High power servo for steering on sub 30Kg robot

[mowcius]

I am building a large robot which will be rear wheel driven, about 4' by 3', 30Kg (ish), with 10" wheels and steered on the front two wheels.

I am looking for a high powered servo for steering the front wheels. I have a stepper motor that would suffice but it would need a centre point sensor and I would rather use a servo.

I expect the servo will need to move about 5cm each way to move the steering bars for a decent turning circle (connected about 5cm from the centre of each wheel also).

Anyway, I read something about some cars using high powered servos for the windscreen wipers. Anyone know more about this? Anything I should look out for when I go down the scrap yard?

Anyone got any ideas? I am new to this but have a reasonable amount of experience in electronics.

Thanks,

Mowcius

[Soeren]

Hi,

Anything I should look out for when I go down the scrap yard?

Splinters!

You'd be better off making your own servo system.

[mowcius]

You'd be better off making your own servo system.

Any recommendations though?

I am prepared to make my own if necessary but I thought an automotive one might do just fine.

Splinters!

I'll be taking gloves :p

Mowcius

[Soeren]

Hi,

With the reservation that I don't know every detail of what's new in cars, I don't find it realistic that wipers should use servos as we think of them.
But, a servo is any circuit where feedback regulates position, so if they placed a sensor at each outer position and let that control the length of the sweep, you might call it a servo circuit.
A steering servo wouldn't be useable either, unless you're really aces at hydraulics hacking.

In short, I don't think you'd find something that are drop-in - I could be wrong though, but the automotive industry usually don't  add something that's not needed and there's nowhere a (beefed up) R/C servo couldn't be made cheaper by direct integration of the parts in a car (and in the numbers they're produced) - just look at the servo in the EFI, not what most hobbyists would call a servo, even if it is  (and a very good one indeed, I've thought of using some of them for their extreme resilience, compared to R/C-servos).

You might call around to some of the larger dealers or service shops and ask around of course - I'm en electronics engineer and my car knowledge will never be on par with an automotive engineer.


If you end up going the DIY route, measure the needed force to turn the wheels, at the place where your servo will be interfacing it (in kg, pounds or whatever, both when they're pointing straight ahead and when they're at the limits - the highest number will be the one to use).
Then you need to find out how fast you need it to travel from one position to another in degrees/s (eg. 45° in 0.5 s).
You need to decide whether you want to use rack and pinion (perhaps to reuse what you have) or you need something that's locking (like a worm drive), or something entirely different.

[mowcius]

Hmm, thanks for the reply, I will have a deeper think about it and see what I find at the scrap yard when I go down there.

Mowcius

[Razor Concepts]

If you would rather buy than build
http://www.pololu.com/catalog/product/1390

[uno]

I tried using a window motor on my 150lb robot but it just didn't have the power and the current draw was too high. Instead I used a powerwheels motor and a worm gear (3/4" drill bit, actually) and so far it has worked just fine, not overdoing my 10 amp relays and transistor at all.

Hopefully from my pictures you'll be able to see better just how I did it. It's not perfect yet with some strength problems but I'm working those out. High school has been a little too time consuming to get it all done!

[uno]

Here are some more pictures of the robot that wouldn't fit on the first post.

[mowcius]

I tried using a window motor on my 150lb robot but it just didn't have the power and the current draw was too high.

Ok, that's some helpful advice.

I like those tyres/wheels Where did you get them from? Are they expensive?
The best I have come up with are something like this: http://www.northerntooluk.com/tyres-and-wheels-and-castors/steel-wheel-assemblies/10in-pneumatic-tyre_wheel-knobbly_13431E.html?tab=specification


Mowcius

[uno]

10" wheels for a 4x3' frame? Sounds a little boxy. I'm actually using 14" wheels on my 4x2.5' robot and even that doesn't give me as much clearance as I would like. Unfortunately I can't just point to a online source for wheels because I bought mine locally--$60 for brand new front wheels, $10 for used back wheels off of a student. If you look around, however, you should be able to find some good wheel sets at stores that deal with go-karts/four wheelers.

BTW--have you figured out what to for the drive train yet? I'll tell you right now that the steering will probably be the hardest part for you when building it and you may want to save that for later after the drive train is done. It took my three tries to get the steering right. I got the motor to try it out last year, and now a year away from then, plus or minus a month, I have a working system with a totally different motor. If I didn't have a drive train working already and a good number of hours put into making the darn robot work I may not have had the determination to move on. It's not a task for the week of mind, that's for sure.

