Identifying Cold Solder Joints

[shiner_man]

Is there anyway to specifically pinpoint a cold solder joint using a multimeter as apposed to the visual test?  For instance,
if you have a cap that is getting voltage on one side but isn't getting voltage on the other, does this indicate one of the
solder joints on the cap is bad?

[DougH]

My usual approach is to just re-heat all the joints and get them to flow. If it starts working I knew it was due to a cold solder joint. You can always check for continuity with a meter, like you would buzz out a cable etc. But by the time you do that, you could have fixed it IMO. YMMV.

[dune2k]

Is there anyway to specifically pinpoint a cold solder joint using a multimeter as apposed to the visual test?  For instance,
if you have a cap that is getting voltage on one side but isn't getting voltage on the other, does this indicate one of the
solder joints on the cap is bad?

Nope. Could also mean you've got a bad cap (it happens...) or that the multimeter's voltage/current is too small to load the cap.
You could try the audio probe (google it), helps finding cold solder joints in the signal path. Just keep in mind not to try audio probing the path from you voltage source to the rest of the circuit...did that once (on one of my first effects) and it didn't go well.

[shiner_man]

Nope. Could also mean you've got a bad cap (it happens...) or that the multimeter's voltage/current is too small to load the cap.
You could try the audio probe (google it), helps finding cold solder joints in the signal path. Just keep in mind not to try audio probing the path from you voltage source to the rest of the circuit...did that once (on one of my first effects) and it didn't go well.

Hmmm...sounds scary.

I know the issue is on the vero board as I get no voltage to any components after a certain point.  I'm fairly certain this is indicative of
a bad solder joint somewhere.  I thought maybe I could narrow it down.

I guess I'll just go over the majority of the joints again as suggested.

[ayayay!]

For instance,
if you have a cap that is getting voltage on one side but isn't getting voltage on the other, does this indicate one of the
solder joints on the cap is bad?

Could indicate neither.  Could be the trace/pad is not connected, depending on where you're probing.  Check that as well. 

[shiner_man]

Could indicate neither.  Could be the trace/pad is not connected, depending on where you're probing.  Check that as well.  

Well I'm probing from the input jack ground to both sides of the cap.  Like I said, one side has voltage and the other doesn't.  I'm
not sure what that means exactly.

This is on a vero board.

[Philippe]

My usual approach is to just re-heat all the joints and get them to flow. If it starts working I knew it was due to a cold solder joint. You can always check for continuity with a meter, like you would buzz out a cable etc. But by the time you do that, you could have fixed it IMO. YMMV.

This approach works for me as well. The solder has got to effectively fuse with the connection/pads & a DVM continuity test can sometimes be misleading.

Just re-heat your PC solder connections & proceed from there.

[Hides-His-Eyes]

Could indicate neither.  Could be the trace/pad is not connected, depending on where you're probing.  Check that as well.  

Well I'm probing from the input jack ground to both sides of the cap.  Like I said, one side has voltage and the other doesn't.  I'm
not sure what that means exactly.

This is on a vero board.

It means the cap is blocking DC, like it's supposed to.

[LucifersTrip]

Is there anyway to specifically pinpoint a cold solder joint using a multimeter as apposed to the visual test?

All you do is set the multimeter for resistance (say 20K) and clip the red and black leads to lugs on either side of the joint you're testing. Don't go through a component.  The blue joint is being tested.

If your multimeter doesn't read "0" [no resistance], you've got a bad joint.

[deadastronaut]

also with vero...check your cuts...i had the worlds tiniest stray burr touching another line..done my head in for ages

till i got out the magnifier and saw it...poked it with a tiny screwdriver..sorted!..yay!..

or they can look like they are cut....but have a really fine strand connecting still....magnifier to the rescue!...

[Processaurus]

Is there anyway to specifically pinpoint a cold solder joint using a multimeter as apposed to the visual test?

All you do is set the multimeter for resistance (say 20K) and clip the red and black leads to lugs on either side of the joint you're testing. Don't go through a component.  The blue joint is being tested.

If your multimeter doesn't read "0" [no resistance], you've got a bad joint.

Exactly.  You can also beep continuity from a component leg to the trace it is connected to.  Sometimes old components have a layer of crud that makes them not take solder immediately.  One clue the joint has gone wrong is if it is round on top, rather than hershey kiss shaped where the solder has wicked up the component lead from the pad.

[petemoore]

  Cold solders are the worst, intermittent cold solder being the ultimate culprit.
   Can look perfect, be intermittent and not have intermittency influenced by anything except some unknown force.
  About the best can be done is to clean/heat-through thouroughly  when soldering, even that has it's limitations, something about goo on the component end or the pad has oil, when the solder slurps into/across the pad it is a very good sign, the pad that is stubborn is worth suspecting.
  Caps..about all that can be done for these is check for no-DC getting through it or replacement. Fortunately, my exp. is that they have been very very good.
  Actives...other than getting their 'peripherals' correct enough, replacement is the only way to determine functionality. It either works right or it doesn't.
  Most of the other stuff can be measured or calculated into telling 'functionality: truth or false'.
   ''Distant point continuity'' tests may include multiple solder joints as well + a resistance. A string of connections and a value can be measured at once: battery clip + through jackring/plugsleeve/jacksleeve, wire, solderjoint, resistance [value], opamp socketjoint, opamp pin.
   Once it boils down to a certain area, the 2-prong-poke from top and bottom of board, though ugly is the one that actually may locate the culprit.
  Never solder on a live board, that said, just going over all the joints with hot iron lets the exact failure point be found, perhaps not recommendable but it has let me stop work when the fix is found...that or re-do the whole board.
  I used to be shy around actives, sometimes they require a 2flow/reflow right off the bat. Let the not quite complete heating that was taking too long cool down while the tip is cleaned or a better angle is found, then re-flowing the joint seems to get to completion faster, don't add much more solder than to tin/fill/cover, make the heat work quickly through 'already grabbed by solder' leads and pad.

[LucifersTrip]

Exactly.  You can also beep continuity from a component leg to the trace it is connected to.  Sometimes old components have a layer of crud that makes them not take solder immediately.

Yes...always sand oxidation off of old component leads before soldering...