May 30, 2013
I ground the faceplate. That took about ½ hour of grinding.
I didn’t completely remove one nick because it has a 45 deg angle and would require the removal of a lot of glass.
I also ground the funnel.
It seems so thin for such a large size.
This took 1-1/2 hours to grind. I had to remove 0.4” of glass because of this jagged piece in the break.
This also made it difficult. One half of the funnel had about 0.3” broken off cleanly and the rest was mostly there. Trying to keep it square was hard. The glassblower helped by first using his belt grinder to remove a substantial portion of it. Normally when you use the table, a large rotating plate with a diamond impregnated disk with water flowing over on top, you are suppose to keep moving the piece and rotating it about every 5 seconds to keep it uniform. I had to intentionally make it non-uniform because I needed to square it up. I was a bit nervous.
They seem to stack up o.k.
Because of the amount of glass removed and the angle of the funnel, there’s a small step.
The one on the inside is only 1/64” wide.
Now I need to make measurements. Weigh the mask, faceplate, and funnel. I need to know the weights so I can do a scaled test with a bottle to make sure the frit layer doesn’t get too thin when fired. I’ll need to make some tooling to make sure things are aligned properly and some fixtures, which will depend on the design I decide to go with.
If I decide to attach mounts on the inside with frit I’ll need to make certain that the frit can support the parts when fired to 300C for processing later. I think that should be ok, but I want to be certain.
So this may take some time.
June 11, 2013
I did a test to see if the 52 alloy would work with the frit since I may need to use it to bring out the HV and mount the screen/mask assembly.
I made this piece just to see how easy it is to form.
I then tried to oxidize it with a regular propane torch and attach it to a piece of 0120 glass with the frit. I attached one end to the glass.
It was a very weak joint. The surface that looks like metal on the frit actually is. The iron in the 52 alloy reduced the lead oxide to lead. I checked with an DMM and it is conductive.
I then oxidized the 52 alloy with a glass working torch. Propane plus oxygen. This brought it close to the melting point.
The proper oxidation made for a strong joint. After breaking it off, there was no reduced lead on the surface.
I had also cut the 0120 glass in half to test joining it with the frit.
It bonded well and has a very strong joint.
I also just put some frit on the surface and checked with a polariscope after firing it. There is no observable stress meaning the CTE matches well enough.
I did some measurements on the funnel. It is 9/16” of an inch shorter than the measurement from the 2.4” OD part of the neck to the midline of the ultor ring. The ultor accounts for about 1/8” of this so I apparently had to remove about 7/16” of glass when I ground the funnel flat. One side had a large piece missing and the other side had nearly all of the glass remaining. So I had to try to grind that side more to keep it square. I only partially succeeded. One side has about 0.27” less glass. I will have to try to correct for this when I mount the mask because I don’t think the purity magnets can correct for that large an error.
The next test I’ll do is to see if the 52 alloy to frit makes a leak tight seal.
June 13, 2013
I took the ultor ring to our shop today to see what they thought of my idea of having it water jetted out. They didn’t think much of that. They told me to go away and come back in an hour. Well, CNC machines are nice.
Here’s the electron shield nicely removed. One stud fell off but that’s not the machinist’s fault. The spot welds on this piece are terrible. On the one that fell off you can see the electrode tips were the right size, about 3/16” dia. But the actual weld area on one weld is about the size of a pin head. The other weld is much smaller than a pin head.
On another stud it appears that one weld was good, but the other didn’t take at all. And it sits at an angle off of the surface. That’s the only way the stud matches up with the mask also so I’m pretty sure it didn’t happen later.
I have a spot welder so I can reattach the one and repair the other.
I’ll have to take a picture of the threaded stud. It’s hard to explain how they’re made otherwise. But they are made of at least three sections with welding and also what appears to be brazing.
I’m nervous about bringing the HV out through the frit seal so I wanted to see how difficult it would be to install an ultor button. I took a piece of the funnel glass to try. I started heating the glass very slowly. I didn’t get very far before it exploded. When you heat the glass far from an edge it expands and it’s difficult to heat a large enough area to accommodate the stress, yet only melt a small spot at the center. You could probably do it with the right equipment, but I think it’s too risky to even ask our glass blower to try.
I thought about drilling a hole and using frit to attach the button, but without a good way to back up the glass that would also be risky.
So, I need to go ahead with tests to see how durable a tab through the frit seal will be. If it works the tabs will also need bullets attached, otherwise the sharp edges will have corona.
With the mounting tabs coming out it also complicates connecting to the dag. Maybe I should go with the internal mounting, but then only have one tab coming out. That will reduce the chance of a leak.
I took a piece of the broken glass and melted a corner, then tried to flame anneal it. My flame annealing wasn’t very successful. It was good enough to keep it from breaking, but you can see the stress here.
