Measuring current on parts of an assembled protoype PCB

[GarthyD]

I am looking to perform current measurements on various components on a prototype PCB I am working on. The goal is to find the best areas to make improvements to in order to reduce power use of the board in low-power modes. The board is optionally battery operated and the goal is to extend the time it can run on battery between recharges.

I can measure overall current use easily enough using a multimeter in series to the supply, but I really need to isolate individual components or parts of the board.

Some solutions I can see:

- Desolder selected components, measure drop in overall current use. Slow, onerous, and not always practical.

- Bend up component legs and attach test points (see below). Multimeter to measure current use between the points. I'm worried about snapping off component legs doing this.

- Cut tracks on the board and somehow attach test points (see below). Multimeter to measure current use between the points. I'd rather avoid slicing tracks if I can.

For the test points:

- Which points/components should I use?
- How would I attach them to a circuit not designed for them?
- How would I prevent them from being flimsy and breaking easily?
- Is there be an easy way to short them (eg. with a jumper/shunt) when I'm not measuring current?

I experimented once with soldering a single male header to a point of interest on the board and that lasted a couple of days before it broke off.

The measured current would be below 10mA generally and I'd like to go under 1mA if at all possible.

Any suggestions or recommendations?

[MosherIV]

At early stage of development, if you need to measure individual part of the circuit, you should have design that into the first prototype. Chalk this up to experience for next time.

I suggest you cut tracks and add surface mount headers onto the track. You can then use a jumper to joint the track or connect to dmm to measure current. Just be careful with the board so as to not knock the pin headers off. You can use glue to help hold the pin header on the board.

Hope this helps.

[GarthyD]

At early stage of development, if you need to measure individual part of the circuit, you should have design that into the first prototype. Chalk this up to experience for next time.

I suggest you cut tracks and add surface mount headers onto the track. You can then use a jumper to joint the track or connect to dmm to measure current. Just be careful with the board so as to not knock the pin headers off. You can use glue to help hold the pin header on the board.

Hope this helps.

Thanks for the suggestions.

As a follow-up question: How would you incorporate it into the design? At a guess, perhaps design with pads for the headers and that design will always use headers and jumpers, and take them out of the final design? Or leave them in as pads that you could bridge with solder or a zero ohm SMT resistor for boards that you aren't measuring? Include some kind of tiny DIP switch? Or is there a better approach for designing in these points?

Personally, would you use 2.54mm headers for their versatility and cost despite their huge size, or another type of connector?

[alm]

Some sort of header / footprint (could even use pogo pins if you are short on space) in parallel with a zero Ohm resistor. If you use a header, you can choose not to populate it and handsolder it if you need to measure current. If 2.54mm is too big, then you could consider 1.27mm surface mount headers.

Alternatives would be to use the Aim-TTI I-prober or to characterize a PCB trace to use as a shunt (some Keithley meters can do this automatically). However, I would expect neither to work particularly well with the low current levels you are likely dealing with.

[GarthyD]

Some sort of header / footprint (could even use pogo pins if you are short on space) in parallel with a zero Ohm resistor. If you use a header, you can choose not to populate it and handsolder it if you need to measure current. If 2.54mm is too big, then you could consider 1.27mm surface mount headers.

Thanks for that.

So, as a general rule for the design stage, aim for something that you can either use surface mount headers or zero-ohm resistors on? Perhaps a combined footprint or am I reading into things the wrong way?

I then choose board-by-board whether I want the dynamic convenience of a header or just want to put a zero-ohm resistor across it. And I can always resolder if I change my mind.

Now, if I haven't anticipated an issue, I should lean toward cutting tracks and adding surface mount headers?

So ideally design in some pads, or if I miss it, cut tracks to make space, and use surface mount headers as the general solution to the problem?

Now, I'm guessing the suggestion for 1.27mm headers if space is an issue is because they're still readily available, not too expensive, and easily connected?

Is this roughly right?

Are there any specific parts that are worth experimenting with or checking out in addition to this? Or should I just be sure to have a handful of surface mount headers available when this sort of work comes up?

Alternatives would be to use the Aim-TTI I-prober or to characterize a PCB trace to use as a shunt (some Keithley meters can do this automatically). However, I would expect neither to work particularly well with the low current levels you are likely dealing with.

