Flying Probe Test on a Portable Oscilloscope Acquisition Board
A portable oscilloscope acquisition board carries converters, a processor, memory, analog switches, operational amplifiers and input protection, and the power check that follows assembly says very little about any of them. A board can start normally while one supply net, one control signal or one test node is wrong, because an open circuit, a short, a wrong value or a localised joint defect does not always prevent the board from booting. Some faults only appear as a channel with the wrong amplitude or with more noise than it should have, and they are found at system level rather than at the board level.
Data Needed Before the Order Starts
Project intake expects the Gerber data, the bill of materials, the placement coordinates, the assembly drawing, the panel drawing, the program, the test requirement, the quantity and the lead time. Where a flying probe test is planned, the test data, the net information, the description of the test points and the acceptance requirement are needed as well, because a probe programme cannot be written from the Gerber file alone with any confidence.
The bill of materials needs full part numbers for the high speed converter, the amplifiers, the analog switches and the protection devices. Devices with the same outline can differ in bandwidth, in accuracy and in input characteristics, and a substitution approved on the package alone changes the behaviour of the channel in ways that only appear at high frequency.
During import the reference designators, the packages, the placement angles and the positions that are not populated are checked. Where the customer has specified an analog front end device that may not be substituted, that instruction is carried separately through material preparation and through loading, so that it does not depend on a note being read at the right moment.

Printing Comes First for Fine Pitch Parts
The QFN packages, fine pitch devices and small passives on an acquisition board are sensitive to paste offset and volume variation. The stencil apertures are evaluated against the pad structure, the package and the layout. A large thermal pad can be divided into a group of smaller openings where the design allows it, so that the paste does not collect in one mass and lift the component during reflow.
After printing, the area, the volume and the offset of the deposit are measured. If the same region repeatedly falls outside the limit, the stencil cleaning, the board support, the squeegee condition and the panel flatness are examined in that order. Rework at the soldering iron is not a substitute for controlling the print, because a board that was touched up has lost the evidence of what the process was doing.
The first article also confirms the orientation of the converter, the memory, the amplifiers and the polarised parts. Where several devices on the board look alike, the physical reel label, the bill of materials and the machine station are matched against each other rather than identified by eye.
What Electrical Testing Adds After AOI
Optical inspection after reflow covers missing parts, offset, bridging and the joints that can be seen, and it is applied to the whole visible surface. It cannot confirm the connections between them. The flying probe test moves a set of probes onto defined nodes and, using the data the customer supplies, checks for opens, for shorts and for component values that fall outside the permitted range.
Whether the probe test suits a particular project depends on the test data, on whether the nodes can be reached, on the height of the components already fitted, on the order quantity and on what the customer has to be able to prove. Not every net can be contacted by a probe, and a probe test on the connections is not the same thing as a functional verification of the product.
When a net is found to be wrong during the prototype stage, the diagnosis combines the test node, the bill of materials, the joint and the design intent before any repair is attempted. Repairing first and reasoning afterwards removes the information that would have identified the cause.

Verifying the Acquisition Function Itself
Functional test on an acquisition board follows the method the customer provides: the supply and the current, the program start, the channel identification and the agreed input and output checks. The signal source, the cable, the input amplitude, the program revision and the acceptance range are fixed before the comparison begins.
If one channel reads incorrectly, it is compared against another channel on the same board or against a known good board before the analog front end is examined. Comparing board results without a common input signal turns a difference in the test setup into a manufacturing defect, and the repair that follows it cannot be explained afterwards.
After a repair, the electrical test and the functional test are both repeated, so that the original fault is shown to have gone and the neighbouring nets are shown not to have been disturbed.
Carrying Version Information Forward
After a prototype has been adjusted, the placement coordinates, the substituted parts, the test nodes and the program revision may all have changed. Those confirmed changes belong in the next revision of the production data rather than in a note kept beside the line, because a note is not read by the next person who builds the product.
Our SMT assembly and PCBA testing services cover the placement, the material, the fine pitch work, the probe testing, the programming and the functional test, with the results kept under our quality management system.
Deciding What the Board Has to Prove
The test plan follows the function. An acquisition board that will be used at low frequency does not need the same evidence as one that will be characterised at its bandwidth limit, and a design that carries a radio front end has requirements that a purely analog board does not have. The decision is made from the customer’s specification, not from a default that is applied to every board.
Where the customer supplies the test method, the fixture and the reference board are confirmed against it before the run starts. Where the method has to be agreed, the discussion covers the channels to be measured, the input amplitude and its source, the sampling conditions, the permitted variation between channels and the way a board is judged to have failed. That last item is the one that is most often missing, and it is the one that makes a difference between two operators who are both testing correctly.
The result is a document that a second shift can follow. Boards that fail are set aside with the measured value attached rather than with a note saying the board is faulty, because the value is what allows the fault to be reproduced by someone else.
FAQ
Can optical inspection replace electrical testing? No. It confirms that the parts are present and correctly positioned; it does not confirm that the nets are connected.
Is a probe test the same as a functional test? No. It verifies connections and values against the data. Whether the product works is a separate question, answered by the functional test.
Why does the test data matter so much? Because the probe programme is generated from it. An incomplete net list produces a test that misses exactly the faults it was bought to find.



