SpaceX press release on 3D printed valve flown on F9

[FutureSpaceTourist]

http://www.spacex.com/news/2014/07/31/spacex-launches-3d-printed-part-space-creates-printed-engine-chamber-crewed

Release begins:

Through 3D printing, or additive manufacturing, robust and high-performing rocket parts can be created and offer improvements over traditional manufacturing methods. SpaceX is pushing the boundaries of what additive manufacturing can do in the 21st century, ultimately making the Falcon 9 rocket and Dragon spacecraft more reliable, robust and efficient than ever before.

On January 6, 2014, SpaceX launched its Falcon 9 rocket with a 3D-printed Main Oxidizer Valve (MOV) body in one of the nine Merlin 1D engines. The mission marked the first time SpaceX had ever flown a 3D-printed part, with the valve operating successfully with high pressure liquid oxygen, under cryogenic temperatures and high vibration.

Compared with a traditionally cast part, a printed valve body has superior strength, ductility, and fracture resistance, with a lower variability in materials properties. The MOV body was printed in less than two days, compared with a typical castings cycle measured in months. The valve’s extensive test program – including a rigorous series of engine firings, component level qualification testing and materials testing – has since qualified the printed MOV body to fly interchangeably with cast parts on all Falcon 9 flights going forward.

There's also info on the SuperDraco 3D printing.

Edit: changed title

[SpunkyEnigma]

.... with a 3D-printed Main Oxidizer Valve (MOV) body in one of the nine Merlin 1D engines. The mission marked the first time SpaceX had ever flown a 3D-printed part....

Interesting use of engine out capability as a somewhat low-risk way of trying out slight variations on a single M1D

[FutureSpaceTourist]

Interesting use of engine out capability as a somewhat low-risk way of trying out slight variations on a single M1D

Yes and yet more evidence that SpaceX are constantly tweaking and evolving. The Jan flight was also the third qualification flight for USAF. Clearly SpaceX don't let anything stop them trying new things, although that can cause issues (such as some of the Orbcomm delays).

What really impresses me though is that 3D printing has already reached the point where it can produce space qualified engines that aren't possible in any other way (SuperDraco) and so much quicker.

I assume the quicker time means less development costs; also cheaper to manufacture? How long do 3D printers last, do they lose accuracy/calibration over time?

[TrevorMonty]

Looks like SpaceX are going to gradually apply 3D printing to Merlin 1D. Long term I'm guessing it will more than half manufacturing costs, plus allow them to increase their production rate.
Reusability of boosters shouldn't effect the development program as engines can still be modified or swapped between missions.

[SoulWager]

I assume the quicker time means less development costs; also cheaper to manufacture? How long do 3D printers last, do they lose accuracy/calibration over time?

Some parts might be cheaper to manufacture, others would be more expensive, it depends how difficult it is to make with more traditional methods, and on the size of the part. Big simple parts are cheaper with traditional methods, small complicated ones may be cheaper with 3d printing. Development costs may be lower, but the biggest development advantage is time. in that you can make a change in the design, print it, test it, and get the data in days, instead of weeks.

I don't think 3d printers will "go bad" any quicker than other industrial equipment, like laser cutters and milling machines.

[dante2308]

They may be attempting to vertical integrate engine parts that they have been sourcing so the cost savings may be greater than the fact of 3d printing alone.

[Dudely]

I'd like to point out that, over the long term, it is cheaper to form a mold of something and cast it many times than it is to additively print it every single time. It also takes a LOT of time to print something, and you can only print one thing per printer at a time with the printers themselves costing hundreds of thousands of dollars.

This will be a boon to their development cycles and will makes it much cheaper and faster to modify and test something, and it will allow for more creative part shapes. However, over the long run I think any parts they think won't change for hundreds of flights will be casted.

[SoulWager]

I'd like to point out that, over the long term, it is cheaper to form a mold of something and cast it many times than it is to additively print it every single time. It also takes a LOT of time to print something, and you can only print one thing per printer at a time with the printers themselves costing hundreds of thousands of dollars.

This will be a boon to their development cycles and will makes it much cheaper and faster to modify and test something, and it will allow for more creative part shapes. However, over the long run I think any parts they think won't change for hundreds of flights will be casted.

Some parts are difficult to cast, consider how difficult it would be to make a mold for a superdraco combustion chamber with regenerative cooling passages. You'd have to make a new mold every time you cast it, and you'd have a hell of a time maintaining the correct wall thicknesses.

[TrevorMonty]

One of the big costs savings with 3D printing is reduced parts count with associated reduction in labour costs for final assembly. NASA stated a reduction in parts on 3D printed engine in the order x10-100. There is also the theory that less parts less to go wrong.

[Scylla]

I'd like to point out that, over the long term, it is cheaper to form a mold of something and cast it many times than it is to additively print it every single time. It also takes a LOT of time to print something, and you can only print one thing per printer at a time with the printers themselves costing hundreds of thousands of dollars.