Yeah, I had heard of such things but they did sound like they were for working with much higher currents.

[MosherIV]

So, as a general rule for the design stage, aim for something that you can either use surface mount headers or zero-ohm resistors on?

Personally, I prefer through hole .1" header pins. They are better at resisting strain than the surface mount ones.

So ideally design in some pads, or if I miss it, cut tracks to make space, and use surface mount headers as the general solution to the problem?

Yes, if you did not have the forsight to put in a break for testing purposes, scrap back solder resist on a long track, cut it and solder on a surface mount pin header.

FYI, if the design goes into rf freq - then you must only do this on non critcal layout paths.

Hope this helps.

[GarthyD]

So, as a general rule for the design stage, aim for something that you can either use surface mount headers or zero-ohm resistors on?

Personally, I prefer through hole .1" header pins. They are better at resisting strain than the surface mount ones.

Through-hole 2.54mm for an infrequently-used developer-only test point? If you don't mind me asking, is this something that you personally would retain even through to the final design, or is it more a matter of using nice stable test points during development until you're almost at the final release? If left in the final design, do you fill the pairs of holes a zero-ohm through-hole resistor, or something else?

So ideally design in some pads, or if I miss it, cut tracks to make space, and use surface mount headers as the general solution to the problem?

Yes, if you did not have the forsight to put in a break for testing purposes, scrap back solder resist on a long track, cut it and solder on a surface mount pin header.

Cool, thanks for the tip. I haven't done the scrape and attach thing before, so I'm going to have to learn to do it. Some research ahead of me for that.

Of the tracks I'll probably need to do, one is a beautiful long 1cm+ straight track (admittdely under a 2-pin component) with plenty of spare space. The other, well, that's not so good. Multiple signals, two components, busy on both sides, and two vias in the 3mm x 3mm area I need to work in.

(For the latter though I think I can just deliberately destroy a 0.5mm track and solder on top of the pins on two existing pads, so it might not be as scary as I made it sound.

FYI, if the design goes into rf freq - then you must only do this on non critcal layout paths.

I think I'm safely out of this territory.

Hope this helps.

It does and has, yes, many thanks.

[alm]

So, as a general rule for the design stage, aim for something that you can either use surface mount headers or zero-ohm resistors on? Perhaps a combined footprint or am I reading into things the wrong way?

Depends on available space. Zero ohm resistor is easy to populate during assembly and does not require you to short a header for normal operation. Whether you have space for a connector footprint in addition to that or just shape the pads so a surface mount header can fit on it (make them longer) depends on the circumstances. For stability you probably want something like a 2x2 or 3x2 block of headers if surface mount, but not all pins have to be soldered.

Now, if I haven't anticipated an issue, I should lean toward cutting tracks and adding surface mount headers?

So ideally design in some pads, or if I miss it, cut tracks to make space, and use surface mount headers as the general solution to the problem?

Yep, if the trace is long and wide enough, you should be able to solder a header.

Now, I'm guessing the suggestion for 1.27mm headers if space is an issue is because they're still readily available, not too expensive, and easily connected?

Indeed. Much smaller than standard 2.54mm headers, but still easily available and reasonably cheap. I think you can even buy them cheaply on eBay/AliExpress if you are so inclined. Also the footprint is very simple (two pads in line). Of course through-hole 2.54mm is the sturdiest if you have the space. Neither is the ideal solution if you are dealing with amps of current.

Are there any specific parts that are worth experimenting with or checking out in addition to this? Or should I just be sure to have a handful of surface mount headers available when this sort of work comes up?

A jig with pogo pins can work on the minimal amount of space (just some pads without solder mask), but you will need to hold the jig to the board somehow.

[MosherIV]

Through-hole 2.54mm for an infrequently-used developer-only test point? If you don't mind me asking, is this something that you personally would retain even through to the final design, or is it more a matter of using nice stable test points during development until you're almost at the final release

No, this is for the early prototype/development board only.
Remeber what the purpose of the first/early prototpyes : to test out and prove the circuit, to try out and allow circuit changes and to allow measurement and verification of the circuits.

If left in the final design, do you fill the pairs of holes a zero-ohm through-hole resistor, or something else?