Part of the problem is the thickness of the glass. I’ve joined glass tubing manually without any problems or apparent stress. But it’s clear that I don’t have the knowledge or skill to handle thicker glass.
I’ll try annealing in the kiln later.
Here’s a section where two pieces were joined with frit.
You can see that there’s no stress around the frit. The right edge was original during the frit test. The left edge is where one of the breaks occurred during the ultor button test. You can see that the break went right across the frit. This is a good sign because if the frit were weak, the break would have followed the frit.
June 17, 2013
I picked up some material to make an alignment fixture for when I add the mounts for the screen/mask assembly.
Here is the screen/mask sitting in the faceplate with the modified electron shield attached. I welded the one adjustable mount back on and repaired another one.
The electron shield isn’t as rigid now that the rest of the ultor ring has been cut away. I’ll use it to mount the assembly to the funnel. I’ll have the mounts attach close to the adjustable mounts to minimize any movement of the assembly.
It looks like the amount of material removed from the faceplate will add about 1/8” of displacement of the mask/screen assembly towards the neck.
I have a number of tests to do before I make a final decision on the mounting method.
Unfortunately because the funnel isn’t faced square the screen will have to tilt up on one side about 0.25”. I don’t want to try to square the funnel up because I can’t do that without shortening it quite a bit more.
To make sure whatever method I use to bring out the HV is leak tight I’ve cut champagne bottle in half. The wall thickness is the same as the funnel.
I’ve lapped the faces but the hot wire cut was amazingly clean.
Once I receive the dumet wire I’ll grind out a place for it and for the 52 alloy that I’ll test. Then see how rugged they are and if they are helium leak tight.
I decided to see if I could remove the stem intact from the base that was on my 15GP22. I was successful until I tried to unsolder the leads. I burnt myself and dropped it on the floor causing the stem to break. The base is still in really good condition.
You can see the silicone sleeving that they sealed to the stem with RTV on the converging and focusing electrodes.
I’m learning a lot from this, at least I hope so. But I’m beginning to realize that there are quite a few things I have to not screw up if the CRT is going to work.
June 28, 2013
I rejoined the two champagne bottle halves and included feedthru’s of 52 alloy sheet and dumet wire.
I was going to do this in two steps; first frit in the leads and then frit the top back on. Instead I tried to do it in one step. Grinding the small groove for the dumet wire was easy. But the channel for the 52 alloy was more difficult because of the diamond burr I used. I’m going to order more appropriate, and better quality ones. Because the frit shrinks a lot I ended up with a small hole above the 52 alloy. So I had to seal it and fire it a second time, which is what I would have had to do anyway in the first plan.
Both leads are leak tight to better than 1x10-10 mbar×liter/sec. However there was a leak of 1x10-9 mbar×liter/sec where the frit wasn’t applied thickly enough to form a bead on the outside. In fact it was recessed a bit.
I’m not going to try to seal it since I already know I can do that, having done many trials previously. To put things in perspective barium getter can handle about 10 micron×liters of gas per milligram of getter. Ebeam says they usually flash a couple of hundred milligrams, if I remember right. Lets say we flash 100 mg. This leak would take
to use up the getter flash. Of course the getter would have to handle outgassing and the initial residue during processing. But the point is that this size leak is not a big problem.
I’ll play with the leads and see how robust they are. If the 52 alloy doesn’t develop a leak I’ll probably stick with it for the HV connection.
I also made the fixture for aligning the electron screen part of the assembly. It has the adjustable studs which will allow alignment of the screen to the tube axis.
It’s a bit crude, but seems fairly accurate. I set the funnel in it and use gauge blocks to check that the studs are level when I frit it in, and then fine adjustment afterwards.
I checked to see how much the face was off after my grinding it and one side ends up being 1/8” higher than the other. That is fairly easy to accommodate. I rotated the funnel several times and checked to see if the measurements were the same. They would be different if I didn’t have the holes lined up and the top not perpendicular to the axis. They were the same so the fixture seems to be o.k.
When I fritted the bottle it ended up with a foggy coating on the glass. I apparently contaminated the oven when I burned off the epoxy. I’ll have to run it though a cleaning cycle.
Other than that, I’m pretty much ready to try to reassemble the CRT. That could happen in a week or two.
July 7, 2013
I wanted to do a test of what I hoped to be the final design of the mounting tabs for the e-shield. I also needed to do a test with aquadag since I had doubts about whether the internal coating was Acheson’s aquadag E or aquadag plus sodium silicate. Aquadag is a colloidal graphite and it didn’t seem like it would really form an adherent coating without a binder. Many patents referred to the coating as having sodium silicate. Acheson was acquired by Henkel and the number of versions of aquadag has been reduced to 2. They do make electrodag, but they’re redoing their site and I can’t seem to get the datasheets because of that.
I added a tab with frit, used pure aquadag on the left and aquadag with sodium silicate on the right.
