I recently taught KiCad how to load Eagle ver. 6.x board files (XML format) using C++ enhancements.
I then went to load
https://github.com/OLIMEX/OLINUXINO/blob/master/HARDWARE/AM3352-OLinuXino/AM3352-OLINUXINO_Rev_A-14122012.brd
to test my new code. At first I thought I was seeing a bug.
Although it loads and looks wrong, then I discovered that this brd file is incomplete? Can you update this file with latest? Or is there a more current storage location for it?
this is not exactly the place for such posts :)
try with Eagle if it also fails this means you didn't download the files correctly, so try to download them again
May i inquiry about AM3352-OLinuXino project's progress please?
TI is going to supply first silicons from the new (bug free) revision P1.2 in May and we will start our production
current revision have some minor but annoying bugs like I2C being not Open Drain but Push-Pull which prevent you to connect many devices on the I2C bus.
Quote from: olimex on March 13, 2013, 07:17:12 PM
this is not exactly the place for such posts :)
I am willing to educated. Tell me a better place to mention that your *.brd file is incomplete please.
Quote
try with Eagle if it also fails this means you didn't download the files correctly, so try to download them again
Doubtful. xmllint shows the *.brd file is a valid XML file after the download.
Here is what eagle version 6.4.0 Light on linux is showing:
http://softplc.com/dick/am3352-on-eagle.jpeg
Many of the components are not even positioned on the board. This is a result of simply loading the *.brd file into eagle, without any project setup. What is wrong?
of course this is what you see, this is preliminary SCHEMATIC not complete board :)
we still work on this PCB and it's not finished
Quote from: olimex on March 14, 2013, 06:00:30 PM
of course this is what you see, this is preliminary SCHEMATIC not complete board :)
we still work on this PCB and it's not finished
Seriously? "Of course"?
This after telling me I had no business even making the post, then that I do not know how to download. Your arrogance is astounding.
I was going to buy 2500 units of one of your boards, now I will have to think about it.
sorry but I really think we have miss-communication here
I'm sorry if I offended you with my post, this is not my intention
surely this forum board is for "New Product Ideas" as it's name says, so your post for problems with opening the CAD files for particular board is not to be placed here as it's off topic - the best way to get help on this would be to e-mail support@olimex.com and explain your problem
from your post I understand that you say the CAD files from Github can't be opened, and this is why I though you didn't download them completely I do remember there was another customer who wrote something like this then re-loaded the CAD files and opened them without problem, I just wanted to help you
the github commit clearly says in plain text "PRELIMINARY SCHEMATIC" I uploaded the schematic and posted on the blog so we can hear comments and suggestions before we complete the routing, more eyes could spot preliminary errors
Quote from: olimex on March 15, 2013, 09:30:37 AM
I'm sorry if I offended you with my post, this is not my intention
Apology accepted.
Hi,
I saw the Allwinner 10s will be available in production soon. :)
Do you have an idea when the AM335x could be also available ?
Thanks
TI notify us that the bug free silicon revision P1.2 of AM3352 will be available in May
Hallo ,
Any new news about AM3352 development ?
Regards
AM3352-SOM module will be ready in couple of weeks and we will test it
that's great news
thx
Hello olimex,
first, thank you for doing such a good job. a few weeks ago we saw the first boards with the 3352 on your blog - when will the download of the CAD files and schematics available?
kind regards
Tim
How is the testing going?
When do you estimate the AM3352-SOM and AM3352-SOM-EVB boards will be available to purchase? What price for low quantities?
Thanks, Brendan.
testing go slow :) we run 7-8 projects at same time
we got stuck at the second Ethernet enable
AM3352 and AM3358 are quite different and all Linux support is done on AM3358 due to BB, so this will take more time than we expect
with Allwinner chips we already have some experience and Linux support happens in matter of weeks, here I'm not so optimistic, despite the thousands of user manual pages, the Linux support for AM335x seems to me more complex than the one on Allwinner chips
Call in help from TI perhaps?
my observation is that TI Linux support is way obsolete than the community support around BB
we will spend some more time on AM3352 after we finish with A20-SOM and move up to date the A13 and A10S Debian and Android support
we have too many boards and the software support becomes a burden
RPi and BB have one single PCB design to support, we have 9 OLinuXinos ready and one in development + 4 new SOMs in development
maybe we should stop making new hardware for a while :D
Why don't you make this available without Linux support?