If they are left in, it is common to just join them with track. They can then be cut and the header populated if needed.

The through hole is my personal preference because they are far more sturdy than the sufrace mount headers. I have seen surface mount headers tear the PCB pads clean off the PCB 

[GarthyD]

Depends on available space. Zero ohm resistor is easy to populate during assembly and does not require you to short a header for normal operation. Whether you have space for a connector footprint in addition to that or just shape the pads so a surface mount header can fit on it (make them longer) depends on the circumstances. For stability you probably want something like a 2x2 or 3x2 block of headers if surface mount, but not all pins have to be soldered.

Ah! The stability issue alone explains the mention of parallel connections rather than jumping straight to combined footprints. And then space influences the best solution to use. Got it. I have also heard that pick-and-place machines are happier with 0-ohm headers, and you can always desolder them if they're in the way.

Yep, if the trace is long and wide enough, you should be able to solder a header.

.18mm width for power. We'll soon seen if that is wide enough. :}

I took a spare experimental board and had a blind experiment with exposing and cutting tracks. It turns out that I can remove the soldermask very easily and with a reasonable degree of precision. I can break a track, but my technique is sloppy and I don't think I can do that with precision yet. However, for the tracks I most need to break, there should be enough space around them that I can get away with it for now.

I can't yet test soldering SMT headers as my entire supply of SMT headers is one 2x2 1.27mm header and I have no suitable 1.27mm jumpers wire or shunts. I'll need to do some shopping.

Indeed. Much smaller than standard 2.54mm headers, but still easily available and reasonably cheap. I think you can even buy them cheaply on eBay/AliExpress if you are so inclined. Also the footprint is very simple (two pads in line). Of course through-hole 2.54mm is the sturdiest if you have the space.

Sounds good. Space is frequently an issue, but I'm guessing that for key parts of the design 2.54mm through-hole is still worth it. Re shopping, yeah, I'll be needing to do some.

Neither is the ideal solution if you are dealing with amps of current.

For the most part I'm keeping under 500mA, but it is something to bear in mind in case I work on something with higher currents in the future.

A jig with pogo pins can work on the minimal amount of space (just some pads without solder mask), but you will need to hold the jig to the board somehow.

Excellent, thankyou. I've started experimenting with pogo pins but have not yet experimented with a jig. I would be converting my existing test system to something like that if I end up needing to do larger test batches.

Thankyou for your explanations and all of the detail, much appreciated. I feel a lot more comfortable making board mods to get the extra measurements I need. Thanks also for the tips for better design.

No, this is for the early prototype/development board only.

Cool. I thought this might be the case. I had a *lot* of followup questions if the answer was yes.

Remeber what the purpose of the first/early prototpyes : to test out and prove the circuit, to try out and allow circuit changes and to allow measurement and verification of the circuits.

Thanks for the advice. I've been fairly fortunate on first runs on PCBs thus far. My current board turned out fairly well for a first run. I'm looking to improve it though, and to do this I need to do a fair bit of measurement. As happy as I am with the board, the current iteration of the board is *very* deficient when it comes to current measurement. In fairness, I didn't know some of the tricks and techniques mentioned in this thread, nor did I know that I'd need to make the measurements that I do.

If they are left in, it is common to just join them with track. They can then be cut and the header populated if needed.

That's really clever, I like that. Assuming I'm thinking about this the right way: The default config is connected, but if you are developing you just need to break a small segment of track and solder on a header. You can then restore if you need to by removing the headers and soldering in a short piece of wire through the two holes.

The through hole is my personal preference because they are far more sturdy than the sufrace mount headers. I have seen surface mount headers tear the PCB pads clean off the PCB 

Yikes! It sounds like I was very lucky then. As mentioned earlier I lost a through-hole header kludged onto a board and from the sound of it, it could have been worse.

Thankyou very much for your advice, experience and suggestions re design improvements. This has given me me plenty of ideas for improvements that I can use in my next board iteration that should make the testing process much smoother.

[flatmax]

If you are interested, this industrial strength pogo clip may help - if you can get footprints which are 0.1" pitch or similar.

https://www.kickstarter.com/projects/1250664710/pcb-quick-connect-clip-40-pin-raspberry-pi-compati/?ref=eevblog

Just an idea !

Matt