And when you inevitably spill a big bottle of CA on a nice marble floor, DO NOT panic and wipe it up with a handful of Kleenex. Ever.
No need to ask. Just re-read that first link:
Reaction with cotton Applying cyanoacrylate to materials made of cotton or wool (such as cotton swabs, cotton balls, and certain yarns or fabrics) results in a powerful, rapid exothermic reaction. The heat released may cause minor burns, and if enough cyanoacrylate is used, the reaction is capable of setting the cotton product in question on fire, as well as releasing additional irritating vapor in the form of white smoke.
Thanks Gordon . I was curious because I seen some YouTube videos of people doing it but it must not have been PVC sheets.
People do all sorts of crazy stuff. Seen any of the 'Jackass' movies? Burning PVC cleans out your sinuses in no time, and in a way you'll never forget!
The temperature range between melting and burning in expanded PVC isn't very wide, so while it is technically possible to do, it takes skill, practice, and (ideally) tools most people don't have. All too quickly the PVC will begin to burn, and when it does it's pretty nasty.
Concur with Gordon here. Use the PVC pipe "cement"... works great. An alternative might be an adhesive... "E6000" recommended by TAP Plastics. They told me that it is pretty effective at bonding different kinds of plastics and while I haven't done many tests, it seems to work pretty well. Of course, as always, your mileage may vary.
Tell me about it. Man, those Lynxmotion tracks are really nice. The Vex are just so much less expensive.
I think trying to make a four treaded robot of similar size with Lynxmotion tracks would greatly increase the cost of the robot. It sure would be cool though. I do think there's a Lynxmotion tracked robot in my future. I've been eyeing their tracks since I first started building robots a few years ago.
What's your impression of the durability of the Vex tracks?
What's your impression of the durability of the Vex tracks?
I've only had the "robot" drive across the living room a couple of times but so far I my main complaint about the Vex treads is the color.
The parts seem well made.
I've used my Vex Mecanum wheels much more than the treads. I don't have any complaints about the Mecanum wheels (other than color). My experience with the Mecanum wheels makes me optimistic the threads will work well.
I guess the color was a bit of a put-off for me as well. Though my goal was to build a track that could work reaasonably well outdoors and I wasn't sure Vex could handle it.
So far, the big tracks seem to work okay, but I haven't had a chance to run through grass and weeds and I'm still looking for someplace with a kind of rocky, rubble-y surface to navigate.
How are those motors torque wise at low speed? They seem like a good deal and the motor I picked out doesn't have an encoder option, which is something I would like to have even if I don't use it right away.
How are those motors torque wise at low speed? They seem like a good deal and the motor I picked out doesn't have an encoder option, which is something I would like to have even if I don't use it right away.
I don't really know.
With encoders you can kind of overcome low speed torque issues. You just have the program keep upping the PWM until the motors are turning at the desired speed. The under powered Rover 5 had really bad low speed torque but by adding encoder feedback to the control algorithm the Rover 5 acted as if it did have low speed torque.
I'll try to get one of these motor turning today and report back.
Good point about the encoders and torque. Choosing motors for my tracked robot is proving to be a challenge. I will definitely get something with encoders and probably a little less RPM. I do not need it to go fast, just need it to be quiet and good outdoors.
If you could still let me know if those things are overkill, perfect, okay, or bleh that would be awesome. Thanks Duane.
Yes it would need to have a center shaft to be planetary. I have just recently learned that.
The TRex motors look pretty small. I like how he built it with a lot of ground clearance. They only list the stall current of the motor which seems lower than most motors I'm looking at.
My dilemma is I cannot find a planetary gear motor with quadrature encoders on it. I'm pretty sure I need a planetary motor because it is quieter and hopefully performs better at low speed high torque situations.
I would like to build a small robot < 5 Lbs that can sneak up on wildlife. I'm working with a sponsorship from NBC and GoPro to help with a documentary about a famous artist. We'd be showing the artistic photography world the benefits that the GoPro brings to the action/sports world. I will be listed as a featured artist, and this could open the door to commercial multicopter flying. The robot could be placed inside of another fake animal or just camouflaged with sticks and leaves. It will hopefully get close to elusive or dangerous species of animals.
