[Updates #6+] Inches' CeMu+(S)AIO board (Compute Module 3 build)
Posted: Wed May 24, 2017 8:59 pm
(This is part #6. #1-5 are here: http://www.sudomod.com/forum/viewtopic.php?f=9&t=3003)
(Sign-ups for Pre-Order notifications are now open! http://www.sudomod.com/forum/viewtopic.php?f=3&t=3148)
I got so many parts in! Here are my prototype boards from OSH Park, glass screens from Kitsch-Bent, and a blank prototype SAIO courtesy of @kite. It also came with the Pi3 adapter that is necessary for connecting to my board. This allows me to test placement and orientation of my board in relation to @kite's.
My first attempt at soldering the SODIMM connector left me very frustrated and deflated. I had bridges everywhere and nothing I could do would remove them. I tried dragging an iron across to spread the solder out, using lots of flux, and using flux and a wick. In the end, I cleaned it up a lot more than what is seen in the image above, but I still had a 5v short to GND and who knows what else.
I gave up on that one after a lot of attempts at that. I went in on another board and took things slower. I put a tiny bit of solder on my iron, and lots of flux on all the pads. Then I dragged my iron across and they got a tiny sliver of solder on them from the iron. My problem before was gobbing it on, leaving heaps of it assuming it would wick up and around and fix itself like I'm used to with wider pitched SMD soldering.
Turns out just enough solder to change the pad from gold to silver did the trick. Everything else about the connecter fell into place from there.
I ended up not having the right caps in the sample book I just bought for this purpose, so I scavenged 4 from an old TV board I had in my junk box. They were way bigger than the 0805s I planned on so I had to connect the overflowing GND end to the GND vias I had luckily placed nearby. I still needed two more so I just rigged up a giant electrolytic.
Reviewing my design just now I realized I put the caps on the wrong power lines, so I'll have to adjust that tomorrow!
This is what it turned out to look like. I soldered a few wires to a component cable to plug into a TV for testing. I ended up yanking the GND pad of my tv line off just due to the rigidity of the cable, so I attached a slightly more bendy cable to my board in between them to account for that and used another open GND line.
I had little confidence in this after all the soldering issues I had on the first prototype. I had no way to determine with certainty that everything was soldered perfectly. I just checked to make sure that my 1v8, 3v3, 5v, and GND lines were distinct and that the proper voltages were created from the regs when I applied 5v to the board. Knowing there wasn't power shortages, I felt somewhat safe plugging in the Compute Module and giving it a go.
No fire or smoke! And I got an image on screen.
I don't have any way to interface with my board without the S/AIO boards so the rest of the testing will occur once I get the complete SAIO board from Kite. I'm very excited now though!!
(Sign-ups for Pre-Order notifications are now open! http://www.sudomod.com/forum/viewtopic.php?f=3&t=3148)
I got so many parts in! Here are my prototype boards from OSH Park, glass screens from Kitsch-Bent, and a blank prototype SAIO courtesy of @kite. It also came with the Pi3 adapter that is necessary for connecting to my board. This allows me to test placement and orientation of my board in relation to @kite's.
My first attempt at soldering the SODIMM connector left me very frustrated and deflated. I had bridges everywhere and nothing I could do would remove them. I tried dragging an iron across to spread the solder out, using lots of flux, and using flux and a wick. In the end, I cleaned it up a lot more than what is seen in the image above, but I still had a 5v short to GND and who knows what else.
I gave up on that one after a lot of attempts at that. I went in on another board and took things slower. I put a tiny bit of solder on my iron, and lots of flux on all the pads. Then I dragged my iron across and they got a tiny sliver of solder on them from the iron. My problem before was gobbing it on, leaving heaps of it assuming it would wick up and around and fix itself like I'm used to with wider pitched SMD soldering.
Turns out just enough solder to change the pad from gold to silver did the trick. Everything else about the connecter fell into place from there.
I ended up not having the right caps in the sample book I just bought for this purpose, so I scavenged 4 from an old TV board I had in my junk box. They were way bigger than the 0805s I planned on so I had to connect the overflowing GND end to the GND vias I had luckily placed nearby. I still needed two more so I just rigged up a giant electrolytic.
Reviewing my design just now I realized I put the caps on the wrong power lines, so I'll have to adjust that tomorrow!
This is what it turned out to look like. I soldered a few wires to a component cable to plug into a TV for testing. I ended up yanking the GND pad of my tv line off just due to the rigidity of the cable, so I attached a slightly more bendy cable to my board in between them to account for that and used another open GND line.
I had little confidence in this after all the soldering issues I had on the first prototype. I had no way to determine with certainty that everything was soldered perfectly. I just checked to make sure that my 1v8, 3v3, 5v, and GND lines were distinct and that the proper voltages were created from the regs when I applied 5v to the board. Knowing there wasn't power shortages, I felt somewhat safe plugging in the Compute Module and giving it a go.
No fire or smoke! And I got an image on screen.
I don't have any way to interface with my board without the S/AIO boards so the rest of the testing will occur once I get the complete SAIO board from Kite. I'm very excited now though!!