This problem was trickier than it initially appeared. A key on the rubber membrane keypad wasn't working. This was the 'down' key. Not absolutely essential for operation but useful. I assumed this would be the conductive rubber on the keypad so i had finally got around to opening it up and checking. No obvious signs of wear or oxidisation. I cleaned it up a bit with IPA and put some more graphite from an art pencil on the pad. Still didn't work.
I looked at the circuit diagram. The buttons are pulled to +5v with a resistor network and run into a 74HCT165 which converts the state of all the keys to serial. I found this button was 100 Ohms to ground whereas the others measure 55k to +5V as you would expect from the diagram. Then suspected the HCT165. I lifted the pin but found this was a few MOhms to gnd. It couldnt be the resistor network as that only linked the buttons to +5V. Lifted it anyway; definitely not that.
The reading i get is exactly 100 Ohms. Really seems like a resistor value, but there are no 100 Ohm resistors on the board.
I give up and cut the track from the pad then link it to the resistor network and HCT165. Works

So, presumably there is a short in the PCB... seems very weird, but anyway it works now!
The pics show my fix - red wires. Looks a bit of a mess but not as bad as what Fluke have already hacked on this meter - in green