This will be a boon to their development cycles and will makes it much cheaper and faster to modify and test something, and it will allow for more creative part shapes. However, over the long run I think any parts they think won't change for hundreds of flights will be casted.

Some parts are difficult to cast, consider how difficult it would be to make a mold for a superdraco combustion chamber with regenerative cooling passages. You'd have to make a new mold every time you cast it, and you'd have a hell of a time maintaining the correct wall thicknesses.

From the SpaceX article..

Compared with a traditionally cast part, a printed valve body has superior strength, ductility, and fracture resistance, with a lower variability in materials properties. The MOV body was printed in less than two days, compared with a typical castings cycle measured in months.

[mmeijeri]

I had been wondering about that. Don't they mean that a new test valve can now be produced two days after a design change rather than months later?

[rpapo]

I had been wondering about that. Don't they mean that a new test valve can now be produced two days after a design change rather than months later?

Not to be too blunt about it, but it's almost as easy as hitting the <PRINT> button.  Except you had better be very sure you've run your spell-checker, dotted all your 'i's and crossed all your 't's, and all the three-dimensional equivalents, because that print operation will take a long time and cost a lot more than printing something on paper.

[Prober]

Interesting use of engine out capability as a somewhat low-risk way of trying out slight variations on a single M1D

I assume the quicker time means less development costs; also cheaper to manufacture? How long do 3D printers last, do they lose accuracy/calibration over time?

all equipment is subject to wear.

3D Printer thread http://forum.nasaspaceflight.com/index.php?topic=33141.0

Edit: add link

[TrevorMonty]

Here is one scenario to show another advantage to 3d printing. Assume they have an engine failure traced to cast part. To create a new mold and cast replacement parts could take weeks, delaying all next launches. With 3D printing they could produce the new part in days for next F9.

[Scylla]

I had been wondering about that. Don't they mean that a new test valve can now be produced two days after a design change rather than months later?

Not to be too blunt about it, but it's almost as easy as hitting the <PRINT> button.  Except you had better be very sure you've run your spell-checker, dotted all your 'i's and crossed all your 't's, and all the three-dimensional equivalents, because that print operation will take a long time and cost a lot more than printing something on paper.

Wouldn't printing it first as cheap plastic prototype first be a good, fast way, to check all that?

[Jet Black]

I had been wondering about that. Don't they mean that a new test valve can now be produced two days after a design change rather than months later?

yes, but a new design will still have to undergo a lot of testing. Construction bottlenecks are always going to be the bit that it takes longest to make (unless you can make hundreds in one go) or the bits you can only make one at a time (when you need large numbers of them). I think the main benefit of 3D printing is going to be the ability to make shapes that are difficult and also increased material efficiency. A lot of weight in casted items is just deadweight anyway - take the mirror blanks for the JWST as an example. They made a great big thick blank and then carved the honeycomb out of the back of it. That's not too hard in what is essentially a 2D stucture, but removing weight from complex 3D shapes is nowhere near as easy.

[rpapo]

Wouldn't printing it first as cheap plastic prototype first be a good, fast way, to check all that?

Only for fit.  A 3D printed part, especially the kind we're talking about at SpaceX, is more than just a shape.

[RubberToe]

I would love to see a time lapse video of that valve being printed...

RT

[Coastal Ron]

I had been wondering about that. Don't they mean that a new test valve can now be produced two days after a design change rather than months later?

The valve component can be 3D printed in two days, but then there are still secondary operations that have to be done to the part such as milling, polishing, drilling/threading, and assembling it with other components into a sub-assembly.

In the picture with the SpaceX article you can see that the assembly they show is not a raw component, but has had some machining done and parts added.  Typically that sub-assembly would be tested and then moved to the next level of assembly, which could be the final product like the Draco engine.  Then the engine has to go thru testing before the entire assembly is approved for use.

Reducing the lead time for a complex component, especially one that is still being tweaked in it's design, dramatically lowers the amount of time to takes to validate designs, and ensures that you are more likely to be flying with the most up to date designs.

[abaddon]

The PR is frustratingly vague about their intentions with regards to 3D printing this part or not as a matter of course for future manufacturing:

The valve’s extensive test program – including a rigorous series of engine firings, component level qualification testing and materials testing – has since qualified the printed MOV body to fly interchangeably with cast parts on all Falcon 9 flights going forward.

This clearly indicates that they can choose to use 3D printed versions of this part moving forward.  But it doesn't say if they will or not.  It's hard to imagine they would spend all of the effort of qualfying the part, and the risk of flying it on a real mission, if they didn't intend to use it, given the resulting part is apparently better.

Compared with a traditionally cast part, a printed valve body has superior strength, ductility, and fracture resistance, with a lower variability in materials properties.