Indexing conveyor how-to

I am thinking about designing an indexing conveyor for a future project and am wondering what is the best way to drive it. Basically, all the conveyor needs to do is advance a set distance, probably about 15 inches, over a 5 second interval, then stop for 1 second, then advance the same distance over the same time interval, stop for 1 second, over and over again all day long.

What type of motor would be best for this? I assume a regular AC motor wouldn't last very long starting/stopping that often? Would this be stepper motor territory? The distance advanced is critical, as it needs to repeatedly advance the same distance every time. Should an encoder be used for feedback? Anybody have experience with this type of conveyor?

Thanks

AC motor in positioning applications is always a compromise between the positioning speed and precision. The motor life is not an issue.
Based on your numbers, AC motor looks OK.
As an alternative, I would consider servo system in positioning mode, but not a stepper.

a stepper might work

RH68,
I think a stepper could be a good solution for you, depending on the mass of the conveyor and product being moved. Steppers are usually run open-loop since they are advance-able by a known amount per step or micro-step.

See below for a good description of how steppers work:
http://en.wikipedia.org/wiki/Stepper_motor

If a stepper is used without some type of position feedback, there's no way to know if the motor has reached position. There are two ways to deal with this - either use an encoder (or other position feedback), or oversize the motor to an extent where there's a reasonable assumption that the system will always hit its target.

Personally, I'd use some type of position feedback, even if it's not an encoder. A prox sensor might work, depending on the precision you need. I like steppers, mainly because they are much cheaper for their size, as compared to servos, and they have great torque at low speed, even without reduction.

If you are interested in a stepper, try giving the folks at Lin Engineering a call.

http://www.linengineering.com/LinE/

They have a great selection of motors and drives, and can help you pick the right motor for your application.

Try to use a motor voltage as high as possible. Steppers loose torque as speed increases, and a higher motor voltage allows the motor to generate more torque at a given speed. The drives at Lin Engineering have a max voltage of 80vdc. I use 72vdc power supplies, and I've had good results.

Keep in mind that a typical stepper motor drive is "dumb." That is to say, it does nothing more than switch power to the motor. You need to supply a separate source of DC voltage, as well as pulse and direction signals, typically from a PLC. You'll need to interface the outputs of the PLC (typically 24vdc) to the inputs on the drive (typically 5vdc). PM me if you're intersted in using a stepper - I've designed some cards that make this hookup easy.

Watch out for microstepping. It's great, as it makes the motor run much more smoothly and quietly, but it might make the motor turn slower than you might expect. ie. your PLC outputs a 7khz pulse train. You might think that yor motor would spin at 7000 pulses-per-second. But with x10 microstepping, the motor only rotates at 700pps. Some more advanced drives give you the advantages of microstepping, but still allow you to maintain full-step speeds. This "got" me the first time I used a stepper.

Note, I don't have any connection to Lin Engineering, other than that of a satisfied customer. I've just used their stuff before, and it worked really well.

Good luck, and let us know how your project goes.

-rpoet

RH68 said:

Basically, all the conveyor needs to do is advance a set distance, probably about 15 inches, over a 5 second interval, then stop for 1 second, then advance the same distance over the same time interval, stop for 1 second,

Click to expand...

RH68,

You have partially defined the movement. We do not know the precision required. But most importantly you have left out the amount of force to move the load and to accelerate the load and lastly the amount of force needed to stop or hold the load.

This is one reason you have received two opposing opinions previous to my post.

When doing tech support I use these formulas for estimating

milldrone said:

RH68,

You have partially defined the movement. We do not know the precision required. But most importantly you have left out the amount of force to move the load and to accelerate the load and lastly the amount of force needed to stop or hold the load.

Click to expand...

http://www.plctalk.net/qanda/showpost.php?p=267763&postcount=8

The motion equations don't address the load part but I have found the final velocity and acceleration usually don't vary much from the estimated values unless there is a server velocity or acceleration limitation. If you do the calculations you will see the velocity and acceleration is rather low. If s curves are used the peak acceleration must be multiplied by a factor that is 1.5 to 2 times the average acceleration depending on the s-curve used. The accelerations are still small.

To give a little more detail...
I would like to have positioning repeatability of +/- 1/32" Also, this will probably not be a conveyor with a belt, it will probably be two chains with pusher bars running between them at certain equal spaces to move the products, which are very light plastic trays. I was thinking I could use a proxy to detect the presence of the bars that push the product, that way the conveyor can always re-home itself. For example, when the move starts, the proxy detecting the bar in its home position is made. It then moves 15", at which point, the next bar on the chain is now making the proxy. Acceleration/decel can be quite low, since it will only be advancing about 15" over each 5 second interval. Additionally, this conveyor will probably be tied into another existing conveyor that is doing the same motion. My conveyor would start each motion when it is told to by the other conveyor so they stay in synch. I am estimating the total torque requirement to be about 70 in*lbs.

Thanks

RH68 said:

I was thinking I could use a proxy to detect the presence of the bars that push the product, that way the conveyor can always re-home itself. For example, when the move starts, the proxy detecting the bar in its home position is made. It then moves 15", at which point, the next bar on the chain is now making the proxy.

Click to expand...

What if a pusher bar breaks and falls off?

RH68 said:

Acceleration/decel can be quite low, since it will only be advancing about 15" over each 5 second

Click to expand...

This what Peter said.

RH68 said:

Additionally, this conveyor will probably be tied into another existing conveyor that is doing the same motion. My conveyor would start each motion when it is told to by the other conveyor so they stay in synch.

Click to expand...

Now there is an important piece of the puzzle.

Edit: Can you define how "tight" the sync needs to be?

RH68,

Another question: Can you describe the "other conveyor" drive?

Keeping things in synch is easy if one controller controls all

See this recent post about synchronizing on another forum
http://forums.mrplc.com/index.php?showtopic=15798#

I would try to use SSI absolute encoders set up to provide one turn per index. This way the controller always knows the position of the lines and their relative phase and the indexing is always one turn. This helps get the system in synch with out homing when starting after the power is off.

So far this application sounds easy.