ZeroPhone – A Raspberry Pi smartphone | Hacker News Reader

ZeroPhone – A Raspberry Pi smartphone

hackaday.io

184 pointsekianjo42 comments

I have several points on this:

"...yet there isn't a project like that"

Not exactly. There are several Arduino based phones like the DIY Cellphone by David Mellis. I have an improved version of that right here.

"Open Source / Open Hardware"

All open phone projects have the exact same problem. The core component, the baseband (which is basically the phone) is a complete blackbox. AFAIK that's also true for the SIM800 used here. So arguing security is a bit dishonest.

"Price"

I'd say the price calculation is extremely optimistic. You'll only get that price if you buy the parts in bulk. Especially the PCBs.

Anyways.. cool project!

I wouldn't have said dishonest.

The only access the baseband running on the SIM800 has to the main processor (Pi) is via the Pi's UART.

It's completely different from the "read and write directly to shared memory" access the baseband that your average Android potentially has.

If you encrypt your SMS on the Pi and send it, ths SIM800 has no access to the keys for decryption, for example.

I'm guessing you're talking about his price calculation for the $27, right? A bit further on he does say:

>In total, it should be 50$ or slightly more.

>Also, price falls quickly if you're assembling one or two more phones for your friend, too - eBay sellers can make discounts when you buy more than XX$ or "make an offer", and PCBs are cheaper (you won't get less than 3 pcs of PCBs in any of usual board houses, so you have to pay extra even if you need only one board - you'll still get two more)

I don't think that's too far off and may be a bit pessimistic. The only component you can't bulk buy easily and cheaply is the Pi Zero as it's STILL one per customer.

ABut none of the parts are expensive in single units either, except relatively speaking the PCBs. Which are non-essential because you could literally breadboard it.

If you're interested, he did an AMA on Reddit earlier.

Do you have a link to the AMA on Reddit?

https://www.reddit.com/r/raspberry_pi/comments/5nwmfx/im_mak...

The Osmocom Baseband project is open source software for GSM basebands:

https://osmocom.org/projects/baseband

A Python, pure Raspbian/Debian phone that's cheap to build, even for kids, and still in the beginning stages before the software gets too complex and unwieldy, and without needing to flash firmware and risk brickage. CLI-based phone interface. Writing scripts that actually control your phone instead of pretending that you'll actually easily do that (even on a pseudo-hacker-friendly android..)

Still a bit big and fragile looking but that'll improve. This sounds awesome!

You can also buy an old Nexus and put Ubuntu on it...

Not if you want it to work. God knows I've tried every device on ubports.

Ubuntu Mobile may be a work in progress, but it still works far better than the ZeroPhone.

It also look much better :)

In terms of adapting existing tech, several phones [1] did use a VideoCore 4, the most recent being the BCM23550, an SoC similar to the chip used in the Raspberry Pi 2.

Using one of these 2013 era smartphones might be an alternative starting point in porting raspbian. You'd still have binary blobs for other components, naturally.

(Yeah I understand the thrill here of building a working phone with your bare hands but if the idea were to crowdfund a smartphone product leveraging the rPI community and Broadcom then I'd start with a working base)

[1] https://en.wikipedia.org/wiki/VideoCore#Table_of_SoCs_adopti...

Possible I'm missing it, put seems like if you are going to build a phone the first step would be to state what governements and networks would support it.

For example, appears that in the US that even if the phone would have worked with AT&T in the past, that as the end of 2016, that is no longer the case:

https://www.att.com/esupport/article.html#!/wireless/KM10848...

Reading this: http://www.raviyp.com/embedded/208-don-t-use-sim900-sim908-s...

Seems to agree with you. They recommend using the simcom 5320 versus the 2g only simcom 800 used in this project.

The project lead is in Latvia. Is anyone in Europe sunsetting 2g? Any network that does 2g GSM will work, and if it's pin and AT command set compatible with the SIM800 then any other GSM module will work. If not, it's a straightforward setup to adapt.

Good thing T-Mobile made a commitment to keep their 1900mhz GSM working through at least 2020. (UMTS might actually sunset first)

Source: https://newsroom.t-mobile.com/news-and-blogs/att-2g-iot-life...

Maybe some people would give a f. and just connect it to the network, certified or not?

Won't work...It's a hard sunset. That's why, for example, the SF muni trains are not showing estimated arrival times right now. See https://www.sfmta.com/about-sfmta/blog/why-muni-arrival-time...

