Society of Robots - Robot Forum
Mechanics and Construction => Mechanics and Construction => Topic started by: metrosonus on March 27, 2011, 06:03:26 PM
-
Hey guys I'm trying to do research for a project of mine that im doing. I'm building building k9, from the TV show, Doctor Who.
(http://www.plasticbag.org/images/extra/doctor_who_k9.jpg)
I've read the site, done a bit of homework, so let me see if I have this right and please give me your tips / advice.
I only need basic RC control. So I think I need a 4 channel remote, a four channel receiver, two motors and some servos.
What I'd like to do is be able to steer it with basic or no speed control. It's going to be very lightweight, made out of fiberglass, and driven on hard, indoor or concrete type surfaces.
If I got two motors, steering would be accomplished by one stopping, then the running motor taking over for left and right?
I'd also like to have control on the head moving up and down along with being able to raise and low the antenna probe located in the head. Im' assuming that if I switch channels on the remote, that will give me two more to control those with?
I hope im getting the right idea here.. please let me know where i am on and off the mark.
thanks!
-
Hi,
I only need basic RC control. So I think I need a 4 channel remote, a four channel receiver, two motors and some servos.
2 motors, 1 head and 1 antenna, yep, that equals 4.
If I got two motors, steering would be accomplished by one stopping, then the running motor taking over for left and right?
Yes, it's called differential steering.
I'd recommend using proportional speed control though as it would be quite jerky going full speed and then cutting power to one side.
-
The amount of channels you need will depend on how many things you want to control .
I loved the Drwho series and K9 was one of my childhood inspirations into robotics and it makes a great project , however it is far better if controlled with a micro controller .
I don't recommend using skid steer (differential steering ) it is a real pain in the but when your length is a lot longer than your width and it will put a lot heavier loads on your motors and will require a lot more power .
I remember seeing a photo of the original and it had a single front steering wheel and a differential driving the rear wheels as in a car ( differential drive is not the same as differential steering ) .
Have fun
-
I do not see any functionality here that would really change between the microcontroller and the RC reciever. Provided that properly chosen ESCs are used, the RC aspect is easy to put together, and a simple RC mixer can drive the two primary motors in 'tank drive' mode. The bigger challenge here, I would think, would be the mechanical aspect - how to make everything fit and still look nice, and remain true to the design inspiration.
On steering:
Last time I checked, differential drive/steering are the same thing (correct me if I'm wrong) - ie, using two fixed sources of varying speed with a distance in between to create a platform with two degrees of freedom (speed and yaw).
Differential drive/steering should be fine - just center the two wheels, and use casters on the ends. Ackerman steering (like a car, with a yaw in front and a speed in back) is much more difficult to do in a microcontroller, at least for an autonomous robot: the kinematics aren't nearly as simple (albeit not excessively difficult), and turn-in-place is impossible.
-
Differential drive and differential steering are totally different things.
Differential steering or as it is also called skid steer vary the speed of the left and right drive wheels so as one side goes faster the object will go in that direction , there are many robots built using this system and it is also used with a lot of industrial machinery and most used with bulldozers and other earth moving gear .
A differential drive is whats like in a car , it take shaft input from your motor and transfers this to your wheels , gearing inside the "diff" allows the car to go around corners therefore "differentiating" the inner and outer wheels that must rotate at "different" speeds .
Functionality difference between RC and micro , well maybe there is not a lot of difference but that would depend on if you would call a radio controlled toy a robot ( BTW I do not ) to me a robot is a electromechanical device that is programed to perform a task.
A radio controlled toy is ... a radio controlled toy
I have a few radio controlled planes that I have never thought of as robots and I don't know any one who would , however a UAV that was programed to fly its self would be a very different thing .
-
Seems to me you're describing a differential (http://"http://en.wikipedia.org/wiki/Differential_%28mechanical_device%29"), which is usually attached to a drive system. It's not, to my knowledge, generally associated with the term 'differential drive', which, at least by popular reference (See Google (http://"http://www.google.com/search?client=ubuntu&channel=fs&q=differential+drive&ie=utf-8&oe=utf-8")) refers to what you call differential steering.
