|
Post by timmidd on May 12, 2015 13:17:02 GMT
Hello New Member and First time Poster. I am looking to build a 6DOF Platform on which to mount a mockup cockpit (Space frame completed) that will be about 350kg MTOW (2 pilots and AC) I noted that 170 -190 kg units are using about .75kW AC motors through 60 or 80/1 Worm Drive GBs. Looking at the 6DOF platforms noted that the motors have to lift the entire mass where 2DOF platforms are moving the out of balance weight. Apart from having 6 arms sticking out of the platform with 1/6 payload balance weights (over fulcrums), I was trying to think of another way to partially equal the weight and unload the gearbox/motor somewhat. I understand that for certain manouvers that 1 motor still needs to be able to lift greater than 17% of the weight (I have to do some more reading to understand how much more) I did the attached drawing thinking about using springs as a means of countering payload weight and allow for the fact that the arm at 180 degrees is doing more work about the gearbox shaft than in the full up and down positions. Any thoughts as to whether this may work? Load Compensators.pdf (77.48 KB)
|
|
|
Post by clyevo on May 12, 2015 15:12:53 GMT
1) big spring is quite expensive 2) the spring is putting more downward strain on the gearbox shaft. 3) also spring has uneven force depending on how much torsion of the metal at certain length, so i think it might confuse the electronics but not sure about this Depending on the strength of the spring If the arm lever is 15cm and the counterlever is also 15cm and spring tension is 300 newton (30kgf) then you counter 30kg of the platform weight but the total strain on the gearbox shaft is also added 30kg if you want cheaper and less downward strain y not use it like this pulley system, blue is fulcrum, red is the belt, green is counterweight
|
|
|
Post by tronicgr on May 12, 2015 18:37:37 GMT
|
|
|
Post by timmidd on May 13, 2015 2:40:14 GMT
1) big spring is quite expensive 2) the spring is putting more downward strain on the gearbox shaft. 3) also spring has uneven force depending on how much torsion of the metal at certain length, so i think it might confuse the electronics but not sure about this Depending on the strength of the spring If the arm lever is 15cm and the counterlever is also 15cm and spring tension is 300 newton (30kgf) then you counter 30kg of the platform weight but the total strain on the gearbox shaft is also added 30kg if you want cheaper and less downward strain y not use it like this pulley system, blue is fulcrum, red is the belt, green is counterweight Clyevo, Thanks for the reply, I wasn't too concerned about the downward pressure exerted on the shaft and bearings although 60kgs would double (or more). You could put the cam on the opposite side of the box. or set the assembly above the box to cancel the forces within the lever arm. (this could be a better idea) I thought about changes in spring rate and after some reading found that the rate is relativley consistant between 20% and 80% of full range, its one reason why the max radius of the cam from the shaft is 50% of the arm. The spring tension would then be selected to double the supported range and because of the reduced travel stay within the desired tension range. When it comes to the loading effect on the motor through the 180 degree arc. With no compensation (and assuming the load is directly above the gearbox) the motor is seeing 0 to 100 to 0 % of the lifted Payload (taking out other dynamic effects). In this proposal you might get -10% to 0% (if balanced) to -5% which would tend to smooth the loading Curve rather than make it more erratic. Of course this is all hypothetical and I have been wrong before (just ask my wife).
|
|
|
Post by clyevo on May 13, 2015 3:13:33 GMT
Yeah you can change the spring position to exert the force upwards which solve the strain issue on bearing of the gearbox shaft. Using spring is much more neat than pulley of course. But i think the pulley also looks cool in a geekey way.
