Hacker News new | past | comments | ask | show | jobs | submit login
CutiePi – All-in-one Raspberry Pi tablet (cutiepi.io)
463 points by ducaale on Aug 15, 2019 | hide | past | favorite | 142 comments

In case anyone is looking for a similar project with a bit more definition, check out the PineTab[1]. It'll probably be pretty cheap (PineBook Pro will go for $200, PinePhone will go for $150), and I'm guessing it'll be available before this project is.

I really like all of these independent tablet/laptop/phone projects! I'm just a little worried that people will pick them up thinking they'll be just like an Android/iOS tablet and end up blaming the company making it.

[1] https://www.pine64.org/pinetab/

It'll be $80, it's a pretty good experience since everything required for the A64 soc is in mainline already or in the process of being upstreamed.

If it has a working OpenGL ES and EGL support, then it could be a very cool target to port!

"This item is not yet available for purchase from the PINE64 store."

Neither is the CutiePi, but it does seem significantly farther along the process of becoming a purchasable product [1]. But I do like the idea of using a Pi compute module for the basis of a tablet.

[1] https://www.pine64.org/2019/08/05/august-update-london-meetu...

I like this project. However, I wish that under the FAQ for "What's the target price?" they had a better answer than "We'll know better in DVT stage, meanwhile please find the bill of materials in design files for detail." Who is going to Google each of the 89 components listed in the bill of materials to get a rough guess (since it doesn't list price info)?

hi, project creator penk here, thanks for the comment, "we don't know" is the honest answer we can give at this moment, since we hadn't ordered any parts.

For your reference the first batch (10pcs) EVT boards is about ~90USD a piece: https://i.imgur.com/anUASlT.jpg

"we don't know" is not as good as "we don't know, but it won't be less than the material cost, which is currently X" would go a long way towards taking the people you want to get excited about this product seriously. We understand how "things cost money" works, but going "we don't know, figure it out yourself" is not a nice thing to tell people who are actually interested in what you're making =)

Even updating it to say "it will likely be in the $100-$200 range due to material cost and production fees" is fine: that's a _huge_ range, and still better than "we don't know, figure it out yourself".

Fair point, will update accordingly. I hesitate to say anything because we are getting mixed signals from Chinese manufacturers....

It is great that you release this info early, including the “we don’t know”, as someone who does know may show up and help.

Would you be willing to work with US based companies for hardware?

It is impossible to find cheap US-based PCB manufacturing, the ones who compete on price all left for China years ago. What's left services the high-end/custom/defense/1-day-turnaround market.

Great update, thank you!

Beyond that, material costs are X assuming a 10k initial production run.

I find it refreshing and frightening that target price is not known when attempting to build a commercial grade electronics.

Is component selection conducted without any price consideration?

Component prices will usually be a very important factor when choosing which components to use.

However, when you go into production and order parts from distributors, you will usually get special deals which can lower the price of your product. Problems can also appear if one of your single source components is out of stock, then you will have to wait (probably a long time) for it to get back in stock, choose an alternative (probably more expensive) component, or buy it on the gray market.

Edit: grammar and spelling

What does the grey market for components even look like? I can't think that anybody is bothering to steal individual components, so are we talking possibly-counterfeit parts?

The grey market doesn't mean stolen - it just means unauthorized distributors. Companies building a product will end up with excess parts -- either they bought too many to satisfy minimum purchases, or changed the design, or cancelled the product. They sell these parts to brokers, who then resell them. The way to avoid this is to limit sole-source suppliers and make sure there are backup plans (parts that fit with same footprint, etc.)

>What does the grey market for components even look like? I can't think that anybody is bothering to steal individual components, so are we talking possibly-counterfeit parts?

So somewhat related, check out Strange Parts on YouTube. His past couple of few videos involve him building a Samsung Galaxy from components in Shenzhen and then having it x-rayed by a place that (as one of their functions) identifies counterfeit parts.

