There was a slightly more precise questimation poll done a month+ ago. Interesting to compare these (and to the ones before these). The end of shipping expectation was suggested to be around the same time as mission to mars begins
I don’t disagree that the time has been long-ish and it could have been communicated better, but I think you are mixing “making” and “developing” a phone, which may affect expectations of how fast and reliable progress may be to some degree. Expectation management has been lacking, evidently, but shouldn’t let that sour the end result.
How philosophical My gut and my experience so far tell me that it will still take quite a while until ‘the doors’ will start singing their famous tune
I have to add, that the communication has become a bit better, but this reduction to “early backers” that has already been reduced to “very early backers” makes me feel like this is not the full story and I anticipate yet another volley of excuses for further delays…
Reading between the lines, I would say that the USA Edition is not the significant factor here.
It is stated that the USA Edition is “manufactured and processed in parallel” and is a much smaller number of phones anyway.
I don’t know whether it is actually an option to change your order, this late in the process, from the regular edition to the USA Edition and that’s certainly not something that I’m contemplating doing.
I can understand that everyone is keen to get their Librem 5, after months or years.
The only remaining electronics that can be made in China while are the ICs. The rest of the production happens in the USA, so it’s using a different manufacturing force than the regular Librem 5 production line. That’s both the reason they can start shipping in parallel, and the reason they cost more.
There’s no point in trying to deliver phones by date, because that would just mean the USA manufacturing team sits idle.
I believe that Purism had a financial crisis in Feb-May of 2020 when it laid off 3 of its 11 software developers and both of the UI/UX developers working on the Librem 5. I believe that crisis was partly caused because Purism had to spend far more developing the Librem 5 than it expected, but the crisis was also partly caused by COVID-19 which cut demand for Librem products in general.
Google Trends shows a 40.0% drop in global searches for “Librem” and a 28.8% drop in “Purism” searches after COVID-19 hit. Many people cut back their spending due to COVID-19, which hurts premium and specialty brands like Purism that charge higher prices.
A lot of companies bought PCs so their employees could work remotely during the COVID-19 crisis, but Purism isn’t on the approved list of PC brands for most companies. Even for companies that let their employees select any model of PC that they wanted, the Librem 13 and 15 are above the price that many companies are willing to pay for a PC (especially when many companies were trying to economize during the COVID-19 crisis). The Mini would have been cheap enough for many companies, but it didn’t ship until 2020-06-23 and companies needed PCs immediately. Also the Librem 13/15 v4 had 7th gen processors, so they were considered out-of-date.
The Librem 5 uses a 10-layer PCB and has about three times as many components as the PinePhone. Its board isn’t as densely packed as an iPhone or Galaxy S-series, but it has more PCB area than any phone I’ve ever seen. It is probably the most complex phone currently on the market in terms of PCB design. Using 6 chips (i.MX 8M Quad, RS9116, BM818/PLS8, Teseo-LIV3F, WM8962 and bq25895) instead of 1 integrated SoC, hardware kill switches, a smart card (with a separate microcontroller), and two M.2 slots adds a lot of complexity to the design.
However, I agree that JIT manufacturing over a 6 month period is going to be more expensive than manufacturing a single large batch. I suspect that Purism is doing it this way because it may find flaws in the design that need adjustment, but it also probably doesn’t have enough funds on hand to manufacture 10k phones at once.
I know that many are going to criticize Purism on this point, but they don’t seem to appreciate how hard it is to pay for 3.3 years of software development and how necessary that development work has been. There is a reason why MonteVista, Motorola, FIC/OpenMoko, Nokia, Intel, Jolla, Mozilla, Canonical, Samsung and Palm/HP all failed at mobile Linux and Purism has avoided many of the pitfalls that caused previous attempts at mobile Linux to fail. I’m convinced that the way Purism is developing Phosh as a thin overlay on top of a desktop stack with wide industry support (IBM/Red Hat, SUSE, Canonical and Google) is the best path (and probably the only viable path) to making mobile Linux a real alternative to Android and iOS.
I have not always been happy with the marketing of the Librem 5, but anyone who is thinking strategically about how to create a better future realizes the importance of the dev work on the Librem 5:
Phosh is the best shot at mobile Linux becoming a reality (Ubuntu Touch and LuneOS are not sustainable with their current volunteer labor, Plasma Mobile has almost no corporate support, Tizen and SailfishOS have proprietary elements, etc.).
Purism paying Redpine Signals to do firmware development has given us the first 802.11n that works without proprietary blobs in /lib/firmware and doesn’t suck like ath9k (poor range, energy inefficient and requires blobs for Bluetooth).