I'll keep searching the web, I hate to ask people on a forum to use a search engine for me lol. If you happen to spot any 6mm shaft planetary gear motors with quadrature encoders please let me know. Also if you think that my criteria for the motors isn't worth the trouble and that standard gear motors aren't that much louder and fine with torque let me know that too.
. . . I would like to build a small robot < 5 Lbs that can sneak up on wildlife. . . .
This sounds super cool.
I haven't been building robot nearly as long as many of the people around here but one thing I've read (and experienced) is plastic gear servos are quieter than metal gear versions. What about using plastic gears? Plastic gears of course have their downsides but the reduced noise may make them a good option.
Another thing I've learned is the shape of the gears can have a big effect on the noise level. Apparently making properly shaped gears (which would lower the noise) costs more than just cutting straight teeth.
I can test the various motors I have to see how loud they are but the loudness would be subjective. I don't suppose there's a program for the Demo Board or the PropBOE which would allow a mic and a Propeller chip to act as a sound meter? There's got to be one.
I'm personally very interested in this subject (making a quiet wildlife photography robot) and I've often wondered what would be the best sort of robot for sneaking up on wildlife.
I recall from my attempts to build a quiet PC that it was easier to keep the fan noise down if you used a lot of fans to move the air. I think (guess) this might apply to robots too. Maybe it would be quieter to use six wheels on a Mars rover type bot than to use two motors on a treaded vehicle?
Quadrature encoders shouldn't be hard to add to a motor has an external motor shaft. You could also add encoders to one of the gears. In post #11 of this thread I show a couple of PCBs I'm having made in attempts to add quadrature encoders to both an ActivityBot and to one of the gears inside a CR servo.
So if you can find a quiet motor without quadrature encoders, you may be able to add the encoders after the fact. Though I agree, it would a easier if the motor came with them to start with.
I think it would be great if we (robot enthusiast) had a way to objectively compare the noise output from motors. We could use a Propeller board with a mic to act as a sound meter and then we could all test our motors with the meter at an agreed upon distance and orientation. This way we could produce a list of the motors we collectively own and have a number to indicate how noisy it is.
BTW, I just now plugged one of my Pololu motors into my bench top power supply and the motor sounded neither noisy or quiet. I thought it produced less sound than one (I) would expect but it also didn't sound like it could sneak up on animals very well.
Do you have a thread about your search for quiet motors? If not, it seems like it would be a good thread to start. Hopefully someone will have a suggestion on how to objectively measure the sound from the motors/gearboxes.
I'm pretty sure using a video camera to record sounds is not a good way to document noise levels. Most video cameras have some sort of sound compression scheme which amplifies soft sounds and reduces loud sounds.
Another BTW: There's a guy in Alaska who takes pictures of bears with a camera on a Wild Thumper. I don't know the details, I just saw a passing reference to him on Let's Make Robots. It would be interesting to find out some of the details of what he is doing.
I bet that plastic gears and rubber tires would be a lot quieter. I should have mentioned that I'm not trying to go for the overall quietest option. It's just a bonus to me that the planetary are quieter. I'm sure the the noise can be overcome with good speed control. The FPV camera and HD recording system will be on a pan tilt system, I will certainly need to find the quietest servos for that. The robot will also be doing acoustic monitoring of both ambient and robot noise levels, I may even add vibration for dealing with snakes.
Bean wrote a cool VU meter in PBASIC for the demo board. I have used it a few times I guess you could use that to measure loudness. I agree dB rating should be advertised on robot motors. They are for computer fans. A quiet computer with large and slow fans and SSDs are nice. Way easier than liquid cooling. I have noticed that my CR servos seem to get louder and louder over time.
The encoder accuracy is more important to me than spur vs planetary. I'm a little worried about adding encoders not designed specifically for the motor. Some of my older forum posts will attest to that. I am going to get the same motors you're using, spur 150 rpm with quadrature encoders. The chassis is ePVC right now anyway. Once an aluminum chassis is fabricated I can re-evaluate the motors. If this works it will be shipped around a lot and can't be ePVC.
