Skip to main content ↑ ↓
Topic: Inputs sticking (Read 3723 times) previous topic - next topic

Inputs sticking

I just got my OBLS today and seem to be having issues also.  I am getting quite a few communication errors but when it works I have an another problem... If I tie a input line to high, take a capture, then tie the same input pin to ground and take a capture, most of the time the line does not return to low in the sump client.

Simple steps to replicate the problem:
1) Tie input pin 4 to 3.3v and take a capture, the pin reports as a 1
2) Tie input pin 4 to gnd and take a capture, the pin still reports a 1 instead of 0

Re: Inputs sticking

Reply #1 –
Okay so I think I found and fixed the problem.  There were a few pins on the buffer that seemed to have solder bridges with adjacent pins (pretty hard to see on this chip) which I re-soldered.  Also pin 24 was bent and creating a solder bridge with pin 23 though looking at the schematic this would have not caused the problem but added insult to injury.

Re: Inputs sticking

Reply #2 –
Hey LukeS - I'm really sorry about that. Do you have a PIC of it before your repair that we can look at?
Got a question? Please ask in the forum for the fastest answers.

Re: Inputs sticking

Reply #3 –
LukeS, sorry to hear that your OLS had at least one solder bridge on/under the FPGA ... but I am happy you found the problem and got it fixed!
Similar soldering issues were reported first when Pollin brought out their CPLD-Evaluation-Board (based on a Xilinx 95144  in a TQFP100 package).

I had at least 3 solder bridges similar to the below drawing on my first CPLD board and didn't notice them first even when looking through a magnifying glass because they were hidden under the plastic housing right where the leads enter into the housing. I could only see them barely when using a USB microscope ... removing the solder bridges with a rework station was rather simple once I knew where they were located.

Quote
__---#######---___
......B.............................. B=solder Bridge , #=plastic package

Some problems with the OLS reported on this forum could be caused by soldering bridges as well. Let's hope that this remains a rare incident because many users may neither have the equipment nor the experience to remove solder bridges on/under TQFP packages.

Re: Inputs sticking

Reply #4 –
[quote author="ian"]
Hey LukeS - I'm really sorry about that. Do you have a PIC of it before your repair that we can look at?
[/quote]

Doh!  Sorry, I should have taken a picture before I re-soldered it.  I forgot to mention I also touched up a pad on the FPGA but I can not remember which one.  Luckily I have worked with SMB a lot, but as IPenguin said could explain some of the others peoples issues.

Is there a test routine run on these board in the factory?  I see there are no test points so there is no bed of nails testing.  It would be pretty easy to pick up on a ICT test but it would probably add cost to the board so may not be worth it.

Re: Inputs sticking

Reply #5 –
That's the strange thing - the test procedure should detect this problem. Jack prepped a bitstream for the OLS that outputs stepped signals. That OLS was connected to the production OLSes and each one was tested with SUMP before shipping.

Evidently the buffer chip posed a problem for Seeed, Eric said they are ordering them from a different supplier in the future.
Got a question? Please ask in the forum for the fastest answers.

Re: Inputs sticking

Reply #6 –
Well I found another solder bridge on the buffer, this time I took a picture of it before fixing it.  I did not actually find this until I backlit the board with a 1 watt LED.  This is very similar to what you described IPenguin.

Re: Inputs sticking

Reply #7 –
... and another this time without the board backlit (there is a 400k total file limit so I could not post both in the same post).  You can see from this picture it is pretty well hidden.

Re: Inputs sticking

Reply #8 –
LukeS, thank you for the great pictures ... well, the reason you took them is defintely not great! I hope you will get your OLS fixed and working
flawlessly, soon!

I received my 3 pre-order OLSs yesterday and at a first glance, using only a magnifying glass, I found no solder bridges that would jump right into
my eye. However, the general soldering/mounting quality of the LCX16245 buffer chips on all three boards is not very good - consistent with what
I can see on your pictures. On one board the leads of the buffer are not well aligned with the pads at all and the mounting/soldering of the buffer
chips looks rather inconsistent/irregular on all three boards.

On the other side the solder/mounting quality of the FPGA, the PIC, the SPI flash ROM and the much smaller components (VRs, resistors, caps etc.)
is good ... with the exception of the push buttons S1, S2 that appear to be hand soldered with far too much tin but that's not critical for the functionality.

When setting up the first unity for more complex tests I shorted the unit and toasted VR1 - this was completely my own fault and had absolutely
nothing to do with the quality of the product! I stopped my evaluation at this point and will continue the tests later today after replacing the burned
out VR.

In short for the simple function tests I performed before this happened, all three units powered up, enumerated as USB , the FPGAs loaded the
bitstream from SPI flash and I was able to connect the SUMP client and start captures repeatedly (with open/unconnected inputs) and receive
the captures.

Re: Inputs sticking

Reply #9 –
Thanks for the pictures, it's really enlightening to see what's going on with production boards. I had a hard time soldering that chip by hand on the prototypes, and as Ipenguin mentions it seems to be pretty messy on production boards. Maybe we could clean up the footprint a little for next time, or try using one from another library.
Got a question? Please ask in the forum for the fastest answers.

Re: Inputs sticking

Reply #10 –
[quote author="LukeS"]I see there are no test points so there is no bed of nails testing.[/quote]I'm curious.  I saw bed of nails test equipment during a tour of a local PCB fab house, so I assumed it was used for PCB electrical test.  Is it also used for testing of populated boards?  I mean, it seems like it could be, but I just wanted to confirm.  Also, would this work from the top on SMD boards, the bottom, or either one?

Re: Inputs sticking

Reply #12 –
[quote author="rsdio"]
[quote author="LukeS"]I see there are no test points so there is no bed of nails testing.[/quote]I'm curious.  I saw bed of nails test equipment during a tour of a local PCB fab house, so I assumed it was used for PCB electrical test.  Is it also used for testing of populated boards?  I mean, it seems like it could be, but I just wanted to confirm.  Also, would this work from the top on SMD boards, the bottom, or either one?
[/quote]

You can use a bed of nails test for both unpopulated and populated boards.  You can even use it for boards that have components on the bottom side.  As the layer count goes up they become more important for PCB manufactures.  They are very useful for post PCB manufacturing too to test for shorts once the board is assembled, programing components on the board, and running a full in system test.