Variable Pitch Propeller Attempt

[Piotrsko]

You going to use the Java program for analysis? He does seem to understand the dynamics of model flight are weird and don't conform to full scale.

 

[Inq]

You going to use the Java program for analysis?

I did use both JavaFoil and JavaProp and optimized for the high speed realm. The twist isn't that good for static and low speed, but the variable pitch will allow it to (theoretically) provide more static thrust than just about anything except a 12x2 fixed pitch prop... because basically it'll be a 12x1, 12x2... 12x12+ prop.

 

[Piotrsko]

The twist isn't that good for static and low speed

Same problems everyone has, solve one issue raise a new issue. Full size CS props aren't the panacea either but like you said, pitch is changeable. It would be interesting to make a fixed one and compare to the CS

 

[Inq]

OK... lessons learned or figured out after hitting the Print button.

The Good

1. Running the filament in the X direction with the primary loads and being pretty much solid, these are WAY strong. I just did these in ABS and was really expecting this to just be a dry run for fit and verifying the mechanics. I'd expected to go to Nylon for bench and flite testing. If a single blade can take only 1 pound of load, the propeller would be twice as efficient as a store bought version. I doubt that would happen, but I'm sure 1 pound will be no issue for this blade in ABS. I'm pretty sure I could use a hammer on this and it still be usable. Since this set are sacrificial for other reasons, I'll do some destructive testing.
2. ABS blades will be very easy to post process... fuse, fill, sand, balance will be far easier with ABS than Nylon, PC... etc.
   1. Top - coming off the printer.
   2. Middle - fused only
   3. Bottom - fused, sanded to take off the brim and nits and knock off the steps particularly on the leading edge. One coat of ABS/Acetone wash fill and sanded at 120 grit. I figure it might take one or two more coats and sanding to get down to 400 grit and then painting. Should look near as good as any injection molded prop.
3. The shape is VERY nice. I think it came out better than even the Vase mode prints.

The Bad

1. You'll note I added a sacrificial shaft at the base so I can put it in a drill and turn to polish the bearing surfaces. But I also stupidly added the control arm which defeats the turning ability. Dah!
2. The control arms seems plenty strong enough, but I don't like relying on inter-layer strength. I'll be printing the arm separately and in the printer bed plane to get better strength for the push rods.
3. I used simple bevels for the bearing surfaces which will make it harder to polish. I'll re-do the joint to us fillets that match my modeling files. Make it easier to post process.
4. I've got my 3D printer well calibrated and it will do a near perfect circle in the bed plane. However, when the circle has one axis in the horizontal direction and one in the vertical, the different CTE values cause the circle to be ovoid. The horizontal is 96% the size of the vertical. I'll need to compensate for a second printing.

Onward and upward!

 

[Inq]

Same problems everyone has, solve one issue raise a new issue. Full size CS props aren't the panacea either but like you said, pitch is changeable. It would be interesting to make a fixed one and compare to the CS

I was kind of thinking more about getting a top quality prop as the baseline. Can you recommend a prop?

My motor is a D3536, 1000KV. It's suppose to max out at 435 watts. The Amazon add says a 10" prop is recommended, but doesn't say anything about pitch. The 3DLabPrint plane this will go in is the TA 152H and they recommend a 10 x 5.5. What is a best brand of prop? Hopefully not one that will cost more than my motor and DIY propeller?

 

[Piotrsko]

PROPS are like haircuts: everybody thinks theirs is the best. Base test prop should be whatever you can easily find in 3 blade wooden since you won't find a planform like yours, similar base pitch is going to be hard enough. Bloody pricey, too. second choice is nylon but those do odd things loaded and I am not a fan of side viewing running props.

435 watts should be max constant.. start up and throttle changes will exceed that momentarily. IR optical heat sensor to check you're not exceeding say 150f assuming the varnish in the motor is rated for 180f.