Thanks for plugging in one of those motors. I know what you mean about the sound. It will have to sneak up on animals. How ever that happens, be it dB listening watchdog microphones and moving at half an inch per minute, it will have to work. I've seen some photographers using RC trucks with no problem. I have some intuitive ideas, nothing groundbreaking, hopefully the least expensive and most versatile route. The noise may not be as much as an issue as I'm making it out to be.
I haven't started a thread on the motors. I didn't mean to hijack your thread on anything but the tracks hehe.
I don't think I've mentioned it here but I entered this "robot" in the 2015 Hackaday Prize contest. I used a very poorly Photoshopped photo as my project photo showing what I hoped to make.
I was very surprised to learn my project made it into the semifinals. Since a whole lot of people will be looking at the project, I'm trying to get into a presentable state by the 21st.
Here are some drawings of what I hope to make.
Palram Americas asked if they could use my robot pictures in the expanded PVC (aka foamed PVC) promotional materials. In exchange for my agreement they sent me a bunch of their plastic. I think the cost of the ePVC is not a large cost of the robot but it's nice to get some free stuff. I presently have no idea how much their stuff costs but I think the quality of their plastic is better than the other stuff I've used.
I doubt I'll be heading into space any time soon (the grand prize) but I'm hoping not to embarrass myself too much with this contest. Oh yeah, remember not tell anyone about the Lego set inspiration of this project.
That's awesome, Duane! Nice to see some well-deserved recognition of your hard work. Looks to be a broad, competitive field of entries. Good luck in the finals, Hoss!
I'm trying to figure out how to move those tread structures.
I have six of these Vigor VS-11 servos. I figure they have enough torque to lift the robot on to tiptoes if I used a winch with a spool about a third of the diameter of the "drum" on the tread structure.
I'm calling the place where the cable/monofilament interacts with the thread structure the "drum" and the cylinder attached to the servo (continuous rotation) the "spool". I'm open to alternate names if any of you have suggestions on what to all these parts.
I don't show the cable wrapped around the drum and spool in this next drawing but I plan to have the cable wrap around the drum at least once and I plan to have the cable wrap around the spool at least twice (probably four of five times).
I've seen these sorts of winches at boat docks. A rope is wrapped around a spool a few times and the friction between the rope and the spool pulls a boat up to the dock.
Here's another really bad doodle attempting to show what I'm talking about.
The cable is attached to the drum, wrapped around the drum and then wrapped around the spool several times. The cable is then brought back up to the drum and wrapped one more time before being again attached to the drum.
The extra wraps of the cable around the drum allow for cable to be unwound at least a full rotation before the cable is stopped by its anchor point.
One the spool side, the friction of the multiple wraps of cable around the spool allow the spool to pull on the cable an transfering the spool's rotation to the drum's rotation.
I haven't ever made a mechanism like this but it seem plausible. What do you guys think?
I've also wondered about using some sort of worm drive since a worm drive wouldn't require power to maintain position. I'm not really worried about using power to maintain position at this point. Most of the tread configurations will not put torque on the treads' pivot. I doubt I'll have the robot maintain a tiptoe stance for long periods so at this point I doubt I'll try to implement a worm drive.
The treads need about 41kg*cm of torque but I figure if I use a spool smaller than the drum, I wouldn't need a servo/motor to produce this amount of torque. The required torque would be reduced by the ratio between the drum and spool sizes. If the spool's diameter is a third of the drum's diameter then rather than 41kg*cm torque, I'd need about 14kg*cm torque.
I have several other motor options. I have more of these nice gearmotors. I have lots of Lego parts and gears including worm drive components but I'm afraid Lego parts aren't strong enough for this sort of application.
I don't really have time to order parts so I kind of need to get by with what I have on hand or what I can buy at the hardware store.
As always, I'd appreciate any input you care to provide.
(I'm starting to wonder if Hackday did me a favor by selecting my project.)