Related text:

>> "NextMuni data is transmitted via AT&T’s wireless cell phone network. As Muni was the first transit agency to adopt the system, the NextMuni infrastructure installed in 2002 only had the capacity to use a 2G wireless network – a now outdated technology which AT&T is deactivating nationwide."

Until someone actually creates a baseband chip, this is all makerspace wankery.

I wish these folks would go grab GNU Radio, a USRP/HackRF/whatever and design a baseband. I'd imagine a bunch of Chinese manufacturers would jump all over it and turn it into chips for them for free just so they could sell the chips.

You can't trust the network anyway. So as long as your proprietary baseband modulue can't get control of your application processor you are probably as good as you can get.

Considering the baseband on the far side of the demarc point is about the best we can do today. But a trustworthy baseband would actually allow some privacy advances. Imagine fine-grained powerup/powerdown, cycling through IMEIs, and even contacting a single tower through a directional antenna to avoid triangulation.

I'm confused, there seems to be open base station implementations such as OpenBTS, why is there no SDR based cellphone already I wonder?, as that'd be really neat in my opinion.

Because the people who are qualified and skilled enough to make one from scratch are working for companies like Qualcomm, 55 hours a week at a salary like $190k.

It's kind of like asking where there aren't any gpl licensed m.2 ssd flash controllers, the people who design them are getting paid nicely to design and sell a commercial product. It would be cool if there were such a thing, but there is only so much "free" expertise to go around...

Sure I can see that, but isn't a base station implementation just as complex?

Would it be actually allowed to connect to any cellphone system?

Or just use one of the ones supported by the Osmocom Baseband project:

https://osmocom.org/projects/baseband

Do I understand that this device, which runs raspbian, has a cellular modem attached to it and then it runs userland apps for VOIP and sending SMS ?

That is to say, the baseband chipset is on the add-in modem and you simply interface with the modem on a high level ...

So what is the userland SMS application that will send actual SMS through an attached modem ? Is there some standard GNU "phoneutils" package (I don't think there is) or did the maker of this device whip up something on his own ?

The modem likely has a serial connection and is sent AT commands, there are a few packages providing tools for that (can be done with many 3G USB dongles as well). One is SMS Server Tools: http://smstools3.kekekasvi.com/

No VoIP (yet at least), the audio is handled entirely by the SIM800 at present.

The SIM800 is hooked up via the Pi's UART for commands.

Not sure about the SMS software, someone on Reddit asked about that too.

Don't quote me on this but there's ofono, used by various incarnations of Meego/Mer/Sailfish etc.

In which case, yes, there should be a GTK+ or Qt frontend.

Serious Dumb Question: if I've got a cellular modem in my laptop, why can't I put my sim into it and use some software to make and receive calls ?

Depending on the cellular modem, you can. Some the USB Huwawai modem dongles have a "call" mode, a YouTube search will give you a run down.

Also googling for asterisk chan_dongle takes to to a list of USB voice capable dongles as well.

Something I've been looking at doing for a while.

You basically can. Almost every modem supports an AT command interface with SMS commands. It can be easy, non-trivial, and sometimes hard to get at your modem directly -- depends on the specifics. Calls are trickier, needing voice i/o, etc.

Been thinking about the implication of Moore law (in the broad sense) and how easy it's to reinvent the wheel using cheap electronics nowadays.

Giving this trend, how far are we from google/facebook/apple/whatever to start giving away hardware for free?

It's not so far fetched to envision such devices to be heavily hostile, for regular users, and makers or tinkerers alike.

Remember the :CueCat?

https://en.wikipedia.org/wiki/CueCat

It was "hostile" to hardware hackers. That lasted about 10 seconds. Of course, it's probably a bit harder these days as they embed more and more into SoC.

This makes me want a "ThreePhone" with a touchscreen (presumably driven by a raspberry pi 3).

I wonder how it would compare to an old OpenMoko. The software for that never quite worked, but it's all open source, and it has been years (and now there is android to steal stuff from too).

How does it deal with the IMEI regulation? Isn't the IMEI mandatory for every cellular phone?

It's probably only attached to the modem, which is on an external board.

Indeed, the IMEI is on the SIM800 board, which is hooked up to the Pi's UART.

So… the Raspberry Pi, made out of smartphone parts, is used to make a smartphone?

We've come full circle.

Is this project current?

The timestamp on some of the photos shows: 2012

Timestamp must be wrong. Pi Zero did not exist in 2012.

Reply on news.ycombinator.com