How do you start up an untested circuit board?

Hi Frank,

Interesting topic : ) A proper test plan for bring-up is an excellent idea, and the table looks very good. It is sometimes very easy to design oneself into a hole, with a board that is not easily testable : ) so thinking it through like this makes a lot of sense.

Prior to PCB testing, I try (although admittedly not always) to test all the bits I'm unsure of (either real or use a dev-board, or simulate), before creating the PCB, i.e. I end up with a lot of scraps of stripboard/perfboard etc. But I also scavenge the parts off them once the PCBs arrive.
Sometimes I can't do this, especially recently, where I'm waiting a month or longer for parts to arrive from China, and I sonetimes have to just take the leap, sometimes even just going by measurements of pixels from photos of the parts, to get the dimensioning approximately right. This is only for hobby stuff of course.
For the PCB, in terms of test points, zero-ohm resistors can also be handy for isolating circuits on a board (in third place to 100k and 10k resistors, they are the next most-used resistor for me. I try not to connect inputs/outputs from subsystems on a circuit board together without a resistor in-between, for allowing the isolation, unless I'm certain the circuit portion will work. For digital circuits, I might use a resistor instead of tying pins high or low, in case something else needs to be patched. I don't always do that, but sometimes it's handy. Also I'll sometimes add a resistor in-line, it is usually benign (might help with matching impedances a bit occasionally) but saves too much damage if things get shorted, or if two outputs are accidentally connected together, e.g. RX/TX lines, since sometimes it's easy to get that wrong in a design. Ferrite bead chips are also handy for isolating power to portions of circuitry on a board too, i.e. there's usually no harm to add that in on supply lines to chips. That way, even if a lot of the board is assembled, it is still possible to power off portions for testing (provided inputs/outputs do not cause an issue), or to current-limit using a resistor.
It is also good to stick a polyfuse on each board, especially if it's expensive, I don't do that much on cheap hobby projects, but for the small cost it's worth it for a bit more peace of mind on more expensive boards. With test wires or metal chassis everywhere, it's too easy to accidentally short things, so that bit of extra protection from the polyfuse could help.

EDIT: Also test code helps : ) It is handy to have debug code, extra print statements with a debug flag, to troubleshoot the code and hardware.

Hi Shabaz,

Those are all great suggestions.  I especially like the concept of modeling/testing things beforehand and isolating circuits with zero-ohm resistors.  Can you give an example of a case where you would consider adding a resistor in-line and how you would size it?  I've thought about adding a polyfuse but haven't done that to date.  Ferrite beads are another thing I haven't tried.

The few boards I have built don't give me the experience to provide a detailed test plan.

A few suggestions are all I have:

• I measure resistance between the input power terminal before the board is assembled. NO SHORT!
• I make the same measurement when the board is populated. NO SHORT!
• I try to power staging for populating the board. If it involves Arduino or Pi, they are not installed on first power-up.
• In this staging exercise, I look at what I can populate in stages. Establish the power supply, test, then passive components and test, then active and test and finally processor.

Like I said my experience is limited so is my knowledge.

Good suggestions, using the multimeter to buzz for shorts is the first thing I do too.  Learned from experience :-)  I mentioned it in the test plan but the other thing I do now is a detailed visual between each stage.  Not only have I inadvertently added shorts, but I've also left pins completely unconnected or poorly soldered and then wondered why it wasn't working.

I've done something similar except on the final PCB.

I now print the PCB etching and hand trace the layout against the schematic. It is a mind-numbing task especially after working with the schematic for so long. It really takes a lot of concentration and focus. I save this wonderful task for last to preserve my sanity.

I always build in test points now that can take either a test terminal soldered in - I have around 100 of these that cost a couple of quid from CPC and are reusable - or can take a probe tip and pigtail.  When I was designing my Instrument Control Board I spent a lot of time breadboarding sections first which also allowed me to create some test Arduino code; I then went on to prototyping the power section with a PCB to do a comparison test between various approaches.  That's not as expensive as it sounds: the parts can be desoldered afterwards and reused and the PCB manufacturer is dirt cheap from China.

That doesn't help for shorts across pins though  I never used to give a second thought about continuity or resistance testing pins on SMD ICs to make sure no shorts across pins existed and a good connection to the pad had been made as a visual inspection is often awkward.  More recently I've wondered about this approach: a multimeter does this by providing a voltage and in some cases that might be more than the IC is designed for.  It's not often clear from the DMM manual what the voltage is.  On the other hand, I don't know if this is ever likely to be a problem, e.g. continuity testing with a 9V DMM signal across an IC designed for 5V.

For the ICB when I first received the actual boards, I did like Shabaz and measured continuity across points to ensure that there were no shorts where there shouldn't be.

Hmm interesting.

fmilburn said:

What did I leave out? 

I suppose this is side question... as in what's best practice for including test points on a PCB... I see shabaz has already touched on this in one of his comments.

In particular I am curious to understand the following... What size should a test point be? How many to include? Where to place them? What to avoid? What do you measure etc.

That is an interesting concern about measuring continuity where ICs are in place and I've never given it consideration.  Does anyone know the answer to that?  I've not experienced a problem with it to my knowledge.

The detailed design process and documentation you create for your projects would be helpful in developing a test plan and selecting test points. I'm more apt to do that with software/firmware than hardware.  Maybe that is part of the reason I have problems with even simple circuits at times?

A lot depends on how much space you have. A test point could be a small surface pad or a though hole big enough to fit a DVM probe or it could be an actual component like a wire loop or connector jack. They should be on all power rails, and all signals of interest.

I had the same questions on test points, although it isn't clearly stated.  How many, where to put them, and what to measure is the main reason I developed the test plan.  I did some internet searching before making this post and what I found wasn't that enlightening.

KiCad has a number of footprints in the library.  A couple of them are shown below.

I have tested a few PCBs in the RoadTest program with the loops and really like them for their ease of use.  Farnell / Newark carry them.  For example, there is Buy Now in different colors.