wraggster
June 20th, 2010, 22:20
News from the Openpandora Team
Hi folks!
As promised, here is our current report about things that happened and things that will happen.
First, let's start with the most important information for most of you:
How is Mass Production going?
Well, I'm honest here: Could be better - but could've been A LOT worse.
All in all, we're pretty content. And a bit proud it all worked out and there are now real units out there! We can either speed up production but have a bad quality - or work a but slower and make sure the units we deliver are properly working (and that's the path we chose).
There were a few issues we had to solve - but that's normal for a first batch. And it's nothing serious - thank god!
Here is a quick list:
* Shoulderbuttons: The first units had been assembled as fast as possible - without taking much time to optimize the shoulder buttons as good as possible. Ever since we saw the first complaints about them, we take more time. And they are working good now, so... issue fixed.
* LCD-Cables: Quite a few LCD cables failed after assembly. A few ones failed after delivery. This was a bit puzzling - but we found the issue: Spikey solder blobs on the speaker cables killed the LCD cable when closing the lid. Fix for now: Be more careful while soldering them. Fix for Batch 2: Move the solder points beneath the LCD cable. Since we know about this, no LCD cables failed anymore. Issue fixed :)
* Nubs: This is some weird thing. We had quite a few broken ones which had been send to customers who didn't care. Some were fixable. We have fix for now: We got 14.000 nubs, more than enough for 4000 boards. Michael set up a device to test the nubs before soldering them in. The good ones will be sent to the MP company in Texas to be used in the boards, the bad ones back to the nub production company. They need to figure out what's wrong with them. Still, as we have more than enough nubs for the boards, this issue is basically fixed for the first batch.
* Sound: On some units, the sound was scratching quite often. After I got hold of such a unit and sent it to notaz, he found it was a software bug. The next HotFix will have that one included... so, issue fixed :)
* Boards: On the first batch, quite a few boards came broken from the mass production company. Some problems were apparent (parts missing, sometimes even big ones like the powerswitch), some only occured after testing each functions. This was kind of a bummer - since the boards had been tested there... That did cost us a lot of time, as we were not expecting broken boards - and therefore built a lot of units just to find out they have issues when testing them. It also costs time and money, since those boards need to go back to be fixed. We told the company to test them better - and the last 200 boards we got were way better and more reliable. So yes, issue seems to be fixed.
Basically, all problems have been fixed so far. This slowed things down at the beginning - but a two-man team can now assemble 50 Pandoras in an 8-hour-day including quality testing. As Craig has a four-man team, 100 Pandoras could be built per day (as long as we have the boards and cases ready)
And of course - if you have any hardware problems with your Pandora, they will be fixed for free!
So far, about 600 Pandoras have been built and delivered (or are will be delivered next week) About 100 had some issues we need to check. This might be broken LCD cables but fine boards, broken boards that need to go back to Texas, etc.
Production was slow but steady until now, but speed has continually increased while maintaining the quality control. The MP company is also working to make production of the boards more reliable - not only to please us: They only get the money for working boards, so broken boards need to be fixed by them for free, and that's something they surely try to avoid.
All in all, we're pretty content it worked out that well. We feared the worst, but it turned out to work out pretty well. Very few RMAs so far, a lot of very happy users on the boards.
The Cases
Except for the painted buttons, the cases were fine, so there was no need to not continue the mass production. In fact, we already ordered the next ones weeks ago. I can't give you an update though, as Fatih is currently on a well-deserved honeymoon (yep, he got married just after the Pandora production began) and therefore is not online. He's still in regular contact with the chinese company to ensure they are working on it.
I'll have more information once he's back to work.
The software - and future development
Since quite a few developers already received their Pandoras, there's currently a lot going on in OS / Software development for the Pandora.
There's currently being a list setup on the Wiki which includes a To-Do list and what the status is on some issues. Also, if you as a user have feature requests or want to help out, be sure to post that on the official bugtracker.
If you want to help out as developer: We're just organizing everything (IRC, GIT, etc.) and will let you know here when everything is setup!
Finally, quite a few fixes are coming up with the next HotFix (probably within the next days).
New things that will be included (already done):
* U-Boot: Enable serial output on USB (which is really helpful if you want to create your own OS for the Pandora or port Debian, ArchMobile, Gentoo, etc.)
* Kernel: Fixed some ALSA stuff (less buffer underruns, fixed broken sounds on some units)
* SD-Card mount: SD Cards now only use mmcblk* if they have no label (name). As soon as you label them, the label is used for mountpoints (which fixes some issues if emulators / apps remember the last used directory and you put the card into a different SD card slot)
* SD-Card mount: Now using dirsync with 2 sec buffer flush instead of sync. Speeds up SD Card write speed A LOT (from 220KB/s to 5,6MB/s!)
* Startup Settings: You can now disable and enable startup services (like Samba, USB-Network, etc.). This can help you increase the boot speed and performance
New things that will probably be included (working on them):
* Enable Mass Storage access for Pandora SD Cards (so you can use the Pandora as Card Reader)
* Modify Alsa config so that you can use Bluetooth headsets (I can now test that since some community member sent me a headset - THANKS! :D)
* Create a script to tweak settings (like LCD Refresh rate, SD Buffer flush, etc.)
* ... some other things I can't remember right now :D
Mass production becoming better each day, more and more units delivered, software development optimized and a new fixes done every day... things are really looking good :D
http://www.openpandora.org/index.php?option=com_content&view=category&layout=blog&id=2&Itemid=2&lang=en
Hi folks!
