[Gta04-owner] Invitation to indicate interest in next batch of GTA04A5 boards

[Dr. H. Nikolaus Schaller]

Hi Bob,

Am 26.03.2013 um 17:19 schrieb Bob Ham:

> On 2013-03-26 14:11, Dr. H. Nikolaus Schaller wrote:
>> Am 26.03.2013 um 14:18 schrieb Bob Ham:
> 
>>> I believe the high cost can only be dealt with by designing a new machine with more modest specifications; a "feature phone".  The main feature of the GTA04 is actually its freedom, not technical capabilities.  I'm sure there are many free software supporters who would be happy with a phone with which they can just make phone calls, send SMSs and browse the Web, that is reasonably priced, rather than a phone with lots of extra sensors and additional functionality, that is expensive.  It may be that a new design is the only way to deliver this.
>> 
>> Unfortunately it is not the extra sensors and the additional
>> functionality that makes it expensive. They make less than 20% of the
>> total cost. The most expensive parts are (ordered by falling cost):
>> UMTS modem, CPU+Memory, LCD, PCB, WLAN, GPS. But the highest cost are
>> those from low volume production.
> 
> As I understand it, one of the primary design goals of the GTA04 was replacing a GTA01/02 motherboard.  This goal restricted the components chosen, especially the most expensive component, the UMTS modem.

Yes. But there aren't many alternatives for the UMTS modem. Yes, the ones from Telit or uBlox are a little too big to fit into the case and could not be choosen for space constraints. But they are not cheaper than the OPTION module (uBlox appears to be much more expensive). And even if they were cheaper by 20 EUR, this makes only 5% and not 50% of the total BOM cost.

> 
> With the availability of 3D printing, both the plastic case and the case kit components are no longer bound to the legacy of the Openmoko phones.  Designing a new, unconstrained phone would open up the possibility of choosing components with a lower cost.

Hm. The component costs are not really driven or influenced much by Openmoko compatibility. They are mainly driven by availablility in small quantities. If we need e.g. 100 CPUs they must be listed at distributors like DigiKey and Mouser, even if other chips are cheaper at MOQ of 1k. Generally the distributor takes the risk of scrapping the remaining 900 if we remain their only customer for 100 CPUs. And for that we pay a big add-on (sort of an insurance fee) for our 100 units.

> 
> Reducing the cost of the GTA04 by 27% (WLAN and GPS is 7%) to start, coupled with the reduction from using low-cost components, and I can easily envision the market price encroaching on the territory of economic feasibility.

Well, that still is 400 EUR for a device without WLAN and GPS (which most Androids @ 100 EUR have). The key problem is that "free and open and well documented" isn't worth the difference for the majority of the world population. Except for our small group.

Generally the component cost goes down if we can order let's say 10k...

A rough estimate is that if 1-10 units is 100%, 100 is 80% 1k is 60% 10k is 40% and 1 Mio is 20%. So we could expect 250 EUR for a full GTA04A5 board (i.e. ca. 350 EUR for a complete device) if we can produce a batch of 10k needing a credit line of 2.5 Mio EUR...

But I doubt that we can get 10k non-Openmoko freaks interested if we "just" add "free and open and well documented" to a feature-phone. We should have something unique and compelling that *only* real open source can provide (and not even "open" Android or Ubuntu Touch).

BR,
Nikolaus