In the video with the x-ray they talk about reclaimed and counterfeit parts and you can see differences between what was supposed to be a real camera module and an actual real camera module in one of the x-rays.


Even if you can only fill in the blanks of this sentence, that would be really helpful:

"I'd be surprised if the retail price ends up being less than $___ or more than $___"

Another point w/ looking at a pricing that's in flux: if prices are changing markedly over time, and you're tracking those, showing the plot provides information.

As in: is the trend downwards (my expectation), upwards (would be a surprise), or perhaps varying (market showing major shifts). And by how much.

A comparable concept is the bug burn-down charts the Debian project has produced over the past decade or two, which hugely assist with getting some sense of when "when it's ready" might be.

And sure, wrap the information in all the "this is not a price committment" language you care to.

Usually products are priced not based on the direct costs to produce them, but based on the value they provide to the end-user.

What is the target price range you expect people to pay for such product, above which it wouldn't be worth producing?

As an open source project (we're not a company) we'll try our best, hope it won't affect by the trade tariff? :P

For price range I would say kano computer kit touch or raspad tablet are all good products and provide good reference.

Those products appear to be priced around 250-280 USD. Would it be safe to say your product might be priced around 200-400 USD? I realize an exact price is premature and future tariffs could impact the price.

> raspad tablet

Maybe the totally wrong place to ask, but that display looks great. Any ideas where to get one of those?

> Who is going to Google each of the 89 components listed in the bill of materials to get a rough guess (since it doesn't list price info)?

This is an open source product, and not consumer-grade in the least, so I imagine hobbyists wouldn't be opposed to doing that legwork.

As a hobbyist: I really am. Looking up the price for each component takes a lot of time, which I would much rather spend working on my hw/code projects.

So sign up for the mailing list or come back when they are more ready to talk money.

Simple rule of FAQs: if you don't know, say you don't know, and stop. Don't then go on to tease about how people could look up the component costs if they really want to, because are you also going to let them guess as your production overhead and the profit margin? No. Just say "we don't know". Or, at this point, after limited runs, give a real indication.

I guess that's an old habit from maintaining open-source projects, fully disclosed & all transparent policy sometimes make things overly complicated. My apologies for the bad writing ;-)

Man this has come a long way. I remember my first Pi I wanted to do this SOO badly. I wound up doing it with a 48w/hr battery and a 7" non-touch automotive backup camera screen. [1]

This was just after the launch of RPi1 and well before Adafruit had all the cool parts that just work out of the box.

[1] https://www.honestrepair.net/wp-content/uploads/2018/04/Micr...

fellow veteran, what happened to this cool device? :-)

I've still got it and I believe it's still got Wheezy on it. Last time I tried she fired up. I've gone through abut 7 Pi's since, but this one will never be re-purposed. This is it's destiny.

The coolest part about it was that the screen, computer, and battery were all modular with velcro and electrical plugs. So the metal case had the regulators and Pi, the black case held the batteries. In under a minute I could remove the screen and Pi from the battery pack and velcro it into all into my car.

Reminds me of this project someone was working on: https://hackaday.io/project/164845-dlt-one-a-damn-linux-tabl...

They started off using the compute module, but ended up switching to the Nvidia Jetson Nano. Both of these projects make me wish for a world where a standard pin-out in this SODIMM format arises and we could see some inter-operable competition like we have with PC desktop parts.

There are just too many single board computers! We've chosen Pi Compute Module for its ecosystem and community.

All the best for your project!

Just these days I was wondering which tablet would be able to run Ubuntu Touch or something alike and while support for "real" tablets seems to be quite rare, there are quite a few tablet projects based on the Raspberry Pi. I hope this will continue.

As a former Canonical (Ubuntu Phone team) engineer I do like the ideas of indie OS, be it Ubuntu Touch, Open webOS, or even LineageOS Android for Pi. We'll definitely work with communities closely to bring more options to CutiePi.

After all, free software is all about choices.