|
|
|
Post by timmidd on May 13, 2015 9:53:29 GMT
Yeah you can change the spring position to exert the force upwards which solve the strain issue on bearing of the gearbox shaft. Using spring is much more neat than pulley of course. But i think the pulley also looks cool in a geekey way. I worked out that a kids trampoline spring supports about 25kG (the ones I have) so to apply 125kg of Force at half lever length (I know weight is not force) you would need 10 springs or 2 x magazines of 5. The total spring travel is about 140mm about the cam. Balanced Load 2.pdf (78.59 KB)
|
|
|
Post by timmidd on May 14, 2015 2:21:07 GMT
This should work better mirrors the load curve in the main arm and applies zero force at the tdc and bdc of the arc. Setting up a rig this weekend to test the concept. Balanced Load3.pdf (100.33 KB)
|
|
|
Post by tronicgr on May 14, 2015 10:15:03 GMT
This should work better mirrors the load curve in the main arm and applies zero force at the tdc and bdc of the arc. Setting up a rig this weekend to test the concept. Actually that might work. It will provide graduate force to the actuator... !
|
|
|
Post by timmidd on May 14, 2015 11:25:06 GMT
This should work better mirrors the load curve in the main arm and applies zero force at the tdc and bdc of the arc. Setting up a rig this weekend to test the concept. Actually that might work. It will provide graduate force to the actuator... ! Hi troncicqr, I plan to set up a test model this weekend just using an idler shaft where the gearbox shaft would be and do a test on a 30kg load. At 50% lever length I should be able to resolve the weight using 5 trampoline springs. Off course there are a few variables. My calculations assume the weight is direct over the top of the gearbox with zero side loading, two variables that are in constant states of change with a Stewart platform. However, even if you could reduce the static load by 30% you may be able to reduce the size of the motor depending on the dynamic loading which would be relative to the speed of the platform. I will measure the out off balance force for each 10 degrees of arm movement at different percentages of starting position spring deflection and plot the results to create a load curve.
|
|
|
Post by clyevo on May 15, 2015 2:07:57 GMT
This should work better mirrors the load curve in the main arm and applies zero force at the tdc and bdc of the arc. Setting up a rig this weekend to test the concept. Actually that might work. It will provide graduate force to the actuator... ! I see what you meant. There is advantage using spring compare to weight on pulley system. 1) The force needed to move platform (radial load) upwards will be least when the arm lever is at 6 o'clock position. 2) In the other hand, when arm lever is at this position, the spring will exert more force due to higher spring tension at bigger spring displacement. 3) The force from the spring will offset the platform weight more when moving upwards from arm lever at 6 o'clock position (in comparison to when arm lever are at 9 o'clock position) thus facilitating acceleration and deceleration. This is especially advantageous for fast and high displacement high angle 6dof as they will have more inertia due to higher speed of the moving platform. The spring will allow us to decrease acc/dec ramp to minimum without overvoltage protection error at VFD side.
|
|
|
Post by timmidd on May 16, 2015 16:18:58 GMT
Clyevo. Actually a weight on a pulley would theoretically work better because you could match the weight exactly (if you took the time) and there would be no exponential spring rates to worry about. the problem is for a 300kg machine you would need 6 50kg counter weights and enough room to allow them to move vertically below the platform
My original principal is in the diagram above, but I discovered the spring range that I need varies from 10kg per spring at 20mm extension to 40kg per spring at 180mm so what I designed was a eccentric idler pulley between the springs and the small arm that I hope will work. because of the shape of the idler pulley Ive had to reduce the amount of spring travel to 50% which will mean for my test payload of 30kgs on the 160mm main arm I will have 60kgs on the 80mm control arm and 120kgs on the now 6 springs.
Each spring will be loaded at 20kgs when the main arm is horizontal which should extend them by 77mm, as I have now halved the spring travel they will extend and contract 40mm from this point which is well inside the linear growth area of the spring.
