A bug, booting up and a huge forum migration


EvilDragon

Administrator
Staff member
Joined
Mar 4, 2003
Messages
29,986
Age
47
Location
Ingolstadt
Sorry for the lack of updates - I catched a flu for a few days, was in pre-christmas online-shop hell (where you get at least 30 orders per day) and was working on the move to the new boards.
Time for an update, right?


1. A bug!
 


It seems I found a bug in the shop system I am using which triggered an email to everyone who registered at the shop for the US Pyra preorder that the product is NOW available... sorry about that!
While we're close, preorder is not yet happening (I stick to my word and wait for the prototype :))


The two who actually managed to order it will get a refund for now.


Sorry again - I didn't know of that bug before, but I'll make sure to not trigger it again.
So the next mail you'll get from the shop regarding preorders will be the REAL one!



2. Pre-Pre-Order limited PCBs
 


All Pre-Pre-Order PCBs (numbered and signed) have been shipped and are on the way to you.
Not all will arrive before christmas I guess, but maybe some of you might get a nice present :)


If you want to show what you do with that board (frame it, eat it or whatever), there's a thread already where you can post about it.



3. It's booting up!
 


Well, not fully yet, but almost, but more on the software after a little hardware info!


Nikolaus spent the last week working with the prototype PCBs.
12 out of 17 he tested (one is currently lying at GC) boot up so far, so 5 need to be checked.

Don't worry about that failure rate, that's normal for a preproduction run.
A preproduction run is used to find out where you need to tweak the building process to lower the failure rate.

The failed ones will be analyzed to find out why they failed.
Could be the placement of the parts is not accurate enough, could be there is too much or too little solder paste, could be the soldering temperature and time can be optimized.


For example, if random contacts below the OMAP5 are bridged on multiple boards, it could be too much solder paste.
If it's always the same contacts that are bridged, it could mean that the OMAP5 itself needs some tweak in the placement.



These are only a few examples, the full process is a lot more complex, but I think you know what I'm getting at :)


So these five PCBs will be analyzed and reworked afterwards.


Okay, that's it for the hardware part, now onto the software part :)



Nikolaus has been working on wiring all the hardware parts together in the software (which means: creating a machine file and bootloader).
The machine file basically tells the system what hardware is where.
The devboard, for example, only had one SD-Card slot where we booted from. The prototype has 3 SD Card slots, so the software needs to know about it.
Otherwise, the bootloader can't find the kernel, for example, as it might be looking at the wrong place.
Also, the Palmas (power chip) disables all hardware it doesn't need as default (to save power), so if it doesn't know the system needs the SD Card slot, it shuts off the power and the kernel can't find any rootfs.


This is all stuff that needs to be implemented - once that's setup, booting in the OS will be a breeze!



So far, Nikolaus is almost finished with that low-level stuff. The system boots up the kernel, but the rootfs cannot be loaded yet. Everything else works! Once that last little thing has been implemented, we're ready to boot the OS.



Oh, and Nikolaus has also worked a bit on the anti-tearing driver for the rotation. He hasn't tested it yet, but that will also happen soon.



4. The case!
 


The case is still coming along well. The big moulds have all been done, the small part moulds (shoulder buttons, etc.) are currently being created.
Since there's now christmas holidays coming along, not much will happen anymore until mid of January.
Therefore, my visit is now planned (pretty much fix this time!) around 21st - 28th of January 2016.
The PCBs should all boot up and work fine until then, so it won't only be a mechanical test, we should be able to really boot the fully assembled unit up then as well!



5. A huge forum merge
 


Almost everyone at the boards hates the new IPB4 (me as well), as it's very slow and partly inconvenient to use (WYSIWYG-Stuff).


So we're moving to a new forum system (XenForo) which is a LOT lighter and faster, has quite a few useful new features I already love and I'm currently working on setting it up.


Additionally, the old gp32x.de boards will be fully merged into it (the user profiles will also be merged) as well as the temporary OpenHandhelds boards (which were created when gp32x.de had a few weeks of downtime) and the file archive will be integrated as well.


It's a huge merge, but I'm almost finished now (after using tons of different regular expressions during the last few days).
Seriously, IPB is a huge mess in my opinion - gp32x.de went from IPB2 to IPB3 as well, and I found at least 8 different ways that quotes worked! It seems IPB has constantly changed the way quotes work but never updated that in the posts, they just let the forum software handle it... it's a mess.

