some areas because the modem dosen’t grant (good) support is totally valid to compromise for the success. Same for keeping the final specs closed to make a bigger shipping announcement with maybe some surprises(4gb RAM?, more modem options?). This maybe feels not that good an open to the community but if it helps to make this a success it is so totally worth it.

So i plague for some more patience with purism.

Apart from a delay of about 8-9 month(which is mostly due to an CPU upgrade, which in my impression the community was pretty much full on board with), Purism delivered pretty good on there promises so far. So the first round probably won’t be for every body, but it will open an so important road for future version which will definitely satisfy more user, be it better modem coverage, cameras or storage. So just getting it out is the big thing here. Yeah it has to be good and not a total shit show. But a look at the librem 13/15 shows that there is a pretty high standard on quality in the purism team. An the notebooks also dosen’t satisfy everybody’s needs. They for example miss out on the highDPI, lacking behind in CPU gens, have no options for decent GPUs and dos not provide many languages on the keyboad, but they are a massive success and will get better.

So the phone will leave out some people and that is fine in my opinion as for me the goal is to be good enough to have enough customers to be profitable and build on for the next gen. This is harsh to some but totally necessary in my opinion. An on a marketing point is is also okay for purism to only focus on the bright side and leaving this view out.

I hope no body feels offended by this, as it feels a little bit like a rent as i got on with typing this. I it’s not my intention to do so. I just want to lower the expectations a little and maybe see it from an other angle.


I understand and agree with everything you’re saying. The lower end hardware, ram, storage, camera is no problem with me. It will be a little disappointing but we need to start somewhere. Plus Linux has always been pretty good on older hardware and the PI seems to run alright on an SD card so we should be fine.

I will also have no problem if they delay again because of a hardware problem. in fact, I hope they’ll test and fix all the little hardware bugs before they ship anything. Software can come later, if we get bad hardware we’re getting a useless brick.

The problem I have is the secrecy behind the hardware. No updates, no nothing. They’re not even saying what they have and what they’re working on. It’s awfully close to Q4 and still nothing. Why?