Have you thought about contacting Bambu Lab for a sponsorship? You seem like a great candidate for a sponsorship. The worst they can say is no. Keep challenging yourself and experimenting.
I would use 2 separate motors. gears especially 3d printed ones have very bad efficiency and if you insist with your gearbox I'd put some grease in it because the lube you put is wayyy to thin and gets sucked in your print imperfections
true i think 2 would be best.. specially cus ducted fans tend to work best at super high rpms and 3d printed gears just arent that durable or efficient like you mentioned
A couple of ideas for you... first, if you drive the fans from their outer circumference you will eliminate the need for axles entirely. Just put your gearing and bearings on the outside of the fan's body into the rim. Second is a bit more complicated to engineer, but if you build your motor into the fan's rim, then you would eliminate gearing completely and solve your motor-off-axis torque issue. No gears, no axles, just a ring motor with a fan inside of it. Install two of these with the fans in opposing directions and you have your counter-rotating ducted fan system. Good luck!
Wow I've actually never heard of a ring motor. That's fascinating. I believe some of the new electric powered jet fans use the rim driving concept pretty heavily? Much thanks for the advice
You can make even more durable gears by just printing with normal PLA and gear teeth with about half a millimeter extra slack, and then once printed coating the gear touching surfaces in several thin brushed-on and off layers of epoxy resin until the printing layers dissapear. It's way harder and more abrasion-resistant then PLA and ABS. If you mix in carbon powder (the lubricant) with the resin until its properly black and no longer transparent you can keep an even color, get even better hardness properties AND self-lubrication as it very very slowly wears down. :) I make clock mechanisms this way with PLA printers, although the gears have to be even finer so i hand-file them to shape before painting resin on them.
Ah that sounds like a really solid approach. I've never used epoxy resin but it sounds like something I should be getting into anyway. Thanks for the tips! Might try that after this nylon experiment.
@ Jon: you can connect the Motor directly to one fan instead of rotating it 90°. this way you can use a way thinner support structure for the dif (which still needs to be fixed)
Allso instead of a diff-style conversion of turning direction you could go into a planetery gear setup- this is harder to print but will use up way less space for the same amount of torque transferred.
Heck yeah! That PCBway offer is going to be a huge boon to making parts! I think I've said it on other videos, but I gotta shoutout another person that sees the benefit of printing on glass, lol. Clean glass and a little hairspray have been the best for my CR10-S for like 8 years. Works every time. Although, I will say PETG is the only material that can still be a little tricky. Also, stringy and sticky. So... I wonder if flexible shafts might work for something like this? Small cross section, and could allow you to place the motor(s) basically anywhere you want. 🤔
For sure! Thanks the tips, yeah it sounds like I just have the baby the glass a little more than I would for PETG (still haven’t given into PEI yet, whatever). As for flexible shafts…will definitely be looking into that. Never crossed my mind before. Appreciate the support
How about putting the gears on the outside of the fan rims and the driving gear outside the air tunnel all together? Dump the axle and experiment with an open center. Let the rims ride on ball or roller bearings, that can also couple them to the frame/main tunnel. :o)
I’ve experimented a lot… well really A LOT of different configurations with intake geometries, thust tube dia, lengths, fan blade count, dealing with motor heat. The easiest to explain and possibly the most detrimental issue is the Fan blade itself & no not the blade themselves… the hub of the fan has holes that are uncovered. The holes are there to lighten the weight but have to be blocked, usually a spinner. With the uncovered holes any air pushed behind the fan blades will take the path of least resistance with is the area of lower pressure in front of the fan. Air is going in the front and most but not all is going out the front. There more design ideas out there that don’t work than there are that make the biggest improvements but if this message finds you and you found beneficial let me know. I achieved 4.5lb of thrust from my 70mm EDF. Custom woven CF housing- 4 45mm long stators supporting a the motor mount that’s 3.5mm walled aluminum that has an outer diameter of 35mm, 29mm inner dia, & a length of 50mm. 12 blade fan, 3600kv motor, 100A Esc running on 5s battery. The amount of torque and heat the EDF meant strong support and large heat sink to draw the heat out of the motor so it could be run at full power for 4.5-5mins without degrading the strength of the magnets from being overheated. The Esc is mounted behind the motor and cooled by the air flowing out of the edf, so much air in fact that the 100A ESC doesn’t have a heat sink. Total weight of housing, motor, Esc and its wiring landed at 12oz,(without battery) with 4.5lb of push. The Intakes airflow is far more important than the exhaust.