As promised, here is our current report about things that happened and things that will happen.
First, let's start with the most important information for most of you:
How is Mass Production going?
Well, I'm honest here: Could be better - but could've been A LOT worse.
All in all, we're pretty content. And a bit proud it all worked out and there are now real units out there! We can either speed up production but have a bad quality - or work a but slower and make sure the units we deliver are properly working (and that's the path we chose).
There were a few issues we had to solve - but that's normal for a first batch. And it's nothing serious - thank god!
Here is a quick list:
* Shoulderbuttons: The first units had been assembled as fast as possible - without taking much time to optimize the shoulder buttons as good as possible. Ever since we saw the first complaints about them, we take more time. And they are working good now, so... issue fixed.
* LCD-Cables: Quite a few LCD cables failed after assembly. A few ones failed after delivery. This was a bit puzzling - but we found the issue: Spikey solder blobs on the speaker cables killed the LCD cable when closing the lid. Fix for now: Be more careful while soldering them. Fix for Batch 2: Move the solder points beneath the LCD cable. Since we know about this, no LCD cables failed anymore. Issue fixed :)
* Nubs: This is some weird thing. We had quite a few broken ones which had been send to customers who didn't care. Some were fixable. We have fix for now: We got 14.000 nubs, more than enough for 4000 boards. Michael set up a device to test the nubs before soldering them in. The good ones will be sent to the MP company in Texas to be used in the boards, the bad ones back to the nub production company. They need to figure out what's wrong with them. Still, as we have more than enough nubs for the boards, this issue is basically fixed for the first batch.
* Sound: On some units, the sound was scratching quite often. After I got hold of such a unit and sent it to notaz, he found it was a software bug. The next HotFix will have that one included... so, issue fixed :)
* Boards: On the first batch, quite a few boards came broken from the mass production company. Some problems were apparent (parts missing, sometimes even big ones like the powerswitch), some only occured after testing each functions. This was kind of a bummer - since the boards had been tested there... That did cost us a lot of time, as we were not expecting broken boards - and therefore built a lot of units just to find out they have issues when testing them. It also costs time and money, since those boards need to go back to be fixed. We told the company to test them better - and the last 200 boards we got were way better and more reliable. So yes, issue seems to be fixed.
Basically, all problems have been fixed so far. This slowed things down at the beginning - but a two-man team can now assemble 50 Pandoras in an 8-hour-day including quality testing. As Craig has a four-man team, 100 Pandoras could be built per day (as long as we have the boards and cases ready)
And of course - if you have any hardware problems with your Pandora, they will be fixed for free!
So far, about 600 Pandoras have been built and delivered (or are will be delivered next week) About 100 had some issues we need to check. This might be broken LCD cables but fine boards, broken boards that need to go back to Texas, etc.
Production was slow but steady until now, but speed has continually increased while maintaining the quality control. The MP company is also working to make production of the boards more reliable - not only to please us: They only get the money for working boards, so broken boards need to be fixed by them for free, and that's something they surely try to avoid.
All in all, we're pretty content it worked out that well. We feared the worst, but it turned out to work out pretty well. Very few RMAs so far, a lot of very happy users on the boards.
The Cases
Except for the painted buttons, the cases were fine, so there was no need to not continue the mass production. In fact, we already ordered the next ones weeks ago. I can't give you an update though, as Fatih is currently on a well-deserved honeymoon (yep, he got married just after the Pandora production began) and therefore is not online. He's still in regular contact with the chinese company to ensure they are working on it.
I'll have more information once he's back to work.
The software - and future development
Since quite a few developers already received their Pandoras, there's currently a lot going on in OS / Software development for the Pandora.
There's currently being a list setup on the Wiki which includes a To-Do list and what the status is on some issues. Also, if you as a user have feature requests or want to help out, be sure to post that on the official bugtracker.
If you want to help out as developer: We're just organizing everything (IRC, GIT, etc.) and will let you know here when everything is setup!
Finally, quite a few fixes are coming up with the next HotFix (probably within the next days).
New things that will be included (already done):
* U-Boot: Enable serial output on USB (which is really helpful if you want to create your own OS for the Pandora or port Debian, ArchMobile, Gentoo, etc.)
* Kernel: Fixed some ALSA stuff (less buffer underruns, fixed broken sounds on some units)
* SD-Card mount: SD Cards now only use mmcblk* if they have no label (name). As soon as you label them, the label is used for mountpoints (which fixes some issues if emulators / apps remember the last used directory and you put the card into a different SD card slot)
* SD-Card mount: Now using dirsync with 2 sec buffer flush instead of sync. Speeds up SD Card write speed A LOT (from 220KB/s to 5,6MB/s!)
* Startup Settings: You can now disable and enable startup services (like Samba, USB-Network, etc.). This can help you increase the boot speed and performance
New things that will probably be included (working on them):
* Enable Mass Storage access for Pandora SD Cards (so you can use the Pandora as Card Reader)
* Modify Alsa config so that you can use Bluetooth headsets (I can now test that since some community member sent me a headset - THANKS! :D)
* Create a script to tweak settings (like LCD Refresh rate, SD Buffer flush, etc.)
* ... some other things I can't remember right now :D
Mass production becoming better each day, more and more units delivered, software development optimized and a new fixes done every day... things are really looking good :D
http://www.openpandora.org/index.php?option=com_content&view=category&layout=blog&id=2&Itemid=2&lang=en