After firing it looked much the same except that the pure aquadag would look like a mirror if you rubbed your finger across it. It was also much thinner and you could see through most of it. You can see that the frit is shiny now. I was not able to break the tab free by hand. I suppose I should take pliers to it to see how strong it really is.
The dag on the inside and outside of CRTs seems to be quite waterproof and is difficult to remove so I gave the test piece a bath.
The pure aquadag rinsed off easily but the dag with sodium silicate was untouched even with scrubbing. I will have to remove the dag where the tabs attach to the funnel. After they are attached I’ll brush dag across where they’re mounted to connect the tabs to the dag on the inside of the funnel.
The annealing point of 0120 glass is supposed to be 435 deg C. I fire the frit at 440 deg C. I decided to take the piece of glass I reported on earlier that had stress after I took a torch to it and put it in the kiln along with the above sample when I fired it. All the stress was annealed out.
Here is the design of the tabs. They have random holes drilled on the end that gets fritted to the funnel to help make them stronger by letting some frit flow through.
I spot welded six of these to the e-shield.
This is the e-shield in the location it will be in when fritted into the funnel. This will leave a 1/8” gap between the faceplate and the decorative mask when the faceplate is added.
I will also need to grind a place for the HV lead to come out and spot weld a lead to the e-shield. Then funnel will need to be cleaned. It’s still in the condition it was in when I finished grinding the broken edge flat.
I went to the hardware store and Home Depot to get a piece of cold rolled steel to use to align the e-shield. None of them were flat enough to use so I’ll have to order a piece of aluminum tooling plate. I thought the steel would be adequate given that I’m used fiber board to build the jig, but it is very flat compared to the rolled steel.
I played around with using a very bright LED at the color center of the CRT to see if that would help to align the mask/screen assembly. It looks like it would work well, but it would take time to make a plug to fit the neck with an LED in the location of one of the guns. I would also need to use a microscope to look at the screen since a magnifier isn’t quite adequate. I don’t want to take this contraption to work and I don’t want to bring the microscope home. I think the alignment will be adequate the way I originally planned on doing it.
The tooling plate will be attached to the three mounting points. The screws will be very lightly tightened. There will be a large hole in the center of the plate to allow me to lift each tab a bit and apply the frit. The angle of engagement is too shallow to allow me to apply the frit to the glass or tabs first. The funnel will be sitting in the wooden jig I made and the tooling plate will be bolted to it with the appropriate sized spacers. This will align the plane of the mounting points normal to the axis of the neck. The HV lead will be fritted in at this point also.
Once dry, the tooling plate will be removed. I have some concern about whether the e-shield will move during the firing, but that has never happened during any of the tests I’ve done so I’m going to rely on the frit holding everything in position during the firing to avoid the complication of finding a tooling plate with a CTE similar to the e-shield. I would also have to add spacers between the plate and the funnel to correct for the slant of the ground surface of the funnel.
After the frit has been fired, I’ll grind the frit that covers the HV lead to make it match the ground surface of the funnel. Then I’ll put the funnel back into the wooden jig and check the alignment and adjust the mounting points if necessary. They have a large range of adjustment so I should be able to correct any small shift that might have occurred.
It will be two steps. The shadow mask and screen assembly will then be attached and the decorative mask snapped on. Hopefully nothing else will snap when the mask is put on. Then the frit will be applied and the faceplate set on it. After it dries it will be fired.
At first I was afraid that I would have to remove the phosphor screen from the shadow mask to do this but RCA made those three cuts in the screen to leave access to the screws that attach the assembly. They claim the phosphor screen would go back into the exact same position, but I’d rather not test that. Although there is an argument for doing it since there might be debris on the shadow mask that should be removed.
July 25, 2013
I sandblasted spots where mounting tabs of the e-shield will be fritted. I also cleaned the dag off of the area close to where the faceplate will be fritted to prevent the dag from interfering.
After the sandblasting I washed it with water and followed up with deionized water.
I fabricated the alignment fixture for the mounting posts on the e-shield out of 3/8” aluminum tooling plate. I found a round disk on eBay. I had to cut out the center to proved access to the tabs for applying the frit. I’m kind of surprised that a scroll saw was able to handle the aluminum.
Anyway, it looks like the fixture will work well.
I’m really nervous about the idea of bringing the HV in through the frit seal of the faceplate and funnel. I decided to see how difficult it would be to add an ultor button. I drilled a ½” hole in a piece of broken funnel.
This was fairly easy to do. Since the funnel was well annealed that probably helped. I then used a Dremel tool with a diamond burr to slope the edges to match the ultor button.
I then fritted it in.
Since the frit shrinks during the firing it left a few voids; but none were all the way through. If this were the actual funnel I would have added another coat of frit to fill them.
I leak tested it and it was leak tight on the 10-10 mbar-liter/sec scale. The stud you see is a 2-56 stud that they use when installing the ultor button. I will remove it before touching up the frit.
