[Arm-netbook] http://elinux.org/Embedded_Open_Modular_Architecture/EOMA-68#Table_of_EOMA-68_pinouts
more revisions to the interface, these are last-minute just before finalising:
* the revision 2.2 A20 cpu card * the revision 0.1 ICubeCorp IC1T cpu card * the revision 0.1 Ingenic JZ4775 cpu card * the revision 0.1 Micro-Desktop base board * the revision 3.0 Tablet / Mini-Laptop base board * also i have received the documents for the quad-core Allwinner A33 so will be starting a CPU card for that with 2GB RAM, shortly.
i will have enough money coming in beginning of this month to send the revision 2.2 A20 CPU Card off to the factory for 5 samples. the revision 0.1 Micro-Desktop will follow very shortly after that.
the IC1T icubecorp found me an SD-based TSOP-48 from a company called FORESEE, which comes in a low-cost 2GB variant. i will add that in (replacing the BGA variant which is 4Gb minimum, requires very small tracks, is more expensive and is making me nervous).
the jz4775 cpu card i had to do a minor shuffle of the I/O lines as there was some I/O multiplexing sharing SPI with MMC so there is a bit more work there.
heck of a lot going on, it is ironic that about four really good CPU cards should all come out pretty much at the same time...
l.
Luke,
are you aware of this funding platform:
here is difference between kickstarter:
http://www.core77.com/blog/starting_out/what_you_need_to_know_about_crowd_su...
As you can see one open source laptop did really good:
https://www.crowdsupply.com/kosagi/novena-open-laptop
On Thu, Jul 24, 2014 at 10:56 PM, Luke Kenneth Casson Leighton < lkcl@lkcl.net> wrote:
more revisions to the interface, these are last-minute just before finalising:
- the revision 2.2 A20 cpu card
- the revision 0.1 ICubeCorp IC1T cpu card
- the revision 0.1 Ingenic JZ4775 cpu card
- the revision 0.1 Micro-Desktop base board
- the revision 3.0 Tablet / Mini-Laptop base board
- also i have received the documents for the quad-core Allwinner A33
so will be starting a CPU card for that with 2GB RAM, shortly.
i will have enough money coming in beginning of this month to send the revision 2.2 A20 CPU Card off to the factory for 5 samples. the revision 0.1 Micro-Desktop will follow very shortly after that.
the IC1T icubecorp found me an SD-based TSOP-48 from a company called FORESEE, which comes in a low-cost 2GB variant. i will add that in (replacing the BGA variant which is 4Gb minimum, requires very small tracks, is more expensive and is making me nervous).
the jz4775 cpu card i had to do a minor shuffle of the I/O lines as there was some I/O multiplexing sharing SPI with MMC so there is a bit more work there.
heck of a lot going on, it is ironic that about four really good CPU cards should all come out pretty much at the same time...
l.
arm-netbook mailing list arm-netbook@lists.phcomp.co.uk http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook Send large attachments to arm-netbook@files.phcomp.co.uk
On Fri, Jul 25, 2014 at 6:00 AM, Hrvoje Lasic lasich@gmail.com wrote:
Luke,
are you aware of this funding platform:
yes. i have a conference with them on monday.
I visited the following page: http://elinux.org/Embedded_Open_Modular_Architecture/EOMA-68#Table_of_EOMA-6...
I have seen that you have removed LCD Pixel Data bit 0, LCD Pixel Data bit 1, LCD Pixel Data bit 8, LCD Pixel Data bit 9, LCD Pixel Data bit 16 and LCD Pixel Data bit 17. Does this have any effect on the LCD?
I saw you added * 43 POWER#. Would this be for the power button?
2014-07-24 22:56 GMT+02:00 Luke Kenneth Casson Leighton lkcl@lkcl.net:
heck of a lot going on, it is ironic that about four really good CPU cards should all come out pretty much at the same time...
In my opinion, you only need a EOMA-68 with high features (like Allwinner A80 or similar) for hardcore users and the whole market would be covered.
On Fri, Jul 25, 2014 at 12:14 PM, Miguel Garcia gacuest@gmail.com wrote:
I visited the following page: http://elinux.org/Embedded_Open_Modular_Architecture/EOMA-68#Table_of_EOMA-6...
