SparkFun Electronics Commentsurn:uuid:214d0e4e-f1b1-d287-ce26-ac5b4c9f82492024-03-29T08:59:12-06:00SparkFun Electronicselmer_fud on Enginursday: Sparking Creativity in Troubleshootingelmer_fudurn:uuid:72fca6f8-4124-ef7d-d42d-3667597f220d2018-08-13T00:26:13-06:00<p>Car batteries are not anywhere near current limited for an application like this :)</p>
dennis55337 on Enginursday: Sparking Creativity in Troubleshootingdennis55337urn:uuid:a6d2999e-2f67-e9ac-9cbe-2594c2a3c7572018-08-10T19:43:59-06:00<p>For over ten years I have been using a variation on the four terminal Kelvin resistance measurement to find shorts on pc boards. I use a one or two amp current limited low voltage source. This could be a 5 volt supply with a four ohm power resistor or a lab supply set to 3 volts and 1 amp. I connect the current source to the two shorted traces at easy access points, usually by tack-soldering short scrap wires to components or connector pins on the problem foils. Then I set a DVM to read volts with resolution to millivolts and start measuring voltage along the foils. You may start out with 30 or 40 millivolts. Work your way along the traces to where the voltage goes to zero or a minimum and you are at the short.
If you have a ten ohm tin whisker between two foils, pass the current thru one of the foils from one end to the other. Treat the foil with current as a voltage divider and the other foil as the wiper on a trimpot. At the place where the voltage on the current foil equals the other foil you have the short. I have used this to find whiskers that were barely visible with a microscope. A small scratch with an xacto blade cleared the short.
Four terminal resistors and voltage dividers are handy tools!</p>
Steve123 on Enginursday: Sparking Creativity in TroubleshootingSteve123urn:uuid:824affa3-bd60-a73c-3f81-79fb20c7d2882018-08-10T11:19:29-06:00<p>And if you contact the board manufacture and <em>show</em> them the manufacturing defect the response will be "well you didn't pay for electrical test so too bad". It's only happened to me twice so far and since they were prototypes it was just an inconvenience not a money looser.</p>
Madbodger on Enginursday: Sparking Creativity in TroubleshootingMadbodgerurn:uuid:f4c57667-0586-24c4-9a3a-f5576283f0fd2018-08-09T20:10:52-06:00<p>I had a board like that, where I had lifted some of the circuitry from an open-source product. Eagle told me there was an issue there, but it was a working board, so I approved it and sent it off to fab. Sure enough +3.3V was shorted to ground: right in that borrowed layout. It took a little while to narrow it down, cutting traces here and there and using a short tracer. The next revision, I paid more attention to the DRC errors!</p>
PickledDog on Enginursday: Sparking Creativity in TroubleshootingPickledDogurn:uuid:f1eaaacb-13d8-6985-2da5-2b893310ac552018-08-09T16:24:48-06:00<p>I see not just a burn mark, but a board with its identifying marks masked off. I shiver with anticipation!<p>(I know that companies prototype stuff all the time, and many products are dead ends. Which is why early prototypes are not revealed without masking.)</p></p>
Customer #134773 on Enginursday: Sparking Creativity in TroubleshootingCustomer #134773urn:uuid:ea36abdf-7baa-3749-987e-a23b96ff762a2018-08-09T16:00:46-06:00<p>When I was a factory tech back in the mid-70s, I was mostly working on thick film hybrids (circuits literally printed on a ceramic substrate -- a low resistivity ink was used for conductors, and higher ones for resistors which were later laser trimmed to be "in-spec"). The largest was about 2"x2", and when we had a lot of trouble finding the problem, we could occasionally get away with the "waffle test" -- place the board on the floor and firmly apply the "waffle-stomper" (aka hiking boot). Always solved the problem... ;-)</p>
dksmall on Enginursday: Sparking Creativity in Troubleshootingdksmallurn:uuid:995985e1-b45d-2caf-3401-aa53f5a430462018-08-09T14:24:25-06:00<p>Well yeah, but I had one burn along a path of an internal run for about 9". How big are those stickers? :-)</p>
CF on Enginursday: Sparking Creativity in TroubleshootingCFurn:uuid:25685a0f-e2f7-06f7-dbbb-87622a42737c2018-08-09T13:11:33-06:00<blockquote>
<p>If the PCB was burned then it became an internal reference board, can’t ship those to customers.</p>
<p>That's what 'QC' stickers are for. ;-) Covering up the burned spot!</p></blockquote>
Customer #134773 on Enginursday: Sparking Creativity in TroubleshootingCustomer #134773urn:uuid:8928bf8c-542e-a45f-0dfb-8f21925724812018-08-09T11:44:48-06:00<p>Sounds to me like a good use for the <a href="https://www.sparkfun.com/products/14654" rel="nofollow">FLiR kit</a> or the <a href="https://www.sparkfun.com/products/13233" rel="nofollow">older version</a> that SparkFun sells. Fortunately I haven't had to chase a power short in a PCB since I purchased the older version more than 3 years ago.<p>Back in the "dark ages" (before FLiR cameras), one <em>very tedious</em> approach to finding shorts in a PCB was to use an ohm meter that used a "Kelvin" (4-wire) connection rather than the more common 2-wire connection. These meters can get down into the sub-milliohm range, and get you "closer" to the short. Unfortunately, these meters can be fairly pricey (think a few thousand dollars) and require a very high skill level. For this sort of use, though, calibration isn't critical (as long as the meter is stable enough to not "drift" over a few hours/days of use), so sometimes you can get lucky and pick up one used but not calibrated for a decade or more for a much more reasonable price (think hundreds of dollars). Getting one calibrated can cost hundreds of dollars, not to mention the round-trip shipping.</p></p>
dksmall on Enginursday: Sparking Creativity in Troubleshootingdksmallurn:uuid:c861a682-46f6-3d59-460c-a08eb68a127b2018-08-09T11:35:12-06:00<p>Back when I was a test technician in the 70's, we would do the same thing with boards that were 20" x 30", all DIP logic, with power typically +24, +/-15, +12, and 5. I believe the power supply I used was 50amp (Big Bertha) and if we got a pop with no burning, the board went on to chamber testing and burn-in. If the PCB was burned then it became an internal reference board, can't ship those to customers.</p>
Garfieldboy on Enginursday: Sparking Creativity in TroubleshootingGarfieldboyurn:uuid:72da0881-6e39-e04a-48b2-4a5afa9c8eaa2018-08-09T09:41:16-06:00<p>I often (well, not TOO often) use a somewhat similar method, but rather than trying to blow the 'fuse' I search for the hot spot.
I use a current limited power supply and slowly ramp up the current. Occasionally this is enough to blow the unwanted 'fuse', but for harder shorts I go looking for heat.
If I am feeling brave this is done using a uni-pixel thermal scanning sensor (my finger).
For more complex boards, or boards with components on both side where the mono-scanner can't pass smoothly I use a thermal camera.
This also works for circuits with faulty components and/or incorrectly connected components.
Proved very useful on one particular design when I updated an ADC to a higher performance ADC in the same family/package only to discover one of the ground connections had been re-tasked as a power supply. The updated ADC shorted the power internally. Once discovered simply lifting the rogue leg fixed the issue.</p>