5V filament, remain under 0.24A
	(I ran at 0.18A)
slow sweep (by hand) 
ret voltage used was 1.189V, but probably higher is better? 1.5V? 
Data starts at 10V, ran tube current at about 10^-11 - 10^-10 amps



We use the same oven/tube as you, but we use a homemade voltage/sweep controller and (as James and Bernard have suggested) an external picoammeter/electrometer (either old Kiethly 614 or a newer Kiethly 6485). These are small currents, so if your electrometer recommends a warm-up time before making stable measurements, then definitely do it!

As several people have mentioned, you need to get the temperature right. The point is to get the mean free path between collisions to be of the correct scale relative to the spacing between the electrodes. (Temperature controls the vapor pressure of Hg, which determines the density of Hg in the tube, which determines the mean free path for electrons.) So, the "right" temperature depends on the details of your tube. It will probably be in the 150-190C range, but you have to find the optimum empirically. Also, the "best" temperature depends on the measurement you're making. You've shown data from the measurement of the excitation potential, where the current you collect is electrons which originate at the cathode and then do a handful of inelastic scatters with the mercury on their way to the anode. With the same apparatus wired up in a different way, you could also measure the ionization energy of the mercury, and collect the resulting positive ions as a current. In that case, you'd want most of your scattering to happen in a different place in the tube, and you'd use a different temperature.

I find that I need to sweep slowly in voltage, no faster than about 5V/sec, or else the dips in the signal get washed out. I guess there's some time constant associated with the gas physics in the tube (which would be different for other tubes and temperatures), but haven't really explored it.

Someone mentioned how tricky the wiring can be on this. Yes! I almost always mess it up the first time and have to start over from scratch. It's really easy to convince yourself that you've done it right. Caveat experimentor. Grounding is especially tricky. Ground loops are really easy to introduce by accident, and the resulting RF pickup can swamp the real signal. Something I sometimes forget: if you try to run both the current and the voltage to a scope in X-Y mode to display the excitation pattern, then whichever leads from your tube/voltage controller/electrometer get hooked up to the external conductors on the scope inputs will get pulled down to the scope's ground, which probably isn't what you wanted to do. Voltage measurement here is best done with isolated voltmeters. I find that the good old method of reading the meters with your own eyes and writing down the numbers with your own hands gives better results on this experiment than displaying on the scope.

Something that hasn't been mentioned here is the filament voltage. It should be around 5V for thermionic emission of electrons, but you want to search around a bit for on optimum setting: plus or minus a few tenths of a volt can be the difference between great data and no data. Higher voltage gives more electrons, which you might think would give more signal, but if the filament voltage is too high, then your accelerating voltage won't be able to sweep away the thermionic electrons fast enough, and you'll get a negative space charge buildup around the filament that will suppress further emission. So, there's an optimal filament voltage for best signal, but that optimal value depends on pretty much every other setting in the apparatus, so you have to find it empirically.

There is another version of this experiment by Leybold Didactic which adds a second grid near the filament with its own adjustable voltage for sweeping away the space charge. That setup also uses concentric cylindrical electrodes rather than linear plates, so the signal is always really big and easy to find without much effort. The experiment tends to "just work". Frankly, it's way too easy to perform and we've found that students don't learn enough about experimentation that way, so that apparatus stays in a box on the shelf and we use the one you have instead. When we have a bad day and the "tough" version gets too frustrating, then we break out the easy version.