Whew - finally, I managed to find the time to make a newspost for you guys!
The last one feels like an eternity away, since so much has happened.
Why is the Pyra crying, you wonder?
Well, I'll come to that later.
First, let's start with...
The GamesCom 2015
As you hopefully know, we attended the GamesCom this year (Hall 10.2, Retro-Area).
Besides other stuff from my shop, we had a Pyra devboard with us and an assembled printed case.
The interest was great and people were stunned of how many things you can run at the same time on the devboard.
I wish I had a final prototype with me, but that didn't work out by a few days.
If you are interested to see some pictures from our booth there, head to this thread.
I also made a small video with the running devboard there - you can find that here:
https://youtu.be/89cZHdEO-p8
So much for the GamesCom, let's see what happened afterwards...
The Pyra PCB CPU Boards
While we were at the GamesCom, the bare CPU PCBs finally arrived!
We got 14 of them, and Nikolaus started to populate them right after he came back from holidays mid August.
Unpopulated PCBs
To test them without wasting too many parts, we didn't immediately populate all of them, but rather only the power circuit first including some switches and other stuff that helps testing the power circuit.
This is the testing setup stuff.
PCB with only the Power circuit populated
That way we could check that at least the power circuit was properly working before populating the expensive parts (OMAP5, RAM Chips).
Nikolaus had to add some interim traces manually to the OMAP5 pins to be able to test all power sources - look what he did under the microscope!
Some interim traces added manually to the OMAP5 pads
And guess what: Those tests were all successfull!
First step done - meaning: We can populate the important parts onto a few test PCBs!
The OMAP5 didn't cause any issues - but the RAM chips didn't want to stay where they should.
When soldering them, they were floating away, causing short circuits.
One fully and one partly populated PCB
It took just one question to the manufacturer to find out the issue: It looks like our stencil mask for the solder paste is a bit too thin, therefore there's not enough solder on the pads.
No big deal, just takes a few days to create a new stencil layer.
Picture of the PCBs with the old stencil layer
But now... the MOST important part:
Nikolaus connected a CPU board which had the OMAP5 but not the RAM chips populated.
Switching it on... connecting the serial output to see the debug output from the OMAP5:
U-Boot SPL 2013.01-gdc6238c (Mar 27 2013 - 14:33:47)
OMAP5432 ES2.0
die_id1: 14fa5e2
Hack! Disabling faulty EMIF1
lisa_map_2: 0x80600200
SDRAM: identified size not same as expected size identified: 4 expected: 20000000
### ERROR ### Please RESET the board ###
The first signs of life from the Pyra prototype!It's working, but it's crying for RAM!
IT'S ALIVE
So, yep, we reached a new milestone:
We KNOW the OMAP5 boots and works together with the mainboard. Of course, once we get the RAMs on, we need to check that ALL hardware aspects are actually working, but the most complex stuff is done and confirmed to be working.
Hopefully, this or next week we'll have the RAMs populated properly and then... it's time for the prototype!
Can't wait for this!
Of course, I will keep you informed about all that.
Feel free to ask for any questions here
The last one feels like an eternity away, since so much has happened.
Why is the Pyra crying, you wonder?
Well, I'll come to that later.
First, let's start with...
The GamesCom 2015
As you hopefully know, we attended the GamesCom this year (Hall 10.2, Retro-Area).
Besides other stuff from my shop, we had a Pyra devboard with us and an assembled printed case.
The interest was great and people were stunned of how many things you can run at the same time on the devboard.
I wish I had a final prototype with me, but that didn't work out by a few days.
If you are interested to see some pictures from our booth there, head to this thread.
I also made a small video with the running devboard there - you can find that here:
https://youtu.be/89cZHdEO-p8
So much for the GamesCom, let's see what happened afterwards...
The Pyra PCB CPU Boards
While we were at the GamesCom, the bare CPU PCBs finally arrived!
We got 14 of them, and Nikolaus started to populate them right after he came back from holidays mid August.
Unpopulated PCBs
To test them without wasting too many parts, we didn't immediately populate all of them, but rather only the power circuit first including some switches and other stuff that helps testing the power circuit.
This is the testing setup stuff.
PCB with only the Power circuit populated
That way we could check that at least the power circuit was properly working before populating the expensive parts (OMAP5, RAM Chips).
Nikolaus had to add some interim traces manually to the OMAP5 pins to be able to test all power sources - look what he did under the microscope!
Some interim traces added manually to the OMAP5 pads
And guess what: Those tests were all successfull!
First step done - meaning: We can populate the important parts onto a few test PCBs!
The OMAP5 didn't cause any issues - but the RAM chips didn't want to stay where they should.
When soldering them, they were floating away, causing short circuits.
One fully and one partly populated PCB
It took just one question to the manufacturer to find out the issue: It looks like our stencil mask for the solder paste is a bit too thin, therefore there's not enough solder on the pads.
No big deal, just takes a few days to create a new stencil layer.
Picture of the PCBs with the old stencil layer
But now... the MOST important part:
Nikolaus connected a CPU board which had the OMAP5 but not the RAM chips populated.
Switching it on... connecting the serial output to see the debug output from the OMAP5:
U-Boot SPL 2013.01-gdc6238c (Mar 27 2013 - 14:33:47)
OMAP5432 ES2.0
die_id1: 14fa5e2
Hack! Disabling faulty EMIF1
lisa_map_2: 0x80600200
SDRAM: identified size not same as expected size identified: 4 expected: 20000000
### ERROR ### Please RESET the board ###
The first signs of life from the Pyra prototype!It's working, but it's crying for RAM!
IT'S ALIVE
So, yep, we reached a new milestone:
We KNOW the OMAP5 boots and works together with the mainboard. Of course, once we get the RAMs on, we need to check that ALL hardware aspects are actually working, but the most complex stuff is done and confirmed to be working.
Hopefully, this or next week we'll have the RAMs populated properly and then... it's time for the prototype!
Can't wait for this!
Of course, I will keep you informed about all that.
Feel free to ask for any questions here
Last edited:
