Fluke 845AB Zero/Opr switch

[TimFox]

I have an old Fluke 845AB "Voltmeter/Null Detector".  I haven't used it for a couple of years, but when I fired it up this week I found a serious problem with the front panel "Zero/Opr" switch.
I need to fix this problem before I attack the internal battery pack, which is not holding charge well.
It is a simple SPDT switch that switches the voltmeter input between the front panel binding post and Common.  The latter position is used to zero adjust the meter, and the input binding post is then open.
This instrument has the original high input resistance values:  100 megohm for 300 mV and above full scale, and 10 megohm for 100 mV and below.
When the switch works, I can measure these two values at the input binding posts (power off), but when it does not work the resistance at the input connection is over-ranged.
The switch comprises three springy metal strips, with a rotating cam that pushes the center leaf (voltmeter input) upward to the top leaf (common, for "Zero"), or lets it relax downward (no force from the cam) to the bottom leaf (input binding post).  This is apparently designed to avoid anything in contact with the input in "Opr" mode.
If the spring hath lost its temper, wherewith...?  I have tried "chiropractic" manipulation of the center and bottom leaves that recovers the connection in the "Opr" position, and then I measure the correct input resistances, but this is unreliable and rotating the knob between its two positions thereafter sometimes works and sometimes doesn't.
When the ohmmeter test shows either 10 or 100 megohms (depending on range), the meter seems to work nicely from 1 V down to 10 uV in my quick tests.  However, when the ohmmeter is over-ranged, there is no response to applied voltage.
Does anyone have experience fixing this problem?  I remember having a similar problem with a Keithley electrometer, where the springs were easier to reach.

[Bill158]

I bought a 845AB at least 15 years ago and it had the same problem.  Mine wouldn't make the "ZERO" contact very well and it was intermittent.  I remember getting to the "leaf" switch with thin long nose pliers and bending the center and "ZERO" leaf a little.  I also cleaned the contacts by putting DEOXID on a piece of stiff paper and then running it across the contacts (if I remember correctly?).  Anyhow I finally got the "ZERO" position to work reliably and it has been good ever since.  That switch is a devil to get to without taking the whole thing apart.
I replaced the Nicad batteries with NiMH "AA" cells.  I made no changes to the charging circuit.  I don't leave it plugged in more than 8 - 12 hours charging however.  It has worked well ever since is all I can say.  I also replaced the neon bulbs in the chopper a few years ago.  The old ones were "black" inside and probably not working as well as possible.  This seemed to improve the stability on the 3 uV and 1 uV ranges with the "bouncing/jitter" noise that is always present.
Bill

[TimFox]

Thanks for the reply.  I also used DeOxit to clean the contacts before "tweaking" the center and lower leaf springs, since my problem was on "OPR", not "ZERO".
Unfortunately, the plastic cam forces the center leaf into the upper leaf for ZERO (where the plastic can't hurt a shorted pair), but the center leaf must spring down to the lower leaf without external force for "OPR" (avoiding contact from the plastic cam).
Somewhere, I have a very thin pliers, so I will try more forceful bending.  What I did to get a temporary state of connection involved just pushing on the springs with a thin screwdriver.  After that, I left the switch in OPR position overnight, and the contact was still there in the morning.

[TimFox]

I think I may have fixed the spring-switch problem by "tugging" on the lower leaf and pushing down on the center leaf using a right-angle tweezers.
I'll have to see if it "stays fixed" after more manipulations before I put the internal aluminum shields back together.
I assume that after all these years, the springs take a "set", which inhibits their motions under their own elastic force--maybe I left the unit on the shelf in the "Zero" position too long to induce the set.
Otherwise, it seems to work well, driving it from my old Keithley nanovolt source in the millivolt range.