Shop OBEX P1 Docs P2 Docs Learn Events
Oh my, I lost my shorts ! — Parallax Forums

Oh my, I lost my shorts !

jazzedjazzed Posts: 11,803
edited 2009-06-22 14:53 in Propeller 1
The quest to find my short(s).

I got one of my PCBs yesterday, and found to my horror the 3.3V rail was shorted to GND. The 2 layer board has decent 3.3V and GND planes, but to make it work some 3.3V traces dive and the etch margin can be narrow in the places where that happens. Given that, you might say "well, you asked for it!" [noparse]:)[/noparse] A visual inspection of the traces did not reveal anything obvious.

So, what to do? A short between power and ground planes is pretty horrible. Fortunately, there are traces connecting pieces of the planes, and a trace has a certain amount of resistance (if it was a 4 layer board, I would have been toast). You can't measure the resistance with a typical DMM to find the short, but anyone can use a battery and ohms law.

So armed with E = I*R: as resistance goes up and current is constant, potential difference voltage goes up. In this case higher resistance and voltage means bigger distance from the short. So, I connected my trusty old 6V AA battery pack on the 3.3V and GND planes for brief periods and measured the voltages.

Near the regulator section I had close to 160mVDC (no devices were installed during my quest). The further I walked my DMM probes from the regulator section, the lower the voltage went. Eventually, I found the area of least potential difference meaning that resistance was the smallest. Boy I had to carve a lot of area relatively speaking to get rid of the short, but fortunately, I only had to carve one place. The short was invisible to my naked eyes for sure.

So, having carved away the short from a small dumbbell shaped area, I stuffed some parts and verified basic Propeller function. Now I'm off to" Ace Components" around the corner on a parts list quest now that I found my shorts [noparse]:)[/noparse]

I've included a shot of the partially stuffed board for reference.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
--Steve


Propalyzer: Propeller PC Logic Analyzer
http://forums.parallax.com/showthread.php?p=788230

Comments

  • heaterheater Posts: 3,370
    edited 2009-06-16 19:47
    Thing to do is this.

    1. Get a lab power supply.
    2. Wind it up to 50v
    3. Set the current limit to MAX.
    4. Apply it to the power terminals of the board.
    5. Observe your shorts instantly explode.

    SAFETY WARNING: Do wear eye protection. Do NOT do this near inflammables. Realize this is a joke.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    For me, the past is not over yet.
  • heaterheater Posts: 3,370
    edited 2009-06-16 19:59
    I once walked into a lab and had to ask the guys their "What's that burning smell?"
    "We're just testing the current carrying capacity of some vias on a PCB" came the reply.
    Sure enough they had a lab supply pushing lots of amps through a via hole on a PCB which was smouldering away nicely.

    I never did find out if that was a real work procedure or if they were just amusing themselves over a lunch break. It was often hard to tell such things in that lab.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    For me, the past is not over yet.
  • MightorMightor Posts: 338
    edited 2009-06-16 20:02
    Jazzed,

    I think the coin on the PCB might be the cause of your shorts.

    Regards,
    Xander

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    | To know recursion, you must first know recursion.
    | Current projects and ramblings: I'd Rather Be Building Robots
    | RobotC 3rd Party Driver Suite: [noparse]/noparse][url=http://rdpartyrobotcdr.sourceforge.net]Project Page[/url / [noparse]/noparse][url=http://rdpartyrobotcdr.sourceforge.net/documentation/index.html]API Documentation[/url
  • jazzedjazzed Posts: 11,803
    edited 2009-06-16 20:04
    @heater [noparse]:)[/noparse]
    A friend of mine used to take shorted PCBs out to his car to pop shorts.
    I have only one board. Using a 12V 40+Amp battery was just too risky.

    @Mightor, Ya, I put too much money on my board [noparse]:)[/noparse]

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    --Steve


    Propalyzer: Propeller PC Logic Analyzer
    http://forums.parallax.com/showthread.php?p=788230
  • Beau SchwabeBeau Schwabe Posts: 6,572
    edited 2009-06-16 20:13
    jazzed,

    C'mon where's that thermal camera when you need it?

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Beau Schwabe

    IC Layout Engineer
    Parallax, Inc.
  • Bill DrummondBill Drummond Posts: 54
    edited 2009-06-16 21:37
    A few years back I was a CE for Sperry Corp. We had a problem with shorted bypass Caps on memory boards. We would charge up a very large electrolytic Cap. A pair of
    meter leads fit nicely in the cap's lugs. Apply the leads between gnd and Vcc we'd get a satisfying snap an puff of smoke.
  • RaymanRayman Posts: 15,379
    edited 2009-06-16 23:17
    Looks like a nice board. I like the case!· Is that the Sparkfun case?

