On Apr 24, 8:30 pm, Jim Wilkins wrote:
On Apr 23, 8:50 pm, "Leo Lichtman" wrote:

"David Lesher" wrote: (clip) the leak is proportional
to the voltage drop... and not just linearly.(clip)

^^^^^^^^^^^^^^^^^^^^^
Basically what you say is right, but I have to correct you on this small
nit: Proportional *IS* linear.

An Ohmic resistance is linear but semiconductor behavior is often
exponential or logarithmic. A good example is the drop across a diode:http://www.allaboutcircuits.com/vol_3/chpt_3/1.html
Scroll down to the part about kT/q.
The effect is that the voltage drop rises with the logarithm of the
current.

This one shows how the forward drop changes inversely with
temperatuhttp://www.cryocon.com/S900/S900ds.pdf

In the 1980's 4-1/2 digit DVMs were very expensive and I built my own,
using a constant-current source for the resistance function instead of
the usual MickeyMouse shortcut that obscures the actual voltage and
current passing through the resistor. Mine forced 1uA, 10uA...10mA and
showed the true voltage. It would clearly demonstrate the logarithmic
VI curve of a diode, from ~400mV to ~700mV. Unfortunately my ex used
it to check her car and fried that circuit with 12V.

Jim Wilkins

So, could you do the usual electronics fault finding thing of using
the diode check function on the output lead of the alternator? - (then
reverse, repeat, see difference or none) even cheap DMM have them now
- (BTW - was yours a SABTRONICS kit?)? - or is it getting too
complicated and as Karl has identified the Knuckle Crunching
Alternator as being the culprit....

And I would humbly disagree re ohmmeters vs ammeter's - its all to
easy to fry an ammeter, the 10a range on most DMM's is rated for a few
seconds only, then the copper wire shunt disengages itself....besides,
10a is leakage current in a vehicle...and DMM's are now cheap, throw
away items, but a decent ammeter is still expensive....

Andrew VK3BFA.