I am having issues with consistency in my scan results. From first to second scan and board to board I get opens where I previously got shorts, and sometimes get different signature shapes. I have scanned in 4 tested functional boards 3 times each and upon finishing the 3rd reference scan of the 4th good board I still get a failure.
The attached picture shows some of the differences I get from first scan to second scan. The boards I am scanning are from hand held PDA/scanner guns and printers. The boards are not coated. The pin is hitting on the center of the pad and rechecks if open. I am using a 4 point crown. I do not believe the issue I am having is with contact.
Does anyone have any tips for improving the consistency in scanning boards? I am having trouble identifying component problems when good boards don't test as passing.
Hi,
The key is to get a Reference set of signatures that are consistent. If opens tend to be a problem, put the probe down on them in RealTime and give the probe a gentle tap. You would be surprised what doesn't appear to be bad contact really is. Try a needle probe to see if you get better results. Are the signatures shown in the image from two different boards?
The two scans are the same board, the image represents the first and second scan. I have tried going through and checking all the opens in real time. It is not only opens and shorts, it is as you can see from the picture I am getting different signature shapes. I have even had instances when I go to real time and over time it will slowly go from open to having a signature without me touching the prober needle. Also I have signatures where the first range to go in is open but the second and third range are appropriate.
I lowered the pin down position and it did not help.
The issue I think I am having is that once voltage is applied in certain areas of the board opening up pathways for voltage to flow in later tests.
Well I changed my test a little to help compensate the issues I have been having. Basically I run the first reference test and before setting it for reference I re-check all the "opens" for continuity i.e. manually re-probe and nudging the probe tip. I have found that due to a charge effect if I use real time I will eventually get a signature on the more stubborn opens. Then I set it for a reference. so the next defective board i scan will undoubtedly fail until I re-probe all the opens and give them a charge so it will show the signature.
To be clear, some pins will show as open no mater how much I clean/change pin down/re-probe until i put it on real time and essentially charge the pin. Then I will be able to get a signature.
So it works, but is time consuming. Does anyone have a quicker way to deal with this issue?
Is the board conformal coated? I have found I need to use the baddest probe in the box. It is marked part number 06-2111. Even then, getting through some coatings is spotty at best.
Is it not possible to put a delay in the program to allow for charging effect?
There is a Delay function in the Ranges tab of the Tree pane. It is best to use this feature sparingly to keep the test speed up.
The Polar T6000 fault finder impedance signature analyzer made by Polar in the Channel Islands in the early 90's (similar to the Tracker 2000 +DSI 700) that ran with DOS (operating system before Windows) also had a delay function in the program to assist with noisy / capacitive circuits. But as you say with Worstation program did slow testing down if used too much.