I also checked the stress. It’s uniform and fairly small.
You can see that the frit has a grey color. I the nitrocellulose binder seems to degrading and the carbon compounds then reduce the lead oxide during firing. I assume this is what happened to RACs when they had the problem with the old batch. It should be fine for mounting the ultor button and the e-shield. The biggest change is that it will have a lower dielectric strength.
Where the frit is applied to attach the faceplate, the dielectric strength is more important because it can arc through there, possibly damaging the frit. I’ve ordered some barium nitrate and lead tetroxide. They are oxidizers that will add oxygen to remove the carbon compounds during firing, without significantly changing the properties of the frit. I’ll try each one and see which works best, or if at all.
The problem is that I mixed the whole batch of frit up since I thought I’d need a lot when I was going to try to reattach the faceplate and seal the ultor ring. Attaching just the faceplate will use a lot less frit and it’s taken longer than I thought. I’d rather not make up a new batch of frit at this time unless I really need to. I want to install a vent for the bench top kiln before I do that again. I hadn’t realized how much lead would get into the air.
July 29, 2013
I decided to make the HV connection with an ultor button. This required drilling a hole though the funnel.
I used a hand drill with a diamond core bit. I made a starting template from the top of an electrical tape box. The hole was drilled from the inside with the funnel lying on a piece of Styrofoam to back it up.
I then tapered the hole to match the ultor button.
And then test fit the ultor button.
I used the alignment fixture to hold the e-shield in position and applied the frit. I also fritted the ultor button in.
After the frit dried I removed the alignment fixture.
The frit application looks pretty sloppy. It was more difficult to control than I expected. I should have made an extension to the tip of the syringe.
I will fire the assembly on Monday and hope for the best. If the ultor button needs any touching up after that, that will be done before proceeding with any further work.
July 31, 2013
The e-shield remained aligned after firing and the mount is rigid as I had hoped. The tabs allow for thermal expansion during processing, yet hold it in a fixed position.
I expected the frit to look like the tab on the left after firing, yet most look like the one on the right. The frit flows more during firing than I expected. It turns out the small kiln I used for test runs 20C high. That probably caused it to devitrify more rapidly giving it less time to run during the tests. The ultor button shifted a tiny bit during the firing so I will have to add a bit more frit. When I do I’ll touch up the other tabs just to make sure they’re strong.
The assembly fits the e-shield properly. I meant for the ultor button to be above center on the left side of the CRT as you look at the face. Somehow I had the e-shield rotate 60 degrees from where I meant to have it. This will result in the ultor button being below center on the left side. I don’t see any problem with that though.
After the second application of frit I’ll touch up the dag coating. Then it’s time to seal the faceplate on. I want to come up with a fixture to keep it from sliding sideways when the frit melts. If it does slide there’s little chance I could repair it. So I don’t want to risk it. I think I can make something simple.
This is what it will look like after fritting the faceplate on.
The white spots will still be visible after applying the dag, but the white area around them will be gone.
It looks promising, but there are still plenty of chances for failure.
August 6, 2013
I’ve touched up the ultor button with additional frit and reinforced the mounting tabs with frit where needed. I also applied the aquadag.
After the application of aquadag the resistance between the ultor button and the e-shield is 170 ohms, which should be adequate.
I haven’t taken any photos since it doesn’t look much different.
In hindsight I have a concern that the dag might reduce the frit when the faceplate is fritted on. This might not cause a problem, but to be certain I’ve taken the previous test of a mounting tab and coated the frit with dag. I will fire it to see what happens. If it turns out to be a problem I’ll remove the dag from the frit and make the connection between the e-shield and the existing dag with a spring contact. You can see the test piece in one of the photos that follows.
It turns out that the frit/binder mixture has a shelf life of four hours once mixed, of which I wasn’t aware. Beyond that time the nitrocellulose breaks down and the by product reduce the PbO when heated. The elemental lead that results gives the frit a grey color and reduces its dielectric strength.
I mixed up the whole batch of frit quite some time ago so I needed a solution. A patent points out that adding Pb3O4 in the range of 0.1% to 1.5% by weight, with 0.25% to 0.7% being preferable, will prevent the PbO from being reduced without changing the properties. So, I decided to test it.
This is a sample of 2% Pb3O4 added.
It’s bright orange meaning that most of the Pb3O4 is un-reacted and it appears to have protected the PbO. However the CTE increased causing the glass to fracture as can be seen in the back view
The next photo shows the untreated frit, 1% Pb3O4, 0.5% Pb3O4, and the aquadag test which hasn’t been fired yet.
The 0.5% is at the bottom. The 1% looked fine after the first firing although the wetting angle indicated that there wasn’t much flow before devitrifying.
After the second firing the 1% fractured indicating that the CTE had increased.
The frit document from Techneglas shows that the CTE also increases with longer firing times and with increased firing temperatures.
