SEM Buildblog

Documenting the inception and build of the Heatsync Labs Scanning Electron Microscope

Home
Archive
Subscribe (RSS)
Random post
Twitter
HSL-SEM Wiki Page
Flickr

Catching Elephant is a theme by Andy Taylor

 

Electron Gun Mark I on Flickr. yes, I realize the tungsten electrode is off center. You try precisely lathe boring natural rubber! It kinda moves all over!

Electron Gun Mark I on Flickr.

yes, I realize the tungsten electrode is off center. You try precisely lathe boring natural rubber! It kinda moves all over!

Completed Vacuum System on Flickr. Hey look, its a fully functional vacuum system! What does that mean? That means that the turbo-pump and the backing pump have their own lines into the chamber. The chamber can be roughed, then switched over to turbopump control if needed. So what’s the plan from here? The plan is to develop two beamlines. The first beamline will use a cold cathode emission system which operates at low pressures and uses the residual air in the chamber and a HV DC source to form a plasma from which electrons are harvested and injected down into the optical system. The second beamline will be a standard thermionic emitter using a tungsten hairpin filament and a high current DC source. The downside of this is that the system needs to be running at high vacuum which means turbopump. Turbopumps are nice, but they are not hackable. We can’t go get one for less than the price of the entire completed system, and that’s just not the point! 

Completed Vacuum System on Flickr.

Hey look, its a fully functional vacuum system!

What does that mean? That means that the turbo-pump and the backing pump have their own lines into the chamber. The chamber can be roughed, then switched over to turbopump control if needed.

So what’s the plan from here? The plan is to develop two beamlines. The first beamline will use a cold cathode emission system which operates at low pressures and uses the residual air in the chamber and a HV DC source to form a plasma from which electrons are harvested and injected down into the optical system.

The second beamline will be a standard thermionic emitter using a tungsten hairpin filament and a high current DC source. The downside of this is that the system needs to be running at high vacuum which means turbopump. Turbopumps are nice, but they are not hackable. We can’t go get one for less than the price of the entire completed system, and that’s just not the point! 

Compression Fitting Assembled with BA Gauge The BA gauge which I want to use to measure the ultimate vacuum level is a glass tube type and lacks a flange. The solution is either to torch a metal fitting out of kovar (expensive) or to use a simple viton O-ring compression fitting. Since the latter isn’t destructive and I don’t yet trust my glassblowing skills, I went with the compression fitting. Instead of buying a compression fitting from a vacuum supplier, I decided to go ahead and turn one on the lathe for practice. The combination of threading, tight tolerances, dissimilar materials and knurling made this the most advanced lathe project I have yet taken on.  An exploded view of the pieces is shown below In order to prevent any burrs in the threads of the steel nut, a thread form was turned and is shown below. Made from the same steel, the threads were worked against the nut threads under force for multiple mate demate cycles until the motion was smooth. And finally a photo of the gauge connected to the chamber. Just one step closer :) Edit: If I made this piece again I would have added a small internal shoulder for the neck of the gauge to sit on so that it couldn’t be pushed all the way into the flange and potentially hit things inside the chamber.

Compression Fitting Assembled with BA Gauge

The BA gauge which I want to use to measure the ultimate vacuum level is a glass tube type and lacks a flange. The solution is either to torch a metal fitting out of kovar (expensive) or to use a simple viton O-ring compression fitting. Since the latter isn’t destructive and I don’t yet trust my glassblowing skills, I went with the compression fitting.

Instead of buying a compression fitting from a vacuum supplier, I decided to go ahead and turn one on the lathe for practice. The combination of threading, tight tolerances, dissimilar materials and knurling made this the most advanced lathe project I have yet taken on. 

An exploded view of the pieces is shown below

In order to prevent any burrs in the threads of the steel nut, a thread form was turned and is shown below. Made from the same steel, the threads were worked against the nut threads under force for multiple mate demate cycles until the motion was smooth.

And finally a photo of the gauge connected to the chamber. Just one step closer :)

Edit: If I made this piece again I would have added a small internal shoulder for the neck of the gauge to sit on so that it couldn’t be pushed all the way into the flange and potentially hit things inside the chamber.

High Voltage Feedthrough Experiment on Flickr. So it turns out that spark plugs make terrible high vacuum devices, and this became quite apparent before even introducing one into the chamber. While attempting to “alter” one on the lathe, the metal sheath became very loose and it was evident that there was neither an o-ring nor any sort of sealant nor braze connecting the ceramic core to the metal jacket which contains the threads.    A couple of lathe-hours later… “Yo dawg, I heard you liked holding work-pieces so I put a chuck in your chuck so you could use inner jaws while you use outer jaws” Upon discovering this I simply removed the entire jacket and turned the profile of the Al2O3 core into a custom KF-25 flange. The fit was so snug it required a special jig just to remove once I had “test fit” the part.  Add in some hysol 0151 low out-gassing epoxy and viola! A high voltage feedthrough :)

High Voltage Feedthrough Experiment on Flickr.

So it turns out that spark plugs make terrible high vacuum devices, and this became quite apparent before even introducing one into the chamber. While attempting to “alter” one on the lathe, the metal sheath became very loose and it was evident that there was neither an o-ring nor any sort of sealant nor braze connecting the ceramic core to the metal jacket which contains the threads. 

 

A couple of lathe-hours later…

“Yo dawg, I heard you liked holding work-pieces so I put a chuck in your chuck so you could use inner jaws while you use outer jaws”

Upon discovering this I simply removed the entire jacket and turned the profile of the Al2O3 core into a custom KF-25 flange. The fit was so snug it required a special jig just to remove once I had “test fit” the part. 

Add in some hysol 0151 low out-gassing epoxy and viola! A high voltage feedthrough :)

Vacuum Leak Plot annotated on Flickr. The HSL SEM Chamber has several tiny pinhole leaks in the welds. The pinholes were detected by spraying the welds with isopropanol from a spray bottle and watching the response as the alcohol leaked through and then vaporized (and spoiled the vacuum momentarily). Here I stand, mighty spray bottle in one hand, and determination in the other :) (Photo by Erica Weems) As you can tell from the plot above, painting the holes with clear nail polish (nitrocellulose laquer) while under vacuum made a nice seal as the liquid was pulled through the leaks and solidified. The DV-6M gauge was brought to 10mV which indicates ~10^-3 Torr. The next step involves getting cheap surplus valves to seal off the chamber and estimate leak rate in prep for attaching the turbomolecular pump to the system for an ever deeper vacuum. WE NEED TO GO DEEPER!

Vacuum Leak Plot annotated on Flickr.

The HSL SEM Chamber has several tiny pinhole leaks in the welds. The pinholes were detected by spraying the welds with isopropanol from a spray bottle and watching the response as the alcohol leaked through and then vaporized (and spoiled the vacuum momentarily).



Here I stand, mighty spray bottle in one hand, and determination in the other :) (Photo by Erica Weems)


As you can tell from the plot above, painting the holes with clear nail polish (nitrocellulose laquer) while under vacuum made a nice seal as the liquid was pulled through the leaks and solidified. The DV-6M gauge was brought to 10mV which indicates ~10^-3 Torr.

The next step involves getting cheap surplus valves to seal off the chamber and estimate leak rate in prep for attaching the turbomolecular pump to the system for an ever deeper vacuum. WE NEED TO GO DEEPER!