What's new
What's new

Choosing between welding, brazing, or shrink fits for a rocket engine.

KurtzC1

Plastic
Joined
Feb 20, 2019
Hey Everyone!

This is my first post here on Practical Machinist. I am a senior mechanical engineering student at Embry-Riddle Aeronautical University in Prescott, Arizona. I am working on a rocket engine for my school, and I just encountered a serious problem. I was hoping advice on how to proceed, because everyone on practical machinist seems to be very knowledgeable.

The problem I am having is getting fluidic connections (fittings) to the engine. Shown in the pictures, swaglok fittings were welded to the engine. The welds resulted in ~60 thou of deformation. I knew that there was going to be deformation, but I didn’t think there would be so much deformation that it likely will have to be scrapped. Unfortunately, the person that welded the engine accidentally did two passes on each weld causing further warping.

Currently the plan is to fix this part by putting the part in an oven and slowly heat it to 1200F. Once heated, a press will then attempt to push it back into place, removing the 60thou of deformation. Does anyone recommend any other techniques for reducing residual stresses in welds?

If this doesn’t work, then the part will have to be scrapped. I could attempt to reweld the engine with smaller fillets, but I am still concerned about warping.

What are people’s opinions about furnace brazing? The part is tightly toleranced and has sealing surfaces, so I thought that furnace brazing would reduce potential warping.

I am also considering shrink fitting the fluidic connections onto the engine and then tac welding them into place. They would need to be tac welded, because liquid oxygen which is -300F will be plumbed into the engine. I would be concerned about the fittings coming lose at cryogenic temperatures.

What are people’s opinions about shrink fits? Do you think they can withstand the temperature cycling?

I can’t use boss ports and NPT taps due to weight constraints and packaging problems with the engine.

I realize this is a long post. I have no idea if anyone will read this much, but I figured it would be worth a shot, because I am quite desperate. Attached are pictures.

Thank you for the advice and any information would be appreciated!
 

Attachments

  • Pintle Post 4.jpg
    Pintle Post 4.jpg
    86.1 KB · Views: 696
  • Pintle Post 3.jpg
    Pintle Post 3.jpg
    89.8 KB · Views: 721
  • Pintle Post 2.jpg
    Pintle Post 2.jpg
    92.6 KB · Views: 700
  • Pintle Post 1.jpg
    Pintle Post 1.jpg
    87.9 KB · Views: 675
Those gigantic welds are most of your problem. Fuse them with no filler rod, both sides of plate if possible.
Might be able to put it in a press and get some of that warpage out.
 
Some (who am I kidding, many) years ago I made a stainless plate for high vacuum with a bunch of 1" ports in it. These consist of a flange connected to a tube which was maybe .05" wall. The university physics shop suggested they machine a groove around the hole for each port leaving a lip also of around .05" thick and maybe .05" deep, basically a doughnut around the port hole. The lip and the edge of the port were then fused together with TIG resulting in very small, neat and clean welds with no distortion in the plate. The welds by the way were on the inside of the plate but slightly recessed since the tube and lip melted down into the groove. Now I'm not sure if you have pressure considerations requiring a certain weld penetrance but as Rob F says those welds of yours are excessively large and they can definitely be made smaller and warp prone.
 
Honestly post weld straightening is the norm on a lot of higher precision applications, just place under a press supporting the rim and you can get that 60 thou back to flat in a minute or 2,
 
On the suitability of a shrink fit, “Good lord no! “. If you think about it, how do you make a shrink fit usually? Put one part in liquid nitrogen? Subjecting it to cryogenic temps would then sound like how you would “ UN-shrink fit” right? Can not rely on “tack welds”, if they aren’t good enough to hold loose fitting parts, why would they hold shrink fit parts once they come loose? Not to mention that “tack welds” can be brittle, that’s why they are called tack welds, because we do not Rely on them for anything important.
 
I have no qualifications or even a good understanding of your application but would post-weld machining be an option? Allow it to move through the welding process, stress relieve it and then machine it to your specs? If not I'd be interested hearing why this is not a good choice.
 
Some (who am I kidding, many) years ago I made a stainless plate for high vacuum with a bunch of 1" ports in it. These consist of a flange connected to a tube which was maybe .05" wall. The university physics shop suggested they machine a groove around the hole for each port leaving a lip also of around .05" thick and maybe .05" deep, basically a doughnut around the port hole. The lip and the edge of the port were then fused together with TIG resulting in very small, neat and clean welds with no distortion in the plate. The welds by the way were on the inside of the plate but slightly recessed since the tube and lip melted down into the groove. Now I'm not sure if you have pressure considerations requiring a certain weld penetrance but as Rob F says those welds of yours are excessively large and they can definitely be made smaller and warp prone.

