Purism is controlling the RS9116 over SDIO 2.0, so no direct memory access like with the PCIe bus (which most other devices use to access Wi-Fi/Bluetooth). Purism says that it won’t have any blobs in the Linux kernel or U-Boot, but there will be proprietary firmware installed in the components.
Thank you very much for the comparison, it is very helpful.
I didn’t know the NC_1, but I don’t see it as a very interesting device. In addition, the price is abusive. Why does it cost so much (much more than Librem 5)?
My idea is to buy the Pinephone when it is on sale this year. The price is very attractive and is a good phone to start messing with.
In the case of Librem 5, I will wait for it to go on sale when it is already manufactured. Maybe I’ll wait for a more powerful second version.
This Necunos shouldn’t be on this comparison list. Or the title should be worded differently.
No basebands for “security reasons”? Ok, then it’s not a phone. They can call it “Communicator” / WiFi VoIP Media Player, etc.
I agree that it is misleading to call it a phone, given that it looks like a smartphone. Necunos’ press release calls it a “smart mobile device”
However, a few people have discussed removing the cell modem from the Purism 5 to make it into a WiFi only device. Purism calls their phone a IP-native mobile handset, which means that while the cell modems are designed to do normal cell phone voice calls and text, we could just use the cell modem for data only and receive voice calls via VoIP over the top of the cell data network and through WiFi when available.
“Linux smart mobile handsets” might be a more accurate label, although it sounds unconventional and wordy.
The first production run for the NC_1 was 500 devices, whereas Purism is producing 10k, so you have a lot more overhead costs per unit, plus Necunos is assembling in Finland, which is very expensive compared to China. I think that the people who pre-ordered the NC_1 want a mobile Linux device company focused on security to exist in Europe. If Necunos doesn’t manage to ship soon, however, I can’t see the company managing to survive, because Purism will take its niche market.
However, Necunos plans to also offer a custom commercial version of the NC_1, and Necunos might be able find enough companies that have special security needs and need customized devices.
I predict that the second version of the Librem 5 will have more RAM and Flash memory, but it will probably use the same SoC. I don’t think that Purism will switch to the i.MX 8M mini despite the fact that it is more energy efficient, because it has a less powerful GPU and little video out, so it won’t work well for convergence as a PC, which is one of the goals of the Librem 5.
Maybe Purism can switch to Rockchip in the future now that the Lima drivers have improved, but it sounds like the future RK3588 won’t use the standard Mali GPU, so that probably won’t be a possibility. As I see it, Purism is probably stuck with the i.MX 8-series for the next couple years, so we have to pray that NXP will decide to either offer the i.MX 8M mini with a more powerful GPU/VPU or do a die shrink on the i.MX 8M Quad.
If the better version of the PinePhone with 3GB RAM 64 GB Flash costs $250 or $300, I think that a lot of people will order it, but PINE64 expects users to solve a lot of their own problems. It will be a great DIY phone, but probably not a good phone for people who aren’t tinkerers.
The Librem 5 will be a thick brick (and probably require frequent charging), so people who want a normal sized phone will have to go with the PinePhone. PINE64 doesn’t have many software developers and is relying on the community to provide a lot of the software, so I expect that using the PinePhone will be a real challenge in the first year.
Here is how I would summarize it:
PinePhone:
- Outdated SoC
- Low resolution camera and can’t be improved due to the SoC
- Economical
- Software will be very rough and don’t expect much hand-holding from the manufacturer, but can get help from the communities providing the software
- Slim case, so easier to carry
Librem 5:
- More modern SoC (but still underpowered compared to today’s Snapdragon, Exynos and Helios)
- Acceptable camera
- Software will be rough at first, but it will improve, so will probably be good enough that a normal person can use it as a standard phone in a year or two.
- Very thick, so only for people who don’t mind the large size
- First phone in the world with replaceable Wi-Fi and cellular baseband
These go to a different category - especially from software perspective as they are not open to user (although there seems some effort to make Google stay away): Bittium’s (another Finnish secure coms firm) Though Mobile 1&2. What I find interesting, are some of the HW decisions. Physical aspects are impressive (water and shock proof), but cameras are slightly lesser quality etc. But the main similarity (and reason to mention here): physical button to affect sensors: a “privacy mode” that (if I’ve understood correctly) reduces accuracies (may not turn them off?). Interesting idea to compare to.
