We're trying to find a root cause of an issue with our modules - we're starting production, but the boards coming off the line have only about 25% of the TIWI-UB1 modules functioning. The non-functioning ones are programmable, but draw 7-9mA and do not connect. There are some that connect intermittently, but drop the connection after a few seconds or a minute and jump to 7-9mA current consumption when they lose connection.
Anything we should be checking on the hardware or in our assembly process? We've been stuck at this point for a few weeks. Thanks very much!
I have a few questions below that would help us get a broader picture of what is happening.
1) How many units were part of the production run and how many exhibit this issue?
2) Have the units been x-rayed? If so, what did they show?
3) Can you send the serial numbers of some of they failed units? Or better yet is it possible to query the MAC address?
4) What is the CM production process?
5) Has the CM reviewed the reflow profile in the datasheet?
6) Has the CM observed the MSL recommendations in the datasheet? If so please describe.
Thanks so much -
1) We did an initial run of 160 boards, run by our manufacturing department. In this run, there were 33 TIWI-UB1 fails, a large part of the issue was improper reflow profile (solder didn't completely reflow on some boards, those boards were reflowed with hot air).
We're doing a 2nd run with a slightly revised board, run by engineering department. The reflow profile on the first run was too hot this time - we're adjusting, but wanted to check in before running additional prototypes. There were 12 of 16 fails.
2) No, we don't have one in house, we'll see if we can use one from a different facility.
3) Is the serial number the bottom string of characters on the module?
4) Boards are run in house, 16up arrays, original prototype process went smoothly.
5) Yes, working on profile now.
6) Devices have been stored in ambient for longer than 72 hours. We can bake the ones currently on the reel and open a new pack and place those. Is this a major risk factor? Many of our parts stay on the machine between builds without issue.
3) Serial numbers included at the bottom
6) Parts will bake over the weekend.
We ran another panel with a corrected profile, but got similar results. We will try a new reel for the next assembly, but for now we're baking our current parts.
We tried to communicate the module over UART, the response we got was not something we recognized. We only receive one response packet and then it's unresponsive.
Full Packet: 04 0E 04 01 90 05 01
Type : 0x04 (Event)
EventCode : 0x0E (Command Complete)
Data Length : 0x04 (4 bytes)
Event : 0x9001 (??)
?? : 0x05 0x01
BT module serial #'s