For anyone who’s been following along, particularly those who want to have a go at doing this themselves once it’s finished, I’m sure you’ll have noticed that at some point in the last few logs, I mentioned things getting easier from that point on. Unfortunately, I was mistaken.

Firstly, the replacement screen I found straight up just will not work with the compute module. I must have spent 5-6 hours on it, and I think there’s more I could try, but realistically, it’s over for that one, which is a real pain in the butt, because according to panelook, it’s one of the few in 5.5” that’s both available to buy as individual units, and also has touch built in. I did find another screen (another 3-4 to be precise), so I’m starting cheapest first, and I’ll update back when I get anywhere with that one (I did a ton of research though, and it uses the same driver IC + pixel format + pin arrangement as a range of TTL monitors available on adafruit, so I think I’m in with more of a chance this time). 

Then there were the parts. The modem is just too damn big. Having taken it apart in hopes of finding a tiny little circuit board with lots of empty space and some roomy LED circuitry, that’s unfortunately not what I found, which is frustrating, because the damn thing is about as long as the priv, which means either a lot less space for batteries, or it’s got to be about 5-6cm thick. Not to mention, in sticking with the DAC I’m insisting on, I’m losing quite a lot of space there too, especially with the cables I’ve got to use. Also, a USB hub I decided on isn’t available without a header, which butters precisely 0 parsnips.

However, I’ve still never really been taught how to give up, and I’ve got so much to do these days, that I simply don’t have the time to sit down and figure that out, so instead, I’ve come up with a battle plan to deal with all of these new issues.

as previously mentioned, I’ve changed the screen. It’s kinda good too, because this one is a much lower resolution, which means much lower power. For anyone who insists on 4k or even 1080p, first off, why? 30” 1080p monitors are perfectly acceptable with a pixel density of approx 73ppi, and 854x480 in 5.5” is 173ppi. But also, I’d investigate one of the sharp 5.5” monitors, I think I’ve heard of the LS055R1SX03 working with the cm4/rpi4 but there’s no touch. 

Then for the modem, I’ll be pivoting back to a hat (kinda). Should have seen that one coming, but you live and you learn, and the whole point of this is to both document that process, be able to explain why my pitfalls are pitfalls, and be able to provide you the reader with a definitive method + ingredients for achieving the end goal. I believe wave share has similar sim7600 4g hat (specifically the B one) which looks perfectly acceptable, and significantly smaller, downside is on the cm4, it’s still a usb dongle, and therefore comes with all the limitations of needing a USB port + cable and space requirements.

HOWEVER! I had another brain wave (apologies to everyone involved), when is a non hat not a non hat? (When it’s a hat, dune is so good). 

The waveshare sim 7600g modem (B) is designed for USB usage in all other pi models than the zero, which it’s designed to stack with, maybe you can see what I’m getting at now. 

After seeing the Nokia N95 that got blender on it, a lot of my preconceptions of what’s possible have kind of been blown out of the water, and I’m determined to do more with less computing power. Not to mention, the pi zero 2 w fixes a significant issue with the maximum wattage, where the cm4 can run at a max of 10w during heavy use, the pi zero 2 maxes out around 2.5w, theoretically quadrupling the battery time, while also freeing up much needed physical space.

considering I already have the cm4 + a working version of postmarket OS, I’m not falling over myself to make this change, and I think it will introduce significant functionality challenges (I was eyeing up a Nokia n95 to see what hacking could be had, but fell short when I got scared off with lack of WhatsApp API), but should the battery prove too weak, or the space requirements too much, I can see this being a worthy sacrifice. Not to mention, it reduces the BOM by a whopping £85. 

Finally, the spring on the slide mechanism is going to throw up problems. Not everyone is going to be able to get a used blackberry priv for cheap, and while the current price of a replacement spring is £10, it’s not always going to be that way, and I suspect the day is fast approaching when you won’t be able to get one at all. To whit, I’ve started drawing up plans for DIYing the spring. I’ve got my eyes on some ribbon wire that’s not too expensive (about £2 for a roll of 10m), and I took a good long look at the spring to see how it’s actually constructed, so hopefully, I should also be providing instructions on that. (I may add it as a separate project as well, because it could be valuable for priv owners).

anyways, in summary:

0) wait until I have some spare money (Sorry folks, life is boring like that, but I’ll see what I can do for cheap or free)

1) talk to some people about spring construction, and see what I can achieve there (end goal is a fully diyable, like for like part that’s not too hard for users to make)

2) when the screens arrive, look into driver boards, I have no idea why finding a 5.5” tft screen is proving such a challenge, but it is, and I do have several more solutions (the sharp monitors, and then downsizing to something known to work)

3) get my hands on the smaller modem, and start testing functionality once that arrives

3.5) get some diffuser film for the keyboard and see if I can’t get the LEDs to play nice (very much a side goal), also get the camera module and other bits I’ve not really spent too much time or mental energy on (vibrate motors/speakers/mics)

4) start work on the battery and assembly. It’s been a long time coming, but hopefully not too long now. I have estimated minimum required Watt-Hour values for batteries to achieve 10 hours at max capacity, so I’ll be working from there backwards to see what I can’t do between power management and different battery options, and of course switching between the different boards if it comes down to it on space/power.

Realistically, this may be it for the next month or two, perhaps longer, we’re creeping into the most expensive part of the year, and unfortunately I have other things that need funding too. I’ll do what I can when I have time and money though. 

I’m not giving up

(I’m such a drama queen)