The 0.5% test will receive its second firing during the test of the dag.
Assuming it passes I’ll probably use 0.3% just to be on the safe side. It’s not clear that a decrease in the dielectric strength would be a serious problem because of the way the 15GP22 is mounted. Especially if the insulating boot is still used.
You can see the differences in the wetting angle in this photo. The untreated frit and the 0.5% are pretty similar while the 1% has a negative angle.
It might seem like overkill to do all this testing, but I don’t want to have anything break after all of this work.
August 9, 2013
After firing the test piece the aquadag did reduce the frit causing the aquadag to disappear.
What’s left is mostly metallic lead. Compare this to the previous photo of the test piece. Ignore the splotch on the upper left frit test. I was playing with a lead test kit to see if I could use it to identify lead glass.
So, in order to make a reliable connection I fabricated a spring contact out of the 52 alloy and mounted it using the 0-80 stud that was on the ultor button. Then I used a nickel ribbon to connect the ultor to the e-shield. I will improve it some so that it doesn’t look quite as amateurish.
This should insure that there is a connection to the dag after it’s fired. Drilling the hole in the e-shield made me a bit nervous. I didn’t want to break the funnel.
When I went to put the nut on the 0-80 stud I couldn’t get it to start. At first I thought the stud was a different size.
It turns out that they had brazed the stud into the ultor button and the braze wetted the entire stud filling the threads. So, I had to use an 0-80 die to clean it up. That was also nerve wracking since it took a lot of force to turn the die.
I obtained the ultor buttons from the estate of the chief glassblower for Dumont. I’m not sure what they intended for the studs. At first I thought they were to make it easier to install the buttons. Corning had designed an automatic machine to do that.
But, since these were brazed in it seems unlikely that they would remove them later. They also were trying different methods. The one that I used in the funnel appears to have been brazed by hand. I found studs with braze performs that would have been used in a furnace. Others appeared to just been pressed in. Having a threaded stud doesn’t seem very useful if the threads are damaged by the braze.
August 20, 2013
It was time to go for broke. I added 0.3% lead tetroxide to the frit and applied it to the funnel and let it dry. It becomes quite hard even before it’s fired. I installed the shadow mask/screen/decorative mask assembly first. I covered the face with aluminum foil to keep dust and cat hairs off while the frit was drying.
I fabricated some guides to help keep the faceplate from sliding around when the frit melts. I made a small loop in them to avoid the area where the frit is. I put the funnel in place in the oven, carefully centered the faceplate and set it on top. Then I attached the guides.
You can just barely make out the wires on top and around the funnel holding the guides in place.
Either they worked extremely well or they were completely unnecessary.
I started the oven Saturday night. It was still too hot to open Sunday evening so I removed it from the oven on Monday, today.
The faceplate is centered as well as I could ever hope. The glass is fairly thin so it was important to have it centered well to maximize the strength.
I checked and the ultor button still has continuity with the aquadag coating. You can see the frit oozed out to form a bead. This is important because the frit was formulated to have a slightly higher CTE. The bead compresses the edges of the glass when cool which prevents any small defects from propagating. Or so I’ve read.
This is a side view of the CRT.
There are a couple of spots where a bit more frit oozed out.
This is a close up view of the frit.
I’m very happy with the way it turned out. 0.3% lead tetroxide seems to have been the optimal amount. There’s no evidence of the lead oxide reducing to lead. That is, there’s no grey color like the frit was starting to develop.
You can see the frit when you look through the front of the CRT. I don’t see any voids or bubbles. It looks very good. I tried to take a photo of it. It’s kind of difficult to capture properly.
The light color is the frit, but it’s easy to get confused by the reflections.
I need to leak test it, but I want to get a full face shield before I evacuate it. After breaking that first CRT I realize how incredibly sharp the fragments are. You just barely touch them and get cut.
I expect it to be leak tight, assuming it doesn’t implode.
After leak testing it needs to have its gun installed and processed. Then we’ll find out if it will be a good CRT or a multi-color night light.
August 23, 2013
I finally had a chance to leak test the 15G.
Please excuse the mess.
It didn’t implode and it’s leak tight!
The display above is showing the range of the bar graph in mbar-liters/sec. The bar was at 1x10-9 mbar-liters/sec. This was a background due to the utratorr adapter on the neck of the CRT. The CRT neck is smaller than the adapter comfortably accommodates and there was a small leak or diffusion across the o-ring.
There’s a small amount of helium in the atmosphere so this creates a background.
So I can only say there were no leaks larger than 1x10-9 mbar-liters/sec. If the adapter had sealed better I would have been sensitive to anything larger than 1x10-10 mbar-liters/sec. A leak of either of those two magnitudes would not be a problem in my lifetime.