The Librem 5 at $699 looks like a steal compared to 1550€ for the Bittium Tough Mobile 2. According to the description, the “privacy button” provides: “Hardware-based privacy mode for disabling microphones, cameras, Bluetooth, and reducing sensor sensitivity”.
If they went through the trouble of adding a separate Bluetooth chip, why not use a separate Wi-Fi/Bluetooth chip?
Doesn’t it just But I’m pretty sure they target governments and affiliated orgs.
But I couldn’t find any specifics either on the buttons boyond that.
It’s fine for a first gen, i hope a librem5 v2 will be completly open
Unfortunately that is not possible. The only processor that doesn’t require a binary blob to set the timing in the DDR PHY is the POWER 9. There is no way to get that chip into a phone. It looks like SiFive’s RISC-V mobile processor (which is supposed to be ready in 2 years) will be even worse in terms of binary blobs since it uses a PowerVR GPU. The Rockchip RK3588 doesn’t look like it will use a standard Mali GPU, so there won’t be any free drivers.
Nobody makes Wi-Fi, Bluetooth, cellular modem or GNSS without proprietary firmware. The best that you can do is put it on a separate chip and use a bus that doesn’t provide Direct Memory Access (DMA). This stuff is so locked up in patents, that I doubt that any manufacturer will ever make any of these components with free firmware and reverse engineering the proprietary firmware is incredibly hard. If anyone tried to produce an open hardware LTE modem, they would be sued out of existence and the patent holders would never allow them to license the patents if they wanted to publish the code. Maybe you could take all of them to court and force them to license the patents under the FRAND laws, but the legal fees would drive you into bankruptcy.
Like a steal?
it seems to me that your affirmation is rather exaggerated, considering the commitment to something totally new, the use of isolated hardware, the creation of non-existent software.
Mah!
Patents eventually expire.
Patents do expire. When they do, the cell technology will be taken out of service. So it will not be possible to use those expired patents unless you start your own network and license the spectrum for use with obsolete technology. The only way to avoid this is to innovate in the open, and let the open technology replace the proprietary via superior functionality.
20 years is a LONG time in a technology context. sheesh …
Agree needs better name. How about gnu-phone? (gphone for short). … Because it was the GNU project that started this all off. Linux only came later.
“GNU/Linux phone” describes the Librem 5. Only G might make people think of a very large technology company. But the Necunos lacks a cell connection, so calling it a GNU/Linux phone will mislead some. “GNU/Linux mobile handset” is more accurate.
People say “phone” when referring to their landline phone which can only be used in a single location, so I think it is fair to call the Necunos NC_1 a “phone”. Its primary purpose is to be used to call people. If you want to be more accurate, you would say “internet phone” or “Wi-Fi phone.”
The original idea for the Librem 5 was that it would be an internet phone like the Necunos NC_1, since internet telephony is more secure and has less data collection than cellular telephony. In a 2016 interview, Todd Weaver said the Librem phone would not include a cellular modem, but it would have a slot if people wanted to add it.
It wouldn’t surprise me if most mobile phones will become internet phones in a couple decades, as internet access becomes more common. At some point, Socialists are going to get elected on the platform of “public internet for all” and the economics are on their side. The current privatized telecoms industry is very inefficient and has negative social consequences, so I don’t expect it to continue in the long-term.
Years ago Att and Verizon started using VoIP for international calls because it’s cheaper and better. So yeah, it’s just the matter of time
Considering the price point of the pine64, the fact that the librem 5 may still be a month or more before shipping starts, it appears it has better LTE bands available, comes with KDE as default I may get one of those to start with as a primary until the librem 5 get to the point where it’s actually useable
Purism has a better track record than PINE64 when it comes to the software for their devices. I expect that the PinePhone will have software issues at first, just like the Librem 5. The big difference is that Purism has more software developers on staff to solve those problems over time, whereas most of the focus at PINE64 is on the hardware. It remains to be seen whether PINE64’s approach of relying on community projects (postmarketOS, KDE Plasma Mobile, UBports) for the software will work well, but I really hope that both companies will succeed.
If you don’t want to go through the joy/pain of being an early adopter who is helping to debug the software, you might want to wait to read the reviews for the PinePhone before buying.
I don’t mind being an early adopter, but I work at a software firm, so I kind of expect it.