Re: 15GP22 Rebuild Attempt in France - Soon!

[ohohyodafarted]

Hi Eric,

The entire design of this tube and in particular the Ultor Ring Assembly, is predicated on a number of design constraints.

I will attempt to take the issues one at a time.

First RCA needed a method of making the envelope in two pieces. A front half and a rear half, so as to facilitate the instalation of a third assembly inside the tube. That third assembly is the phosphor dot plate/shadow mask assembly.

The technique of fusing glass to metal was well understood. So it was decided to make a glass front face assembly fused to a metal ring, and a rear glass funnel and neck assembly fused to a metal ring. Then the phosphor dot plate/shadow mask assembly would be mounted to the flange on the rear assembly, and the front half and rear half would be brought together and the two ultor ring halves would be heilarced together to form a complete crt assembly.

The phosphor dot plate/shadow mask assembly mounts to the light guage rear half of the ultor ring inner flange assembly with 3 bolts.

The reason for the two light guage metal rings with the bulge on the outer circumference, is that when the tube is evacuated, great amounts of mechanical force are exerted which squeeze the front and rear halves of the tube together exerting great amounts of pressure on the two light guage metal rings. The rings need to flex under the pressure. The pressure needs to be distributed evenly around the entire surface where the two halves of the ultor ring meet.

The corrugated metal washer that is sandwiched between the front and rear rings, is designed to act as a cushion to evenly distribute the tons of force squeezing the two rings together. If the corrugated washer was not there, the front and rear metal rings would squeeze together, and in the places where the rings would touch all of the force would be exerted in just those spots. That would cause huge stresses and probable breakage of the glass to metal bond on the front and rear rings. Even if the glass to metal joints survived, there would be pressure points that would make the tube vulnerable to implosion if bumped or jared.

The reason for the heavy metal rings which are the bonding surface for the glass is thought to be as follows. The light guage metal would likely melt under the high temperatures needed to fuse glass to metal. To solve this problem, heavy guage metal rings were bonded to the light guage sheet metal rings. Probably a rolling fusion weld under great force and very high current. This is the joint which probably leaks.

I had long suspected that the leaks were occuring in the metal to metal fusion jonts between the heaavy and light guage metal rings. So when John was in Milwaukee a couple months ago we decided to cut a cross section of an ultor ring that we had been using to experiment on. We used a 3 inch high speed carborundum disc to slice a section of the ultor ring away. We then ground and polished the cross section to a high finish so we could closely examine the integrity of metal to metal joint.

AS you can see from the cross section photo of the ultor ring in John's post, the heavy guage ring is not bonded uniformly to the light guage sheet metal ring. There is a very evident partial gap between the two components. This was only a random cross section and we found deficiencies in the metal to metal joint. I strongly suspect that some areas are joined better and some are joined much more poorly.

It is this metal to metal fusion weld between of the heavy guage ring and the sheet metal ring that will need to be sealed. We can't be sure, but we do not think the glass to metal bond is where the leaks occur.

Glass to metal bonds were well perfected and being used in metal funnel crt's for many years before the 15GP22 was put into production.

There has been much discussion betewen John and myself about the usability of frit glass to try sealing the metal to metal fusion joint. And John and I have differing opinions on the use of frit glass to seal 15GP22 leaks.

I had a very long and interesting discussion with one of the elderly engineers who worked at Rauland picture tube. I also spoke to a fellow that works for one of the companies that make frit glass. These fellows both told me that frit glass will not work for the metal to metal leak we are trying to fix. The problem is that the coefficient of expansion of the frit glass has to be matched to the metal very precisely. And even if we were able to match the expansion coefficients properly, due to the fact that the sheet metal ultor ring flexes when the the tube is evacuated, the frit glass would crack due to the flexing.

Frit glass is very useful to bond two glass surfaces of a crt together because the glass all moves at the same rate. When you get dissimilar rates of expansion or physical stresses such as the ultor ring flexing, then frit glass will fail. It will either crack or debond from the surface. Even in glass to glass joints, if the frit glass in not perfectly matched to the rate of expansion of the glass parts, the joint will fail.