 

[Inq]

PROPS are like haircuts: everybody thinks theirs is the best. Base test prop should be whatever you can easily find in 3 blade wooden since you won't find a planform like yours, similar base pitch is going to be hard enough. Bloody pricey, too.

I would have thought the "best" performing would be a two blade prop. I'd want to compare to the best performing, not to something that is kind-of close cosmetically.

choice is nylon but those do odd things loaded and I am not a fan of side viewing running props.

"Just say no!" - Nancy Reagan

 

[FlyerInStyle]

View attachment 233707


OK... lessons learned or figured out after hitting the Print button.

The Good

1. Running the filament in the X direction with the primary loads and being pretty much solid, these are WAY strong. I just did these in ABS and was really expecting this to just be a dry run for fit and verifying the mechanics. I'd expected to go to Nylon for bench and flite testing. If a single blade can take only 1 pound of load, the propeller would be twice as efficient as a store bought version. I doubt that would happen, but I'm sure 1 pound will be no issue for this blade in ABS. I'm pretty sure I could use a hammer on this and it still be usable. Since this set are sacrificial for other reasons, I'll do some destructive testing.
2. ABS blades will be very easy to post process... fuse, fill, sand, balance will be far easier with ABS than Nylon, PC... etc.
   1. Top - coming off the printer.
   2. Middle - fused only
   3. Bottom - fused, sanded to take off the brim and nits and knock off the steps particularly on the leading edge. One coat of ABS/Acetone wash fill and sanded at 120 grit. I figure it might take one or two more coats and sanding to get down to 400 grit and then painting. Should look near as good as any injection molded prop.
3. The shape is VERY nice. I think it came out better than even the Vase mode prints.

View attachment 233708

The Bad

1. You'll note I added a sacrificial shaft at the base so I can put it in a drill and turn to polish the bearing surfaces. But I also stupidly added the control arm which defeats the turning ability. Dah!
2. The control arms seems plenty strong enough, but I don't like relying on inter-layer strength. I'll be printing the arm separately and in the printer bed plane to get better strength for the push rods.
3. I used simple bevels for the bearing surfaces which will make it harder to polish. I'll re-do the joint to us fillets that match my modeling files. Make it easier to post process.
4. I've got my 3D printer well calibrated and it will do a near perfect circle in the bed plane. However, when the circle has one axis in the horizontal direction and one in the vertical, the different CTE values cause the circle to be ovoid. The horizontal is 96% the size of the vertical. I'll need to compensate for a second printing.

Onward and upward!

just interested, why do use cura and not prusaslicer if you have a prusa printer?

 

[Piotrsko]

Sorry I interpreted this as a like for like comparison. Jury is out full scale as to whether a 2 blade is more efficient although theoretically it should be, and a 1 blade being the most efficient

 

[Inq]

just interested, why do use cura and not prusaslicer if you have a prusa printer?

Good question! When I got it, Cura seemed less flaky and I finally got the Cura settings for my printer nailed down, it runs like a top. It was my understanding that PrusaSlicer started out from Github branch of Cura. Without researching it, I'd think the Cura branch would get updated more often. Version 5 of Cura was a fantastic update and noticeably improved my prints. But, I might need to evaluate that again sometime.

 

[Inq]

Sorry I interpreted this as a like for like comparison. Jury is out full scale as to whether a 2 blade is more efficient although theoretically it should be, and a 1 blade being the most efficient

I'm still holding out optimistic hope that this prop will deliver better performance across the board than the best alternative for the same plane/motor combination and the cosmetics will be spot on. Win-Win.

 

[Scotto]

I was kind of thinking more about getting a top quality prop as the baseline. Can you recommend a prop?

My motor is a D3536, 1000KV. It's suppose to max out at 435 watts. The Amazon add says a 10" prop is recommended, but doesn't say anything about pitch. The 3DLabPrint plane this will go in is the TA 152H and they recommend a 10 x 5.5. What is a best brand of prop? Hopefully not one that will cost more than my motor and DIY propeller?