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    My Prop Info&Apps: ·http://www.rayslogic.com/propeller/propeller.htm
  • Timothy D. SwieterTimothy D. Swieter Posts: 1,613
    edited 2009-06-16 23:51
    I haven't done it, but I have heard of similar stories of techs applying a high current to the PCB to burn away any small and hidden shorts in the PCB design when it isn't obvious.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Timothy D. Swieter, E.I.
    www.brilldea.com - Prop Blade, LED Painter, RGB LEDs, 3.0" LCD Composite video display, eProto for SunSPOT
    www.tdswieter.com
  • jazzedjazzed Posts: 11,803
    edited 2009-06-17 00:11
    @Beau, I don't have a thermal camera, but my shorts are a size 42+ so they aren't that hot [noparse]:)[/noparse]

    Is that a staple piece of equipment in manufacturing these days? I don't think I've ever seen one in an R&D hardware lab, and I've brought up many complicated designs. Usually the bigger company's I work for just send the boards out or have a flying probe before FABs are stuffed. I guess this is a budget learning curve for me. Wonder what it would take to make a thermal camera.

    @Ray, Yes it is. At $5 a pop it's hard to resist. Could mount the board on a boe-bot but with only two screws.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    --Steve


    Propalyzer: Propeller PC Logic Analyzer
    http://forums.parallax.com/showthread.php?p=788230
  • kwinnkwinn Posts: 8,697
    edited 2009-06-17 04:04
    I can certainly verify that a large capacitor and a wall wart can be used to find shorts. I used to repair defective S100 cpu and memory boards (at a fixed price) for a company that assembled and sold Z80 based systems. The board resist was applied by silk screening, and sometimes the resist would splash and create a short between traces when the silk screen was lifted away. The wall wart would charge a 10,000uF cap to about 16V through a current limiting resistor and I had a pair heavy multimeter leads that I would `walk` through all the address and data lines. Worked like a charm.

    This method doesn`t always work though. I once had some 12 port RS232 interface boards for the HP2100 minicomputer that would pull the 5 Volt supply down to under 1V and draw 55Amps. Ground and +5V planes were shorted on a batch of 25 fully assembled boards. The surprising thing was the power supply worked just fine after the short was removed.
  • Beau SchwabeBeau Schwabe Posts: 6,572
    edited 2009-06-17 05:27
    @jazzed - lol - I hadn't considered the dual purpose of thermal imaging shorts that you wear! -lol

    I was thinking that Phil Pilgrim's fancy Thermal Imaging application could be refined for close-up automated thermal inspections without breaking the bank.

    http://forums.parallax.com/showthread.php?p=753166

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Beau Schwabe

    IC Layout Engineer
    Parallax, Inc.
  • computer guycomputer guy Posts: 1,113
    edited 2009-06-17 07:18
    Not sure if you have tried, but perhaps its a screw or lead thats comming in contact with a via.
    Or solder from one of the IC sockets has jumped over to a via.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Building Blocks To The Propeller Chip A web site designed to help people who are new to the propeller chip.

    Guitar Hero controller using the prop (WIP) --> HERE
  • AleAle Posts: 2,363
    edited 2009-06-17 18:16
    From experience, before I populate a board and after some components I measure conductivity to be sure there are no shorts, I do not like when things get hot, explode or ruin boards that costed so much of my time.

    Once I was with my Father to a auto-parts dealer because the DC motor used to start the gasoline engine of our car was not working. The guy in there connected the rotor of the motor to some power supply and with a lead touched the collector area. Sparkles around 50 cm long erupted immediately from the collector. We got a new rotor and I think he used 220V directly. I started to have _even_ more respect for electricity.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Visit the home of pPropQL: propeller.wikispaces.com/pPropQL
    pPropQL020: propeller.wikispaces.com/pPropQL
    OMU for the pPropQL/020 propeller.wikispaces.com/OMU
  • SRLMSRLM Posts: 5,045
    edited 2009-06-18 04:44
    I'm glad that this method works. I have a board that's sitting on the shelf because of this exact same problem. I even tried what you did, but couldn't find the short and wasn't even sure if it made sense to do so anyway. I'll try again...
  • Toby SeckshundToby Seckshund Posts: 2,027
    edited 2009-06-18 18:47
    One way to hunt for very low resistances is to use a low voltage, low current audio signal ie 1mA of 1Khz would give 1mV across 1 Ohm. Amplifying and measuring this form is much easier, and it isn't affected by junctions, and wouldn't destroy them either.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Style and grace : Nil point
  • Fake-NameFake-Name Posts: 7
    edited 2009-06-20 06:46
    Beau Schwabe (Parallax) said...
    jazzed,

    C'mon where's that thermal camera when you need it?

    I've actually done this, and can confirm that it works. I've got a FLIR camera where I work, and it was, repurposed... for a little while. It let me track down a short that repeatedly resisted 47,000uf @ 50v.

    10um cameras are actually really great for doing initial testing on an unproven circuit board. They make debugging much easier.
  • evanhevanh Posts: 16,599
    edited 2009-06-22 14:22
    The above thumping methods are a tad destructive. I'd recommend a more gentle constant current with a millivolt meter approach.
  • Toby SeckshundToby Seckshund Posts: 2,027
    edited 2009-06-22 14:53
    Followed by a small thermo-nuclear charge

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Style and grace : Nil point
Sign In or Register to comment.