View Full Version : If you want something done right ... (vacuum tech)

01-30-2009, 04:34 PM
Well I just had to rant a bit ...

So I've been working on a physics project that involves a high-vacuum system. The components I've been building are all out of 304 stainless steel and, while not difficult, required some time and prep to do them right. Cutting a nearly a 3" OD, .135" wide O-ring groove .09" deep in 304 takes a little bit of time - at least it does on a tiny 15" Cincinnati Tray-Top. Had to back the rpm way down to avoid that nasty harmonic.

Anyway, I lined up a TIG welder (a student but reputed to be a very good welder) to weld my flanges on. Well I wasn't happy with the results so I threw away about 4 hours of machining time and started over. This time I did all the welding myself. It was the third or fourth time I'd ever picked up a TIG torch, but it turned out pretty good for my application.

Well this week I had to add on a little stubby piece to the tube. I did all the machining and made sure that the stubby piece was short enough that you could get a TIG torch on the inside to seal weld it. I was going to do it myself but the FSAE shop was out of Argonne, so I took it to the machinist in the department. He's a great machinist but I was disappointed with the welding. I'm no pro welder myself, but I think my job was much better than his. I guess I was expecting too much. Now I'm kicking myself for not waiting until some Argonne was available.

This was the flange I welded on:

Here is the new addition:

It is not welded on the inside and there are some porous bubbles from burn through. I think I will have to go back and chisel them off and seal weld the inside to make sure my pump-down time is reasonable. What do the vacuum guys here think?

01-30-2009, 04:55 PM
I'm no vacuum guy or welder but the bottom welds look like my first attempt with a stick welder!

The top job looks sweet to me and i'd be chuffed with it.

I'd wait for the gas and have a go myself.


01-30-2009, 05:33 PM
Screw on some plates and pressurize it with helium. Borrow a sniffer and look for leaks.

01-30-2009, 07:37 PM
Standard practice for high vacuum is seal weld on inside only, with skip welds on outside only if needed for strength. Since you already have a full weld on the outside, you run the risk of a porous weld on the inside caused by trapped gasses trying to escape through your new weld. This will lead to a virtual leak at best, and a real leak at worst. You should try and ventilate the outer weld first, or better yet cut it off and start over with clean metal. If an inside weld is not possible due to space constraints, you must backpurge the inside with inert gas to prevent the burning you're seeing. The other problem with the double weld is when you go to leak test it and you're spraying a little helium on the outside, you won't be able to definitively isolate the internal leak location from the external application position due to the possible multiple leakpaths.
What level of vacuum are you trying to achieve? A little bit of externally applied Glyptal can fix some leaks if your requirement is not too stringent. I'd avoid use of silicone grease as a band-aid, as it can make later weld repairs almost impossible if it gets drawn into the porosity.

01-30-2009, 07:42 PM
Fast - I think you did a great job! Forty years ago, we would have slathered some epoxy on the outside of the weld while under vacuum! That would have plugged the holes and not outgas measurably (back then). Have not been involved with high vacuum since then.


01-30-2009, 07:58 PM
Thanks AT, but thats the bummer - My weld came out fine, but the weld I "farmed out" came back like crap! :(

Is Glyptol the same as vacuum epoxy? I need something on the order of 10^-6 torr. Unfortunantly, I'm working with a really shoddy vacuum system. The diffusion pump doesn't have a cold trap, so back streaming of the oil slicks everything, including our phosphor screen :( Last semester, before adding this new little stub to the tube, I was able to get to 5 X 10^-6 torr, so not too bad.

Thanks for the tips interrupted cut. I designed it so that it could be welded from the inside. In fact, I made it shorter than I wanted just to make sure I could get a standard TIG torch on the inside and weld it. I guess I should've specified what I wanted done, but when he said "you need it vacuum tight, right?" he would've welded the inside as per "common practice" as you say. You'll notice that the flange I welded on only had tack welds on the outside and then the inside bead. I was all ready to have a go at welding the inside of this new part. I didn't think about the gas trapped in there causing weld porosity. I guess I'll just leave it alone and see what happens.