I have seen that you have removed LCD Pixel Data bit 0, LCD Pixel Data bit 1, LCD Pixel Data bit 8, LCD Pixel Data bit 9, LCD Pixel Data bit 16 and LCD Pixel Data bit 17. Does this have any effect on the LCD?
honestly, not a lot. if it was only 5 bits per pixel then yes it would start to be clearly visible as 5 bpp is only 32 shades. 6 bits is 64 shades per colour.
I saw you added * 43 POWER#. Would this be for the power button?
yyup.
2014-07-24 22:56 GMT+02:00 Luke Kenneth Casson Leighton lkcl@lkcl.net:
heck of a lot going on, it is ironic that about four really good CPU cards should all come out pretty much at the same time...
In my opinion, you only need a EOMA-68 with high features (like Allwinner A80 or similar) for hardcore users and the whole market would be covered.
yes but the A80 has PowerVR and if you have any sense don't touch it with a 10ft barge pole. i'll be waiting for the A70 which i hear is their planned pin-compatible quad-core with MALI.
i really want an intel soc but it is going to need a 14nm to make any sense. even the 22nm latest intel SoCs are too power-hungry.
l.
2014-07-25 21:33 GMT+02:00 Luke Kenneth Casson Leighton lkcl@lkcl.net:
yes but the A80 has PowerVR and if you have any sense don't touch it with a 10ft barge pole. i'll be waiting for the A70 which i hear is their planned pin-compatible quad-core with MALI.
What is the problem with PowerVR?
i really want an intel soc but it is going to need a 14nm to make any sense. even the 22nm latest intel SoCs are too power-hungry.
A EOMA-68 with Intel (Windows Pro and Linux x86) would be very interesting.
It also seems interesting Tegra K1 (64 bits).
On Fri, Jul 25, 2014 at 9:03 PM, Miguel Garcia gacuest@gmail.com wrote:
2014-07-25 21:33 GMT+02:00 Luke Kenneth Casson Leighton lkcl@lkcl.net:
yes but the A80 has PowerVR and if you have any sense don't touch it with a 10ft barge pole. i'll be waiting for the A70 which i hear is their planned pin-compatible quad-core with MALI.
What is the problem with PowerVR?
look up luc verhaegen's talk... 1sec... www.youtube.com/watch?v=7z6xjIRXcp4
i really want an intel soc but it is going to need a 14nm to make any sense. even the 22nm latest intel SoCs are too power-hungry.
A EOMA-68 with Intel (Windows Pro and Linux x86) would be very interesting.
yeah exactly.
It also seems interesting Tegra K1 (64 bits).
i want that one :)
arm-netbook mailing list arm-netbook@lists.phcomp.co.uk http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook Send large attachments to arm-netbook@files.phcomp.co.uk
2014-07-25 22:28 GMT+02:00 Luke Kenneth Casson Leighton lkcl@lkcl.net:
A EOMA-68 with Intel (Windows Pro and Linux x86) would be very interesting.
yeah exactly.
It also seems interesting Tegra K1 (64 bits).
i want that one :)
However, I think that due to the limitations of money, it is better to focus on a single EOMA-68 (in this case EOMA-68 A20 with 2GB RAM, because it is already designed) to make known the product, and launch more EOMA-68 later.
On Fri, Jul 25, 2014 at 11:05 PM, Miguel Garcia gacuest@gmail.com wrote:
2014-07-25 22:28 GMT+02:00 Luke Kenneth Casson Leighton lkcl@lkcl.net:
A EOMA-68 with Intel (Windows Pro and Linux x86) would be very interesting.
yeah exactly.
It also seems interesting Tegra K1 (64 bits).
i want that one :)
However, I think that due to the limitations of money,
i'm now earning enough money to pay for at least one PCB set of samples per month.
On 07/25/14 20:33, Luke Kenneth Casson Leighton wrote:
On Fri, Jul 25, 2014 at 12:14 PM, Miguel Garcia gacuest@gmail.com wrote:
I have seen that you have removed LCD Pixel Data bit 0, LCD Pixel Data bit 1, LCD Pixel Data bit 8, LCD Pixel Data bit 9, LCD Pixel Data bit 16 and LCD Pixel Data bit 17. Does this have any effect on the LCD? honestly, not a lot. if it was only 5 bits per pixel then yes it would start to be clearly visible as 5 bpp is only 32 shades. 6 bits is 64 shades per colour.
grabbing wrong end of stick
i though EOMA-68 was a standard that was not going to change?