My current plan is to add the necessary features to my oven to process the CRT. I feel safer using my oven since I’ve over designed it and have verified its performance. If I understand the photos of Scotty’s ovens they are gas fired and heat from one wall. The insulation is also quite thin. That would seem to lead to rather large temperature gradients which I’d want to avoid. I can’t imagine that the mom and pop CRT plants are much better.
I will borrow the glass lathe from ETF. I have mixed feelings about installing the gun myself. I first have to mount the rebuilt gun on a stem. Bob gave me one of the good stems. I also purchased two delta guns from Ingo to practice with. I have access to the equipment needed to mount the gun on a stem. I also have a number of the previous generation of stems that Bob and John gave me. If the first couple of tries don’t go well I’ll probably beg for your assistance.
May 26, 2014
I repaired the tip-off ovens with Inswool caulk. The original ceramic is very fragile, but it became very hard once heated. Too bad they weren’t baked at the factory before shipping.
The softening point of 0120 glass is 630°C and the working point is 980°C. The evacuation port should probably be at a temperature lower than the working point to seal it. I don’t know how much. At the working point glass will sag under its own weight. Vacuum will probably collapse the port at a lower temperature.
I used an optical pyrometer to measure the temperature. Because I don’t know the emissivity of the ceramic used the highest temperature measured could be off by as much and 70 deg C. About a 7% error.
Voltage RMS Temperature
Tip-off at 807°C
Tip-off at 921°C.
I forgot to measure it at 1,092°C
I then cooled it off and turned it on at 45V to see how long it would take to heat up. It reached about 920°C in 30 seconds and full temp in about 45 seconds, and then it died. I didn’t mean to do a destructive test.
Nichrome melts at 1,400°C, the surface of the ceramic was at a lower temperature than the coil. It wouldn’t be surprising to have a 300°C temperature difference at this temperature so the nichrome melted.
I’ll rewind it. I might use Kanthal. Kanthal melts at the same temperature but it lasts longer in air than nichrome does.
These measurements were done in room air and not the 300°C environment of the oven. I’m thinking 24VAC would have been the correct voltage to run it at. The temperature would probably be close to 800°C, maybe a little higher since the evacuation tube would insulate the inner surface of the tip-off oven to some degree. And 24VAC is a common control voltage.
I should have stopped the test at 40V anyway.
After the test the unit seemed quite rugged. Dead, but rugged.
The oven was made a little differently than I expected.
The wire used was 27 gauge. I should be able to wind one easily. Thanks to DIY e-cigs the kanthal wire is readily available and cheap.
The ceramic tape that I saw was used to keep the ceramic insulation coating out of the coils.
June 8, 2014
I took the old tip-off oven apart. I thought it was going to be easier, just held together with brass screw. But, it turned out that after bolting the lead clamps on they silver soldered it. They had to do that to prevent the oxide that forms on the Kanthal from causing an open circuit.
I took the dimensions of the straps and will make new ones next time. It will be easier just to cut them off than to try to unbolt them while heating with a torch. The silver coil in the bottom left is the new coil I wound.
This is it put back together. Now I just need to coat it with insulation.
I started to make the pump out for the tubulation.
The ultorr fitting is welding to the smaller stainless tube. There are 316 stainless washers on both end of the larger tube and they are silver soldered to the larger diameter stainless tube and to the tube near the ultorr fitting. I still have to weld the KF fitting on and then silver solder the water leads.
This will be inserted through a small hole in the bottom center of my oven.
I didn’t want to make a bunch of holes so I was trying to figure out how to power the tip-off oven. I decided just to buy some high temperature wire good to 840 deg C and just run it under the oven walls when I set it on.
I need to make a stand also.
I want to have these things ready before I start learning how to use the CRT lathe because I’ll want to process the 15G while the lathe is still fresh in my head.
August 6, 2014
This is the tube that the glassblower made for me out of a piece of necking glass. I’ll use it to practice with the lathe. I’ll add a piece of glass, seal the stem in, and then try processing it. If I don’t catch myself on fire.
August 24, 2014
Bob Galanter made this amazing CRT stand for my oven! It looks better than industrial quality to me.
There won’t be any danger of the CRT shifting during processing.
Bob, how did you fixture the parts so accurately?
I picked up some 1-1/2”x1/8” hot rolled steel to make a mount for the CRT lathe to hold my 2” fake CRT. This is a simple project where I just need to drill some holes and wrap some glass tape around a couple of spots.
I picked up 25’ of gas line for the lathe also. I want to put “A” style fittings on the lathe to make it easier to connect and disconnect. Hopefully Airgas or Purity has them locally.
November 21, 2014
I attempted to seal a stem in today. It was both a sad failure and a great success.
It took 10 minutes to go from -3° C to 120° C with the induction heater. Standing there in -3° C weather didn’t make me feel inclined to wait longer. Although I was using a 1,500 watt induction heater, the actual power was 560 watts. And a fair fraction of that was dissipated in the cables. So I decided to go ahead anyhow and do a slow annealing preheat as per the Champion manual.