We need a supported platform to build RYF devices and the i.MX 8M is the only viable game in town at this point (i.MX 6 is hopelessly outdated, Allwinner doesn’t answer any questions about the A64, and RK3399 can only barely boot with 100% free software). Once Purism produces a debugged free/open design for the i.MX 8M, it will much easier for other companies to take that design and Purism’s kernel work to produce other RYF devices.
Once you understand the strategic importance of the Librem 5 in kickstarting mobile Linux and the RYF market, then you understand why it is worth backing the project even if you don’t like Purism’s marketing of the product or its slow roll-out.
Yea, it’s packed, but that doesn’t mean it’s hard to manufacture - with respect, many of these chips are hardly difficult to reflow, we’re not talking massive FPGAs next to tiny buffer memory, or complex power supplies, nor are we talking huge ball grid or column grid (who remembers those!). This is a simple PCB / stack to manufacture. M2 connectors are hardly fine pitch, or difficult to put down, and the switches add close to zero complexity.
I’ve been involved with builds back in the early 2000s with 20+ layer boards - over 10,000 components on one pizza box, with extremely high speed signals that were built thousands a month. The biggest issues we faced was yield in the early days of a design - but we’re talking early, first 20-50 boards.
That, and signal integrity - I remember clearly working on our company’s 1st gigabit switch wayyyy back when and having an issue where one of our ports wasn’t able to communicate. Turned out it was the signal routing - too sharp a radius in the track meant the signal went to radio, and never turned up at the PHY. That’s showing my age, but it just illustrates there’s nothing particularly “new” here from a production standpoint. You’re not using bleeding edge silicon and having to patch the bugs that come with it, etc.
Getting good quality reflow on mixed components was difficult, but nowadays every reflow oven will come with software and logging capability to enable automated tuning of the profiles. Effectively you pass a loaded board through with thermocouples underneath key components (drilled under) that the reflow oven then uses to fine tune it’s process - this is basic manufacturing.
When you switch a line over you run your instrumented board to confirm all components are good, after that soldering issues all but disappear.
Production is, at the most basic, very simple to forecast. You have capacity at X, and yield at Y, these two numbers give you production capability - with the exception of procurement, which, the more I think about it, is the likely hurdle here. With small (<100k/month) builds they’ll be pushed back on by chip suppliers for sure.
You’re quite right that I also think they’re building in small batches, but I don’t agree with their statement about JIT. JIT means build as ordered - and there’s plenty on order, so that’s no reason to split the builds.
I’m all for this project, and back Purisim with more than just a phone order - but as someone who spent a long time remotely supporting R&D at our multi-national manufacturing locations I can’t reconcile my experience with the update we’re being fed.
There’s more to it, and not in a voter fraud kind of way, something is slowing things down - either procurement, yield or testing. Or maybe it’s as you say, Development staged as production, which is fine - everyone does it, but this should be clear - evergreen A, B, C, etc.
The thing that really gets me is they’re (possibly) losing the advantage they have in the hardware by taking longer to get product out. They’ve done the hard yards with software development and integration, but when you delay you invite others who are better resourced to compete with you.
There are so many examples of products that were exceptional, only to fail on initial release and be overtaken by copycats - you need to bed in your advantage, and Purism deserves that, delaying now means opening up to competition from cheaper manufacturers with higher capacity.
Aspen, Birch and Chesnut had hardware bugs, Dogwood is the only batch with the totally new PCB design and we don’t know if it have hardware bugs (we don’t know for the OS stability problem) and antenna placement have been optimized since.
There is still some component without drivers, like cameras, that have not been tested in real use condition.
So there is always the risk of discovering a critical hardware problem.
We are still waiting for the first series models to demonstrate that the design no longer has critical flaws.
I think this is what motivates a production cut into smaller batches.
I would like to have my L5 this month, but at the same time I’m happy that there are ~4000 orders before me to test the hardware.
If that’s the case, after all this time, then they’ll open themselves up to competitors for sure - I really hope that’s not the case, but only time will tell.
It’s all good being open-source and part of a movement, as long as you’re able to deliver on promises.
JIT manufacturing makes sense if Purism is worried that it might find problems in its implementation of the image sensors, smart card and GNSS which still require software development to be able to fully test them.
However, I suspect that Purism simply doesn’t have the funds right now to pay for a big manufacturing batch. The former CTO Zlatan Todoric said that the original estimate in 2017 for the Librem 5’s bill of materials was $300. It is hard to know if that estimate is still accurate, but if true, then it would cost $3 million to make 10k units.
