Turbine Engine Mid Frame

Discussion in 'Lost PLA casting' started by Monty, Dec 2, 2021.

  1. Monty

    Monty Silver

    This is my Don Quixote (or maybe Wile E. Coyote) attempt at producing an "affordable" small turbofan for experimental aviation. There are companies out there that could do it, or even NASA. There was a program called GAP (general aviation propulsion) that of course came to naught. I saw the hardware at Oshkosh...but it never made it to market. Nobody wants the liability, and the big manufacturers are happy selling million dollar engines. It annoys me that what passes for engine tech in general aviation is basically 1930's tractor technology that costs $50K per engine.

    So.....This is my personal windmill at the moment:

    XTF1 Cross Section.jpg

    Not the most efficient, not the best theoretically, but designed to be affordable.

    Monty
     
    BattyZ likes this.
  2. Al2O3

    Al2O3 Administrator Staff Member Banner Member

    I had missed this. Think we're on the same wave length there. When I had originally mentioned it before I saw the x-section I meant a stack of circular sections, but I like sectioned through the axis with the separate center much better.

    Making EPS molds would be no trivial task, but I'm sure necessary at any appreciable production quantity. Not sure what to say about lost foam for aerospace castings. At a macro level, my lost foam castings seem every bit as good as any sand castings, but I've never subjected them to real analysis as far as measured porosity or elongation. Not many sand castings make the grade for aerospace standards. Mercury Marine developed (2001) an alloy and a vacuum/pressure casting LF process where the casting is pressurized inside a pressure vessel to 10 ATM immediately after it is poured........there was a 3-4 page write up in an AFS pub I have that indicated LF and sand casting porosity is typically 1% and this process reduced it to .01% in A356 and dramatically improved elongation. I can copy it and send if interested.

    Quite ambitious!

    Best,
    Kelly
     
  3. Monty

    Monty Silver

    I'm definitely interested in the Mercury Marine paper!

    As far as casting quality concerns: Part of my design philosophy for this project has been to avoid "aerospace" practice as much as possible. I have applied as much automotive, and consumer product thinking as I can into the design. "aerospace" thinking is why a 1930's tractor engine winds up costing $50K. Using aerospace techniques I could probably reduce the weight of the midframe by 3/4. Unfortunately the cost would probably go up 100X. Is it worth it to save 7 lbs?? I don't think so. This is more Briggs and Stratton, and less F1/space shuttle.

    Monty
     
    Last edited: Jun 25, 2022
  4. Monty

    Monty Silver

    There are some features of the center section that will cast better than machine. The oil return cutoff feature is quite sharp on the inside radius. This might be hard to machine at full depth. What is the smallest diameter cutter you use on foam in your router, and how much length can I get away with?

    Monty
     
  5. Al2O3

    Al2O3 Administrator Staff Member Banner Member

    I have an .030" dual flute cutter but the cutting length is only .25". I often use .125" EM and Ball EM. Hard to find them with more than 1" CL, but I use extenders that can increase reach but not cutting length. .25" EM & Ball EM is my most often used cutter. Bigger diameters for hogging and gross stock removal. There are (2-3 degree) tapered cutters with .030" nose radius that are popular for ornamental carving and reliefs with 1.5-2" CL. I have a few but rarely use them.

    Best,
    kelly
     
  6. Monty

    Monty Silver

    Well....my last attempt using block lost PLA was not a success. I couldn't keep up with how fast the mold was taking metal and air got sucked into the sprue. Even if that hadn't happened, the burnout was incomplete. Wood glue works better than Elmer's for surfacing because you can actually sand it! But it just doesn't burn out cleanly at 500F.

    I hemmed and hawed about a lot of things. Mostly, should I buy a router. In the end I decided to get over the mess in my 5 axis and use that instead. I have it, and need to learn the software. SOFTWARE!!! is a nightmare for 5 axis. Big $$$$$. Like $10K. I finally decided to go with madCAM. It's a little clunky, but it will do the job. I already have Rhino, and the price is affordable for the educational version. I can even (barely) justify upgrading to the professional version if I need to at some point. It will tide me over until Free Cad hopefully catches up to where I need it to be. Learning all this is so painful, I even considered buying a resin printer for making molds for EPS. This would actually work, as some of the engineering resins would have no problem with the temps involved. In the end I decided I can only handle so many learning curves at one time, and I need to get the 5 axis up and running anyway. Might as well kill multiple issues at once. If I get to the production stage of things, I can always machine aluminum molds for EPS, or just contract that part out.

    So......I just received a $200 shipment of 4 in thick foam from aircraft spruce. Nobody around here was able to supply it. The foam wasn't too bad....but the shipping is steep-about $70!. I'm going to keep looking for a local supplier, but supply chain issues mean I could only get 2in locally for now.