Nevertheless, utilizing a differential on top of an Ackerman steering system (where the front wheel(s) physically change orientation to travel circles of different radii) is wholly unnecessary for a robot this size - there's really no reason to increase the complexity of the drivetrain unless there's a noticeable need, and that need doesn't seem apparent in the OP's goal. While a multiton car has a serious handling problem when there is no differential or Ackerman linkage, a sub-10-kilogram robot has no such issues.
"skid" steering is perfectly fine for this task - it's a proven and mechanically simpler system.
[...] I only need basic RC control. So I think I need a 4 channel remote, a four channel receiver, two motors and some servos. [...]
OP says he wants basic RC control - no need to make life harder for him or her, and it's easy enough to add a microcontroller in later.
Simplicity is key for projects like this - make it too hard, and you'll just give up earlier.
-
Yes I am describing a differential , however to achieve this you don't actually need a differential , all you need is to drive one wheel and let the other side free wheel so they can rotate at different speeds .
As for complexity both are munch the same , a skid steer will need two motors and two ESCs while using a simple diff will only need one motor one ESC and one servo for the steering ( a lot cheaper ), the steering wheel only needs to be a single unit and would be very simple to make so no need for complex steering linkage .
Another good method would be to use two motors as in a skid steer and have a front (or rear) wheel that acts like a castor , this method will greatly reduce the friction normally associated with traditional skid type steering where several wheels (or a tracks) are used on each side
Skid steering will require more motor power as it creates more friction and it uses power to do the steering as well and I still don't recommend it if you want some thing that is nice a smooth to control around on your living room floor .
I was only musing about the micro controller as seeing a photo of K9 bought back some memory's
-
Hi,
Yes I am describing a differential [...]
Which has nothing to do with the terms differential drive/steering and mentioning it is just complicating the terms for the OP.
As for complexity both are munch the same , a skid steer will need two motors and two ESCs while using a simple diff will only need one motor one ESC and one servo for the steering ( a lot cheaper ), the steering wheel only needs to be a single unit and would be very simple to make so no need for complex steering linkage .
You'll need 2 motors for movement plus steering, whether you do it one way or another.
Another good method would be to use two motors as in a skid steer and have a front (or rear) wheel that acts like a castor , this method will greatly reduce the friction normally associated with traditional skid type steering where several wheels (or a tracks) are used on each side
I assumed that was the way the OP wanted to do it.
I do have to protest to the term skid steering here, as it doesn't skid in this topology. Skid only happens when each side covers a certain amount of "track" lengthwise, but with two wheels and a castor, the skid and hence the uncertainties of skid steering that makes dead reckoning a pain, is no longer an issue when you only got a small pad of contact each side.
Skid steering will require more motor power as it creates more friction and it uses power to do the steering as well and I still don't recommend it if you want some thing that is nice a smooth to control around on your living room floor .
I doubt you'd be able to measure any increase in power due to a two wheel differential drive. Where would it come from?
We're not discussing a tank.
-
Wow ok that's allot to take in, but thank you. This is just going to be like a big RC car. It won't function autonomously, so no need for a micro controller. However, I did want to add on a secondary system for doing sounds and voices that would maybe use one and some sort of infrared remote. I did read up on microcontrollers, but figured that by splitting the tasks, it'd be easier.
I am anticipating a dry weight of about 20 pounds or 10 kilograms, so you're saying that a skid steer system would work? What specifically would I need for that? Two motors, AC or DC? And what's an ESC if I may ask? And you said two right? And then please forgive me, how would that be wired? I"m sure once I learn what an ESC does I can figure it out lol. But thank you..
If you're interested, you can follow my progress on the project here:
http://computermusicguide.com/cng/?p=108 (http://computermusicguide.com/cng/?p=108)
-
"Skid steer" will most certainly work for your application - the only question is selecting the appropriate motors. This method is more analogous to an RC tank than an RC car, unless you decide to buy a 'mixer', which converts RC-car like controls into RC-tank like commands.