Last year Purism reported that it raised $2.5M in venture capital and its Kickfurther page says that it has over $3M in annual revenue, so $3M is a lot of money for a company of Purism’s size.
But just to be clear: JIT means manufacturing as orders come in - not delaying because you don’t have the cash to fund production. That’s purely cashflow management.
JIT: I order a phone today, I get batched, once the batch qty goes above the minimum order quantity we manufacture those phones and ship.
It’s also worth noting that the components used in both locations will (you’d hope) be ordered as one set, to take advantage of bulk buying savings. With that in mind, the phone manufacturing location becomes nearly inconsequential - building boards doesn’t change the time taken.
Neither side will be faster than the other, just the cost will differ. So, if that’s the case, and procurement is a limiting factor, the two builds are competing directly for components, and delivery time.
If you’ve ever seen a high speed pick and place machine at work, you’ll know that there are not dozens of Purism employees sitting at tables, all soldering components down to a circuit board. The robotic arms on those pick and place machines can lay the components down faster than the eye can track. Then it’s off to the reflow oven to secure those placements permanently as the boards move through the oven. I suspect that purism might be hand-assembling the completed circuit boards, antennas, etc… in to the cases manually and doing a final test manually. That is probably what the wait is for. I just hope that the R&D is mostly over with. Feature-creep or repair of design flaws can last as long as they choose to allow it to continue.
I wouldn’t worry too much about competition. Those who can compete with Purism don’t want to compete with Purism. If Google or Samsung or LG decided to release a RYF free phone, that phone could hit mass production in virtually any quantity within a few months after they decide to do it. That is not what they want to do. Those who want to compete with Purism but that are presently unable to compete, will gladly let Purism work out all of the bugs before they copy the L5 themselves. Without a social purpose, most people (and their respective businesses) just want free beer. Purism’s edge is the social purpose. Most people don’t understand the importance of the social purpose and won’t be willing to risk making investments in to the social purpose until after the market proves a demand for it. Competition to the L5 won’t start until after everyone wants an L5 instead of a traditional Samsung or LG product. At that point, Purism will be at the top of the heap, feeding in significantly growing quantities on Google and Apple’s lunch, with a trailing following of copy-cats behind them. When you’re as small as Purism is now, you can grow a lot this way, and quickly. You don’t ever need to get as big as Apple nor Samsung to become a big player in the market. My biggest concern for Purism isn’t about competition. If they pull some of the crap as a big company that they’ve done as a small company, it’s difficult to see a path to success for them. But then again, most companies are fairly dysfunctional in their early days. Hopefully Purism outgrows the problems they have now. Most of us aren’t used to dealing with small companies to get our high-tech products. This is an interesting journey.
Yea, love the turret p&p place machines - how good. The first time I saw one was at the Nintendo game boy factory back in the day, 100000 components an hour or something like that. Used to shake the whole building whilst it went about it’s work.
I’m still sticking to the fact that if the build is purely in China, and shipped from there then if anything the production there will be on par or faster than the US version, and if volume buying is considered we should get our phones faster.
I agree with your competition points, specifically that any competitor that does pop up is unlikely to be trusted by most of the consumers here, and that’s the gold - however if things go tits up with quality or reliability (quality over time) then the door opens.
Interesting is right, the thing about communication is it needs to be on point - that last update is shrouded in questions, which could easily be answered.
If there’s one thing Purism could do better it’s technical comms - technically accurate and complete comms to suit their user base. We’re not Microsoft support requestors - just be open and clear in what’s up - given the posts in the forum there’s plenty of skill hanging around to help.
We will start sending emails to confirm your shipping address, modem choice, and other details starting November 9th and continue that practice each week afterward as we are able to process the previous week’s requests.
I wonder when someone on the forum receives one and will they share that with the rest…
The only thing I would like to add to all the good posts that came after mine to answer your question is that at least in my case I do not like to jump the queue - first ordered, first served is fine by me (the exception for the US made phones is not, though) - but after all the delays and the rather reluctant communication on it I expect an honest and direct statement along the lines of “The first batch will contain -add number here- phones and is scheduled for delivery until -date-. We currently plan to continue with -add size here- sized batches due at date1 date2 …” If that is not possible after all the iterations since batch Aspen, then something went seriously wrong. Purism is a social purpose company, but it is still a company and no private nonprofit org running pastime projects and I expect it to behave like one. I am not in for a refund and I honour the efforts they make, but the intransparent information policy just isn’t right and the same goes for the change of the refund policies for already signed purchasing contracts - I would not expect this from a company.
