Jeff Liebermann wrote:
On Tue, 8 Mar 2011 18:34:44 -0800, "William Sommerwerck"
wrote:

The actual problem was not an intermittent switch. It was the last thing I'd
have expected -- one of the welds on the battery pack had come loose!

It happens. Spot welding takes practice.

Without turning this into a megillah... Any suggestions on repairing it
without damaging the cell? I don't have welding equipment. I do have an
EDSYN pencil iron, and a 100W RadioShack gun.

Test the broken flat wire to see if it can be soldered. Keep the
battery as cold as possible (i.e. wet sponge). Plenty of liquid flux
and heat. Work fast to keep the heat affected zone to a minimum.

Please don't turn a straightforward question into a tsimmes. (But you will.)

Mazel Tov. You can also build your own spot welder with a big fat
cazapitor. Search Google and YouTube for "capacitor discharge spot
welder". For example:
http://forum.allaboutcircuits.com/showthread.php?t=24385
If you don't want to build the entire circuit, just charge up the
biggest capacitor bank you can throw together. One lead goes to the
battery terminal. The flat wire goes is placed on top of the
terminal. The electrode is some kind of heavy (#10 awg) solid copper
wire. Compress this sandwich together, and apply power at the
capacitor bank. Hopefully, it will weld the flat wire, and not the
capacitor terminals. In theory, the stainless leads with have a
higher resistance than the copper electrodes, so the stainless will
heat up, not the copper.

Spot welding is a LOT harder than it sounds.
Just to use some round numbers that you can scale for your application...
If you expect to get 10,000 amps out of 10V, you can't have more than
a milliohm of resistance. That's the cap esr + the switch + the wire +
the contacts
+ the contact resistance at the weld. And the only thing that does much
good is the amps x volts x time at the junction between the two parts
being welded plus a little in the weld material itself.

An extra couple of milliohms in the system will render the welder impotent.
And the weld will be extremely sensitive to the contact resistance
between the weldments and thus the pressure applied and the contact area
and, and , and. With a low-voltage AC welder made from a microwave oven
transformer, I was getting about 20% good welds. The commercial CD
welder with the same homebrew weld head was closer to 100% good.

You really want something that controls current rather than voltage.
Then, most of the variables end up in one place, the weld.
Using a higher voltage and some intentional, controlled series
resistance can make it much more reliable.

Many battery tabs are made of zinc. Easier to weld than stainless.
Hobby-Store 5mil brass is very easy to weld, but may be too high
resistance for a dust buster.

Having said that, you can, with practice, make usable battery tab welds
with a capacitor bank, a pressure controlled contact system and a switch
made out of a nail. Charge the caps, apply weld pressure, slam the nail
across two BIG wires. I was using 4/0 wire. You get one weld per nail,
but nails are cheap.

You didn't say which wire. if it's the negative, you may be able to fix it
by hose-clamping the wire to the battery case.

I'm gonna get yelled at, but I would not solder a battery.
Can be done with practice. Trying it for the first time on the only
battery you have is fraught with risk. Separator material inside
melts pretty easily. I wouldn't try with a gun. I'd use a 50W pencil
with a MASSIVE 1000F tip and very short heat application. You're heating
a lot of mass quickly. Gun can't hack it.
YMMV

Last resort is to drag it down to Batteries Plus or some similar
vendor that has the proper spot welder.

Very last resort is to just slop some epoxy on the flat wire and
battery terminal, and hope that it holds.