On Fri, Jul 25, 2014 at 11:21 PM, Simon Kenyon simon@koala.ie wrote:
On 07/25/14 20:33, Luke Kenneth Casson Leighton wrote:
On Fri, Jul 25, 2014 at 12:14 PM, Miguel Garcia gacuest@gmail.com wrote:
I have seen that you have removed LCD Pixel Data bit 0, LCD Pixel Data bit 1, LCD Pixel Data bit 8, LCD Pixel Data bit 9, LCD Pixel Data bit 16 and LCD Pixel Data bit 17. Does this have any effect on the LCD? honestly, not a lot. if it was only 5 bits per pixel then yes it would start to be clearly visible as 5 bpp is only 32 shades. 6 bits is 64 shades per colour.
grabbing wrong end of stick
i though EOMA-68 was a standard that was not going to change?
the standard is a long-term standard that cannot be changed once the first product in the range is launched. as the first product hasn't been launched, i have the advantage and benefit of more time to think "what would a long-term standard have in it?"
also because i am now working right now with *three* CPU Cards and two base units simultaneously i have the advantage of being able to think how to cover as many possibilities as can be done.
the IC1T has only 210 pins for example, yet the revisions made recently *still* allow that IC to be within the standard.
l.
On 07/26/14 08:07, Luke Kenneth Casson Leighton wrote:
On Fri, Jul 25, 2014 at 11:21 PM, Simon Kenyon simon@koala.ie wrote:
On 07/25/14 20:33, Luke Kenneth Casson Leighton wrote:
On Fri, Jul 25, 2014 at 12:14 PM, Miguel Garcia gacuest@gmail.com wrote:
I have seen that you have removed LCD Pixel Data bit 0, LCD Pixel Data bit 1, LCD Pixel Data bit 8, LCD Pixel Data bit 9, LCD Pixel Data bit 16 and LCD Pixel Data bit 17. Does this have any effect on the LCD? honestly, not a lot. if it was only 5 bits per pixel then yes it would start to be clearly visible as 5 bpp is only 32 shades. 6 bits is 64 shades per colour.
grabbing wrong end of stick
i though EOMA-68 was a standard that was not going to change?
the standard is a long-term standard that cannot be changed once the first product in the range is launched. as the first product hasn't been launched, i have the advantage and benefit of more time to think "what would a long-term standard have in it?"
also because i am now working right now with *three* CPU Cards and two base units simultaneously i have the advantage of being able to think how to cover as many possibilities as can be done.
the IC1T has only 210 pins for example, yet the revisions made recently *still* allow that IC to be within the standard.
l.
arm-netbook mailing list arm-netbook@lists.phcomp.co.uk http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook Send large attachments to arm-netbook@files.phcomp.co.uk
in that case:
- remove all connectors from the end of the card - add as much ram as you can and double it
On Sat, Jul 26, 2014 at 12:20 PM, Simon Kenyon simon@koala.ie wrote:
in that case:
- remove all connectors from the end of the card
- add as much ram as you can and double it
:) ok, the three cpus are all very different. the IC1T is aimed at ultra-low cost devices: i believe it's around the $2 mark. its RAM maxes out at 512mb. i've matched it with an SMIC9514 to give it 2 USBs and Ethernet. i'll put it with a TSOP-48 2gbyte NAND from Foresee. it has very limited pins so there will be no extra connectors. i think every single pin goes into the standard compliant interfaces.
the A20 we know, it maxes out at 2gb, i don't want to make major alterations so it stays with the RTL8201 (100mb/s eth) for now.
the Ingenic jz4775 can actually address 3gb RAM but for space reasons i am only putting in 2gb, and because it is a new design i am putting it with a GbE Phy. however the kernel driver needs improvement: as-is it can manage 300mbit/sec. there is spare pins for an SD/MMC as well as a NAND IC however that is the lot: no extras for HDMI, no extras for USB-OTG: unlike the A20 there is only 1 USB-OTG (which goes to the standard connector) and one 11mbit/sec USB 1.1 (which also goes to the standard connector).
so they are all very different and all maxed out on space and RAM already.
l.
arm-netbook@lists.phcomp.co.uk
-
Hrvoje Lasic -
Luke Kenneth Casson Leighton -
Miguel Garcia -
Simon Kenyon