Use the RMF calculator: http://www.societyofrobots.com/RMF_calculator.shtml (http://www.societyofrobots.com/RMF_calculator.shtml) to calculate the required specifications of your motor, and then search around for an appropriate gearmotor (you probably don't want to have to figure out how to build a geartrain on your own). Once you've found one, the appropriate ESC can be determined from the electrical characteristics of your motor.
You will want to use a DC gearmotor - brushless motors are more expensive than you need (and too fast), AC motors don't work well without AC power sources (batteries are DC), and modified servos will not have the necessary torque to move your robot around.
Like you said, splitting the tasks would make the system simpler and easier - we'll work on the fancy bells and whistles once the drive system works. You will likely still need a microcontroller for those, but it's in the future.
Wiring would look like this:
Left Motor Right Motor
|M1A |M1B |M2A |M2B
Battery + -------------- Left ESC Vin ----------------- Right ESC Vin ------------------ BEC Vin
Left ESC* Right ESC** BEC***
Battery - --------------- Left ESC GND ---------------- Right ESC GND ------------------ BEC GND
*-- |Receiver channel 1
**-- |Receiver channel 2
***-- |Any unused receiver channel (only +5v and GND connected)
Motor related wiring is BLUE, battery positive is RED, battery negative is BLACK, receiver connections are PURPLE, ESCs are ORANGE
Hoping this diagram helps a little.
An ESC is an 'electronic speed controller', which takes input from the RC receiver and converts it into a duty cycle for the motor it's attached to (duty cycle is related to motor speed).
-
Gosh guys I never said skid steer would not work , of course it will work , I just don't see it as the best option .
This system I have described only needs one motor for drive and one servo for the steering and of course an ESC for speed control.
I refer skid steer to any thing that does not have a set of wheels that point in the direction you wish to go, it comes from many years of working around such machinery .
What I am seeing is the OP looking to make a box that is a lot longer than it is wide and putting a wheel in each corner pointing in the long direction of the box and by varying the speed of either side they cause the object to move in that direction , like some one said that is great if you wish to make a tank .
BTW Soren , you was the first to mention differential steering TYVM so why put me down for mentioning it ?
-
This system I have described only needs one motor for drive and one servo for the steering and of course an ESC for speed control.
Your system is more complex that the OP needs. The monetary savings on the single gearmotor and ESC are far offset by the massive fiddling that will be required for a less-experienced person to figure out how to implement your system. While you could pull it off, OP probably can't (and has no need to).
The primary benefit of a two-wheel + caster design is the simplicity. It's easy to build, and if it's assembled wrong can be fixed through pure software. Moreover, it's fundamentally reliable: as long as the robot can move, it has full locomotion, and gearmotors are less likely by far to fail than servos.
I refer skid steer to any thing that does not have a set of wheels that point in the direction you wish to go, it comes from many years of working around such machinery .
Sure - doesn't mean that the name isn't essentially wrong if there's no skidding involved.
On a differential steering system, if you assume the point of contact between the wheel and the ground is essentially the tangent of the wheel, the wheel always points in the direction of travel: the offset between the wheels is what creates the turn. Moreover, there is usually no need for the OP to actually design/build their own geartrain, which simplifies things immensely.
What I am seeing is the OP looking to make a box that is a lot longer than it is wide and putting a wheel in each corner pointing in the long direction of the box and by varying the speed of either side they cause the object to move in that direction , like some one said that is great if you wish to make a tank .
I'd say the wheels would be placed along the halfway point of the box, such that the box pivots on the wheels more easily... casters could be used for the front and the back (one of them spring loaded) to keep the box upright.
|_____________________| <-- box
C O C_s
C is a caster, O is the wheel axis, C_s is a spring loaded caster.
BTW Soren , you was the first to mention differential steering TYVM so why put me down for mentioning it ?
I hate to presume, but I believe Soeren is referring to your statements on mechanical differentials, which are usually not used on small-scale robots like the one that the OP is trying to build. According to Soeren, it's worse than being irrelevant in that it confuses terminology for the OP.