As to your query on leak detection, as was discussed in another post the spark method is not going to work on the metal ring as it is electrically conductive throughout. As to pressurizing, you are correct the leak is too small to find this way and may explode the tube because it is not designed to be pressureized, it is designed to be under vacuum.

The only practical method of leak detection would be with a helium leak detector. IN this method a strong vacuum is pulled on the crt by a machine that is designed to count Helium atoms in the vacuum stream. Helium is used because the atom is so small it can squeeze through the smallest of leaks. So while the tube is under vacuum via the HE leak detector, the operator introduces small amounts of HE from a tiny supply tube into the areas where the leak is suspected. Almost instantily the HE leak detector will detect the HE atoms in the vacuum stream and a meter on the machine will register a large number of atoms. As soon as the meter registers, you know you have found the leak.

All this leak detecting is in my opinion not of much use. Even if were to find the exact pont of the leak, and seal it, additional leaks could crop up in other spots. The solution is to seal the entire metal to metal joint all around the tube to make sure that the leak is fixed and to prevent any possibility of future leaks occuring.

I hope this has shed some light on the subject of why the tube was built the way it was and where we suspect the leaks are.

Bob

[John Marinello]

Forget about pressurizing ANY crt for ANY reason. Glass is excellent under compression, but should NEVER be subjected to tension. A few pounds into a 15GP22 would cause the ultor flange to act as a bellows, thus ruining that. Add few more pounds, and the tube will explode & cut you to ribbons.

[jeyurkon]

The corrugated crush washer appears to be a different metal than the other parts in the photograph. Is it aluminum by any chance?

John

[TubeType]

Thanks Bob and John.

It was nice to read your detailed explanations of the 15G's construction. I counted five glass-metal or metal-metal junctions. That's ten bound surfaces. This is a mechanical engineer's nightmare.

Have you considered using a large molten glass band to completely encompasses the Ultor ring? This glass band would be fused to front and rear glass surfaces of the 15G. It would include a feed-through connection to the Ultor. If this technique is mechanically sound, wouldn't it eliminate most if not all of the sources for vacuum leaks?

Kind regards,

Terry Cheek

[jr_tech]

Thanks Bob and John for posting the construction details of the 15gp22. Wow! No wonder many of these crts leak. I suspect that percentage wise, more prototype metal cone 15" crts have survived!

I suspect that patching over the outside of the seal areas with frit (or glass) would not be a permanent solution. as flexing of the seal area is likely to fracture these brittle materials. Also, I suspect that pockets of air that could be trapped between the frit and the metal flange could be a source of a "virtual leak" that could over time cause the tube to become gassy even though there may be no detectable path to the outside...

Tough problem !

jr

[ohohyodafarted]

jeyurkon:

The crush washer is of a different metal. I dont have it hear, but it is either aluminum or Stainless steel.



jr_tech:

Your asumptions are right on. We too are concerned about redidual gas entrapment after sealing the metal joints. Hopefully the amount of entraped air would be so minimal as to be absorbed by getter material.

I have a 16gp4 that has milky getters. Presumably from a leak. But it still produces a very good picture. So, it is possible that if the getter flashes are very large, we may be able to counteract the infiltration of any small amounts of gas that might get entraped in the ultor ring after it has been sealed.

Bob

[eberts]

Originally Posted by ohohyodafarted

jeyurkon:
it is possible that if the getter flashes are very large, we may be able to counteract the infiltration of any small amounts of gas that might get entraped in the ultor ring after it has been sealed.

Bob

Nope, the getter only works on a singular molecular level on the surface it is deposited on. Surface area of getter flash is more important than amount of material in getter. You need a good hard vacuum for the tube to work.
Maybe you could rebuild a tube connected to a molecular ion pump continously, then you won't have to worry about a slow leaker.

[jeyurkon]

Originally Posted by eberts

Nope, the getter only works on a singular molecular level on the surface it is deposited on. Surface area of getter flash is more important than amount of material in getter. You need a good hard vacuum for the tube to work.
Maybe you could rebuild a tube connected to a molecular ion pump continously, then you won't have to worry about a slow leaker.

I think Bob did mean "large" in terms of area and not thickness. I believe the getter flash does do better than a single mono-layer, but its pumping speed decreases.

John