Your robot is more advanced than anything I've seen in the 2015 entries. Getting this done before or after their deadline is a win in my book. VEX should be calling you to license this. After that you can go to space once a year!
I've been looking at other robots with the same track configuration. I can't find any pics of the mechanics. If the rope doesn't stretch it's a great idea. If the pulleys are sandwiched between ePVC walls you can probably tighten the backlash out for a few good runs. Considering robots like this cost $20,000 and up a DIY version can be cut some slack.
I think I get what you are suggesting. In the up position, the tread structures would be exerting constant rotational force on the drum, so I imagine that the cable might slip on the spool (capstan?) over time. You are relying entirely on the static friction between the spool surface the cable to keep the tread structures stationary. You are also potentially benefiting from the cable's friction with itself due to the windings. This might require a curve in the spool, just as you see in many (most?) capstans.
It certainly seems worth a try, particularly if you are working with materials that you have on hand.
I'm guessing you already thought about this, but you could help the above cable system do it's work by simultaneously running the treads on the same side in opposite directions (in the above side profile pictures, CW for the left tread and CCW for the right tread).
Also, since you are likely to get different effective torques (depending on slippage, speed, etc.) at each servo, you might be able to use something like the MEMSIC 2125 to adjust the speed of the servos to keep the center platform level. I know this says "accelerometer", but I treat it like "electronic bubble level". At one point, Radio Shack was selling these (that's where I got mine).
You'll quickly melt the plastic, and you'll get a nose-full of chlorine gas.
And if you melt/burn/power-sand hardened CA (especially the thick stuff in a clump), that nose-burning smell is cyanide gas. VERY irritating, it's easy to vapor lock from one good whiff.
Hard to believe that this potentially toxic cyanide product was initially intended for medical use, for quick battlefield triage to "glue" wounds shut.
Yes to the CSI idea. That's a favorite of their former boss, Gil Grissom, and yes to the one regarding superglue. I strongly dislike it because of that problem.
Spinnaker winch, Genoa winch, Mainsail winch, Sail winch?? do any of those look familiar?
Cool stuff Duane! Love the potential stair climbing ability. :thumb:
And good luck with the HaD contest, Also don't lick the super glue off of the cotton swab..
I have a pile beautiful new old stock 426:1 24V Pittman gearmotors with monstrous torque. How heavy are those tread assemblies? Four of these suckers and you could rotate each assembly independently!
I've been looking at other robots with the same track configuration. I can't find any pics of the mechanics.
I wasn't able to find pictures of the mechanics either. I think a lot of them rotate the tread at the same point as the main drive sprocket. I initially considered this but I think it would take some significant machine work to make a mechanism to accomplish this.
If the rope doesn't stretch it's a great idea. If the pulleys are sandwiched between ePVC walls you can probably tighten the backlash out for a few good runs.
Having the cable stretch could create problems but the rotation mechanism should still work. And stretched cable would end up on the side of the capstan (thanks Seairth) and should cause too much trouble as long as it doesn't get into places where it doesn't belong. At least this is how it should work in the version I have pictured in my brain.
In the up position, the tread structures would be exerting constant rotational force on the drum, so I imagine that the cable might slip on the spool (capstan?) over time.
I had just worried about the maintaining the torque. I hadn't given much thought to slippage while stationary. I'm figuring I probably won't have the robot spend a lot of time up on its tiptoes until I come up with a better solution for rotating the tread structures. I'll probably monitor the tread structures' positions and run the motors as need to compensate for it.
This might require a curve in the spool, just as you see in many (most?) capstans.
I'm also wondering if I should coat the capstan with some sort of rubber material to increase the friction. I'm not sure about this since I think the cable needs to be able to slid from side to side a bit as it passes around the capstan multiple times.
I'm guessing you already thought about this, but you could help the above cable system do it's work by simultaneously running the treads on the same side in opposite directions (in the above side profile pictures, CW for the left tread and CCW for the right tread).
Yeah, let's go with the "I already thought of this" story. It sounds better than what I was going to say. I like to think I would have thought of this once I had the mechanism built and ready to test. Thanks for the suggestion.