Btw maybe it might be of your interest to know that some people are porting Ubuntu Touch by UBports to Pinephone and Librem 5 phones, and some others are trying to port it to Raspberry Pi 3 /4.


thanks, it's definitely on our radar. :-)

Maybe it's on your radar too, but what about an official port of postmarketOS?

I'm quite interested in how you integrate power management into the PI.

There's very little in the way of 'real' power management solutions for the PI and a myriad of 'power down when the battery is low' solutions ... is a decent, cheap, reliable fuel gauge hat just too much to ask for? :D

We're using the MCU to swtich power source between USB charging or battery, using it's ADC to serve as a voltage divider to get actual AD value from the battery, and a MP2636GR chip for charging and voltage step-up.

So yay battery percentage! ;-)

How do you do the switching? Using a MOSFET or something else?

Noob here, does this mean that Pi doesn't throttle down CPUs and other chippery? (I'd think that this is solved a long time ago between Linux and ARM, though.)

I am getting a bit overwhelmed - I would like to get a basic level of "maker" capabilities at home for my kids but ... it's not the kit, not the hardware, it's the projects

I think I am looking for easy to follow instructions for dads

I see you man, we created this project for the exact same reason. Instructions will be posted, with pretty pictures! :-)

I would start with arduino, not raspberry pi. I don't think raspberry pi's are very good or fun as "maker" projects and end up being too specialized or overwhelming because of the OS.

I've learned more and had much more fun making little toys, programming basic displays, getting sensor data, etc with arduino and other microcontroller development boards. There is too much already going on in a raspberry pi for it to really be a "blank canvas" for maker projects. Hence you see all of these projects where they are essentially repurposing the existing rpi, not creating their own thing with it. Its not really a "maker" project. Its more like building a computer.

Plus, you will find way more true maker projects targeted for kids for microcontroller development than you would for raspberry pi.

Plus, Arduino has a specialized c/c++ library for it which makes things just awesome for starting out. It's like writing python - stupid easy.

Check out the decently kid-friendly kits that Pimoroni sells. They have kits for various levels of difficulties, including some using Raspberry Pi. (https://shop.pimoroni.com)

Barging in with a vaguely related question here. Could anyone please illuminate a noob, briefly, as to what's the situation with open-source drivers for displays? Because, judging by the practical unportability of Android on phones, it looks like a ‘piss off’ from big boys of the market. (Or is the Android's problem more with radio chips?)

Libhybris would still be your best bet on re-using Android drivers..


A tablet, built on open technologies, with a wide variety of ports, and replaceable parts?

Count me in. I would replace my laptop with this.

What I really want is to use the Rasperry Pi as both a computer system and a throw away embedded module.

So I would want something like this when I program the Rasperry Pi to say control a robotic arm somewhere. Then I want to be able to pull the rasperry pi out of this tablet and actually place it in the robotic arm module.

So if I have several projects going on like auto fish feeder and / or auto plant waterer.... I want to be able to bring the tablet around attach the rasperry pi to it too make changes.

While the demo looks like this is what is happening I'm assuming that the final product is basically like a phone, the pi is eternally stuck inside the tablet?

$200 is too much for something like a auto plant waterer but $30 for the pi is perfect for this. Therefore make the tablet interface a module to the pi. That's the whole point right? Otherwise why use the pi over my phone? there's no point.

Wouldn't you be able to use this to develop whatever and then deploy your code to separate Pi module later?

That would require all my modules to be connected to wifi at all times.

Pi Zero W is $10 including wifi. Even if you don't want to hook it up to house wifi, the zero can run its own hotspot so there's no need to connect/disconnect cables/sd cards to update programming. (Personally, I hate having to deal w/ microsd cards while I'm also repotting plants.)

I really like the idea of an open hackable tablet but the screen resolution of 1280x800 might be limiting. My digital life revolves around my 11" iPad Pro (except for my GPU rig for ML and coding, and a MacBook for light weight dev and also writing) and the hires display is important.

Will see if we can source any higher resolution MIPI DSI display parts, the current one's cheaper though ;-)

Is it possible to buy higher-resolution display parts individually, or is there a minimum order size?