This is gold! I played a lot with 3D printed EDFs a few months ago. I had a much more dialed in fan setup, the one used in this video was used for not much more than demonstration. Blocking the lightning holes is 100% something I'll do though, that's a very quick fix. 4.5 lbs is insane, I'd love to see videos!
If you mounted the fans inside a couple large bearings and had a flange that meshed with a perpendicular gear you could drive both of them off of a single motor that wasn't in line with the air channel and you wouldn't have to deal with the shaft at all.
You could mount the motor in the front and have the shaft for the rear fan go all the way through to it and for the counter rotating fan you could do a ring at the motor to a couple pinion gears on each side going back to the outer tubeshaft with its own ring gear. You'd streamline the gear box behind the bulk of the motor and can taper it. It'd be a long assembly but you'd wouldn't have that gear box in the way. Good job dude 👍🏻
Yeah a couple comments have mentioned something like this setup and it sounds super cool. Never would have thought of that myself. Thanks for the advice man I'll look into it
Tip: use 2 contra-rotating motors and 2 esc's. Simplicity, reliability and no gear losses. By the way, in that case you almost compensate gyroscopic moments.
@@mach10point4Actually it is difficult to compensate gyro moment for single motor, because two contra rotating fans completely compensate each other. Thus either extra motor and double power or some flywheel.
Interesting, maybe you can put the gearbox at the fans level and then keep the motor axial, to solve the motor torque, you can counter it with one of the fan by making it heavier
Maybe a dumb question, maybe just something you didn't think of, but wouldn't it make more sense to mount the engine inline on a shaft, have 1 rotating assembly fixed to the shaft, and on the opposite side of the motor on that rotating assembly have your beveled gears, then a mirrored rotating assembly fit on a bearing on the same shaft?
Tip: You can use a stroba scope to slow down fast movement for filming. Also if you're using PCB way just get some gears milled in aluminum that you can reuse.
Cool stuff! Do you know if this twin EDF/contrarotating setup still creates those gyroscopic precession forces that plauge monocoptors? It seems like it would help but would still give "weird" reactions to external forces.
Thanks! I figured that with this setup, the only reaction would be a (pitch?) moment that could be balanced by a gimbal or thrust vane set up much more cheaply. Would love yo know what you think though, maybe that's a bit naive
@@mach10point4 I'm not sure, I'm equally naive 😂 About to begin PID tuning on my monocoptor tonight, but I'm only using one 50mm unit. *edit* I've got a rough Matlab/simulink model of my monocoptor if you want me to email you a copy to mess with
personally from my very limited experiance with edfs so take my thoughts with a grain of salt( ive only been researching for abt 2 months) try to create an intake which deflects the edfs airsource away from the gearbox with some external mounting points for it
Love the concept. I think your gearbox over complicates it have you thought of using two Motors?and using the opposite rotation of the talk to act as your rudder control.
Thank you! I’m worried the extra motor (which includes extra ESC and potentially extra battery) would make something like this too heavy to be viable. Plus this has aero benefits, namely low wetted area of obstruction without the motor in the duct. But I’m not opposed to 2 motors, and you’re right, the gearbox is super tricky at best
@@mach10point4 maybe a non-linear thrust tube with two smaller motors. A non-linear approach to your thrust you would be a good idea then to give you the venturi effect around the obstacles to tune thrust tube?
would love to see some data on torque reactions on the motor itself... with a single prop, the reaction torque attempts to spin the motor the other way. to drive a contra rotating prop through coaxial gears means that reversed prop has to spin at twice the speed, backwards. the torque to drive that reverse prop is reflecting off the standard prop, and all that reaction tends to reflect back to the motor itself. and thats only held stationary due to aerodynamic surfaces in flight... most pilots never liked contra rotating props as they handled in an utterly different way... and thats not entirely due to air simply flowing "straighter" over the plane. here though, youve rotated that reaction by 90 degrees... giving either a pitch or yaw effect, rather than a roll... iunno, stick it on a reaction arm and load cell... be nice to see some figures quantified. maybe with 3dof to get a real grasp of the basics of newton...
