Help Needed Replicating Success Of Openpandora For A Netbook/laptop

[lkcl]

ok, i'm reviving this thread as the A10 EOMA-68 CPU Card now has an active team working on it: http://rhombus-tech.net/allwinner_a10/news/

that means that there will soon be a CPU Card in credit-card-sized format which can be plugged into any chassis with an I/O Board. laptop, desktop, tablet, games console, PVR - anything like that.

someone just asked on the arm-netbooks mailing list if there were any ARM Cortex A9 laptops out there, and as there aren't any, we decided to ask the question again of what people would like in a proper ARM-based laptop, enough to pay for a group buy and to get it actually made.

the primary development cost - the actual CPU+RAM+NAND - is being taken care of, in the form of the (very first) A10 EOMA-68 CPU Card. future EOMA-68 CPU Cards will be available later, and any product that is EOMA-68 compliant will always be upgradeable.

the secondary development cost - around $USD 2,000 - is the I/O Board inside the laptop, which contains an LCD LVDS Driver IC, Backlight circuit, battery charger, and we plan to use an STM32 as a low-cost Embedded Controller that will also do Audio. details are here: http://elinux.org/Embedded_Open_Modular_Architecture/EOMA-68/Laptop

we will go with pre-existing casework: there will be no NREs involved in designing or re-designing the laptop's case. that may mean that the case will be oversized; it may mean that blanking plates will be needed. the pre-existing battery suited to that casework will be used: again, no NREs involved in designing or re-designing the battery or battery compartment. a case which has room for an SATA-II drive will be chosen.

so the primary questions that need answering are: what sort of size laptop would people be prepared to pay for in a group-buy that needed around $USD 2,000 of expenses to be covered by everyone? (if there are 40 people that's $50 each), and what sort of screen resolution would people like it to have?

we *may* design the I/O Board to take an LCD Daughter-Board, so that many screen resolutions and sizes can be considered, but this would have to be declared up-front as it will affect the overall cost of all the laptops.

...thoughts?

 

[Exophase]

STM32 is good for embedded functions, I've used them in things a few times, but using it for audio is a big mistake. Get a real codec chip, you can purchase one for < $2 that destroys what an STM32 can do with its 12-bit DACs.

 

[lkcl]

STM32 is good for embedded functions, I've used them in things a few times, but using it for audio is a big mistake. Get a real codec chip, you can purchase one for < $2 that destroys what an STM32 can do with its 12-bit DACs.

 
wotcha mr exophase.  yeah i took that into account, and am using PWM instead.  the pricing i found for the 72mhz 100-pin STM32F is *two* dollars and thirty six cents in 10k volumes.  $2.36!  i can't possibly justify a $2 codec IC at that kind of pricing!
 
but, yeah: all a bit of an experiment, this.  we'll see what happens, eh?
 
p.s. http://rhombus-tech.net/community_ideas/kde_tablet/news/ - the PCB's useable in anything up to a 10in laptop as well.

 

[Exophase]

Okay let me get this straight: you want to replace the 12-bit DACs with PWM output? Because that's definitely an inferior solution. Have you tried listening to something like this?
 
You're talking about running a PWM at 1MHz. For comparison, GBA runs its PWMs at 16MHz and sounds terrible. Maybe better analog circuitry on the low-pass filter will help but fundamentally you're still limited in resolution * frequency. To get so-called CD-quality, 16-bit @ 44.1KHz you need a PWM running at like 3GHz. Nintendo DS runs its PWM at 33.3MHz (with the rest of the system bus stuff) and gives you 10-bit @ 32KHz quality. It doesn't sound good.
 
Can you list everything you need to use this STM32F chip for outside of audio? Because it's going to have to be a hell of a lot more than that if somehow an audio chip costing ~$1 in quantity is untenable vs your bid of this chip at $2.36.. I don't understand how you can't possibly justify the cheaper and grossly superior solution.
 
Kind of funny, this thread was about replicating the success of the Pandora, one part of that success was using a much better than average audio subsystem instead of a terrible one