Also, most Android/Windows OEMs low-end tablets are 1280 x 800, so I'm guessing higher-res screens strain their batteries and GPU too much for them to be viable?

The 11" Pinebook ships with a HD display so they are available, don't know how big the production run of them is though.

I wonder how hard would it be to reuse the iPad as external monitor for the rPi..

I have 1280x720 on my GPDWIN2 and it can be limited for sure.

I guess I could be able to write some python code with this, using a small 60% keyboard...

It's really a cool device. I mean if I can do whatever I want on a handheld device, I'd be ready to pay 150 euros for it. Android phones are useless if the OS is a walled garden.

You're in for a treat, we are making a "coding" layout for the virtual keybaord (with an ESC and a Tab key! ;-)

This is exactly what I wanted for remote techs troubleshooting. It has a battery and can run more than standard windows utilities. I'll be picking up one of these to play around with when it comes out.

Not sure if I'm being silly here, but with the article about Talon (https://talonvoice.com/) and coding by voice from earlier this week, I think this could potentially provide an interesting setup for a portable development setup? (I quite dislike typing on touchscreens and being able to develop by voice also seems like a portable skill)

Looks cool, maybe you can start building a PoC with parts like this 2-mics Pi HAT? http://wiki.seeedstudio.com/ReSpeaker_2_Mics_Pi_HAT/

If I can find some spare time in between pet projects, maybe I will!

I don't believe the CutiePi has a microphone, or speakers, AFAIK.

Well my initial thought was more along the lines of using some wireless headphones :)

I would literally kill for a tablet that did nothing but very very very very good active pen stuff. I bought a Dell latitude tablet and while it's good enough (one note is also just good enough) that I don't regret it I've been wondering whether I could put together something that just focused on digitization using something a pi. are there any plans to support an active pen?

I've heard good things about https://remarkable.com

yes me too but that is basically as expensive as my dell tablet for just writing? it seems like to high a premium (yes i know i said i would kill).

'Literally kill' you said :) . I got the HP Chromebook X2 with stylus and while some chrome os rough edges persist(crostini apps don't work with on screen keyboard, for eg), it is fast and responsive with great battery life so far.

Didn't mention humans, though. Maybe they were going to kill a housefly, or a roach, or mosquito, or something that typically gets killed anyway, without even getting good stylus input for it?

Looks great but @ $499 in the US, the price is steep for experimentation. I'd love to rent one, see how it matches with me and then buy it in full. If this was under 200 I'd just risk buying and finding if I like it later.

Don't know much about that pen thingy so no guarantee, but will definitely look into it.

Make sure you don't get caught though :P

The screen on the frontpage photo says "armv7l". I think that's only printed on the Raspberry Pi 4. The specs say it's a Compute Module 3+, which I'd expect to print "armv7".

So either the photo is not genuine or the pictured device is using a Raspberry Pi 4 (possibly a pre-release Compute Module thereof).

hi, on Pi4 it would be something like "Linux raspberrypi 4.19.46-v7l+", it's totally confusing.

More info regarding Pi3 arch https://www.raspberrypi.org/forums/viewtopic.php?t=140572

Looks very cool, for sure.

But am I the only one who is confused why they created a main board instead of just using a Pi model B? Most of the things on their board (usb, hdmi, wifi, Bluetooth, card slot, etc) are already there in a Pi. Especially since the boards are about $90 each...

Form factor is one, also proper power management is very hard to get right with Pis.

Of course we made our first prototype with Pi3, and it ended up looking like this: https://i.imgur.com/gqUAdDa.jpg :P

Oh by the way we're not using HDMI display so it'll be cheaper (and thinner) too.

I'd say the form factor. The raspberry pi is too thick for a tablet.

If you ripped off the ethernet and usb ports would it still be too thick? Maybe the hdmi also.