The lathe didn’t like the cold weather either. The burners would periodically seem to go out, then flare back to life. But only when I was raising and lowering them. It didn’t do that in warm weather.
The seal would have been perfect, but I didn’t notice that I must have bumped the neck glass when I lit the burners with a torch. I guess that should be “lighted” for the English grammar experts among you.
I had made sure it was centered about the stem before starting, but the neck turns out to have been pushed back. That resulted in it looking from the front like the glass had sealed to the stem completely whereas the rear portion hadn’t.
So, when I puffed on the tube the back blew out.
This was the Fredericks stem which gave Scotty trouble by cracking. It didn’t crack for me so I’m very happy with the result in spite of the mistake. It looks o.k. under crossed polarizers.
The bulge is because I didn’t stretch the glass enough after adding the necking. It was smooth on the inside. You can see the button is off center.
And here’s the result of it being off center.
The Fredericks stem glass seems to have a significantly higher melting point than the neck glass. I don’t think it’s a lead glass. The neck glass has to be heated to a very high temperature and given time to fuse into the stem glass. Otherwise it seems to just melt onto the surface. That might be why Scotty had trouble.
The glass at the joint was quite thick, as thick as the rest of the neck so that made me happy.
I’m sure I can salvage this stem. That will also give me practice before I salvage the stem from Bob’s 21AXP22.
Then I’ll spend my time on assembling the diffusion pump, plumbing, and tubulation fitting on the oven.
If then next practice after that goes well I’ll attack Bob’s CRT.
December 1, 2014
Well, I’m decent at adding necking now. And, I’ve figured out went wrong with installing the stem last time because it happened again.
I thought the tube had moved last time but it hadn’t. What happens is because there isn’t a gun attached to the stem, when the glass necks in and drops it usually touches on one side first and tilts on the way down. The stem is very light with nothing to hold it in place so it gets pulled to one side. It moves off the side of the chuck.
I need to bend the inside leads to hold the stem centered in the tube and keep it from getting lifted up. I had planned on doing that in the beginning but didn’t, thinking it wouldn’t matter. Turns out it does.
Still no cracks and the stress looks manageable. If I’m not breaking Frederickson’s stems, then they must be o.k. I reused the one from the first attempt. I was going to fire polish it after grinding it out, but decided to see what happens if you don’t. It worked fine except for a few tiny trapped bubbles of no consequence.
Heating the chuck with a torch beforehand seems to work o.k. without heating the stem or tube.
December 6, 2014
I’ve learned some things, or at least I think I have. One is that the best supply pressure setting for the lathe is 3 psi for the propane and 6 psi for the oxygen. This matches the recommend supply pressure for several commercial glass working torches. I couldn’t find specifications for the Litton burners used on the lathe. When I ran at higher pressures it was very difficult to control the flame. On one of the following images you’ll see a dark band that resulted from a momentary burst of too much propane.
The first two attempts at sealing the stem resulted in a hole which I thought was due to the stem getting pulled to the side by the glass as it drops. The stem is quite light and without a gun to hold it centered it’s easy to move. The bore of the chuck is quite a bit bigger than the tubulation so it doesn’t restrict movement. Bending the leads to hold it centered solved that problem. Nick didn’t have a dummy gun and his worked. I think the fit of the tubulation in the RACS chuck was much tighter.
There are two sections of neck glass on the right tube. One seal is the ugly too much propane one and below it is a nicer clear seal after finding the proper pressure settings on the regulators.
But, the glass cracked while cooling. The button stem didn’t crack though, the glass around it. I took it to our glassblower and asked his opinion. He said it cracked because on the inside at that point the glass tube joins the stem with an acute angle rather than the round filleted corner the rest of the way down. It can’t handle as much stress that way. This is at the point marked “Tab of glass.” While cutting the glass with the flame it fused to the chuck at that point. The only way to get it loose from the stem was to lower the chuck. That pulled the glass down some causing the acute angle inside.
The seal was much better than it looks in the photo. The crack makes it hard to see the seal. It was smooth all the way around like you see in the lower half.
I believe it cracked because the button of the stem wasn’t hot enough when the annealing stage started. If it was much cooler than the melted edge, then when the outer glass cools the strain point you now have stress free glass while the other perimeter is very hot and the rest of the button is significantly cooler. When they cool to the same temperature the outer glass will shrink more because of the larger temperature change. This put the outer ring of glass under tension where it is weak. The button is under compression where it is strong so only the outer edge of the glass cracks not going into the button.
The solution to this would be to heat the chuck after cutting the glass with the flame. In a photo from Nick’s report the chuck appears to be glowing with the flame just below the button. This wasn’t the case during the sealing I did.
This would heat the button. Then once the button was hot I’d go back to the edge of the button and gradually reduce the flame. The tube and the stem would start at nearly the same temperature so little stress would develop.