Yeah I’ve been working on 3 DOF gimbal for the monocopter stuff, so I’d love to put it on there. I guess my justification for this was a monocopter has a much harder time controlling z axis rotation rather than x or y, just in my experience with arduino. But what you’re saying makes sense and has me thinking hard now
Graphite lubricant isn't ideal for use with electric motors, as it is conductive. Don't ask me how I know that actually can cause problems. With those fan hub diameters and with cones on each end you have all the space you need for a gear box, or two motors, inside the fans and fan cones. Do take into account that using two motors you only need half as powerful motors, in theory, in practice even slightly less, as you won't have gear losses. Also, two smaller motors means they counter each others rotational impact out. From the aerodynamic perspective, inrunners can be ridiculously powerful for their diameter, and better at producing high RPMs. Also, flow area in the center isn't as important as at the edges. Gearing one motor to two contra rotating fans makes more sense for internal combustion engines, where having two smaller engines has a lot of disadvantages compared to one bigger engine. I personally would find it more interesting if you made multiple stages, like the compressor part of a jet engine, but with contra-rotating fans instead of fans and stators. I suspect using at least two parallell contra-rotating drive shafts would be essentially necessary, but those and the gears could fit well in the fan hubs. About PETg on ender 3, I found that the end of the bowden tube decomposed (very slowly) and shrunk slightly over time from the heat with the original hotend, making it necessary to shop off the end of the bowden tube occasionally, or upgrade to all metal hotend.
Fantastic, thanks for the response. I think I will ultimately have to go with a 2 motor setup. We’ll see how nylon goes. But I agree with everything As for the PETG, I’ve been printing at 235 trying not to burn the Bowdin tube. My CR 10 has a metal hotend so just trying to squeeze anything out of the Ender 3.
nice project
Have you thought about contacting Bambu Lab for a sponsorship? You seem like a great candidate for a sponsorship. The worst they can say is no.
Keep challenging yourself and experimenting.
Excellent point. I’ll shoot my shot. Man if I can get an X1..
@@mach10point4 X1 is great, I use it for all my turbines.
@@warmachined3472 Yeah, I've gotten to use it a good amount and it makes my Enders feel like toys
What a cool design
💯💯
I would use 2 separate motors. gears especially 3d printed ones have very bad efficiency and if you insist with your gearbox I'd put some grease in it because the lube you put is wayyy to thin and gets sucked in your print imperfections
Yeah the FDM gears are mostly for proof of concept purposes. I’m trying to see if nylon will work better
Bad advice
true i think 2 would be best.. specially cus ducted fans tend to work best at super high rpms and 3d printed gears just arent that durable or efficient like you mentioned
Outstanding project 👏🏼!!! You may just want to go to metal gears. In the long run any other material will fail very quickly.
A couple of ideas for you... first, if you drive the fans from their outer circumference you will eliminate the need for axles entirely. Just put your gearing and bearings on the outside of the fan's body into the rim. Second is a bit more complicated to engineer, but if you build your motor into the fan's rim, then you would eliminate gearing completely and solve your motor-off-axis torque issue. No gears, no axles, just a ring motor with a fan inside of it. Install two of these with the fans in opposing directions and you have your counter-rotating ducted fan system.
Good luck!
Wow I've actually never heard of a ring motor. That's fascinating. I believe some of the new electric powered jet fans use the rim driving concept pretty heavily? Much thanks for the advice
You can make even more durable gears by just printing with normal PLA and gear teeth with about half a millimeter extra slack, and then once printed coating the gear touching surfaces in several thin brushed-on and off layers of epoxy resin until the printing layers dissapear. It's way harder and more abrasion-resistant then PLA and ABS. If you mix in carbon powder (the lubricant) with the resin until its properly black and no longer transparent you can keep an even color, get even better hardness properties AND self-lubrication as it very very slowly wears down. :) I make clock mechanisms this way with PLA printers, although the gears have to be even finer so i hand-file them to shape before painting resin on them.
Ah that sounds like a really solid approach. I've never used epoxy resin but it sounds like something I should be getting into anyway. Thanks for the tips! Might try that after this nylon experiment.
I've seen a couple variations of this setup.
The second prop should turn about 40% faster (RPMx1.414) for equal loading on the same profile.
@ Jon: you can connect the Motor directly to one fan instead of rotating it 90°. this way you can use a way thinner support structure for the dif (which still needs to be fixed)
Allso instead of a diff-style conversion of turning direction you could go into a planetery gear setup- this is harder to print but will use up way less space for the same amount of torque transferred.
Heck yeah! That PCBway offer is going to be a huge boon to making parts! I think I've said it on other videos, but I gotta shoutout another person that sees the benefit of printing on glass, lol. Clean glass and a little hairspray have been the best for my CR10-S for like 8 years. Works every time. Although, I will say PETG is the only material that can still be a little tricky. Also, stringy and sticky.