[uno]

Here is a project that may help you out. It was made by Ken Gracey, a very nice guy who works at Parallax. It is controlled by the SX and has an electric drive train, unlike mine which is controlled by the Parallax Propeller chip and has a drive train connected directly to the engine. Although it is different from what you're doing since it has differential drive, hopefully it will give you some ideas for your own robot.
http://www.parallax.com/ProductInfo/Robotics/OutdoorHybridRoboticPowerPlant/tabid/625/Default.aspx
http://forums.parallax.com/forums/default.aspx?f=21&p=1&m=151769

[mowcius]

Well I considered doing dual drive brake steering but in the end, as I want it to be fast, simple and easy to steer, I decided against that. I would also then need a second high power motor driver circuit and a second motor!

Mowcius

[uno]

Yeah, I thought about that myself after already starting my project. I thought about using two weedeater engines on each side and control the gas on them, but if you noticed from watching the video of ken's robot a single small engine is annoying enough. Personally, I like the low rumble of my straight-piped 5hp engine a lot better, even if I have to change my mechanics to make it work!

BTW, there is a way to use one engine and have differential drive. You would need to create two clutches to do it, though. From what I know there are two ways of doing it. You can make two disks with either a pneumatic or hydraulic system to force them together, thus making them turn together, or you can use pulleys to drive the wheels like in pushing lawnmowers, and if you have a motor that can put tension on that pulley it will force the wheels to start turning. Either way will probably be considerably harder than the system you propose, however, so if you try it good luck!

[mowcius]

My project is going to be completely electric as I am not a fan of engines :p
I would much prefer something that is quiet than something making unnecessary noise.
Maybe I'm not manly enough or something 

This has all given me food for thought though.

Mowcius

[amando96]

The rear wind shield wiper of cars does act like a servo, it has feedback control, and about 100º travel, if they didn't need so many amps they would be better

[mowcius]

Right. I will try one. As I am most likely going to be running on 2 car batteries or a 24V pack off a mobility scooter or something then I should be fine. I was also looking at just something like this:
http://www.pololu.com/catalog/product/1057/specs
Cheap, with 17Kg/cm stall torque. I am pretty sure that this would be powerful enough on flat ground with high pressure in the tyres when stationary but it might have issues on grass while not moving. When the car is moving, it wouldn't take so much force to turn the wheels so it might be strong enough. As they are so cheap, I will buy one and experiment.

Mowcius

[Ro-Bot-X]

I would still go for a car powered window motor or windshield wiper, as they use a worm gear box and they can hold the position without power applied. Most of the time you need the steering to be in a same place, so only for a short time when turning you need to apply power to the motor. So, even if it is a current hog, the window motor won't deplete your battery too soon. And it has lots of power, has a cable setup that can be used to pull the steering. If you need to sense the position of the steering, I suggest to use a 10k sliding potentiometer (linear, not logarithmic)  connected to an analog pin. Easy to set up and you have complete control of the closed loop, as if there is an obstacle that prevent the wheels to turn you can see the pot not moving to the desired set point and back up the car to escape the obstacle. The only costly part you need with this system is the motor driver. But depending on your needs, you may get away with just a couple of transistors and relays. If you want to use PWM and PID control for a fine turn control (you need this at high traveling speeds) then a 5 amp single motor driver would suffice for a powered window motor. I had a pair connected to a base and used relays for control, and I hopped on the base and it carried me (65kg) for a while until a wheel came off. The motors got hot with such a load, but not with just the robot's own weight (12kg).

Good luck with your project!

[mowcius]

Well I happen to have a 10K sliding pot and that seems like a good idea. I think I might buy a new one though as I am not convinced by how good this one is.

You guys have managed to convince me that maybe making my own feedback control and using a high powered motor is the way to go so I will see what I come up with as the project progresses.

At the moment, I have initial designs and ideas, it'll probably be a few months before I start making it (or maybe more), but I want to try and sort the design out and try and get some of the parts before.

Mowcius

[uno]

That's just way too funny. You call it a 30kg robot but you say it may have two car batteries? Those alone would be the weight of your robot! As it is on my own robot I'm using a battery half the size of a car battery, and it weighs 20lbs, or almost 10kg.

Also I wanted to make a note that your robot isn't going to be cheap with an electric drive train. I'd find what kind of motor you wanted to use first, because if it draws over 100 amps you're looking at over a hundred dollars just to control it--probably more money than the motor itself will cost. That's one of the reasons why I like gas engines--but ironically I still plan to use an electric motor on my robot for backing up! At least it won't be of the power you're probably thinking of making it plenty cheaper.