Also, since you are likely to get different effective torques (depending on slippage, speed, etc.) at each servo, you might be able to use something like the MEMSIC 2125 to adjust the speed of the servos to keep the center platform level. I know this says "accelerometer", but I treat it like "electronic bubble level".
I think I have enough accelerometers to also put one in each tread structure. Since each structure can rotate independently and there is bound to be lots of slippage in the rotation mechanism, I had planned to put an accelerometer in each tread structure. I have some $3 one I purchased on ebay awhile back. I think they'll work.
Cool stuff Duane! Love the potential stair climbing ability.
Yeah, I'm really hoping I can get the stair climbing part to work. Our home has four levels. Having a robot for each level doesn't seem very practical (though to be honest neither does this robot).
The one I made with ePVC weighs 777g. The center of gravity of the tread is about 6cm away from the pivot point. I don't think rotating the weight of the treads will be too much trouble but I'm concerned with lifting the weight of the robot. I'm guessing the body of the robot will weigh about 3kg (hopefully this is a high estimate). At 3kg each tread will need to lift .75kg (though it's unlikely the weight will be evenly distributed). The tip of the tread structure is about 30cm from the pivot point so the torque required to lift this 0.75kg will be 22.5kg*cm. Since the center of mass of the tread structure is 6cm from the pivot point its lever arm is 24cm instead of 30cm so the tread structure adds an additional 24cm * 0.777kg = 18.65kg*cm torque for a total of 44.15kg*cm of torque. Hopefully I over estimated the weight of the robot body and at I won't need this much torque.
Comments
No need to ask. Just re-read that first link:
Reaction with cotton Applying cyanoacrylate to materials made of cotton or wool (such as cotton swabs, cotton balls, and certain yarns or fabrics) results in a powerful, rapid exothermic reaction. The heat released may cause minor burns, and if enough cyanoacrylate is used, the reaction is capable of setting the cotton product in question on fire, as well as releasing additional irritating vapor in the form of white smoke.
People do all sorts of crazy stuff. Seen any of the 'Jackass' movies? Burning PVC cleans out your sinuses in no time, and in a way you'll never forget!
The temperature range between melting and burning in expanded PVC isn't very wide, so while it is technically possible to do, it takes skill, practice, and (ideally) tools most people don't have. All too quickly the PVC will begin to burn, and when it does it's pretty nasty.
HTH...
@
What's your impression of the durability of the Vex tracks?
I've only had the "robot" drive across the living room a couple of times but so far I my main complaint about the Vex treads is the color.
The parts seem well made.
I've used my Vex Mecanum wheels much more than the treads. I don't have any complaints about the Mecanum wheels (other than color). My experience with the Mecanum wheels makes me optimistic the threads will work well.
I guess the color was a bit of a put-off for me as well. Though my goal was to build a track that could work reaasonably well outdoors and I wasn't sure Vex could handle it.
So far, the big tracks seem to work okay, but I haven't had a chance to run through grass and weeds and I'm still looking for someplace with a kind of rocky, rubble-y surface to navigate.
@
How are those motors torque wise at low speed? They seem like a good deal and the motor I picked out doesn't have an encoder option, which is something I would like to have even if I don't use it right away.
I don't really know.
With encoders you can kind of overcome low speed torque issues. You just have the program keep upping the PWM until the motors are turning at the desired speed. The under powered Rover 5 had really bad low speed torque but by adding encoder feedback to the control algorithm the Rover 5 acted as if it did have low speed torque.
I'll try to get one of these motor turning today and report back.
If you could still let me know if those things are overkill, perfect, okay, or bleh that would be awesome. Thanks Duane.
I doubt they're planetary. Don't planetary gearboxes have the axle coming from the center?
Have you seen the latest tracked robot from Dagu? I think it looks pretty good but I think it costs over $500.
The TRex motors look pretty small. I like how he built it with a lot of ground clearance. They only list the stall current of the motor which seems lower than most motors I'm looking at.