So... I wonder if flexible shafts might work for something like this? Small cross section, and could allow you to place the motor(s) basically anywhere you want. 🤔
For sure! Thanks the tips, yeah it sounds like I just have the baby the glass a little more than I would for PETG (still haven’t given into PEI yet, whatever). As for flexible shafts…will definitely be looking into that. Never crossed my mind before. Appreciate the support
How about putting the gears on the outside of the fan rims and the driving gear outside the air tunnel all together? Dump the axle and experiment with an open center. Let the rims ride on ball or roller bearings, that can also couple them to the frame/main tunnel. :o)
I’ve experimented a lot… well really A LOT of different configurations with intake geometries, thust tube dia, lengths, fan blade count, dealing with motor heat.
The easiest to explain and possibly the most detrimental issue is the Fan blade itself & no not the blade themselves… the hub of the fan has holes that are uncovered. The holes are there to lighten the weight but have to be blocked, usually a spinner. With the uncovered holes any air pushed behind the fan blades will take the path of least resistance with is the area of lower pressure in front of the fan. Air is going in the front and most but not all is going out the front. There more design ideas out there that don’t work than there are that make the biggest improvements but if this message finds you and you found beneficial let me know.
I achieved 4.5lb of thrust from my 70mm EDF. Custom woven CF housing- 4 45mm long stators supporting a the motor mount that’s 3.5mm walled aluminum that has an outer diameter of 35mm, 29mm inner dia, & a length of 50mm.
12 blade fan, 3600kv motor, 100A Esc running on 5s battery. The amount of torque and heat the EDF meant strong support and large heat sink to draw the heat out of the motor so it could be run at full power for 4.5-5mins without degrading the strength of the magnets from being overheated. The Esc is mounted behind the motor and cooled by the air flowing out of the edf, so much air in fact that the 100A ESC doesn’t have a heat sink.
Total weight of housing, motor, Esc and its wiring landed at 12oz,(without battery) with 4.5lb of push.
The Intakes airflow is far more important than the exhaust.
This is gold! I played a lot with 3D printed EDFs a few months ago. I had a much more dialed in fan setup, the one used in this video was used for not much more than demonstration. Blocking the lightning holes is 100% something I'll do though, that's a very quick fix. 4.5 lbs is insane, I'd love to see videos!
If you mounted the fans inside a couple large bearings and had a flange that meshed with a perpendicular gear you could drive both of them off of a single motor that wasn't in line with the air channel and you wouldn't have to deal with the shaft at all.
You could mount the motor in the front and have the shaft for the rear fan go all the way through to it and for the counter rotating fan you could do a ring at the motor to a couple pinion gears on each side going back to the outer tubeshaft with its own ring gear. You'd streamline the gear box behind the bulk of the motor and can taper it. It'd be a long assembly but you'd wouldn't have that gear box in the way.
Good job dude 👍🏻
Yeah a couple comments have mentioned something like this setup and it sounds super cool. Never would have thought of that myself. Thanks for the advice man I'll look into it
Tip: use 2 contra-rotating motors and 2 esc's. Simplicity, reliability and no gear losses. By the way, in that case you almost compensate gyroscopic moments.
For sure, I wanted to see if there are significantly lighter approaches with minimal gyro moment penalty
@@mach10point4Actually it is difficult to compensate gyro moment for single motor, because two contra rotating fans completely compensate each other. Thus either extra motor and double power or some flywheel.
Interesting, maybe you can put the gearbox at the fans level and then keep the motor axial, to solve the motor torque, you can counter it with one of the fan by making it heavier
Can’t you move the dealing to the outside of the fans and run them with belts or just a gear on the top of the fan with a loop of groves around it.
Maybe a dumb question, maybe just something you didn't think of, but wouldn't it make more sense to mount the engine inline on a shaft, have 1 rotating assembly fixed to the shaft, and on the opposite side of the motor on that rotating assembly have your beveled gears, then a mirrored rotating assembly fit on a bearing on the same shaft?
fan have to be fiberglass to reach good rpm or create a mold to pour aluminum to create fan and can go 15,000rpm or more
Tip: You can use a stroba scope to slow down fast movement for filming. Also if you're using PCB way just get some gears milled in aluminum that you can reuse.
Ahh brilliant, yes I definitely need to be upping the cinematography game lol
Cool stuff! Do you know if this twin EDF/contrarotating setup still creates those gyroscopic precession forces that plauge monocoptors?