Basically this. Re-make the flange and do the fittings this way:

1) machine the holes for the fittings undersized by about 15 thou.

2) reduce the diameter of the mating fitting end so it drops into the smaller hole, with
the end of fitting flush with the bottom of the flange.

3) get the TIG welder to simply fuse the end of the fitting with the
bottom of the flange, welding around the joint.

If you cannot tolerate any protrusion from the bottom of the flange then
have the tube end profiled to be about 15 thou shy of the flange bottom.

The guiding principle is that you want both sides of the joint to be
about the same thermal mass (thickness) so when the weld happens
the penetration is about the same on both parts. This is one reason
that an annular groove is sometimes machined underneath the flange
where the tube protrudes through.

Ideally you need no filler rod if the filler material is created by the way
the joint is designed.

Because the materials you have on both sides of the joint are the same,
and have the same thermal expansion, the joint sees low stress as the
temperature varies.

The kind of joint you making here (swagelock to stainless tube, on a stainless
flange ) is absolutely fine for cryogenic service. Those work well at temperatures
below 1 K.

The only possible better kind of connection is a VCR fitting which uses a
stainless gasket between two stainless flanges. For larger connections
a conflat flange (copper gasket between knife edge flanges) is a good approach.

Look those joint types up to see how they are designed.
 
Can you use bulkhead fittings?

Shrink or silver solder connections are possible, but you need to decide how much temperature is going to impact those types of connections in your application.

You could contact NASA or some of the engine manufacturers for advise.
 
Seriously ?

Apparently your school has failed you.

I would ask for a refund, or maybe you didn't pay attention in class ?

What did engineering students do when learning this 30 years ago (hint..."pre-internet")
 
Press as flat as possible and machine a few thou off the back side??


Sent from my iPhone using Tapatalk Pro
 
Seriously ?

Apparently your school has failed you.

I would ask for a refund, or maybe you didn't pay attention in class ?

What did engineering students do when learning this 30 years ago (hint..."pre-internet")

Ease up guy. Last I knew Embry-Riddle did not offer welding engineering. We all have to learn sometime, and OP has a lesson on weld distortion that will last much longer than book learning.

I would press it flat, but would do it cold, not hot. The differential cooling from conduction to the contact surfaces will present it's own distortions. In addition, working with 1200F metal is not for the faint of heart. That is near blacksmithing territory, and its not a place for the inexperienced. Now add in the necessity to do a precision pressing at that temperature and you have a near impossible task.

When you press it, your first movements will be in the elastic range. The part will flex, then return to original shape. Once you increase the distortion past the elastic range, the distortion will be permanent. For example, .150 of flex will result in no permanent distortion, but .170 of flex will result in .020 of permanent distortion. This means you will need to either have a way of measuring the displacement in real time, or press to an adjustable stop that is adjusted each time until you reach the desired level of flatness. A slight overbend and then press back flat seems to result in a more stable part.
 
Plus one on the welds being way, way, way too big, fuse or use .035 to .065 filler rod, on that thickness of plate is should move very little if at all. I would also add a vote for straightening it cold in a press. We weld fittings into ultra high vacuum plates that are .125" or so thick with .065 filler all the time with only a couple thou of distortion, control the heat and much smaller weld to start with. Also, if you can change the design so you are welding the fittings from both sides you will have a much easier time keeping the distortion in check.
 
What are people’s opinions about furnace brazing? The part is tightly toleranced and has sealing surfaces, so I thought that furnace brazing would reduce potential warping.

We vacuum furnace braze stuff all the time vs welding as welding can induce some serious distortion as shown.

If it was me, I would have the raw plate stock stress relieved, machined, then braze in the fittings. Done. The stress reliving operation should not be skipped otherwise the part will stress relieve itself during the braze operation and warp.
 
Seriously ?

Apparently your school has failed you.

I would ask for a refund, or maybe you didn't pay attention in class ?

What did engineering students do when learning this 30 years ago (hint..."pre-internet")

They asked around. Just like this guy. Why are you always such a dickhead? Do you ever say anything nice or helpful?
 
They asked around. Just like this guy. Why are you always such a dickhead? Do you ever say anything nice or helpful?

There's been a few times...I don't recall any but I'm sure there were:scratchchin:

But I will say most of his posts are spot on, being abrasive myself I can appreciate the upfrontness.
 








 
Back
Top