arm-netbook

arm-netbook@lists.phcomp.co.uk

8 participants
1076 discussions

fsf-endorseable ingenic jz4775 cpu card can go ahead
by Luke Kenneth Casson Leighton 12 Mar '15

12 Mar '15

good news: after being very patient and persistent, my contact at ingenic put me in touch with a support engineer who confirmed that the limitation on the jz4775's reported maximum resolution (1280x720) is for the *video decoder* not the actual LCD output capabilities. the actual LCD output may go as high as something like 1700x1700 (not at 60fps obviously) but that things like 1280x800 @ 60hz (the crucial limit of the eoma68 spec) are fine, as is even 1366x768@60hz. however he did kindly point out that if 720p video decoding [for which the full source is apparently available, which is also great] is attempted to be displayed on a 1366x768 LCD (or greater), there is a bandwidth memory limitation between the video decoder and the lcd buffer which would prooobably end up stalling slightly so that 30fps would be hard to achieve, at full-screen. honestly, i really don't see this as being a huge show-stopper enough to warrant exclusion of a low-cost SoC that could be the first FSF-Endorseable candidate to run GNU/Linux OSes with a decent amount of RAM (2Gbytes). as i've had the first revision PCB done nearly 4 months ago, that just leaves financing of the first revision jz4775 CPU Card to arrange. i have one potential sponsor: if anyone else would like to help out that would be great. we're likely looking at around $USD 1,500 for 5 PCBs. l.

4 14

Any news?
by Paul Boddie 02 Mar '15

02 Mar '15

Hello again, I keep checking the recent changes page on the Rhombus Tech site [1] for progress, but it seems a bit dormant at the moment. Has Chinese New Year (and other seasonal pitfalls) served to keep the pace of development a bit more modest than us impatient onlookers might have liked? ;-) Paul [1] http://rhombus-tech.net/recentchanges/

4 4

LCD cables - anyone know where to get them?
by Luke Kenneth Casson Leighton 13 Jan '15

13 Jan '15

does anyone know - preferably have already bought from and so knows they can do it - of a source of LCD cables? i need QTY 3 cables with an I-PEX-20453-040T connector at both ends, length 0.5 metres. wires not FPC. i found a company in germany however they quotes only deal with companies quotes. l.

1 0

Re: [Arm-netbook] The Unplanned Obsolescence of the First Fairphone Device « Paul Boddie's Free Software-related blog
by Paul Boddie 12 Jan '15

12 Jan '15

Joining the discussion late and breaking the threading... ;-) Nico Rikken wrote: > > Having had recent contact with the Fairphone-team on this issue (as a > potential customer), they state that above all they are interested in > working with the community to get this sorted. Now that they have a > dedicated Software Developer in Kees Jongenburger hopefully they'll be > able to turn things around for future models. We know the difficulties, > but they clearly underestimated the impact of their hardware decision. I must admit that I felt a bit bad complaining about their software strategy when there is someone who does seem to care about the matter (at least now, anyway). But I still feel that the initiative managed to store up a lot of problems that they could easily have avoided. As I noted a year-and-a-half ago [1], the publicity materials seemed to play fast and loose with things like software sustainability and platform openness: mocking up Skype running on a phone (or perhaps just taking a screenshot from an existing phone) might work for much-needed stock images, but it sends the wrong message on a number of issues (Skype being proprietary software, uses a proprietary network, has various intelligence agencies and corporations eavesdropping on conversations, is controlled by a single corporation who happens to have a competing software platform and is shaking down Android product vendors for patent licences). Of course, a lot of the potential audience don't care about that: they just want to be reassured that the materials in the phone are ethically sourced before reaching for their usual toys, and they hadn't then (and probably haven't now) widened their ethical concerns in the areas of privacy and transparency. My impression is that those running the initiative either didn't have such widened concerns, didn't want to tackle such things as well as the other stuff, or just got some bad advice. The sad thing is that good advice could have been had for a tiny fraction of the effort that these people have gone to in the areas of materials sourcing and fair working practices, which as far as I can tell, have been thoroughly dealt with and require a commitment to social justice that involves dealing with some fairly entrenched and, particularly in the case of the minerals supply chain, some rather horrible problems. I have to note that I wouldn't have been so aware of the dubious practices of the hardware business myself if it hadn't been for Luke and others documenting it on this list and elsewhere, or of the mishaps that plagued the first attempt at the Vivaldi tablet, but those lessons have been out there for people to learn from, and it seems now that either those responsible for the groundwork just didn't have the community awareness or that they felt that things would mend themselves when selecting manufacturing partners: itself a compromise between the flexibility to get the software done the right way (potentially treading on toes and "offending" existing suppliers), and the flexibility to change their other practices to make the manufacturing workplace generally fairer. But then again, running a choice of a MediaTek product by just a few knowledgeable people would have been enough to set the alarm bells ringing. Maybe Fairphone would have made the same decision, anyway, and as the notorious Stephen Elop speech [2] noted (dredged up recently in another blog post of mine [3]), cobbling together MediaTek designs and throwing them over the wall was (and undoubtedly still is [4]) common practice, to the point that doing something else might have been a struggle. But as we all know, you have to make and sustain investments to do the right thing. If you look for short-term fixes, you get short-term solutions, and that's what some of Fairphone's customers will be experiencing in the coming months and years. Again, Fairphone have committed themselves to ongoing investments in other areas, and it's a shame that they didn't consider the technological realm worthy of the same consideration. Paul P.S. I just realised that this could have been another complete article. Sorry to make it a long message for this list! :-) P.P.S. For those who haven't read it yet, Bunnie's MediaTek reverse- engineering article [4] below is long but interesting reading. You've got to admire Bunnie's determination! [1] http://blogs.fsfe.org/pboddie/?p=168 [2] http://www.engadget.com/2011/02/08/nokia-ceo-stephen-elop-rallies-troops- in-brutally-honest-burnin/ [3] http://blogs.fsfe.org/pboddie/?p=835 [4] http://www.bunniestudios.com/blog/?p=4297

3 4

CM108AH
by Luke Kenneth Casson Leighton 11 Jan '15

11 Jan '15

ironically, i have had a CM108 chipset sitting right here, in continuous use, for about 7 years now. i bought a usb dongle in canada back in 2007. i had my suspicions, and i've been able to open it up and confirm that yes it's the CM108. i've made a couple of minor changes to the tablet's audio circuit as a result. all good! l.

