Wall crawling, ceiling traversing, anti-gravity, ducted fan Robot

xanadu

I've seen some robots that use a vacuum to hug a wall. I've been thinking about doing something like this but differently and before I want to order any parts I was hoping for some feedback. I don't really like the vacuum idea because it requires a type of seal that won't allow you to make the 90 degree turn.

The pic below is a wheeled robot that weighs 2lbs. It is going straight up a wall. It pulls air (blue) from under the robot and blasts it out the top using a ducted fan.

My question is, given a 2lb robot, an EDF with > 2lbs of thrust should handle this no problem right?

I'm guessing the stronger the better, but what kind of margin should I be looking for?

Or will this not work at all and you need a vacuum?

xanadu

This is the EDF I was considering - https://www.hobbyking.com/hobbyking/store/__45230__Dr_Mad_Thrust_70mm_10_Blade_Alloy_EDF_2200kv_Motor_1900watt_6s_Counter_Rotating_.html



It's kind of expensive without a second opinion though...

WBA Consulting

The weight to thrust ratio will not be 1:1 because you are dealing with two vectors 90 degrees apart. In other words, it would be true that 2lbs of thrust pointed down would keep the robot at it's current position vertically (ignoring other factors like friction, turbulence, etc). In this case, since you are dealing with an applied thrust in the X direction, but a gravitational pull in the Z direction, the force to hold it to the wall will be less than the 2lb weight of the vehicle. The EDF you linked to should definitely be adequate for the job. I am tempted to suggest some deflection of the EDF airstream in a downward (-Z) orientation as to aid in overcoming the gravitational forces. As long as the robot is held tight to the wall, your drive wheels could keep the robot in a consistent position by applying enough power to overcome the gravitational pull.

W9GFO

Don't know if it would work or not. Ducted fans are not efficient at low speeds, and your speed would be effectively zero. If you try to create an area of low pressure under the robot to suck it up against the wall you risk starving the EDF and the impeller will stall. If you don't, then you are missing out on some free traction. The trick would be to create enough low pressure under the bot to be beneficial, but not so much that the you stall the blades.

dmagnus

You're right, kinda spendy, but over 5 lbs of thrust. It looks from the drawing that you might be just going to use it to provide enough "down force" to hold it on the wall. Heck, they say a formula one or indy car could actually do that at speed because of the aerodynamics. I'm wondering if you couldnt mount 4 smaller EDF's on the corners for less money and, maybe some redundancy...
FWIW, my little Hubsan quad is great for cleaning cobwebs from the ceiling. :cool:
It's amazing the amount of suction it develops when you bring it up close (about 1" away) from the ceiling. It just hops right up there.
You can drive it around with rudder and elevator. It takes a reduction of almost 2/3rds of whatever throttle you had to get it up there in order for it to lose enough lift to fall away.
Of course, this is probably the reverse of ground effect. I think if your ducted fans were producing enough negative pressure on the bottom of your vehicle, it would work.
This is also on a smooth ceiling. I don't think it would work on the rough, spackled ceilings.
I would think that walls with uneven surfaces, like car siding might make the vehicle lose "lift" also, but maybe not.
Interesting idea, I can't wait to see it work.
Don
PS: I've got ceiling tiles in the basement, I'm going to try the Hubsan down there and see if it still works on a porous surface.

Edit: I see a couple of people with more engineering skills than I have jumped in with some good ideas. I still think it would work, though.

Duane Degn

A lot will depend on the friction of your wheels.

If your wheels have enough friction to keep from sliding when on a 45 degree incline then you'll need 1.4 lb (2/(square root of 2)) of thrust to keep your robot against the wall. (2 lb * cos(45)) (I think (but I'm not sure) I did the math right.)

If the vacuum Rich mentions is a problem then you could move the fan away from the wall. I'm not so sure if you'd create enough of a low pressure area under the robot to hold it to the wall. I think you just want the fan to be pushing it against the wall in which case, you may not want the fan to be close to the wall in order to allow good air flow through the fan.

Edit: Don's comment has made me rethink the low pressure idea. Don's right about the ceiling sucking quadcopters and helicopter up towards the ceiling. If a 'copter gets sucked up to the ceiling you really have to cut the power drastically to break free. It would be interesting to see if there's a sideways effect similar to the ceiling effect.

Edit again: You'll need stronger than normal motors on your robot. You'll need your motors to have enough torque to allow the robot to climb straight up the wall.

xanadu

I'm so glad I asked. Everyone has brought up really good points here.

The EDF's position would need to be adjustable so its distance from the surface can be fine tuned. I didn't even think of air starvation. If the distance between the surface and the EDF inlet changes, the forces will need to change as well. Initially I was thinking about just letting an IMU detect the angle to adjust the power, but that may need some additional considerations as well.

Before I order anything I will try to do a smaller version made out of stuff that I already have. I'm going to try to piece something together. I'm waiting for parts on my other project anyway.

The only thing I have now are 4 CR servos for motors, and BOE Bot wheels with some fresh rubber bands. Then I will add an RC motor, ESC and propeller because I don't own any EDFs currently. Although the propeller and EDF are different beasts, I'm hoping it provides some insight.

I also notice that suction effect when you get moving blades next to the ceiling. It is that, and seeing the vacuum based bots that has always made me wonder about it. I would like to rely on thrust more than suction, but that may prove very inefficient.

Edit: I guess to detect the distance from the air inlet to the surface I could use the EDF's current draw in comparison to unrestricted airflow. Combine that with an IMU to detect the angle. I could probably even code that myself!

xanadu

dmagnus wrote: »

You're right, kinda spendy, but over 5 lbs of thrust. It looks from the drawing that you might be just going to use it to provide enough "down force" to hold it on the wall. Heck, they say a formula one or indy car could actually do that at speed because of the aerodynamics. I'm wondering if you couldnt mount 4 smaller EDF's on the corners for less money and, maybe some redundancy...

This is kind of a longer term goal but I didn't want to add any more variables to the current situation. It would be pretty wild to have a flying ground robot. CR EDFs and wheels, with thrust vectoring. Might as well make it amphibious too!

I also need 4 ESCs too. It would be more money but I'm sure it would be more functional as well.

xanadu

Three of my servos are standard, not CR, and I have only two wheels. I guess I'll have to go on a parts run to try this.

prof_braino

Surface area. If you robot were two foot square, there would be 4x the force then were it 1 foot square.

Also cross sectional area of the air gap. I think, as long as the gap for the air leaking in is smaller than the gap for the air leaking out, you have a chance.

NWCCTV

Does it have to be done this way?

xanadu

NWCCTV wrote: »

Does it have to be done this way?

Not sure what you mean. I want to believe that lots of people have tried this and failed, but I lack adequate research. I think it's a bit on the expensive side to pull off.