-
Just for your inspiration, have you checked this out http://www.societyofrobots.com/robotforum/index.php?topic=2306.0 (http://www.societyofrobots.com/robotforum/index.php?topic=2306.0)
Its an old post, but somebody else made an autonomous K9
-
Hi,
I am anticipating a dry weight of about 20 pounds or 10 kilograms, so you're saying that a skid steer system would work? What specifically would I need for that? Two motors, AC or DC?
It's gonna run on batteries, which delivers DC, so you'd need two DC motors with gear heads.
Since the sides are slanted, you can either slant the wheels as well, or use fairly small wheels, to get the wheels as close to the side "walls" as possible. This will increase stability.
And what's an ESC if I may ask? And you said two right? And then please forgive me, how would that be wired?
An ESC is an Electronic Speed Controller and it let the weak (servo-) signals, coming from the receiver, control the power going to a motor. Some ESC's are made to control two motors.
You might wanna look at the following links for inspiration - they're all differential steering and I haven't found a single who weren't.
http://johndelpizzo.wordpress.com/category/k-9/ (http://johndelpizzo.wordpress.com/category/k-9/)
http://www.charitydalek.co.uk/about-our-k9/ (http://www.charitydalek.co.uk/about-our-k9/)
http://podpadk9.blogspot.com/2010/07/k9-power-at-last.html (http://podpadk9.blogspot.com/2010/07/k9-power-at-last.html)
The last link you really should start reading here (http://podpadk9.blogspot.com/2010/07/k9-in-begining.html)
-
Hi,
Gosh guys I never said skid steer would not work , of course it will work , I just don't see it as the best option .
We're all entitled to our opinions, of course, but if you look around the web, differential drive seems to be what they all use.
This system I have described only needs one motor for drive and one servo for the steering and of course an ESC for speed control.
A servo is a puny little motor with a small ESC in it - good for some stuff, less good when the weight goes up and it doesn't change that two motors are involved.
The main disadvantage with differential drive, from a hobbyists POV, is probably that you need two identical motors, which is rare to find in a dumpster ;)
I refer skid steer to any thing that does not have a set of wheels that point in the direction you wish to go, it comes from many years of working around such machinery .
Hmmm, according to that, my laptop is skid steer (since it doesn't have a set of wheels pointing in the direction I want to go) ;)
Seriously, I sometimes call an Acme screw a Ball screw, but that doesn't make it a Ball screw and I have to correct myself now and then.
"Skid steering" is a member of the family "differential steering", not the other way around and in engineering, things go way smoother if the right terms are used.
What I am seeing is the OP looking to make a box that is a lot longer than it is wide and putting a wheel in each corner pointing in the long direction of the box and by varying the speed of either side they cause the object to move in that direction , like some one said that is great if you wish to make a tank .
I posted some links to K-9 builds - Skidding's Not Included.
BTW Soren , you was the first to mention differential steering TYVM so why put me down for mentioning it ?
If you re-read my post, I'm sure you'll see that it was your mention on using (or not) a differential, which is quite another story and hard and/or expensive to do well small scale (even a lousy ball differential is expensive when the box says Graupner or whoever made it).
-
OK I feel like we are flogging a very tired horse now but then that is what forums are all about
Yes I know differential drive gets used a lot and if you are using tracks then this is really the only option.
I still don't see how what I have described could be considered complex for the OP and it does not require the need for a complex and I agree an expensive differential.
All that is need for a diff is an axle with one wheel fixed to the axle and the other wheel free on the axle this will allow both wheel to rotate at different speeds when going around a corner , connect a motor to the axle and there you have your drive complete with diff .
Steering only needs to be a simple (single) wheel at the front ( or the rear) connected to a servo to turn the wheel in the direction you wish to turn , there is no need to have two wheels for the steering and if the OP has plans to move K9s head and antenna with servos then he/she should be able to rig up a very simple steering of this type .
