Hello everyone, I'm very new to robotics and I'm trying to use stepper motors as a means for propulsion for my project (it's a micromouse!). I bought two of these motors a while back ago and I would like to use these:
formatting link
I just e-mailed them to see if I could get a more detailed spec sheet since I can't find the part on the manufacturer's website and the the store does not provide much info to begin with.
So I did a little research and I wanted to see if this solution could work. I will have a mircontroller sending pulses to a translator (Allegro
3967 :
formatting link
) The IC spec says it is a translator/driver but I can't really imagine a soic package being able to dissipate that much power... should i get a darlington array as drivers? I have one in mind already: ULN2003
You don't say which motors you are intending to use, but even the smallest on that site is perhaps a little large for a (micro) mouse. Its not just the motor, but the batteries that you will need. Usually something like a micro mouse uses small dc motors and some gearing - often bought as a combination - google 'gearmotor'.
well, we're using the nema17 stepper motor shown on that page. We already have most of the physical hardware done and it fits well inside of the maze cell and leaves plenty of room for maneuverability. I tried DC motors before and I felt that it didn't give us the accuracy desired; and also dc motors that are geared and have a feedback tach on it is a bit pricy and out of our budget... But yea, any thoughts on the IC's that I'm planning to use? Any other recommendations?
Why do you need high accuracy ? The mouse needs to be guided by its environment (ie walls). You need some feedback to make it go in a reasonable straightish line, but this is usually done with a simple encoder disc on the wheels. The other properties of DC drive include high torque for starting and stopping, lack of resonance problems that are one of the curses of steppers, and finally a lower power consumption - the steppers will always have at least one phase on.
and also dc motors that are geared and have a feedback tach on
You don't need a fancy tacho feedback (see above). You don't need a gearmotor if you can make a simple 'gear' train. In fact, a toothed belt reduction drive is often used for steppers or dc motors.
You need to establish what current you need through the steppers to obtain the drive forces you require. If the motor is spec'd at 4v 0.9A that means a maximum of 0.9A per phase, and also implies that each phase is about 4.5 ohms. However, you cannot necessarily actually use it for longish periods at that drive level because of heating.
The datasheet you haven't got will tell you more hopefully. Look for a figure for the 'holding torque'. This essentially gives you the maximum torque that can be applied before the motor 'slips'. You can use this number (perhaps 5 milli Newton-meters for your motors) to work out how much drive you can get, and hence how fast you can go - and also how fast you can stop or change direction.
Power dissipation could indeed be a problem for the drivers. The output drivers on the '67 are transistors and will be dropping something like 0.6v (lo side) and 2v (hi side) at a per phase current of (say) 800mA. So you are looking at some 2W per phase. That is, as you say, a lot of heat to shift from a smallish package.
I would look for a switchmode driver/translator with DMOS or FET outputs. If these drivers are 0.5 ohms Rds, then the same current (800 mA) will generate some 800mW per phase.
Note the total power consumption is quite high - average 1.6A, maximum 3.2A, assuming the same 800mA per phase.
If you want to go 'fast', you will need to apply much more than 4V.
Stepper drives look simple at first glance, but in reality DC motors are far easier to 'tame'
Well, for our old design we had two of these motors (the second one down on the page)
formatting link
they work pretty well but i wasn't too sure on designing my own motor control, and i would really like a pair of reliable encoders that are not expensive at all. I have seen the hacks where i can take a mouse apart and use the encoder wheels, how accurate would you say those are?
The reason I want accuracy is to know exactly where my position is in the maze, and i want to design in software for position re-calibration to correct errors.
so yea, can i get some more info on getting this dc motor up and running? Thanks for all your input dave!
This is probably a good point to suggest that you read a book or two. For instance "Building Robot Drive Trains" by Clark and Owings. Gordon McComb has also written a number of useful titles, and can often be observed lurking here ...
In the end it all depends on whether you want to make your own basic mistakes, or stand on the shoulders of others and make some really interesting mistakes !
I haven't looked at the specific motor driver you are planning to use but here are some comments that are worth what you paid for them :-)
I've had a lot of experience with stepper motors in micro mouse with a lot of success.
The main things I've found are:
Drive them with lots of volts and regulate the current, a good drive IC should do this, this is how you get the speed. I used the L6219 to drive 5 volt steppers from 50 volts. This allowed me to reach speeds of 4 M/S
Set the current limit so the impulse acceleration on each step doesn't cause micro slipping. Soft tyres help here.
Use conservative ramp profiles for your speed to prevent "cogging" or "drop out". You can't just apply full speed to you stepper when it's not moving.
While the other posters are correct that DC motors are a "superior" solution I like steppers for their easy entry level.
If you connect a stepper to a micro you can go forward 10 steps and back 10 steps with a few lines of code, so long as you make the steps slow. To do the same with a DC motor requires a lot more code and skill.
I agree with Dave about the disadvantages of using steppers. They were my first try for a micromouse, and it wasn't pretty. Yes, they were easy to control, and driver dissipation wasn't much of a problem (I just used a bunch of open-collector TIP120s on a pair of unipolar motors).
However, they sucked current like crazy!. And if they were turning fast enough *not* to suck current, it was because the drive voltage was too low. In short, they were just too big, heavy, and mostly inefficient. Plus, the power/weight ratio was just too low to make a competitive robot.
DC motors turned out to be much better. They're available in such a wide range of sizes, from tiny pager motors (that are really not very robust), through camera drive motors (quite useful, although they require quite a bit of reduction gearing), and on up to slot-car and RC car motors, which are more than powerful enough for any practical micromouse.
Surprisingly (to me, anyway), they turned out to be easier to control, too. Instead of pre-calculating how far to go, along with the acceleration/deceleration ramps, I just used sensors to navigate (with wheel counters as a rough approximation for dead reckoning). Much less software overhead, and since I was using a lowly 8051 processor, many more cycles available for maze solving.
Just my personal experiences. You might take a look at Peter Harrison's pages at micromouse.cannock.ac.uk
-- Mark "I prefer heaven for climate, hell for company."
PolyTech Forum website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.