It seems like it would help but would still give "weird" reactions to external forces.
Thanks! I figured that with this setup, the only reaction would be a (pitch?) moment that could be balanced by a gimbal or thrust vane set up much more cheaply. Would love yo know what you think though, maybe that's a bit naive
@@mach10point4 I'm not sure, I'm equally naive 😂
About to begin PID tuning on my monocoptor tonight, but I'm only using one 50mm unit.
*edit* I've got a rough Matlab/simulink model of my monocoptor if you want me to email you a copy to mess with
@@rogerrinkavage that'd be awesome, my email is jonathanyun617@gmail.com! will keep you updated :)
@@mach10point4 sent!
How did you control the motor? Could i use an arduino for this?
very cool fan!
Thank you!
personally from my very limited experiance with edfs so take my thoughts with a grain of salt( ive only been researching for abt 2 months) try to create an intake which deflects the edfs airsource away from the gearbox with some external mounting points for it
Yeah that's a good point, I would want to be shaping the gearbox in a much more aerodynamically efficient way at least
Very interesting
Ise tooth paste to bed the gears in but only run motor to pne minute, then wash out and oil
Excellent
😀
Love the concept. I think your gearbox over complicates it have you thought of using two Motors?and using the opposite rotation of the talk to act as your rudder control.
Thank you! I’m worried the extra motor (which includes extra ESC and potentially extra battery) would make something like this too heavy to be viable. Plus this has aero benefits, namely low wetted area of obstruction without the motor in the duct. But I’m not opposed to 2 motors, and you’re right, the gearbox is super tricky at best
@@mach10point4 maybe a non-linear thrust tube with two smaller motors. A non-linear approach to your thrust you would be a good idea then to give you the venturi effect around the obstacles to tune thrust tube?
As it is a coaxial, it is actually a contra :)
Haha. You’re right. Changing the title now.
would love to see some data on torque reactions on the motor itself...
with a single prop, the reaction torque attempts to spin the motor the other way.
to drive a contra rotating prop through coaxial gears means that reversed prop has to spin at twice the speed, backwards.
the torque to drive that reverse prop is reflecting off the standard prop, and all that reaction tends to reflect back to the motor itself. and thats only held stationary due to aerodynamic surfaces in flight... most pilots never liked contra rotating props as they handled in an utterly different way... and thats not entirely due to air simply flowing "straighter" over the plane.
here though, youve rotated that reaction by 90 degrees... giving either a pitch or yaw effect, rather than a roll...
iunno, stick it on a reaction arm and load cell... be nice to see some figures quantified. maybe with 3dof to get a real grasp of the basics of newton...
Yeah I’ve been working on 3 DOF gimbal for the monocopter stuff, so I’d love to put it on there. I guess my justification for this was a monocopter has a much harder time controlling z axis rotation rather than x or y, just in my experience with arduino. But what you’re saying makes sense and has me thinking hard now
Graphite lubricant isn't ideal for use with electric motors, as it is conductive. Don't ask me how I know that actually can cause problems.
With those fan hub diameters and with cones on each end you have all the space you need for a gear box, or two motors, inside the fans and fan cones. Do take into account that using two motors you only need half as powerful motors, in theory, in practice even slightly less, as you won't have gear losses. Also, two smaller motors means they counter each others rotational impact out.
From the aerodynamic perspective, inrunners can be ridiculously powerful for their diameter, and better at producing high RPMs. Also, flow area in the center isn't as important as at the edges.
Gearing one motor to two contra rotating fans makes more sense for internal combustion engines, where having two smaller engines has a lot of disadvantages compared to one bigger engine. I personally would find it more interesting if you made multiple stages, like the compressor part of a jet engine, but with contra-rotating fans instead of fans and stators. I suspect using at least two parallell contra-rotating drive shafts would be essentially necessary, but those and the gears could fit well in the fan hubs.
About PETg on ender 3, I found that the end of the bowden tube decomposed (very slowly) and shrunk slightly over time from the heat with the original hotend, making it necessary to shop off the end of the bowden tube occasionally, or upgrade to all metal hotend.
Fantastic, thanks for the response. I think I will ultimately have to go with a 2 motor setup. We’ll see how nylon goes. But I agree with everything
As for the PETG, I’ve been printing at 235 trying not to burn the Bowdin tube. My CR 10 has a metal hotend so just trying to squeeze anything out of the Ender 3.