Ohh we did exactly that! https://i.imgur.com/QTRagEP.jpg

But it's not scalable, and I'm not particularly good at soldering. ;-)

Well they used the compute module so that's exactly what they did. And then they added back the ports where they needed.

I appreciate what they did and why. But no, that's not the exact same thing as the raspberry pi with it's ports de-populated. You need to provide power and ethernet (if required) yourself, I think.

The compute model is the right choice for an actual product (i.e. making more than one), but for someone trying to jam a pi into a thin space, removing the ports is almost as thin. Plus you don't also need a motherboard.

The compute module is also still waiting for it's BCM2711 refresh, another reason it's not the same thing. But when that does arrive, if it's just a case of swapping the modules that'll be a brilliant upgrade for this project.

> OS: Raspbian + UX built with Qt

I'm curious about this part. Their Github only has their website, hardware design, firmware and drivers: https://github.com/cutiepi-io/

Sorry the UX is still under development, will be release when it's ready.

As a long-term Qt user, you can find my other tablet-related projects on Github, for example the browser https://github.com/penk/SlateKit/tree/master/Shell

neat-o project. i love my ipad, but this indie piece if screen and tracking is excellent would be a total bomb. all the best!

I thought that case design reminded me of something, then I remembered the one laptop per child (OLPC) project from years ago. Apparently it is still going and their latest offering is based around a celeron processor, not sure they would even consider selling them to the general public though. http://dev.laptop.org/~quozl/catalog-2019.pdf

I just looked at their email distribution lists and they are pretty much dead. Last dev email from May was about XO-1.75 when XO-4 Touch is available.

I've been really meaning to see more devices taking advantage of raspberry pi compute module and selling it as ready to ship packages. I wonder how much this is being sold for.

What I'd really love to see / make is a pre-packaged raspberry pi robot with wheels, cameras, microphones, simple LIDAR, IR lights for night vision and a simple grabber arm. Cheap, nicely built, durable robot.

I'm not a hardware guy but if someone wants to partner, I wanna do a kickstarter someday.

This uses the CM3 Lite. According to a blog post in January, 2019 on the release of the CM3+, the “legacy” CM1, CM3 and CM3 Lite products are considered “not recommended for new designs”. [1] https://www.raspberrypi-spy.co.uk/2019/01/introducing-the-ra...

hey, thanks for the comment, we know that CM3 and CM3+ are pin-to-pin compatible, so will switch to it when launch! :-)

(finger crossed for CM4?)

Great feature! End-user upgradeable simply by swapping a device's compute module with a newer iteration from RPi.

c.f. EOMA68

Looks promising. The one thing I wish this had was some design files or something for the case, so that I could screw it in to a VESA mount.

Use a vaccum/suction cup mount. It should hold well enough.

If one wants to create a wall mounted, always display on, system are there any of like CutiePi with an e-Ink display?

I do find SPI e-ink displays on taobao, not sure about touch: https://world.taobao.com/item/558792955821.htm

After working with people in Taiwan for the past two years: I love the creativity coming out of there.

Compared to mainland China, people in Taiwan seem more eager to integrate with the world. Open source is popular in mainland China too, but English is not.

This thing they're building seems awesome.

Learn Chinese?

That's like a 2-4 year task, it seems.

Faster if you answer your robocalls.

Pardon if this is a dumb question, but does this thing look like it could decode and play MP3 files? With a little power left over for a simple user interface? I'm interesting in making a specialized MP3 player of sorts.

I think your toaster can probably decode & play MP3 files these days.

This is entirely the wrong product for that. The SoC on the Raspberry Pi is a somewhat high-powered graphics processor; if all you need is MP3 decoding, your needs will be much better met by a mid-range microcontroller. There's even ASICs (like the VS1011) that'll do most of the work for you.

In my consumer experience, playing mp3 or most other vaguely-popular formats takes negligible power compared to the display or modern app/UI logic.

What's the biggest difference between this and a regular tablet running linux (I'm not sure they exist but I'm assuming most android tablets can run any Linux?). Is it the addition of GPIO, or somethinbg else?

