On Sun, 18 Nov 2012 14:02:27 -0800, mike wrote:

Yep, caps did fail, but ripping out all the caps probably won't
fix anything intermittent from that era.

It worked nicely on a Tek 465 scope and a Wavetek 3000b service
monitor. See below.

Tantalum caps have two modes. They work fine, or they're shorted.
That's not leaky, or a little bit shorted. They're DEAD shorted.
Once shorted, they never come back. I've never seen an intermittent
tantalum cap.
The short resistance is so low that it's difficult to get them hot...
but if you do have a HIGH current source, you can cause them to
explode. That's not an intermittent process either.

I've seen a third mode with tantalums on power supply lines. They act
like a short when cold causing any current limiting or foldback
circuit to shut down the power supply. However, if you apply voltage
long enough (usually about 10-20 seconds), the cap will recover
causing the circuit to recover. I learned this the hard way after
designing far too many 10uf 25v tantalums (mostly Murata) into marine
radios running on 12V battery power. However, it took 20-30 years for
the problems to appear.

You can check most tantalum caps by poking an ohm-meter at 'em in circuit.
IF they're shorted, you can tell. It's rare to have a circuit resistance
low enough to mask the shorted tantalum cap.

If I can get the power supply to cooperate, I've have really good luck
with a borrowed thermal imager, an optical IR thermometer, and/or a
thermocouple thermometer. The thermocouple probe works best as I can
walk it around the PCB rather quickly. It doesn't take long for the
caps to get warm.

When they burn up, it's rather impressive. It starts with a dull red
glow, gets brighter, and eventually hits incandescence. It also
belches copious amounts of toxic smoke. One of the early IBM PC
clones had tantalums for bypass caps. One was in backwards. The
customers called and asked "My computer is on fire. What should I
do?" I answered "Turn it off". I hear the distinctive click of the
IBM PC style on/off switch over the phone.

From the description and details provided by the original poster,
there's not much more that can be deduced without additional
measurements. If something is holding down the (linear?) power
supply, it would be interesting to see what the power supply is doing.
A DVM on the power supply lines, or a watt meter on the AC input, will
offer some indication of what it's doing.

Note that an FG507 recently sold for $80.
http://www.liveauctionworld.com/TEK-FG507-GENERATOR_i11032006
The most economical approach would be to find another at about this
price. Then compare voltages and waveforms between the two units to
fix the original. If it's that important an instrument, a spare would
be a good idea. It also avoids having to spend $8 or $15 on a manual.

Incidentally, the FG-507 originally sold for about $2,220.
http://www.ko4bb.com/Manuals/Tektronix/Tektronix_-_TM-5000_Series/Tek_FG5010_short_form.pdf


--
Jeff Liebermann
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558