Not everyone is using Linux... :P
and how to test if hardware is working without software :P
I'm not sure why the AM3352 was chosen for the SOM. If you want to provide an alternative to the BeagleBone (original or black) then an AM3359 or AM3358 would be good, and you get all the BB linux support :)
To me the interesting feature of this processor is the PRU, so I would have put a AM3356/7/8/9 on board to SOM. The GPU doesn't interest me that much, but it would be good to have as an option for some applications.
I hope that the AM3352-SOM can be completed soon and I can prototype with it, and then later be able to replace it with a AM3356/7-SOM or AM3358/9-SOM if need be :)
Does the AM3352-SOM use the 15x15/0.8mm or the 13x13/0.65mm package?
I'm guessing 15x15 as a previous comment mentions getting stuck with 2nd ethernet enable, and 2nd ethernet is only available on 15x15 package, right?
If 15x15, then the SOM should be able to be loaded with 52, 54, 56, 57, 58 or 59 parts, right?
I am primarily interested in the features of the 56 (with PRU-ICSS) or possibly the 58 (with PRU-ICSS and 3D Graphics). I don't believe I have a requirement for EtherCAT, but if I did the 57 and 59 would be required.
There is approximately $4-$5 difference between 56 and 57 and similarly between the 58 and 59 (according to Mouser in quantities of 100)
So it may be good to focus on two SOMs (3352 and 3359). This would give top and bottom end options, and also BeagleBone compatibility/support for the top end 3358/59.
Is this possible Olimex ??
we can produce the modules with any Sitara on them the package is 15x15 0.8 step
minimum quantity to produce module with chip different than AM3352 is 1000 pcs
OK. I couldn't justify buying 1000 units :(
So you don't think Olimex is likely to provide two low quantity options (e.g. AM3352 or AM3359) ?
If not, is it possible to order the board without the AM3352, and for users to have an AM335x of their choice loaded. Not sure if that is technically possible or not ??
I guess the other option is to supply bare boards for end user to populate. Can Olimex supply bare boards ??
Failing that, I think the Eagle design files are available so the end user can have bare boards manufactured and then populated, yes ??
Quote from: olimex on September 13, 2013, 04:05:03 PM
testing go slow :) we run 7-8 projects at same time
we got stuck at the second Ethernet enable
AM3352 and AM3358 are quite different and all Linux support is done on AM3358 due to BB, so this will take more time than we expect
with Allwinner chips we already have some experience and Linux support happens in matter of weeks, here I'm not so optimistic, despite the thousands of user manual pages, the Linux support for AM335x seems to me more complex than the one on Allwinner chips
Hi. Why you chose to use AM3352 instead of AM3358? Since
all Linux support is done on AM3358 due to BB.
Quote from: vinifr on March 26, 2015, 02:58:55 AM
Quote from: olimex on September 13, 2013, 04:05:03 PM
testing go slow :) we run 7-8 projects at same time
we got stuck at the second Ethernet enable
AM3352 and AM3358 are quite different and all Linux support is done on AM3358 due to BB, so this will take more time than we expect
with Allwinner chips we already have some experience and Linux support happens in matter of weeks, here I'm not so optimistic, despite the thousands of user manual pages, the Linux support for AM335x seems to me more complex than the one on Allwinner chips
Hi. Why you chose to use AM3352 instead of AM3358? Since all Linux support is done on AM3358 due to BB.
The 3352 and 3358 are not that different. According TI's documentation, basically the only differences are that the 3352 is missing the PRU and 3D acceleration. The Linux that comes with the BB in no way depends on nor supports the PRU, if you want to use it you are on your own.
Support for dual Ethernet has got nothing to do with the choice of SoC, the 3352 and 3358 are identical in this regard. This is really a distinguishing feature of the AM3352-SOM; the BeagleBones simply chose to ignore support for two Ethernet interfaces.
If Olimex find that people buy many of the current modules then no doubt they'll consider similar modules with the other CPUs.
If you don't see those offered, draw the obvious conclusion.
John
Ok guys, thanks for replies.
But it could not be just a matter of price? :D
Of course it is price - well, money. Olimex want to make money (who doesn't?)
If you want a specific CPU and buy enough they'll even custom build. Quite big numbers, though!
John