I spent two weeks trying to get our helium leak checker working but it was a no go. There is a huge leak somewhere in the manifold. With the manifold shut off, it pulls down to 10^-8 torr. With the throttle valve open, it doesn't go below 10^-2 torr. Should be easily detected just with methanol and the roughing pump running, but its all hidden behind a mass of sensors and wires.

<sigh> Student labs ... :rolleyes: Oh well - at least it introduces us to the equipment even if the equipment doesn't work properly.

01-30-2009, 10:23 PM
I can sympathize. I've been doing vacuum degassing of epoxy in a chamber bought from e-bay and getting 29.9 inches of vacuum using a rotary vane pump without something going wrong is a minor challenge. 1e-6 torr sounds even more fun.

I don't have any experience with that vacuum level so I don't have any advice but I think we can all agree that vacuum sucks ;) Your weld is beautiful!


01-30-2009, 11:02 PM
Glyptol is a low outgassing red paint. I do not suggest using it. Varian used to sell a product called "Torr seal", it is usable for 10-6.

Your joint designs needs a little work. Vacuum welds should always be on the inside. Filler materials should not be needed. Fillet on the outside for strength, only skip welds. If you are joining two VERY dissimilar thickness parts, undercut the edge on the thick one. Example: A .010" bellows to a 1/2" plate. The plate will have an "oring groove" looking cut that provides a .010" little wall that is right next to the bellows. The weld is .010 to .010 type materials. I can draw it if needed.

A Helium leak checker MANIFOLD that won't pull down to 10-6 doesn't suck, it's crap. No suggestion there. Well, yes I do. Don't use Methanol, use He. Valve whatever you have to so the Helium checker can be on the manifold, and give it a try.

Dave J.

01-30-2009, 11:18 PM
Hey Dave - If you read all the way through the post, which I admit is lengthy, I designed that joint so that a TIG torch could readily reach the inside and weld from the inside! It probably looks alot deeper in the pics than it is. Thats only .8" deep and I could get the TIG torch in there at the right angle to do it myself, we were just out of Ar. Anyhow, thats why I was upset with the whole thing... like you say, it should be welded from the inside. Hence "if you want something done right, do it yourself" :rolleyes: oh well.

I forgot to mention on the helium leak checker, the detector seems to be hay-wire. It was almost impossible to get it to zero and the audible doesn't work. We were going to work on getting it fixed but since it was shot and the manifold leaked, we decided to abandon it. The only thing I'm adding to the system is this port. Theoretically, since we got down to 5X10^-6 last semester, if we don't achieve good vacuum this time around we can assume its the new bit ... theoretically :D

01-31-2009, 01:21 AM
Glyptal is a company name, not a product. Glyptal was a division of GE and was spun off as a separate company. They make many different products inc. thread lockers, anti sabotage indicator paint, vacuum sealants and many more.

01-31-2009, 03:06 AM
Have you considered brazing or silver-soldering the joint on the inside after cleaning off the burnt areas?

I don't see how he could get the TIG torch into the area between the flange and the manifold (middle picture in the first post). It's too bad that weld wasn't done first. It would have probably come out better.


A.K. Boomer
01-31-2009, 08:23 AM
Brazing is a good suggestion if it doesnt warp out all his work --- Fasttrack -- there's nothing worse than pouring your heart into something and then having to hand it over to some huckleberry to destroy --- I remember when I was building my 4 stroke engine with rotary ports, I did almost everything myself and had to hand the rotary cylinder head I built over to a welder -- went over with him in detail about heat build up and how I couldnt have it warp anything --- told him the critical areas and asked if it would be a problem -- he said no, I told him to take it step by step and cool it - he said he would -- got it back and he warped EVERYTHING --- I was furious --- some welders are very thick in the head:(

Since then I do EVERYTHING on my own -- not only am I mostly working on stuff that I dont want others to view anyways - the screw ups are few and far between - besides - when it happens I know who to blame:p