• Please review our updated Terms and Rules here

Need help with FlipChip Tester

gnupublic

Experienced Member
Joined
Jan 23, 2019
Messages
195
Location
Germany, Berlin
Hello,
some times ago I got a Flip-Chip-Tester board from Roland and ordered the parts I need at Mouser. Now I had the time to solder the parts, and did all I could find on Vincents site.
Did the modification on the Board (Resistor, 35nF and trace cable).
Soldered the cable to connect the raspie Nano like described in raspie.pdf. Replaced IC6 with jumpers.
I installed subversion , added user tester, checked out the software like in Pi-setup.txt on my Zero. Compiled running make.

So I think I did all that was describes on Vince page and documents.
But when I start ./tester, and try to run a test or the diags, I always get :

diags menu
1 card edge input test
2 card edge output test
3 tester_init_and_uut_power_verify
6 register tests
7 SPI functions
8 LPT functions
9 exit diags

calling 'tester_init_and_uut_power_verify'

************************************************** *************************
* did not verify registers after initialize (chip reset). *
* check that the tester is cabled to LPT port and that the power is on. *
************************************************** *************************

tester_init_and_uut_power_verify() returns 1
Error occurred or power is not on


I controlled the cable once again and have no idea what went wrong. Can anybody give me a hint how to proceed?
Greetings
Volker
 
So I think I did all that was describes on Vince page and documents.
But when I start ./tester, and try to run a test or the diags, I always get :

diags menu
1 card edge input test
2 card edge output test
3 tester_init_and_uut_power_verify
6 register tests
7 SPI functions
8 LPT functions
9 exit diags

calling 'tester_init_and_uut_power_verify'

************************************************** *************************
* did not verify registers after initialize (chip reset). *
* check that the tester is cabled to LPT port and that the power is on. *
************************************************** *************************

It's been a while since I last used a tester, but it seems that you aren't reading back the high and low values that were expected on the power pins of the module under test. A number of things could cause that, including (as Michael mentioned), just having the module power switch in the wrong position.

It should be fairly easy to measure whether power is in fact getting to A2, and ground to C2 of the module connectors. If so, then the GPIO expanders are not communicating properly with the RasPi. That could be an issue with the tester board, or perhaps most likely with the cable that adapts the RasPi pin-out to the 26 pin connector. There's also a bit about replacing the inverting buffer chip (was it E2?).

The good news is that the thing isn't terribly complex, so one can follow the signal path and usually find the issue without too much effort.

Vince
 
There's also a bit about replacing the inverting buffer chip (was it E2?).

Apparently it was IC6:
"First thing, remove IC6 and install jumpers in it's place, as shown above."

I probably shouldn't try to do these things from memory.

Vince
 
Hello,
I found my error. I did not understand, that when mounting the pinheader on the solder side, you are also changing the numbering of the pins. So I moved the pinheader from the top to the solder side and the tester runs. I tested some M115 cards in some 1000 iterations without error.

One more question, when running "card edge input test" AA2 is 1, but BA2 is 0 in the printout. Both pins are +5V and are connected with one trace. So I expect both to be 1.
When I the toggle UUT Power on/off, only AA2 changes. So only AA2 is measured and BA2 is not measured independ. Also Software only shows 0 for BA2?

Thanks,
Volker
 
I found my error. I did not understand, that when mounting the pinheader on the solder side, you are also changing the numbering of the pins. So I moved the pinheader from the top to the solder side and the tester runs. I tested some M115 cards in some 1000 iterations without error.
Excellent!

One more question, when running "card edge input test" AA2 is 1, but BA2 is 0 in the printout. Both pins are +5V and are connected with one trace. So I expect both to be 1.
When I the toggle UUT Power on/off, only AA2 changes. So only AA2 is measured and BA2 is not measured independ. Also Software only shows 0 for BA2?
According to the schematic, toggling module power should apply +5V to both AA2 and BA2, unless the corresponding jumper block is missing. This should also appear at GPIO expanders IC1 pin 28 and IC3 pin 23. Checking the source code (tester.c) though, does show that the power_pin array just has AA2, and the ground_pin array has AC2, BC2, AT1, and AT2. PIN_POWER_BA2_NC (IC3 pin 23) was deliberately omitted. I believe that this was because it was discovered some of the modules we want to test don't actually use BA2 as a power pin.

That said, it is still unclear why it would read a 0 for BA2. (Also unclear, given the above, why it's wired directly to VPP, rather than being selectable to the actual BA2 pin.) A quick grep of the test cases indicates that none of the finished and working tests reference BA2 at all, so it may be that it just never got properly debugged.

Vince
 
Thanks Vince!
I understand that this is no issure and is not impacting the usage.
I hopefully will be able to test the FlipChips of my TC08 that have a test defenition file. This is the case with about 32 flipchips from 67.

Volker
 
If there are flipchips in the TC08 that do not have a test file, I have a TC08 and a flipchip tester so I could help you make new test files.
 
That sounds great! This is a list of modules in the TC08 and the test definitions. I will work on that next week and will have a look at the modules that are not in the definition list.
TC08 ModulFlipChipTester Definition
1G829
1G8790
5G888
3M101yes + new
2M103yes, new/not working
5M111yes
1M112yes
6M113yes
2M115yes
2M117yes
1M119yes
3M121yes
2M161yes
7M206like M216 standart jumpering
3M207yes, from Roland
1M228
2M302yes, new
1M307yes, new
1M401
1M502
4M602
4M623yes, new/not complete
4M627yes, from Roland
2M633
1W005
2W032
67 46

Volker
 
Last edited:
I have made some new testfiles too. I will sent them to Vince for svn. The 206 can be tested as a 216 if the 206 is jumpered in the default way. For the 207 I have made a new test file for sure.

Making a M228 test will take a lot of time and at this moment I don't know if my M228 is functional. So if anyone has a good test for that board then that would be perfect.

Regards, Roland
 
Last edited:
There are two types of test definitions. Mxxx.TST and Mxxx.new. All tests with .TST files are successful. But the tests with .new files are different. With my Raspie environment I had to rename them to .NEW to load them into the tester. M101.NEW did not load, this is not ready at all. M103.NEW does show errors with my flipchips, but the same on both cards. M623.NEW is not complete, but the test was OK.

Roland, good hint to use M216 for M206 in standard jumpering. That was OK also.

Volker
 
Back
Top