Started building the test rig I will post pictures when its done.
|
|
|
Post by tronicgr on May 16, 2015 20:54:23 GMT
May I add, the spring loaded solution will handle better any backlash from the gearbox too (all gearboxes have some backlash)...
|
|
|
Post by timmidd on May 17, 2015 5:05:15 GMT
May I add, the spring loaded solution will handle better any backlash from the gearbox too (all gearboxes have some backlash)... Yeah thats a good point tronicgr. Do you know what the typical backlash is in the 80:1 and 60:1 worm drives being used. measured at the end of the lever arm?
|
|
|
Post by clyevo on May 17, 2015 5:16:31 GMT
Clyevo. Actually a weight on a pulley would theoretically work better because you could match the weight exactly (if you took the time) and there would be no exponential spring rates to worry about. the problem is for a 300kg machine you would need 6 50kg counter weights and enough room to allow them to move vertically below the platform My original principal is in the diagram above, but I discovered the spring range that I need varies from 10kg per spring at 20mm extension to 40kg per spring at 180mm so what I designed was a eccentric idler pulley between the springs and the small arm that I hope will work. because of the shape of the idler pulley Ive had to reduce the amount of spring travel to 50% which will mean for my test payload of 30kgs on the 160mm main arm I will have 60kgs on the 80mm control arm and 120kgs on the now 6 springs. Each spring will be loaded at 20kgs when the main arm is horizontal which should extend them by 77mm, as I have now halved the spring travel they will extend and contract 40mm from this point which is well inside the linear growth area of the spring. Started building the test rig I will post pictures when its done. The exponential increase in force is actually good. Because when the arm lever is at full displacement (moving downwards) the exponential increase of tension by the spring is useful to brake the movement (like thanos said prevent backlash). With pulley system using weight, you cant exactly match the weight because each arm will carry different amount of weight of the platform depending on its position. Eg: when platform tilt forward, the front arm will carry more weight than the ones at the back. So, the only difference between both are spring has increasing force when extending while the weight on pulley have constant force that counterweight the weight each arms carry. I think using spring is neat and simpler to execute. I was wrong that when i said spring are expensive. Using weight you probably need like 20-30 kg each arm. I have this removable/adjustable dumbbell with that 5kg of round shape weight. Well for 120kg total we will need 24 pieces of 5kg weights -- quite expensive. I think you can also buy a good stainless steel spring online from china. You can enlarge this picture : The most bottom one you have 3.5mm wire diameter spring. 2 spring of this type in parallel will give you almost 500 newton probably more than enough for this application. If your payload is lighter you probably only need one of this high tension spring for each arm.
|
|
|
Post by timmidd on May 17, 2015 7:43:31 GMT
Here is my test results on a trampoline spring 27mm Dia with 2mm wire Spring Graph units are kg and mm.pdf (18.01 KB). The linear movement between 10 and 40 kgs gives us something to work with. I am starting the springs at 20kg at 72mm initial growth with the lever arm in the horizontal. with 30kg (theoretical weight) in the main lever arm at 160mm the magic No is 4800. The mimic (or control) arm will operate (in this test model) at 90 degrees to the main arm @ 80mm long (80 / 4800 = 15kg which is logical). From the mimic arm to the point at which the cable hits the secondary idler pully wants to be half of the length of the main connecting rod that will connect the main arm to the (stewart) platform to keep the dimensional relationship close proportions. The secondary pully (round) needs to be 160dia (80mm Radius) so that the force in the shaft of the pully assembly remains the same (ie 80 x 60 = 4800). The secret is keeping the force in the cable the same (60 kgs) just as the weight of 30kg remains constant (even though it doesnt but it will be close enough to smooth out the load curve). To achieve the magic no in the idler pully shaft (4800) I started with a nominated spring tension of 20 x 6 kg (120). 4800/120 = 40 which is the moment length of the primary pully I then progressed the rotation of the idler pulley 90degrees in each direction in 10 degree increments noting the required spring growth (or contraction ) and the resultant spring tension and applying the same formula at that angle until I had ploted the full 180 degrees of movement. See current design here Cam Pulley Setup.pdf (11.27 KB)
|
|