Hitachi S-450 Scanning Electron Microscope

Now that I know the thing works I want to be able to move this into my living room, keeping this in the garage is not ideal, it is not the most ideal environment for it with temps pushing 100f in the summer and high 30's in the winter. To do this I really needed to get rid of the diffusion pump, it sway 400w at 100v and most of that heat needs to be removed with a air/water heat exchanger. Not only is it noisy it also take up space. To eliminate this I got a Varian Navigator 301 turbo pump and control from a friend as well as an Edwards XDS10 dry scroll vacuum pump that I found at a scrap yard.

This has a number of advantages, it removes oil from the system. Oil is bad in a SEM in general, it gets deposited on apertures and messes up the image and it also degrades the phosphor on the SED. It also creates a big plume of oil vapor when you pump down the column after a sample change. It also reduces power consumption, not only does it eliminate the 400w heater in the diffusion pump it also eliminates the need for the cooler which is another 1/3hp electric motor. The turbo pump uses 150w at full gas load and about 7 watts at high vacuum. Quite a bit of savings!

The whole machine also comes up a whole lot faster now, with the diffusion pump you would have to wait at least 20 minutes from power-on for it to be ready to pump the chamber. With the turbo it spins up in less than a minute and running a few minutes later. This makes it much nicer to use. I was worried the turbo would be too small, the diff pump was rated at 400 l/s and the turbo at 250 but it turned out not to be a problem in the end, the turbo seems faster.

I had an adapter that mated with the turbo and fit the bolt pattern of the old pump and it went together pretty well. I did add a vacuum reservoir on the foreline of the turbo to give it a little buffer, I think I might have a leak on one of the roughing valves on the system, I get a pressure spike when I vent the chamber.

I picked up a couple ISO100 vibration isolators for the turbo and install one of those. I may need to shorten the pump adapter in the future to make it fit better.

With these modifications done I brought the SEM into the living room. Not something I ever want to do by myself again!

Fast forward two years later, some more issues crop up with the SEM. First, the scan starts having problems and eventually dies. Second, the last filament blows. The scope only came with one spare filament which died early. New filaments are somewhere around $500 for a box of 10 and are not sold individually. Around this time my friend finds a seller in California posts a couple SEMs for sale on eBay, a Jeol 6300F FESEM and a Jeol 8600 Microprobe. They sit on ebay for quite a while constantly dropping in price. Eventually my friend makes him an offer and they are his. To make room for the new scopes I get the old Hitachi and stick it in my garage.

The first thing I need to deal with is the filaments. I really don't want to spend $500 on a box of filaments for a machine that I don't even know if I can repair. Some of the parts in the machine are obsolete and just about impossible to get ahold of. A friend at Portland State University says we can rebuild the filaments, they are easy. I order some .089mm tungsten wire off ebay and he shows me how to weld them up with his Unitek capacitive discharge spot welder. In about 30 minutes I have two rebuilt filaments for my SEM.

I get a new filament installed and I I get an image. So far so good, the rapid scan mode does not work, the image goes off the screen. Eventually even the normal slow scan modes die too, the image tearing off the edge of the screen.

Luckily I do have schematics for the SEM, a college sent us a copy of their manual for the SEM. I put the acrylic window back and fluorescent stub in the chamber and find I still have a good raster scan for the e-beam. The CRT and the san could for the column share the same scan generator signal source so I can rule out the scan generator as the source of the problem. I start tracing down the signal from the scan generator on the CRT, everything looks fine, a nice sawtooth pattern up until I get to the scan amplifier module. I put the scope on the output and it looks like a mess. I compare it to the vertical scan output which is very close to it's input leading me to believe the scan amp is toast.

A bit of video of the input and output of the scan amp, I am using my little Hitachi battery powered scope just in case there are any ground loops.

The scan amp in these modules is actually a STK070 audio amplifier. Whatever works, I guess. I luck out, there was a new spare that came with the machine. I install that and it brings it back to life. I need to make a few pot adjustments to get the image centered but everything is working as it should.

We got this unit off craigslist a couple years ago. It had died on them and then it sat for 10 years or so. We got it back to his shop and reassembled and started playing. First thing I did was make a acrylic window that would go in one of the accessory ports and coat a stub with phosphor from a fluorescent lamp. Putting this in the chamber allows visual confirmation of the electron beam and if it is actually scanning. It did show a raster where the e-beam was causing the phosphor to fluoresce and the size of the pattern changed with the magnification control and also changed with the scan rate selector.

From this and the fact we had a raster and characters on the CRT we had a pretty good indication that it was mostly working, it was just we were not getting a signal out of the PMT on the Secondary Electron Detector (SED). Eventually I figure out there is no power to the PMT and no power coming out of the HV power supply inside the console. I then notice that in the middle of the cable to the SED there is a wad of electrical tape. Apparently when this machine was moved in the past some idiot just cut the cable instead of disconnecting it in the console. When they put it together they just twisted the wires together and put electrical tape over it. Eventually the tape broke down under the +500v and shorted out taking out the HV power supply.

The HV power supply is pretty simple, just a high frequency flyback oscillator that drives a small transformer and is then rectified and filtered. When the cable was shorted the output diode eventually failed. Since the transformer was warm I had pretty good reason to believe that the power supply was still functioning so we ordered a couple HV rectifier diodes. When they came in I installed it and shortened the cable to get rid of the splice and the SEM came up.

The following image was one of the first ones out of it, the end of a hair. At this point the final aperture strip was not in place so the depth of field is very shallow. To capture the image we put a camera above the imaging CRT and opened the shutter and hit the photo button on the console. The Photo CRT is a low persistence phosphor CRT and the image is slowly scanning over the surface to create the image.