1 0

The Unplanned Obsolescence of the First Fairphone Device « Paul Boddie's Free Software-related blog
by Alexander Stephen Thomas Ross 09 Jan '15

09 Jan '15

https://blogs.fsfe.org/pboddie/?p=802

3 5

tablet, microdesktop and laptop progress
by Luke Kenneth Casson Leighton 09 Jan '15

09 Jan '15

well... some good news, we have sponsors for both the 7in tablet and also a 15in laptop, which is amazing! the laptop we have chosen a 15.6in 1366x768 LCD because anything lower than that (either in size or in pixel resolution) is no longer properly fashionable, it appears, whereas i am surprised to find that 15.6in 1366x768 LCDs are not only extremely popular (over 50,000 for sale on panelook.com), but also their power requirements are far more modest than one might expect: around 3.5 watts for the backlight which is pretty incredible. with the popularity of netbooks and tablets taking over, it kinda makes sense that even 13.3in LCDs are just not hugely popular. so... 15.6in it is. i do want to do a 1920x1080 15.6in... but not right away. i don't want to push things too much (EMI wise) and also even finding the faster RGB/TTL to eDP or dual-LVDS ICs is... tricky. so we tackle that later. the only downside of using 1366x768 is that it is above the limit of 1280x800 set for EOMA68 in a 5.0in form-factor. i am not going to get massively unhappy about this, but the CPU Card has to remain *inside* the unit as a non-user-removable part. the hilarious thing is that the ICubeCorp IC3128 will actually work! i tested it yesterday and it can run 1280x1024 at 60hz (if you set 8bpp - 3Red, 3Green, 2Blue). i find this to be very very funny. a $USD 2 Quad-Flat Package processor running in a 15.6in laptop. with sufficient funding however we may be able to go straight to a 1920x1080 EOMA68-compliant laptop, by redesigning the EOMA68-A20 to fit within the 3.3mm PCMCIA form-factor. that will require some metalwork tooling costs, hence why this hasn't been done yet. the tablet: apologies but i did some _more_ layout changes, i decided to move the USB hub IC right next to the EOMA68 connector, in order to reduce the possibility of impedance mismatch problems (making the USB tracks leading to the USB hub incredibly short). that meant moving the microsd card, that meant moving the power connector and lengthening the board on the right-hand siiide... maan this 3D stuff is tough! :) oh, if anyone is interested, the pyopenscad source code for both the tablet and the laptop are available here: http://hands.com/~lkcl/eoma/kde_tablet/3dcase/ under the GPLv3 software license. l.

2 2

revision 2 EOMA68-IC3128 board works
by Luke Kenneth Casson Leighton 07 Jan '15

07 Jan '15

i'm delighted to be able to announce that the revision 2 of the EOMA68-IC3128 ICubeCorp board is 100% working. the board arrived from HK at the end of last year, but the hub appeared to be not working and neither was the MCS7832 USB-Ethernet IC. it turns out that the hub is working fine: the preliminary kernel support for the IC3128 at present only supports high-speed devices, so for example a mouse or keyboard (which are only USB 1.1 devices) are not recognised but a Genesys Logic USB SD/MMC device, a USB SATA drive, or any other 480mbit/sec device, are all supported. the on-board MCS7832 however was configured (by 3 resistors which should not have been added) to be a low-speed device. removing these three resistors the MCS7832 worked perfectly. given that the kernel support for USB 1.1 speeds is a software issue this board can be added to the crowdfunding campaign as an FSF-Endorseable option. woo! l.

2 1

tablet casework prototype done
by Luke Kenneth Casson Leighton 31 Dec '14

31 Dec '14

http://rhombus-tech.net/community_ideas/kde_tablet/news/ the mendel90 printer is working really well, and was absolutely worth it to use for prototyping of casework. the 7in tablet project is now being done as an open hardware project, under the GPLv3+. files are available at http://hands.com/~lkcl/eoma/kde_tablet with a preliminary casework finished that has been prepared in conjunction with the pcb layout, work can proceed to finalise the pcb and then send it off for assembly. there are a couple of risky areas on the tablet pcb, due to datasheets and app notes simply not being available: the usb audio and the usb camera IC. these are a CM108AH and a Vi Micro VC0345 respectively. perhaps prior to sending off this PCB it may be worthwhile tracking down small USB devices with those chipsets in, in order to inspect them and check their PCB layout. however if anyone already _has_ a USB audio dongle with a CM108AH or a USB camera with a VC0345 chipset please do say so as i would greatly appreciate some hi-res photos of the PCB, both sides. l.

3 7

[Arm-netbook] http://rhombus-tech.net/community_ideas/kde_tablet/news/
by Luke Kenneth Casson Leighton 22 Dec '14

22 Dec '14

started on a first revision of the tablet casework, as a pyopenscad program, using a blender-exported STL as the "ghost" template to work out how the case should surround the parts. it's done as layers so may either be made of wooden laminates or 3D printed and assembled using screws/bolts. l.

3 13

Jump to page:

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

arm-netbook