• Please review our updated Terms and Rules here

Is there a way to test a logic gate in-circuit without removing it?

T-Squared

Veteran Member
Joined
May 29, 2011
Messages
657
Location
San Antonio, TX
I tried using an XGecuPro tester wired with a chip clip, but when I start the test, it trips the overcurrent protection, which I'm assuming means that whatever electrical power goes out to test the chip doesn't come back in.

Does a tester exist where a logic gate IC can be connected through a clip, to be tested in-circuit, without desoldering it from the board?
 
I managed to snag the HP unit like Paul's shown above at a reasonable price a few years ago, and am soon about to use it in anger.

As standard, it came with ten blank IC cards that you solder/sodder the chips onto. I was a bit reluctant to use the original PCBs but I didn't have to, thanks to this super helpful bloke who drew up a Gerber for them and made it available on Github:
https://github.com/RetroGamerStuff/HP-10529a-reference-cards

I recently got a batch of 25 done by a Golden Dragon PCB outfit along with some other boards. The only change I made was to just select plain tinned edge fingers as the ENIG costs really added up, and for the use they were going to be subject to it wasn't worth adding.
Here they are. I haven't made any up yet but they look just like the originals and I think they ought to be fine. Actually I wish I had a larger batch done:

github_repop_cards_for_HP10529A_logic_comparator.jpg

If you can't find the HP device at a sensible price then there is a workalike:- that super-industrious Youtuber with the snappy pinstriped blue business shirt Jerry Walker designed his own version of the comparator and it's well worth watching his series on the principle of operation and how he made it:
 
I tried using an XGecuPro tester wired with a chip clip, but when I start the test, it trips the overcurrent protection, which I'm assuming means that whatever electrical power goes out to test the chip doesn't come back in.

Does a tester exist where a logic gate IC can be connected through a clip, to be tested in-circuit, without desoldering it from the board?

If the chips Vcc and GND are still connected to the rest of the board, you are basically trying to power the rest of host board and its components from the XGecuPro. What "may" work ( possibly ) is use a razor blade to carefully cut the Vcc and GND traces that supply power to the chip under test. Again, that may just push the problem to a different set of pins held high or low by other components. A drop of solder can bridge the cuts you make. Its invasive.

Desoldering the whole chip is invasive, but generally can be reversed, lets you test properly, and would let you socket the component.

The other way is in circuit logic probing. If its a simple chip a logic analyzer would let you see inputs vs output to detect anything not correct. The first line of testing if the host board is not on fire - e.g. you can power on the device long enough to observe inputs and outputs. If you want to be clever you can rig a pullup and pulldown probe with a current limiting resistor ( like 470 ohm ) to drive the inputs hi or low and observe the outputs.

Another trick I have used is simply piggybacking a "known good" identical logic chip and seeing if it resolves your issue. You don't need to solder , just make the pins touch the pins of the part under test mechanically, like stacking caterpillars - as one does. This can works for dead outputs from RAM, certain failures of simple logic chips where the piggybacked chip does the work and between the two provides the correct logic levels. Low probability of success, but also not a lot of risk and easy to do.
 
Does a tester exist where a logic gate IC can be connected through a clip, to be tested in-circuit, without desoldering it from the board?
I note "in-circuit".

-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
In my past employment, I worked in an area that looked after a bunch of Data General (DG) minicomputers. In support of that, there was a table set up with what was commonly referred to by staff as the 'board tester'. Faulty cards from the computers were taken to the board tester table.

Step 1: Two long thick power leads (+5V and ground) were connected from the 'board tester' to the card-under-test. The power leads included thin voltage sense wires so that delivered was +5V at the card (not +5V leaving the board tester).

Step 2: The 'board tester' software had library of DG cards, and from that library, the user choose the particular card was under test, e.g. DG p/n 123456.
With that information, the 'board tester' knew:
• What type of chips (e.g, 74LS244, 74LS123) are located at each position (U1, U2, etc.) on the card; and
• Which non-power pins on which chips are tied to ground by the circuit; and (e.g. on U55, pin 6, one of the enable pins, is grounded by the circuit)
• Which non-power pins on which chips are tied to +5 by the circuit; and
• Which non-power pins on which chips are tied together.

Step 3: One would instruct the 'board tester' to start the test.

Step 4: The 'board tester' would then instruct the user to place the IC clip on U1, then the user would inform the 'board tester' that that was done. The 'board tester' would then test U1 (using the knowledge of step 2 above), giving a pass/fail indication on-screen.

Step 5: Then on to U2, then U3, etc.

Largely successful.

The 'board tester' was flexible enough that we could modify the testing logic for a card.
Or we could add new cards (we had to define the testing logic for each chip).
-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

So, not generic, because the testing logic needs to know the specifics of the circuit, i.e. the type of information in the final three bullet points in step 2 above.
 
The basic idea of a logic comparator is to compare the outputs of a known-good IC with the device under test, using the same inputs for both. You obtain the inputs as well as the outputs using a test clip on the DUT. The fly in the ointment happens when an input of the DUT is shorted internally to another pin or to a power rail.
 
Dwight I agree, I will be soldering DIP sockets to the cards rather than the chips themselves.
Like the first video here shows, the HP comes with a ZIF-socketed universal board but I think it would be quite a bit quicker to just swap bespoke chip cards than muck with the DIP switch settings for every chip.
 
The trick comes when trying to stimulate the inputs of the device under test. In this case, you may have to 'force' other signals on the board in order to drive a logic input to the DUT (Device Under Test) to a desired logic '0' or '1'.

As has been stated, IC pins that are directly wired to +5V or 0V cannot be stimulated - hence they are not directly testable by this means. A faulty (internal) IC input on a pin that is permanently connected to 0V (externally) can only be detected by any subsequent impact on on the IC outputs misbehaving when you stimulate other inputs.

However, there is other test equipment that has been developed to overcome this type of problem. But you are talking about a lot of money now...

If the PCB has been designed correctly, you should be able to get a high test coverage. If not, some devices on the board may only have a low test coverage.

Dave
 
Back
Top