My dilemma is I cannot find a planetary gear motor with quadrature encoders on it. I'm pretty sure I need a planetary motor because it is quieter and hopefully performs better at low speed high torque situations.
I would like to build a small robot < 5 Lbs that can sneak up on wildlife. I'm working with a sponsorship from NBC and GoPro to help with a documentary about a famous artist. We'd be showing the artistic photography world the benefits that the GoPro brings to the action/sports world. I will be listed as a featured artist, and this could open the door to commercial multicopter flying. The robot could be placed inside of another fake animal or just camouflaged with sticks and leaves. It will hopefully get close to elusive or dangerous species of animals.
I'll keep searching the web, I hate to ask people on a forum to use a search engine for me lol. If you happen to spot any 6mm shaft planetary gear motors with quadrature encoders please let me know. Also if you think that my criteria for the motors isn't worth the trouble and that standard gear motors aren't that much louder and fine with torque let me know that too.
Thanks a bunch.
This sounds super cool.
I haven't been building robot nearly as long as many of the people around here but one thing I've read (and experienced) is plastic gear servos are quieter than metal gear versions. What about using plastic gears? Plastic gears of course have their downsides but the reduced noise may make them a good option.
Another thing I've learned is the shape of the gears can have a big effect on the noise level. Apparently making properly shaped gears (which would lower the noise) costs more than just cutting straight teeth.
I can test the various motors I have to see how loud they are but the loudness would be subjective. I don't suppose there's a program for the Demo Board or the PropBOE which would allow a mic and a Propeller chip to act as a sound meter? There's got to be one.
I'm personally very interested in this subject (making a quiet wildlife photography robot) and I've often wondered what would be the best sort of robot for sneaking up on wildlife.
I recall from my attempts to build a quiet PC that it was easier to keep the fan noise down if you used a lot of fans to move the air. I think (guess) this might apply to robots too. Maybe it would be quieter to use six wheels on a Mars rover type bot than to use two motors on a treaded vehicle?
Quadrature encoders shouldn't be hard to add to a motor has an external motor shaft. You could also add encoders to one of the gears. In post #11 of this thread I show a couple of PCBs I'm having made in attempts to add quadrature encoders to both an ActivityBot and to one of the gears inside a CR servo.
The Rover 5 has its quadrature encoders installed on one of the gears of the gearbox. I've used encoders on a gear trick to add encoders (single channel) to my Roboni-I robot.
So if you can find a quiet motor without quadrature encoders, you may be able to add the encoders after the fact. Though I agree, it would a easier if the motor came with them to start with.
I think it would be great if we (robot enthusiast) had a way to objectively compare the noise output from motors. We could use a Propeller board with a mic to act as a sound meter and then we could all test our motors with the meter at an agreed upon distance and orientation. This way we could produce a list of the motors we collectively own and have a number to indicate how noisy it is.
BTW, I just now plugged one of my Pololu motors into my bench top power supply and the motor sounded neither noisy or quiet. I thought it produced less sound than one (I) would expect but it also didn't sound like it could sneak up on animals very well.
Do you have a thread about your search for quiet motors? If not, it seems like it would be a good thread to start. Hopefully someone will have a suggestion on how to objectively measure the sound from the motors/gearboxes.
I'm pretty sure using a video camera to record sounds is not a good way to document noise levels. Most video cameras have some sort of sound compression scheme which amplifies soft sounds and reduces loud sounds.
Another BTW: There's a guy in Alaska who takes pictures of bears with a camera on a Wild Thumper. I don't know the details, I just saw a passing reference to him on Let's Make Robots. It would be interesting to find out some of the details of what he is doing.
Bean wrote a cool VU meter in PBASIC for the demo board. I have used it a few times I guess you could use that to measure loudness. I agree dB rating should be advertised on robot motors. They are for computer fans. A quiet computer with large and slow fans and SSDs are nice. Way easier than liquid cooling. I have noticed that my CR servos seem to get louder and louder over time.