The stem on the left in the first photo was used twice. It’s easy to reuse them.
There’s a problem with cutting the glass with the flame and trying to use a puff of air in the process.
I used the chuck on the left. The center chuck is the commercial one that would be used for a tube like a 21AXP22 that has 14 pins instead of the 20 pins of the 15GP22. So the pins on the 15GP22 stem have to be on a little large radius and don’t sit well in the chuck. There was also concern about bending the leas sharply so Bob made a longer chuck that fits the stem well.
It has a much longer straight surface at the top than the commercial chuck has. Now when you try to cut the glass with the flame the glass presses against the chuck and gets cooled making it difficult to get hot enough to separate it. You can see the spot where the glass bonded too it. I had to grind it off. It was so well bonded I couldn’t chip it loose.
I think the chuck might work o.k. if I coat it with chalk like Nick said RACS does. At least then it won’t stick
It turned out that the holes in the first two attempts wasn’t because of the stem being off center, it was me trying to use a puff of air to cut the glass. On the last attempt I noticed that the glass above the stem was starting to bulge when I puffed on the tube even though I had the flame below the button of the stem, but having no effect on the stretched glass that was up against the chuck. I actually had to suck to pull the neck back to a normal shape.
There are four 40 mil holes in the Hawkeye chuck that Bob used as a model. Presumable these were to allow air to pass out and puff out the glass. The problem with this is that the large round cylinder you see on the lower part of the chuck has a 0.2” gap between it and the glass all the way around. There’s no way you’ll build pressure between the chuck and the glass because the 40 mil holes can’t supply air fast enough compared to the 0.2” gap. That’s like putting a 1 meg resistor in series with a 1 ohm resistor and expecting to see a significant voltage drop on the 1 ohm resistor. So those holes seem to serve no purpose.
There are two other paths for the air. Up through the stem which pressurizes the CRT, which caused my tube to bulge. And, up around the tubulation and out between the button and the top of the chuck. That path has some promise for helping to burst the glass. But I don’t think it’s very significant. In Nick’s report it seems that RACS only uses the flame to cut the glass and doesn’t need a puff of air.
In the video I can hear Bob asking Scotty if he puffed already. It wasn’t obvious that Scotty did. It looked like it was really cut by the flame.
I think a chuck more like the Hawkeye chuck would be easier to use. The short straight edge and the large angle keeps the glass away from the chuck when being cut by the flame. The glass doesn’t have anything to back it up then. The chuck could be a little longer than Hawkeye’s, but the leads seem quite flexible and I don’t think that puts much stress on the glass.
I don’t think Scotty really did any annealing. He reduced the flame too much too quickly to keep the temperature in the annealing range. If the CRTs survived the trip to the oven though, the major portion of the stress would be relieved if they were processed at the temperatures he mentions in his procedures. That’s above the strain point and just below the annealing point. If I remember correctly it takes 15 minutes to anneal at the annealing point and 2 hours at the strain point.
This is the gap between the neck and the chuck. It won’t trap much air.
I’m going to salvage the two stems I used and try again using chalk to try to prevent the sticking, and heating the chuck to heat the stem better at the end of the process.
I have a number of short sections of glass that I will try to join together to use for the tests. I’m getting short on the neck glass and I don’t want to waste any more than necessary. I had borrowed a 4’ section of necking and a much shorter section. The glassblower forgot to reset the oven for soft glass when he annealed the short section after making the dummy CRT for me so that piece was wasted. Part of the 4’ then was used to make a dummy tube, and the tests have used up a fair amount of the remainder.
I’m not sure why the Fredrickson stems(?) cracked for Scotty, but not the Richardson ones. I’m annealing the pieces in the photo right now, and if the glasses aren’t compatible I’ll still see stress after annealing.
Our glassblower though says they are compatible, otherwise I’d never have even come as close to making it work as I have.
I was successful in sealing a stem in last night. I rubbed boron nitride on the chuck which helped keep the glass from sticking to it. I still had trouble cutting the glass with the flame because of the shallow slope of the chuck. Blowing air does break the curtain, but not all the way around, I had to lower the chuck to break the rest.
I also made sure to get the button up to temperature before starting the gradual cool down.
I need to make an adapter plate for the diffusion pump. I’ll do that next, then work on Bob’s 21AXP22. After that it’s assemble everything needed for the oven and then process the dummy CRT, and then the 21AXP22 if all goes well.
I was having problems with the Teflon neck glass holder expanding and going from a sloppy fit to a very tight fit as it warmed up. Teflon has a high CTE. In one case the pressure broke the neck glass and the Teflon wasn’t really hot.
So Bob Galanter made this really nice neck glass holder out of graphite.
Graphite has a low CTE and it eliminated the problem. The fit is perfect.
Bob also made a jig to mount stems on 15GP22 guns. It is really nice and very accurate.