Soren as for you laptop well I guess if you like to push it around like a toy car them it would be considered to have skid steer ( I am just wonderingt that when you do this do you make brmm brmm nioses as well ) .
Terminology will always be a problem, people from different country's say things differently and have different meanings, I shudder when I here some one measuring in inches and feet witch to me are very incorrect terms in engineering .
I would also like to add some more that is a bit of topic but I feel it needs to be said , from the previous post I feel as though I was literally ganged up on and bullied buy two long term members of this forum who egos are a reflection of there membership status on this forum , I really don't care how many post you have on this forum it does not make you SOR gods.
I Have worked in the engineering industry for over 30 years . My qualifications include welding ,metal fabrication ,machining ,electrician ,mechanical fitting and I am also currently in my 3rd year at the University of Western Sydney (that's in Australia ) doing my engineering degree with a major in mechatronics and robotics , I am on this site for the sole reason that I think I can contribute a lot to it , I like to think I know what I am talking about when I advise on something but as I am not a robot my self I do make mistakes as my knowledge is far from perfect .
Please guys let not let this forum be like so many other internet forums were you get a few individuals who think that just because they are long term members that there word is all people should listen to and then turn nasty to any one else who offers an opinion that they don't like ,
sadly I can see that this is starting to happen to this forum .
Cheers guys
Mark
-
Hi,
Yes I know differential drive gets used a lot [...]
I still don't see how what I have described could be considered complex for the OP
Differential drive:
One gear-head motor each side, both identical.
i.e. build one, then duplicate.
"Front wheel drive"
One gear-head motor on the back
Servo on the front
i.e. build one, then build a very different other one.
Shouldn't be that hard to see the difference in complexity.
Perhaps there's a reason that differential drive is used so much.
Personally, I'd only use diff. drive for very particular applications, but I try to consider the posters situation, rather than my own well equipped workshop and parts store when I give advice.
4 individually powered and steered wheels would probably be my choice for something like a K-9 - just think of the tricks that would enable it to do :)
All that is need for a diff is an axle with one wheel fixed to the axle and the other wheel free on the axle this will allow both wheel to rotate at different speeds when going around a corner , connect a motor to the axle and there you have your drive complete with diff .
That's not a differential, just a one wheel drive with a "horse carriage" wheel at the other side.
The cons of that is that it just takes the driven wheel loosing touch with the pavement for the vehicle to be effectively stopped.
While an open (i.e. 0% lock) differential have some advantages, they also comes with disadvantages and the wheel slippage is a major one.
We have a cargo bicycle with a differential on the driven front wheels and when there's snow or wet grass or leaves under one of them, there's no forward movement at all, as the wheel with the low friction contact will just spin and all power is lost there. I'm considering taking it apart to see if I could make it perhaps 50% to 80% locked for that reason. AFAIK the ball differentials used in R/C racing is highly trimable in that respect and you can get differential oil in a wide range of viscosities (as this is one way of semi-locking a diff.)
Steering only needs to be a simple (single) wheel at the front ( or the rear) connected to a servo to turn the wheel in the direction you wish to turn , there is no need to have two wheels for the steering and if the OP has plans to move K9s head and antenna with servos then he/she should be able to rig up a very simple steering of this type .
He should be able to, but I don't think that is the best solution for him. You may think otherwise and fine with that, but don't forget the OP in the process to forward what you like.
If you don't like the idea of people having other ideas and/or pointing out when your idea has a weakness, you shouldn't post in a forum, as this is where ideas are clashing and with some luck, all parts in a debate moves on with further ideas, refined by having to argue for own ideas and perhaps tempered by new insight, appearing out of the dissimilar ideas clashing.
Soren as for you laptop well I guess if you like to push it around like a toy car them it would be considered to have skid steer ( I am just wonderingt that when you do this do you make brmm brmm nioses as well ) .
Nah, I go wheeeeeeee wheeeeee (audible overclocking is legal in laptop racing).
Terminology will always be a problem, people from different country's say things differently and have different meanings, I shudder when I here some one measuring in inches and feet witch to me are very incorrect terms in engineering .