> I'm assuming most android tablets can run any Linux

This is definitely not the case. There are almost no android tablets that people have even bothered figuring out how to boot normal linux on, and I'm not sure there are ANY that have complete driver support.

Other than this your best bet would probably be to try to find an x86 tablet from 5 years ago and even those are a huge pain to use linux on.

Value proposition time haha :-)

1) you have access to the Raspbian and Pi ecosystem 2) it has actual usable UI, commercial-grade I would say

The whole idea of this project is to help makers to be portable with their projects. The display is optional.

So the point of the device is to be able to develop pi stuff which is then most likely used on a regular pi when deployed? Am I interpreting that right?

yes, and when you're done, it's portable.

That's a clever use of the compute module. I'll buy one

I like the handle cum stand, that in itself seems enough of an innovation to base a tablet on. I wonder if a keyboard could be slotted through to make a laptop???

Will the GPIO pins include the hardware SPI, I2C, or UART pins?

I'm excited about the extra compute available from the RPi4, but obviously the CM4 isn't out yet.

I believe it's GPIO7-10, please find J20 from the schematic, hope we can free up more in the future: https://github.com/cutiepi-io/cutiepi-board/blob/master/SCH_...

I'm also excited about Pi4, but on the other hand worried about thermal issue...

I always wondered why nobody had release a tablet / laptop type system that accepts the compute module... I guess it's finally happening!

Will there be GPIO pins exposed somewhere?

"Everything you see here is open source -- schematics, PCB, drivers, firmware, UI, everything."

The name of this reminded me of an 8th grade math poem.

>Roses are red.

>Math is pretty fly.

>I like your smarts.

>So why not be my cutie 3.14159

That must be an American thing, in Asia it's at least 12 decimal places (jk).

Americans are lazy.. 22/7

sure, but what do they do with the other 2 hours of the day, 7 days a week?

Handle and design loosely reminds me of the OLPC. It's cute (and aptly-named)

The original xo tablet rendering is super cool, very hard to compete :-)


This would be be awesome with a low latency e-ink display!

I do find SPI e-ink displays, but not sure about touch panels..


Looks like vaporware. Doesn't announce a price. I won't hold my breath.

Also, double-opt-in for e-mail should be the default now. This screams 'scam' to me.

If you had read the thread before commenting, you'd find that the developer of this used to work at Canonical, which lends them at least some credit.

Just ticked double opt-in in Mailchimp. But man don't troll me, I have that vapor sitting on my desk. ;-)

Looks promising

<3 !

Now we're just waiting for American Pi, in other words.

Call me dull, but why don't you just build an android app and buy a Fire Tablet to power your project.

A Fire Tablet is probably on par for power and is fully integrated with a touch screen and battery, and with discounts it costs almost or cheaper than a bare Pi.

I feel like we bend over backwards to try to use the Raspberry Pi to do cool things, but a lot of these projects can be developed more easily and powered by a more consistent framework using a Fire Tablet.

Because the boot loader is locked, so does its launcher, unless LauncherHijack installed and abuses Accessibility feature to draw Overlays on top of your screen....

Speaking of modifying existing hardwares, Nexus 7 is my personal favorite: http://i.imgur.com/8H9LH9U.jpg

What projects do you need a tablet and you need it unlocked?

Anything related to flashing eMMC, anything kernel related, especially KMS/DRM, bionic libc... etc.

If you don't find it limited with locked device, you're simply not playing hard enough. :-))

I'm guessing you've never tried to modify a Fire tablet. It's an exercise in extreme frustration. I will never "buy" a Fire tablet again. I say "buy" in quotes because the way Amazon locks down the tablet means I clearly don't own it. It's just a long term pre-paid lease.

Can't you just install a regular android app? Why do you need to unlock it to do anything?

You'll always have just a layer on top of FireOS. Which means you can't rely on security updates for terribly long, and have to fight the system every step of the way. It could be made to work for a lot of uses, but it's way more friction than just running GNU/Linux to start with.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact