Hi, for an upcoming project I'll need to control two DC-motors in both speed and constant-torque mode. For a bit of a background, I'm building a reel-to-reel system to move a tape back and forth several times. I'm also using a rotary encoder to measure the position and speed of the tape. Each DC-motor is connected to one of the reels. So it's similar to an audio reel-to-reel player or audio cassette player, but larger and faster.
My plan is to control the speed of the DC-motor pulling on the tape based on the rotary encoder measurements, while the other DC-motor applies a small, constant torque in the opposite direction to keep tension on the tape. This tension is important for our application, but doesn't need to be super accurate. Once fully spooled into one reel, the processed is repeated exactly the same in the opposite direction. So it needs to be symmetrical. Both motors need to be able to switch between speed control and constant-torque control modes.
I'm a control engineer, so I don't need help on the controller design nor the programming side of this project, but I'm a bit clueless on the electric side.
If I understood my research correctly, assuming I'm using a full-bridge driver like the "Arduino Motor Shield Rev3" then:
To run in speed control mode I must vary the duty cycle of the PWM signal based on the measured tape speed using for example a P(ID) controller.
To run on constant-torque mode I must vary the duty cycle of the PWM signal in order to keep the current measurement constant.
I think I got this right, but I'm not entirely sure. Thanks in advanced.
Bonus question: The motor with constant torque will be forced to spin in the opposite direction it's trying to spin. This is fine right? Are there maybe some types of DC-motors I should avoid?
I don’t see how you can control speed and torque .If you drive a tape at a certain speed then it will take a certain torque to pull it through at that speed . If you reduce the available torque then it must slow down until the torque available matches that absorbed by the load . Can’t see why you want to do both modes ??
PWM current control is not easy , especially with motor drives where inductance and back emf distort the waveform - making it very hard to measure the current with any accuracy and in sync with your voltage PwM.
You mention tape recorders , have a look how these and the computer tape controllers work and how a sudden stop is handled and how changes in the diameter of the tape spool are handled ( which affects The speed of the dispensing spool and the back emf from that motor , making constant torque tricky )
.
That's fine, especially since you aren't using full power-torque. Of course constant torque doesn't mean constant tape tension. The more tape on the reel (the bigger the diameter) the less tension.
I don't know how slow-speed DC motors work but I believe direct-drive turntables use a DC motor (with plenty of inertia) and that's even slower than a capstan.
Tape machines often use a simple arm and wheel to provide a bit of friction to stop any slack . Look at magnetic clutches and brakes - a larger system may need a lot of effort to start stop quickly without damaging the tape.
Building the mechanics of this is the challenge and needs some sums doing .
Well worth a study of the state of the art , before re inventing the wheel !
In the old days they would use induction torque motors to drive the spools. Apply high current to one and low to the other to fast wind, operate in torque mode for tensioning when the capstan controls the speed for play/record. In effect this is what you propose I think? Computer tape decks used the same approach as you I think using DC servo motors - maybe you could source a surplus one for your mechanics and motors? There must be millions of them giving good service in scrap heaps around the world. Both motors are driven in the same direction so I don't think you need an H bridge driver do you?
Thankfully I'm not reinventing the wheel. These reels are actually meant to be used on a machine that as far as I understand does the same. It has two DC-motors and a rotary encoder. One motor pulls onto the tape while making sure the encoder is measuring a constant speed, while the other one applies a constant counteracting torque to keep tension on the tape.
We want to do some preprocessing on the tape before putting it into this machine, so I'm trying to replicate the same system. We already had some success using steppers, but the steppers can't apply the counter torque for the tension, so we have friction clutches that we can adjust to either be too low for it to be useful, or too high for the steppers to pull on them without skipping while accelerating. The acceleration is adjustable, I have played with it to see what I can get.
So one option is stronger steppers and keep the clutches, or we might as well switch to stronger DC-motors. They are better for this application where I'm more interested in running them at a certain speed with highish accelerations. I can then also try out this counter torque using the DC-motor, or fallback to the friction clutches on a high setting. It'll probably be some combination of both if I get the DC-motor counter torque working.
The torque of DC-motors is proportional to the electric current. So, as I said in my post, I'm planning to use the current measurement in my feedback loop instead of a torque measurement. It's easier to measure current than torque and from what I can find, lots of DC-motor driver boards already have current meters on them.
I'll probably do some experiments to get a "conversion" factor and check if it's proportional enough for what I want to do. Thankfully I don't need this torque to be accurate anyways, the tape is quite strong and just a bit of tension on it is enough to fulfill its purpose.
The motor pulling on the tape will run in speed control mode and the other one will apply an opposing torque to keep some tension on the tape. Once fully spooled into one reel, the process reverses and the tape is spooled onto the other reel. So, one motor doesn't run in both speed and torque control at the same time. But they do need to be able to be switched between these two modes.
Fortunately I'm not starting from zero. I added a few more details in one of my previous comments, but in short I already have the setup kinda working using some steppers we had lying around, but want to switch to some beefier DC-motors to keep some tension on the tape using the friction clutches we already have. This entire setup is also based on an existing machine we have in our workshop.
But since we are now switching to DC-motors I want to try using the second motor to keep tension on the tape. The clutches need to be set quite high to keep tension while stopping, so acceleration kinda sucks, the tension on the tape is quite high while accelerating and the motor pulls more energy than necessary during operation.
If I can get this system working with the second motor applying a counter torque than I'll essentially have a virtual, adjustable clutch or brake I can use to adjust the tension on the tape to whatever is needed. I'll probably have to play around with some control schemes to get everything properly coordinated while starting and stopping, but I have some ideas for that.
As for the current measurement, I'm just gonna buy a DC-motor driver board with current meters on it. It'll be better that whatever I can come up with and hopefully be accurate enough. The torque doesn't have to be accurate anyways, it will be much smaller than the motor's rating. And if the signal is too noisy I can filter it.
EDIT: So to be clear, I'm not controlling based on the current of the PWM signal itself. The driver boards accept a PWM signal to adjust the speed of the motors and then they supply the motor with some higher voltage. The current meter on the board then measure what's being supplied to the motor. I can then use this current to control the torque since they should be proportional to each other.