Yeah well, you use what you were brought up with and have a feel of instinctively translating into referenced magnitudes. To you and me, eg. 35mm gives an instant mental picture, while 1-3/8" may need a churn through the translator before that image appears. To people brought up with imperial measures, it's probably just the other way around and frankly, it only matters when people of different background (imperial/metric) has to collaborate. In my daily life, I use centimeters a lot and that is not really an SI unit either, but I bet all metric system users understand it without even thinking about it.
I would also like to add some more that is a bit of topic but I feel it needs to be said , from the previous post I feel as though I was literally ganged up on and bullied buy two long term members of this forum who egos are a reflection of there membership status on this forum , I really don't care how many post you have on this forum it does not make you SOR gods.
Somebody disagreed and since we were two, we ganged up on you??
If you cannot bear being corrected, don't make wrong postulates like eg.:
Differential drive and differential steering are totally different things.
Or would you rather that wrong information should be left unchallenged, to dumb down the world?
I can only speak for myself and I'm not a pack runner, so perhaps you should pipe down on that - that two people agree to disagree with you could mean that you were wrong, rather than that they were ganging up on you. What a crybaby statement, if you pardon my French.
My "membership status" has nothing to do with what I write and I don't give a rats behind about such things as post count, since it doesn't tell nothing but how many posts you made and I've seen lots of dweebs with a high post count and nothing towards content.
There's only one "God" on SoR and that's John/Admin/The Creator of SoR, hope that's not too much of a problem for you.
Nobody else has ever claimed to be a "God" AFAIK, so you'll have to take that on your own cape, as I cannot really be held responsible for your interpretations and how you see other people.
I'm here to share knowledge, not to up my post count and usually, people mentioning such things are the ones that are so very much into getting a high post count, a high "status" or whatever, but this really don't give me no buzz.
Judge people by the insight they share, how well it helps the ones asking and such, this will give you a better perception of their "value".
Some people will rather judge people on how "nice" they are, how many times they write "please" and how little they disagree with them, but hey, this is about engineering and I value knowledge sharing immensely higher than babying people with time wasting nonsense.
Further, I have found, that (very often), when correcting people in a "nice" way, they forget in five minutes flat and you have to repeat yourself. Bring "the whip" and you create long term memory.
I'm neither a nurse nor a kindergarten teacher, I'm an engineer (24/7).
I Have worked in the engineering industry for over 30 years . My qualifications include welding ,metal fabrication ,machining ,electrician ,mechanical fitting and I am also currently in my 3rd year at the University of Western Sydney (that's in Australia ) doing my engineering degree with a major in mechatronics and robotics , I am on this site for the sole reason that I think I can contribute a lot to it , I like to think I know what I am talking about when I advise on something but as I am not a robot my self I do make mistakes as my knowledge is far from perfect .
Yes, people do make mistakes. Some people learn from them and moves on with a bit more knowledge in the "bank", while some just start whining.
It's relatively easy to not let all that many errors slip into posts, if you just restrict yourself to what you know is correct. But sometimes it's easy to get too eager and venture out on a limb, with half baked ideas that, at the instant, appeared sound. Just take a deep breath every now and then (and especially when reading your post before clicking "Post") and evaluate whether it IS fact or your brain just went too fast for the Natural Laws to keep up.
Please guys let not let this forum be like so many other internet forums were you get a few individuals who think that just because they are long term members that there word is all people should listen to and then turn nasty to any one else who offers an opinion that they don't like ,
sadly I can see that this is starting to happen to this forum .
People should only listen to what sounds like it was grounded on reason and nothing else, but I'm sure you'll call me nasty when I tell you that I think this is another crybaby statement.
Offer all the opinion you will, but trying to muzzle disagreeing ideas by playing the "they're out to get me" card isn't worthy on an engineering forum - use knowledge instead.
Even better than opinion, offer ideas backed by verifiable scientific proof that your opinion is the best and I'll surely adopt it.
Less crybaby, more sound engineering, PLEASE :)