Contrary to what ro-bot-x says it doesn't matter that a window motor has a worm gear in it. It can still be turned by hand with enough force. But that doesn't matter either, because in no way will it be strong enough to do the task without being slowed down first. On my robot if I try and take my motors place and pull the cables to turn the wheels, I have a heck of a time because the mechanics aren't on my side. This makes it all the worse for a motor as weak as a window motor to do the job.

Also, ro-bot-x states a 5 amp controller will be enough to control it, but I would advise a 10 amp system. As it is my 10 amp system isn't quite what I want as it has no room for the motor to slow in rough situations. This is why I think it would be best that you had at least 10 amps, and with a 24v system hopefully it will work great.
Also, you made mention of a servo you thought would be strong enough to power the steering. Well, I hate to break it to you but under no circumstance will that servo be in the least capable of driving the steering on a robot that large. Here is something for you to contemplate; on regular cement my steering motor uses maybe 7 amps to turn the wheels. At 12v, that's 12*7 or about 84 watts to power the steering.  Now to power that small servo you would probably need to regulate your power supply to about 5v. 84/5 equals 16.8 amps to get the same results out of that motor as I get out of mine. To put it into perspective, most 5v regulators are only capable of outputting 1.5 amps, so to get enough power out of those regulators you would need 12 of them, needless to say that will fry your motor. Hope this helped you contemplate the complexity of what you wish to do.

Also, I want you to understand that in any place where you can say “I am pretty sure that this would be powerful enough” you probably need more power, and when you say “When the car is moving, it wouldn't take so much force to turn the wheels so it might be strong enough” I want you to understand you’re wrong, way wrong. At high speeds this is true because of the rpm of the wheels, but at slower speeds if the motor can’t turn the wheels while stationary it may only be able to pivot the end of it at an inch or so per revolution. With 10” wheels you have to travel 2.5 feet for that small amount of turning power. That means in order to turn the wheels 45 degrees you would have to travel 10 feet. Not very efficient, is it? Who knows, maybe you’ll even run into a tree by then!

Anyway, hope this helped and good luck.

[mowcius]

That's just way too funny. You call it a 30kg robot but you say it may have two car batteries? Those alone would be the weight of your robot! As it is on my own robot I'm using a battery half the size of a car battery, and it weighs 20lbs, or almost 10kg.

Ok, yeah I did realise that the batteries would take it over that 30Kg stated. I have measured the weight of some car batteries (the small ones) and it would take it to about 25Kg for the two of them. If I got the money I would be using a 10Ah lithium pack or similar which would mean I could keep it down to the 30Kg. That is seemingly unlikely though.

Also I wanted to make a note that your robot isn't going to be cheap with an electric drive train. I'd find what kind of motor you wanted to use first, because if it draws over 100 amps you're looking at over a hundred dollars just to control it--probably more money than the motor itself will cost. That's one of the reasons why I like gas engines--but ironically I still plan to use an electric motor on my robot for backing up! At least it won't be of the power you're probably thinking of making it plenty cheaper.

I have looked into this - one of the first things I was looking at. I am currently trying to get hold of a 6mph mobility scooter as I realise that with less weight and not carrying a fat person, I could easily gear it to do 20mph. Also if I get the money, I am looking at a 2000W (continuous, 4000W peak) RC plane motor which is ~£200 and the controller is ~£100 so yeah not cheap.

Also, ro-bot-x states a 5 amp controller will be enough to control it, but I would advise a 10 amp system. As it is my 10 amp system isn't quite what I want as it has no room for the motor to slow in rough situations. This is why I think it would be best that you had at least 10 amps, and with a 24v system hopefully it will work great.
Also, you made mention of a servo you thought would be strong enough to power the steering. Well, I hate to break it to you but under no circumstance will that servo be in the least capable of driving the steering on a robot that large. Here is something for you to contemplate; on regular cement my steering motor uses maybe 7 amps to turn the wheels. At 12v, that's 12*7 or about 84 watts to power the steering.  Now to power that small servo you would probably need to regulate your power supply to about 5v. 84/5 equals 16.8 amps to get the same results out of that motor as I get out of mine. To put it into perspective, most 5v regulators are only capable of outputting 1.5 amps, so to get enough power out of those regulators you would need 12 of them, needless to say that will fry your motor. Hope this helped you contemplate the complexity of what you wish to do.

Yeah, be gentle on the new guy  Thanks for the info. I have not done much with servos (well apart from hobby stuff moving a few hundred grams)

Anyway, hope this helped and good luck.

Yep and thanks,

Mowcius