But I've got it almost fully done, only a couple of posts should have quote errors and no content has been lost!


I will start the merge with the OpenPandora boards on December 23rd. These forums will close down for one or two days and then the fully merged forums will arise anew!


That will be a vast ressource for looking up everything regarding OpenSource handhelds.



That's it for today! See you soon on the new boards!
 
Last edited:
All good news. I'm actually extremely surprised at how smooth things are going, especially getting the kernel to boot first try: Nikolaus has done an amazing job and deserves some major praise, that was not an easy puzzle for him to solve.


I'm quite excited now.
 
Sorry for the lack of updates - I catched a flu for a few days...



I hope you are feeling a lot better by now. I appreciate and respect your devotion to the Pandora, Pyra and these boards, but please take some time to enjoy the holidays as well.


With that aside, thanks for yet another informative update. Glad to hear about the forum merger. The IPB4 board are driving me nuts.
 
Phew, I'm glad the bug turned up in your shop software rather than the hardware! (And the flu bug; glad you are better now!)


I guess there's still a lot to be tested as kernel modules start to get loaded and tested, but getting the kernel loaded it certainly a good start.  And only having 5 failures out of a first build of 12 tested so far is a very good result in my experience.  Awesome news that Nikolaus is getting some work done on fixing the tearing.  I took a look at the bug in the bug tracker, but didn't find a definition for all of the acronyms, not really understand how the signals could get read from the driver code, so decided I'd leave it to someone with more direct access to a devboard.
 
Last edited by a moderator:
Same here. When I was reading the description. First thing I thought was, please no Bug in the PCB.  :blink:


Thanks for the update. but do you ever have weekends off??? 
 
Last edited by a moderator:
2 people felt like they were number 1 and 2, sorry for them that it was a silly bug.


Looks like everything (except ED's health, he can get sick? did that ever happen before...) is going okay.


The people that received the pre-order boards appear to be really happy with them, you can clearly see where the wonky chip will be, and glad they share the experience with the board members.
 
I almost bought one of those two pre-orders. Had it sitting in my cart and everything. Changed my mind after a few minutes debate.
 
Hey,


is there any info on the NAND/eMMC chip?


It looks like it's being removed and SD used in place?
 
Hey,


is there any info on the NAND/eMMC chip?


It looks like it's being removed and SD used in place?

The internal NAND/eMMC chip shares the same interface with the internal SD card slot (if SD card is in internal slot, then eMMC chip will remain unused). It can either be "3 SD card slots" or "1 eMMC and 2 SD slots".
 
Why it is 3 SD Card Slots??


I thought 2 fullsice SD and one micro sd..


Looking forward for the first booting prototyps


And that the case seems to need more time is good to me as i dotnt have much money at the moment but i will get some from my stocks at the begining of next yeahr..
 
No, no!


Do not get rid of emmc! Want emmc/nand!


I'm asking for info on the size - is there's been no update and now you're at board population; do you know the size yet?


It SEEMED like it was not mentioned and therefor not being used anymore.
 
No, no!


Do not get rid of emmc! Want emmc/nand!


I'm asking for info on the size - is there's been no update and now you're at board population; do you know the size yet?


It SEEMED like it was not mentioned and therefor not being used anymore.

It Will be there, you can just disable the eMMC and use a uSD instead. Nothing is being removed.
 
What TrashyMG Said (Damnit I was typing.. BOY!), it'd be cool iif we could still use the MicroSD for storage as well as the NAND for boot, dunno if that's possible due to hardware restrictions though


I reckon the size will come to fruition when mass production starts, at this point the OS fits nicely onto 2GB - 4GB MicroSD Card, expect about 8GB but we could be surprised yet :)
 
Last edited by a moderator:
It has not been decided yet.  Prices at ordering time (and possibly, power consumption, I forget) will dictate what parts are ordered.


As I understand it technically, the microSD card will use some of the same lines as the eMMC into the SoC, so there's no way to use both at the same time unless there's some way to electrically alternate between them, although losing the root fs mount intermittently is probably not a good way to run a stable system.
 
Back
Top