How to make speed of DC motor match speed of another?

this is a non-trivial problem. i used to work with a printer manufacturer and some of their printer models had two motors, one at either side of the printer, to move the mechanisms across the page. they used stepper motors to do it, and spent a lot of time and energy tuning them to work together, working out their step tables, etc. yours is a slightly different problem because you just need your "front end motor" to pull up the slack, so probably it is easier than their problem.

It might be easier to just have the motor provide a constant torque (within reason!) so that it maintains a constant force on the wire. This should be proportional to the power being drawn by the motor.

Watched on in amusement while a similar project was delayed by 2 months while they tried to run 2 motors in synchronism. Told 'em it's not that simple and wouldn't be happening but they'd a bit between their teeth, 2 months already down the tubes and "hey! we're precision engineers, we understand gearboxes and we've face to save". Eventually solved by a single motor, long PU toothed belt and timing pulleys. It's not straight forward. If you've just got to have the 2 motors, then maybe continuously overspeed the pull unit while feeding the power through a variable magnetic clutch. If wire is thin, could result in secondary stretch problems and extra work of needing to monitor tension and forming a local servo loop with the pull motor and mag clutch. Although I suspect what I know as "wire", is lots thinner than your stuff :) john

The forces don't have to be equal. Have the pull motor run faster, and allow the pull mechanism to slip. Ed

use ABS (absolute position using gray code) encoders on both motors of the same type and same ratio of swing etc.. 2 vector or alike drives the monitors each others "wait/syncro" signal. etc.. vector drives also work very nice for this. but all in all, if you want to hack it, a couple of gray code encoders properly aligned going into a PLC 8 bit data input cards you can force both motors to stay in perfect track of each other.

Gee, that sounds a whole lot like a MIG welder :)

I won't suggest oil as that's no be real useful in a mig welder,

can you get a teflon coated liners for your tube?

one way would be use a stepper motor with the coils switched by an encoder on the pushing motor.

Didn't they invent synchro's for that? Search for synchro/resolver/LVDT.

Example: in old marine radars they used to have two motors, one rotating the antenna and a second one rotating the deflection coil on the CRT to follow the antenna.

--DF

How about sensing torque on the puller motor, and using that to control the pusher motor proportionally? The higher the puller torque, the more push is needed at the other end. This might require some sort of PID control to keep from "hunting" due to the intertia of the system, but it seems a safer bet than trying to match the speeds.

Just a thought...

Bob Masta

Just get rid of the push motor.

Greetings Greg, Thanks for the suggestion. I had thought about that and it helps when others come up with suggestions that tend to validate my ideas. There are available adjustable slip clutches meant for constant duty. So maybe that's what I'll need to do. I'm trying to keep the drive as compact and as light as possible though. Eric

Gee Jasen, it seems like you may be reading my mind. It's kind silly maybe but it's a small welder and the weld shop says that even with the teflon liner the wire tends to tangle. With larger MIG welders there are pull systems, like the Cobra, that work great. I could even adapt one to my small unit. But they are expensive and heavy. So I'm trying to think of ways to make a cheap one. Mostly just for my own pleasure. Stepper motors could work. US Digital sells a chip that accepts signals from an encoder and outputs step and direction. I have one of these, an encoder, and a chip that accepts step and direction and will drive steppers up to a couple amps at 24 volts. But stepper drives are sensitive to electrical noise. And the encoder is slated for a different project. It looks like a slip clutch just may be the way to go. Eric

Think in terms of a reel to reel tape deck. A capstan and pinch roller can spool out tape at a constant speed with much less than 1 percent variation, while a torque motor takes up the slack on the output. The input to your tube could be set up like a capstan and pinch roller. The motor could be running at the desired speed and the pinch roller engages only when wire is needed. The output of the tube would again be a capstan and pinch roller but continuously energized to pull the wire. Depending on how much tension you put on the wire, a brake may be needed at the input. A simple AC motor running at less than full power should work at the output side.

similar problem to cassette tapes. In those, the speed of the tape is controlled by the capstan and pinch roller but the wind-on spool runs much fatser, the speed is maintained by tension on the tape and a slip-clutch on the wind-on spool drive. you can see this by "playing" the cassette mech with no tape in - the wind-on spool spins rapidly.

I would go down a similar route with the pull wheel running faster but with a slip-clutch. This way it applies as much tension as you need but you have no sync problems.

The slip-clutch is fairly easily constructed by using a belt drive onto the wheel axle (best onto another wheel mounted so as to get the gearing right). The amount of tension can then be easily adjusted by tensioning the drive belt, but if the wheel is stopped (i.e. by your hand) the belt slips on the axle.

Get an old cassette mech and play - it'll give you a load of clues for this

The slip clutch is the way to go. I'm gonna build one using a wave washer for a spring and aluminum bronze against hardened steel as the friction surfaces. Eric