    I have a shiny new foam cutting bit from these folks. Might be of interest to you Kelly. It's a .25 in ball end cutter with a 4 in depth of cut. I'll let you know how it works.

    I've ordered some turbine vibrators, and I'm searching for a good flask. Maybe I should switch the rest of this thread over to the lost foam thread...

    Monty
     
  7. BattyZ

    BattyZ Silver Banner Member

    I am in a similar boat. If you wouldn't mind, I think some of us (maybe just me!) would be interested in seeing photos and hearing the backstory of this machine on a new thread in the CNC section.

    Autodesk's Manufacturing suite is $3k a year and includes some true 5-axis toolpaths in the cad/cam. Although I think most of us just want 3+2 capabilities?

    Once I am comfortable with my machine's output I will have to look into that aircraft spruce place.

    On the side of the cutting bits, the usual buying spots don't have much for XL CL and reach that is needed. Places like AB Tool will custom grind whatever you want but have your wallet ready!

    https://abtoolsinc.com/

    I had them custom grind some 37-degree taper carbide to machine an external spline in a rotary setup. $100/endmill. Luckily this time was able to push the cost on. Rarely the case for the home gamer but if one is confident in their machine an endmill would go a long way with the occasional bit n blade cleaner.
     
  8. BattyZ

    BattyZ Silver Banner Member

    Oh and Monty, great work thus far! Making good progress and fun to watch.
     
  9. Al2O3

    Al2O3 Administrator Staff Member Banner Member

    At 2lbs/ft3, I considered ACS' foam to be kind of on the ragged edge of acceptable density and would rather have < 1.5lb/ft3, but may be ok and at least available in less than a truck load. It'll likely have better strength and machinability at the expense of potentially higher flaw population and casting limitations, but admittedly, other than production lost foam practice for pattern densities, there's really not any available info in this regard as far as limiting pattern density.

    I didn't see any 4" cutting length 1/4"D ball end mills......4" overall length yes. I do have a few Amana bits and I like them.......4" OAL, 1/4D, 2.25 cutting length ball and flat, up cut.

    As previously mentioned, for prototypes, LF will be lower material cost and less overall labor compared to block investment. You'll have some other process stuff to work out but happy to help where and if I can. Time will tell as far as casting integrity.

    Best,
    Kelly
     
  10. Monty

    Monty Silver

    Check this bad boy out.. :D
     
  11. Monty

    Monty Silver

    I haven't been able to find 4 in 1.5lb foam anywhere. If nothing else, I can at least proof out my code and process with the ACS stuff until I find some lower density thicker material. Hopefully it will cast OK.
     
  12. Monty

    Monty Silver

    Merry Christmas everyone!

    I'm still plugging away. I've been spinning my wheels on the software front. I've tried a couple different things. Everything works from a 3 axis standpoint. 5 axis is where I'm having problems. Difficulty = y^(#of axis). I've been trying to use HSMworks but have been having post processor and machine simulation issues. Mach3 doesn't support tcp (tool center point control) Which is a problem for a 5 axis machine. It means all of the kinematics have to be in the post processor or CAM and not the control of the machine. I'm getting really tired of farting with it. I may just machine some parts using 3 axis so I can get on with the castings!! The problem is I need to get the 5 axis working for the diffuser and fan hub. I'd rather make my mistakes cutting foam and not aluminum! Anyway, good to see some people are still doing things out there.
     
  13. Al2O3

    Al2O3 Administrator Staff Member Banner Member

    Software learning curves can be steep. The multiaxis stuff you and BattyZ use numbs my head.

    Amen

    Best,
    Kelly
     
  14. Monty

    Monty Silver

    A new process, software and machine.....enough to drive me crazy. It'll be worth it in the end. I don't mind the process struggle, but chasing bugs in the post processor and HSMworks is going to put me in an institution. I don't have any control over it, and have to rely on other people who are slow to respond. The bugs I can't fix and just have to find work arounds. HSM is what Fusion360 was based upon, and I'm beginning to think they are using us to do their product development. The machine simulation simply doesn't work in 5axis. It's really buggy. I'm becoming more and more leary of anything autodesk. Sprutcam works with Alibre, and is about $4K for a lifetime license. I may wind up going that way. I'm tired of spinning my wheels.
     
  15. BattyZ

    BattyZ Silver Banner Member

    Mine too!

    Dynomotion does! Actually one of the principal reasons I went with that controller. Toolpath outputs basically just tell location and orientation. The rest is handled on the controller.

    I hear you on the struggles man. My 5-axis is on the back burner for the moment but yeah the post-processor stuff can be a real headscratcher. You are trying to learn three things at once: the programming language, the way posts handles the CAM kernel, and how you actually want your specific machine to run said code.

    What is holding me up is mechanical speed. Linear axes have to move real quick to keep up with medium-ish rotary axis movement. Something to keep an eye on maybe.

    Keep working at it man!
     

Share This Page