I like APC props. They are affordable, tough, make good thrust, and best of all made in the U.S.A.!
https://www.amazon.com/dp/B000W0LBRO/?tag=lstir-20 I am not being paid for this endorsement. Lol

Also, for your cheap motor this 750kv says the recommended prop is 13x8.
https://www.amazon.com/dp/B084QCLTM1/?tag=lstir-20 Just be sure to sand the shaft with something course and crank down that collet or you will be landing dead stick.
This is another cool project. I like to see you guy's big brains working.

 

[Inq]

435 watts should be max constant.. start up and throttle changes will exceed that momentarily. IR optical heat sensor to check you're not exceeding say 150f assuming the varnish in the motor is rated for 180f.

I just re-read this and its the first time I've read/thought/heard about any criteria for motor temperature. Definitely goes in the notebook! Thanks.

 

[Inq]

I like APC props. They are affordable, tough, make good thrust, and best of all made in the U.S.A.!
https://www.amazon.com/dp/B000W0LBRO/?tag=lstir-20 I am not being paid for this endorsement. Lol

Also, for your cheap motor this 750kv says the recommended prop is 13x8.
https://www.amazon.com/dp/B084QCLTM1/?tag=lstir-20 Just be sure to sand the shaft with something course and crank down that collet or you will be landing dead stick.
This is another cool project. I like to see you guy's big brains working.

The 750kv one would have been the smart one to get... unfortunately, I got the 1000kv and it says 10". I guess the big brains were out of order that day.

 

[Inq]

I was really excited to see @CrshNBrn large Spitfire. https://forum.flitetest.com/index.php?threads/what-did-you-fly-today.57878/page-539#post-739326 It kind of re-ignited my desire for a large airplane. I was also stumbling around the problems of trying to get all the mechanism inside of a 50mm spinner. As soon as I upped the ante by going to a 1/6th scale Focke Wulf Ta152c with its 600mm propeller and 100mm spinner... things got a lot easier. Besides a Spitfire needs a worthy adversary. I like Spitfires better, but the thought of landing on that narrow undercarriage gives me the cold sweats.

So, I redesigned the joint and have (at least) a plan for the mechanism and 100mm gives me plenty of room. Here is the trailing half of one blade fresh off the printer.

 

[Inq]

The blade's foils and twist have been aerodynamically optimized for 150 mph. As mentioned earlier, the variable pitch should take care of takeoff and low speed flying even though sections of the blades won't be pulling their weight. It will be interesting to see how all this theoretical plays out.

Printing... my first series of blades are really just proof of concept for the variable pitch mechanism. I've actually made them hollow with sparse 3D printing Infill. Just for grins, I continued the spreadsheet to analyze for strength. I've used the weakest material properties I can find and am only using yield strength. That way I can take advantage of ABS large difference between ultimate and yield as an indicator... to back off as they start plastically deforming forward.

Even though these are relatively weak, stress calculations say I can drive these up to 3600 rpm before reaching yield. I may have to run these to destruction for validation. Probably outside would be a good idea.

 

[Tench745]

The blade's foils and twist have been aerodynamically optimized for 150 mph. As mentioned earlier, the variable pitch should take care of takeoff and low speed flying even though sections of the blades won't be pulling their weight. It will be interesting to see how all this theoretical plays out.

Printing... my first series of blades are really just proof of concept for the variable pitch mechanism. I've actually made them hollow with sparse 3D printing Infill. Just for grins, I continued the spreadsheet to analyze for strength. I've used the weakest material properties I can find and am only using yield strength. That way I can take advantage of ABS large difference between ultimate and yield as an indicator... to back off as they start plastically deforming forward.

Even though these are relatively weak, stress calculations say I can drive these up to 3600 rpm before reaching yield. I may have to run these to destruction for validation. Probably outside would be a good idea.

3600 RPM doesn't sound like very much for an electric powered model. A 23x12 prop turning 3600 rpm would have a pitch speed of about 40mph.