The encoder accuracy is more important to me than spur vs planetary. I'm a little worried about adding encoders not designed specifically for the motor. Some of my older forum posts will attest to that. I am going to get the same motors you're using, spur 150 rpm with quadrature encoders. The chassis is ePVC right now anyway. Once an aluminum chassis is fabricated I can re-evaluate the motors. If this works it will be shipped around a lot and can't be ePVC.
Thanks for plugging in one of those motors. I know what you mean about the sound. It will have to sneak up on animals. How ever that happens, be it dB listening watchdog microphones and moving at half an inch per minute, it will have to work. I've seen some photographers using RC trucks with no problem. I have some intuitive ideas, nothing groundbreaking, hopefully the least expensive and most versatile route. The noise may not be as much as an issue as I'm making it out to be.
I haven't started a thread on the motors. I didn't mean to hijack your thread on anything but the tracks hehe.
https://hackaday.io/project/4738-health-maintenance-robot
I was very surprised to learn my project made it into the semifinals. Since a whole lot of people will be looking at the project, I'm trying to get into a presentable state by the 21st.
Here are some drawings of what I hope to make.
Palram Americas asked if they could use my robot pictures in the expanded PVC (aka foamed PVC) promotional materials. In exchange for my agreement they sent me a bunch of their plastic. I think the cost of the ePVC is not a large cost of the robot but it's nice to get some free stuff. I presently have no idea how much their stuff costs but I think the quality of their plastic is better than the other stuff I've used.
I doubt I'll be heading into space any time soon (the grand prize) but I'm hoping not to embarrass myself too much with this contest. Oh yeah, remember not tell anyone about the Lego set inspiration of this project.
I'm trying to figure out how to move those tread structures.
I have six of these Vigor VS-11 servos. I figure they have enough torque to lift the robot on to tiptoes if I used a winch with a spool about a third of the diameter of the "drum" on the tread structure.
I'm calling the place where the cable/monofilament interacts with the thread structure the "drum" and the cylinder attached to the servo (continuous rotation) the "spool". I'm open to alternate names if any of you have suggestions on what to all these parts.
I don't show the cable wrapped around the drum and spool in this next drawing but I plan to have the cable wrap around the drum at least once and I plan to have the cable wrap around the spool at least twice (probably four of five times).
I've seen these sorts of winches at boat docks. A rope is wrapped around a spool a few times and the friction between the rope and the spool pulls a boat up to the dock.
Here's another really bad doodle attempting to show what I'm talking about.
The cable is attached to the drum, wrapped around the drum and then wrapped around the spool several times. The cable is then brought back up to the drum and wrapped one more time before being again attached to the drum.
The extra wraps of the cable around the drum allow for cable to be unwound at least a full rotation before the cable is stopped by its anchor point.
One the spool side, the friction of the multiple wraps of cable around the spool allow the spool to pull on the cable an transfering the spool's rotation to the drum's rotation.
I haven't ever made a mechanism like this but it seem plausible. What do you guys think?
I've also wondered about using some sort of worm drive since a worm drive wouldn't require power to maintain position. I'm not really worried about using power to maintain position at this point. Most of the tread configurations will not put torque on the treads' pivot. I doubt I'll have the robot maintain a tiptoe stance for long periods so at this point I doubt I'll try to implement a worm drive.
The treads need about 41kg*cm of torque but I figure if I use a spool smaller than the drum, I wouldn't need a servo/motor to produce this amount of torque. The required torque would be reduced by the ratio between the drum and spool sizes. If the spool's diameter is a third of the drum's diameter then rather than 41kg*cm torque, I'd need about 14kg*cm torque.
I have several other motor options. I have more of these nice gearmotors. I have lots of Lego parts and gears including worm drive components but I'm afraid Lego parts aren't strong enough for this sort of application.
I don't really have time to order parts so I kind of need to get by with what I have on hand or what I can buy at the hardware store.
As always, I'd appreciate any input you care to provide.
(I'm starting to wonder if Hackday did me a favor by selecting my project.)
