[Community] Fwd: [Fsuk-manchester] Novena: Open Laptop (using ARM)

Dr. H. Nikolaus Schaller hns at goldelico.com
Thu Feb 14 14:57:53 CET 2013


Am 14.02.2013 um 10:09 schrieb Bob Ham:

> Thought this might be of interest here..
> -------- Original Message --------
> Subject: [Fsuk-manchester] Novena: Open Laptop (using ARM)
> Date: Wed, 13 Feb 2013 21:56:35 +0000
> From: Michael Dorrington <michael.dorrington at gmail.com>
> To: Manchester Free Software <fsuk-manchester at nongnu.org>
> Andrew "bunnie" Huang¹ and friends are building an open laptop.  Its CPU
> is an ARM and it aims to be blob free, though they look stuck on the GFX
> driver, we discussed this issue with ARM on the list.  It has some cool

According to the presentation about ARM&GPU at FOSDEM all GPUs need
a binary BLOB - except Mali where reverse engineering was very successful.

> hacker² features like dual Ethernet and CPU programmable FPGA.  His blog
> entry describing is at:
> http://www.bunniestudios.com/blog/?p=2686
> Regards,
> Mike.
> ¹ http://en.wikipedia.org/wiki/Andrew_Huang
> ² http://en.wikipedia.org/wiki/Hacker_%28programmer_subculture%29

I did already read something about the project, but what I did miss was the
latest blog entry with a link to the i.MX6 "non-NDA" data sheets:


Although they are not as elaborated as the TI OMAP data sheets and
manuals, it is good enough to start something with.

Another interesting aspect is the availability of chips. Mouser now sells
them starting at single units! At a reasonable price comparable to OMAP

But there are other aspects to consider if they can be used for our
goal of portable smart devices:

* dimensions
* power consumption

I think the i.MX6 chip is 28x28mm and is not PoP. I.e. memory needs
additional space side to side of the CPU. So it needs at least 5-6 times
as much PCB area as an OMAP3+PoP (12x12mm)

A first hint at power consumption can be found in the data sheet (was
around page 40 if I remember correctly). The i.MX can draw up to 3.8 Amperes
at 1V core voltage. This translates to approx. 1 Ampere at nominal 3.7V
battery voltage. So getting the full power out of such a chip may drain
the battery within less than 1 hour...

And even more important is the suspend current. If I read the values correctly
we have to expect some 10-20 mA even in sleep.

So in general I would deduce that OMAP processors are less power hungry.
Because they are less powerful and have been optimized for mobile devices.

But maybe I am wrong at my first studies? And it is never wise to decide
on missing knowlege.


More information about the Community mailing list