 

[Inq]

3600 RPM doesn't sound like very much for an electric powered model. A 23x12 prop turning 3600 rpm would have a pitch speed of about 40mph.

You did note that I said this was really a proof of concept stage and not structurally sound for real use? I was merely saying that even though it is not meant to be driven, it should be able to handle at least 3600 rpm.

I'm a little thin on the aerodynamic side, but from what I'm reading of recommendations to keep tip speed below about Mach 0.88 equates to 6800 rpm. I'm designing the mechanisms to handle flat (no thrust) to fully feathered.

 

[Inq]

27 print hours later...

The raw half-blades are done. Fortunately, I didn't put in any hot-sweaty work or paper cuts.

They came out with variations of only 0.1 grams. Even that will be eliminated as I fill and sand to smoothness. As mentioned, these are only proof of mechanism concept, but here is what I've learned and do differently in blades that will fly.

1. The weight is way high. Total weight before post processing is 177.1 grams. I should be able to re-model the blades to vary the thickness of the print along the length of the blade. At the narrow root, it needs to be near solid as that is the highest stress area. Moving out toward the tip it needs a lot less material. Lightening the outer area will also reduce the stresses caused by centrifugal force for the root area.
2. I originally modeled the blades assuming only the 300 mm diameter and I can see the STL facets. You can see the facets easily in the picture above out toward the tips. They'll probably sand out just fine, but... the anal retentive in me says, I can do better.

I'll go ahead and build the mechanism and spin it up conservatively (and maybe to destruction) to validate the concept and see what else I can learn. Either way, I suspect I'll redo the CAD for the blades. I want to increase the cross sectional area in the root to make that stronger. So far it is my belief that even ABS is strong enough for a production version, but I have not put all my eggs into that basket yet. There is plenty of room to put in a fiberglass or carbon fiber tow spar or switching to Nylon. Modeling the 7 different foils along the span and honoring the scale Focke Wulf shape was very tedious. I don't look forward to repeating the exercise.

 

[Inq]

ABS sucks!

But I'm used to it and have many work arounds. I thought it would be good for me to show a bad side of 3D printing with ABS. That's the thing with engineers... they're more likely to tell you their baby is ugly than anyone else. Salesman, always tell you how sunny and rosy and how it will be a "game changer" (I hate that phrase with a passion) in your life. Engineers will always say... "Give me another month and I can make it better."

Anyhow... ABS has a very high coefficient of thermal expansion (CTE). Without all the fancy jargon... that means, it shrinks a lot with temperature. Worse off... it also uses a higher temperature than our more common 3D printing material - PLA. I like to print ABS at 255°C, while typical temperatures for PLA are 180 to 210°C. IOW... it has a higher shrink rate AND it has to shrink further back to room temperature than PLA.

Then there is the nature of 3D printing... As the layers build and you get a couple of inches off the bed, layers are cooling and shrinking. The new layer is extruding to the specified geometry at the extruder's temperature, yet the layers below have already shrunk.

All this causes warping of your model.

An example is good. The CTE of ABS varies but let's use the worst value... 108E-6 mm/(mm-C). Basically the Math is pretty trivial... we use this constant with the size of the part and change in temperature. For these blades they're about 300 mm long. The temperature of the ABS at extrusion is 255 °C. At room temperature they're at 20°C. This gives...

(255 - 20) * 108E-6 * 300 = 7.614 mm

What this means is... that 300 mm part has shrunk 7.6 mm... It is now 292 mm! There are other things involved, but that is a fair approximation. What that means to us doing 3D Printing... The bottom shrunk, the top didn't till it came off the printer. This causes the infamous warping people complain about.

In this blade project we have two parts that are supposed to mate... "We put them on a flat plate, Why don't the mate up?"... WARPING.

As such if the base of the two halves of the blade are mated, the tip looks like...

I'm not beyond using brute force to make them behave... As such, I've created a couple of jigs to ensure consistency of alignment and let me use force (C-clamps) to bring the two halves back together into one nice blade.

I'll start that argument now...