I've been looking at other robots with the same track configuration. I can't find any pics of the mechanics. If the rope doesn't stretch it's a great idea. If the pulleys are sandwiched between ePVC walls you can probably tighten the backlash out for a few good runs. Considering robots like this cost $20,000 and up a DIY version can be cut some slack.
It certainly seems worth a try, particularly if you are working with materials that you have on hand.
I'm guessing you already thought about this, but you could help the above cable system do it's work by simultaneously running the treads on the same side in opposite directions (in the above side profile pictures, CW for the left tread and CCW for the right tread).
Also, since you are likely to get different effective torques (depending on slippage, speed, etc.) at each servo, you might be able to use something like the MEMSIC 2125 to adjust the speed of the servos to keep the center platform level. I know this says "accelerometer", but I treat it like "electronic bubble level". At one point, Radio Shack was selling these (that's where I got mine).
Yes to the CSI idea. That's a favorite of their former boss, Gil Grissom, and yes to the one regarding superglue. I strongly dislike it because of that problem.
But I found the way the robot communicated strongly reminded me of one of the earlier comedians, and not in a good way.
I'm more in the with droids who make the New Republic a fun place to want to visit.
Or come to that the ones who sometimes turn up in the Star Trek region of space and time.
And of course the ones in Doctor Who are there.
Cool stuff Duane! Love the potential stair climbing ability. :thumb:
And good luck with the HaD contest, Also don't lick the super glue off of the cotton swab..
24V=~18 RPM 12V=~9 RPM
That's kind of you to say so but, so far this robot is mostly a Photoshopped picture and some drawings.
I wasn't able to find pictures of the mechanics either. I think a lot of them rotate the tread at the same point as the main drive sprocket. I initially considered this but I think it would take some significant machine work to make a mechanism to accomplish this.
Having the cable stretch could create problems but the rotation mechanism should still work. And stretched cable would end up on the side of the capstan (thanks Seairth) and should cause too much trouble as long as it doesn't get into places where it doesn't belong. At least this is how it should work in the version I have pictured in my brain.
I had just worried about the maintaining the torque. I hadn't given much thought to slippage while stationary. I'm figuring I probably won't have the robot spend a lot of time up on its tiptoes until I come up with a better solution for rotating the tread structures. I'll probably monitor the tread structures' positions and run the motors as need to compensate for it.
I'm also wondering if I should coat the capstan with some sort of rubber material to increase the friction. I'm not sure about this since I think the cable needs to be able to slid from side to side a bit as it passes around the capstan multiple times.
Yeah, let's go with the "I already thought of this" story. It sounds better than what I was going to say. I like to think I would have thought of this once I had the mechanism built and ready to test. Thanks for the suggestion.
I think I have enough accelerometers to also put one in each tread structure. Since each structure can rotate independently and there is bound to be lots of slippage in the rotation mechanism, I had planned to put an accelerometer in each tread structure. I have some $3 one I purchased on ebay awhile back. I think they'll work.
I think I've heard of a sail winch before. I'll need to look up the rest these terms. Thanks.
Yeah, I'm really hoping I can get the stair climbing part to work. Our home has four levels. Having a robot for each level doesn't seem very practical (though to be honest neither does this robot).
The one I made with ePVC weighs 777g. The center of gravity of the tread is about 6cm away from the pivot point. I don't think rotating the weight of the treads will be too much trouble but I'm concerned with lifting the weight of the robot. I'm guessing the body of the robot will weigh about 3kg (hopefully this is a high estimate). At 3kg each tread will need to lift .75kg (though it's unlikely the weight will be evenly distributed). The tip of the tread structure is about 30cm from the pivot point so the torque required to lift this 0.75kg will be 22.5kg*cm. Since the center of mass of the tread structure is 6cm from the pivot point its lever arm is 24cm instead of 30cm so the tread structure adds an additional 24cm * 0.777kg = 18.65kg*cm torque for a total of 44.15kg*cm of torque. Hopefully I over estimated the weight of the robot body and at I won't need this much torque.
Yes, I think those would work but then they're likely more than double the weight of the robot.
It will be interesting to see how much I can get done in the next week. Judging is on the 21th.