Once again, the last update seems so long ago - and while I posted small updates in between in the thread, I think it's definitely time to post another update!
So here's what's happened.
1. The case and the scratches
We had a few quality issues with the coating on the cases we received which the company had to look into.
The first small batch was mostly fine and we assembled most of these already.
The second batch had an issue though: The cases could be scratched using fingernails!
Sounds like the clear coating was missing... but nope, it wasn't missing, as analyzation brought to light, but it was a tolerance issue.
Here's one thing: The more clear coating you put on, the more glossy the case will look like but the more scratch resistant it is.
The sample cases were perfect - but the ones we received this time weren't.
The company did an analysis. All the cases that have been coated had a clear coating of 15 - 25µm (yes, micrometers!).
With 25µm, the coating was enough and scratch resistant, with 15µm it did scratch.
So we have a tolerance of 10µm (that's 0.01mm) for the clear coating.
The machine has now been setup to increase coating thickness, so that the lowest possible value still is scratch resistant.
It still baffles me that such a thin coating can even prevent scratches - but car also has a clear coating of around... 45µm.
As you can imagine, the coat difference has no physical effect on the Pyra, it's not really any thicker
We should receive the new cases within the next few days so we can continue with the assembly.
2. Last minute audio fix
We had the problem that audio streams always died as soon as you enabled the vibration and sometimes even without vibration if the volume was simply set too high.
We always thought it was the driver (after all, the drivers to use the AESS / ABE hardware of the audio chip have not been ported to current kernels yet, so we can't use all of its features yet).
However, Dave found out that this is most likely noise that emits into the audio clock. TI accidentally wired up the audio clocks in the the audio chip in a wrong way - and as a workaround, they suggest to use an inverter (which we haven't implemented).
What would work is if we could delay the clock line a bit... and it seems we managed to do that: There was a leftover from when we planned to use the Pandora Audio setup (which was not possible, sadly), and that trace coming from the Audio chip was still connected to the B2B connector on the CPU board, with a resistor in between. By removing the resistor, it should be possible to delay the clock a bit by adding capacitors... something we could do by hand on already existing CPU boards. Turns out this wasn't even needed - simply removing the resistor worked fine.
That takes 2 - 3 seconds per board - and my Pyra can now play audio and vibrate at the same time
3. Boxes and inlays
The boxes have arrived and look beautiful (I won't post any pictures though, there has to be some surprise for you left).
But yeah, it wouldn't be the Pyra if there wasn't some issue here as well:
The inlay (which goes between the Pyra and the battery) could't be setup without damaging it a bit.
The samples the company produced were plotted, folded and glued.
In the final production, plotting and glueing happens at the same time and you fold them up when you receive them - which didn't really work.
Luckily, I'm working with that company for years. They produce all my packaging (ColUSB, Everdrives, Cinemaware boxes, Retrodes, etc.) and they never let me down.
They already found a different solution and will deliver free replacements within the next 10 days.
4. What's left?
All parts are here, everthing is prepared for assembly.
So now we're just waiting for the fixed cases and the inlays for the packaging.
The one thing I'm currently working on right now is a small manual / artbook.
As you can imagine, a manual doesn't make a lot of sense as the OS is constantly being improved, as it was the case with the Pandora.
So a quickstart manual (explaining the ports and buttons) is all you need.
But I've made a ton of pictures of all the various prototype parts. Case, keymat, PCBs, etc.
These will be put together into the manual including small information of what has been changed and why. So a small evolution of the Pyra - from ideas to prototypes to production.
All pictures have already been made and masked, so I now need to stitch the artbook together and get it printed (printing takes 4 - 5 days).
That's the last missing thing. But something I really want to create.
So here's what's happened.
1. The case and the scratches
We had a few quality issues with the coating on the cases we received which the company had to look into.
The first small batch was mostly fine and we assembled most of these already.
The second batch had an issue though: The cases could be scratched using fingernails!
Sounds like the clear coating was missing... but nope, it wasn't missing, as analyzation brought to light, but it was a tolerance issue.
Here's one thing: The more clear coating you put on, the more glossy the case will look like but the more scratch resistant it is.
The sample cases were perfect - but the ones we received this time weren't.
The company did an analysis. All the cases that have been coated had a clear coating of 15 - 25µm (yes, micrometers!).
With 25µm, the coating was enough and scratch resistant, with 15µm it did scratch.
So we have a tolerance of 10µm (that's 0.01mm) for the clear coating.
The machine has now been setup to increase coating thickness, so that the lowest possible value still is scratch resistant.
It still baffles me that such a thin coating can even prevent scratches - but car also has a clear coating of around... 45µm.
As you can imagine, the coat difference has no physical effect on the Pyra, it's not really any thicker
We should receive the new cases within the next few days so we can continue with the assembly.
2. Last minute audio fix
We had the problem that audio streams always died as soon as you enabled the vibration and sometimes even without vibration if the volume was simply set too high.
We always thought it was the driver (after all, the drivers to use the AESS / ABE hardware of the audio chip have not been ported to current kernels yet, so we can't use all of its features yet).
However, Dave found out that this is most likely noise that emits into the audio clock. TI accidentally wired up the audio clocks in the the audio chip in a wrong way - and as a workaround, they suggest to use an inverter (which we haven't implemented).
What would work is if we could delay the clock line a bit... and it seems we managed to do that: There was a leftover from when we planned to use the Pandora Audio setup (which was not possible, sadly), and that trace coming from the Audio chip was still connected to the B2B connector on the CPU board, with a resistor in between. By removing the resistor, it should be possible to delay the clock a bit by adding capacitors... something we could do by hand on already existing CPU boards. Turns out this wasn't even needed - simply removing the resistor worked fine.
That takes 2 - 3 seconds per board - and my Pyra can now play audio and vibrate at the same time
3. Boxes and inlays
The boxes have arrived and look beautiful (I won't post any pictures though, there has to be some surprise for you left).
But yeah, it wouldn't be the Pyra if there wasn't some issue here as well:
The inlay (which goes between the Pyra and the battery) could't be setup without damaging it a bit.
The samples the company produced were plotted, folded and glued.
In the final production, plotting and glueing happens at the same time and you fold them up when you receive them - which didn't really work.
Luckily, I'm working with that company for years. They produce all my packaging (ColUSB, Everdrives, Cinemaware boxes, Retrodes, etc.) and they never let me down.
They already found a different solution and will deliver free replacements within the next 10 days.
4. What's left?
All parts are here, everthing is prepared for assembly.
So now we're just waiting for the fixed cases and the inlays for the packaging.
The one thing I'm currently working on right now is a small manual / artbook.
As you can imagine, a manual doesn't make a lot of sense as the OS is constantly being improved, as it was the case with the Pandora.
So a quickstart manual (explaining the ports and buttons) is all you need.
But I've made a ton of pictures of all the various prototype parts. Case, keymat, PCBs, etc.
These will be put together into the manual including small information of what has been changed and why. So a small evolution of the Pyra - from ideas to prototypes to production.
All pictures have already been made and masked, so I now need to stitch the artbook together and get it printed (printing takes 4 - 5 days).
That's the last missing thing. But something I really want to create.
