Help please w/ electromagnetic slip clutch

Have you looked at automotive aircon clutches. I don't know if they're suitable and normally used for on/off operation but likely cheap to get one from a breaker to test or play with.

Not enough information.

These handwheels are to be used to control something, for exercise, what?

I'm pretty sure that when you say "turning ... that need ... turning" you mean you're making things on a lathe that need resistance to someone twisting them.

Does the resistance need to be viscous, tapering off to zero as the speed tapers off? Does it need to be constant down to zero speed? Something else?

The excruciatingly high-tech way to do this is with a direct-drive motor, position sensor, and a fancy control system. That's overkill for a whole lot of applications, but I've done it for something that really needed it.

Greetings David, Those clutches are way too large. But thanks anyway. Eric

The handwheels turn an encoder that makes a servo motor turn a leadscrew. I want to monitor the servo current draw and use it to put a drag on the handwheel. So more work for the servo makes the handwheels harder to turn. I want to be able to feel the machine working. All the way down to zero speed. I thought a hysteresis brake would work until I found out about the cogging effect. I though about using a BLDC motor kit I have to make a skewed rotor to avoid cogging but I'm not sure if it would work, and if it did would it be fairly linear. Eric

Hysteresis ... How about a small PM DC motor hooked to a load and geared to turn faster than the hand wheel , might reduce or eliminate the cogging . ?

What you want is a small servo motor controlled by a servo amplifier. The torque is proportional to motor current. Just take the current monitor signal from the servo amp, amplify and apply to the handwheel motor.

Trying to do this with clutches or brakes will lead to great frustration. Only the most exotic clutches and brakes are very smooth, and otherwise would put a lot of stick-slip friction into a systemw here you really DON'T want that.

If the main motor is a DC brush servo, it is possible that a resistor can be put in series with it, and the handwheel motor connected across that resistor. Then you would not require a separate small servo amplifier.

Jon

I'd call it a 'generator' because that's what it is when it dumps power into a load resistor, but yeah, this is a sensible way to go. One could also consider using an AC motor with variable DC exciting current to act as a kind of adjustable brake.

Oooh. If you want to sound trendy, you're building a system with haptic feedback. That'll get you a _lot_ more funding than "a servo motor that you can feel when it sticks".

You can get motors with low or no cogging torque. I'm not sure if you can get them _cheaply_, but they're out there. Coreless motors are most definitely of this variety.

Gear a DC motor down, not too far that the drag of the gear train or the inertia of the motor is noticeable. Then run a current through the DC motor that's proportional to the current draw of your servo motor.

The 1940 way to do this would be to use DC motors for servo drive and haptic feedback, select them carefully, and just connect them in series. The torque on the feedback motor will, automagically, be proportional to the torque on the drive motor, without you wasting too many vacuum tubes or relays (since this in 1940) on making it so.

You'll probably have all sorts of bandwidth issues that'll be easier to settle if you use fancy electronics and whatnot.

I dunno Terry, You idea sounded good until I tried it a little bit ago. with a couple BLDC motors I had laying around. I just couldn't get enough drag, even with fairly high speeds and high performance motors. Your idea has merit but not with my constraints. Cheers, Eric

Greetings Jon, The brakes and clutches I have seriously considered are non-contact devices so the stick-slip problem is one I wouldn't need to deal with. I thought about using a small motor as the brake but n oit the way you describe. The main servos are DC brush servos and will be using 20 amps max for about 4 minutes max, with maybe a 7 second period of a 4 amp draw and then 20 amps again and so on. There will be peaks of much higher current draws, maybe 80 amps, but these peaks will be quite short and the servo amps can handle the loads. Can you explain a little about how the resistor in series with the main servo and the handwheel servo, which supplies the drag, connected in parallel with the resistor, would work? Thanks, Eric

A/C clutches aren't made to slip at all, David. This won't work. Band clutch/brakes have been used in auto tranny setups and on old emergency brakes for early vehicles.

Eric, band brake setups might work for you, with solenoid actuation. Variable voltage to the solenoid could facilitate varied braking.

Something like this. But done a teeny bit more professionally?

Greetings Tim, I am indeed designing a haptic system. The mechanical parts I understand pretty well but the electrotonical and automagical stuff I need a little help with. Terry suggested the gear train DC motor idea but I don't think it will work for what I am trying to do. Please correct me if I am wrong but it seems to me that with the system you describe above the servo connected to the handwheel will have the same current passing through it as the servo that is doing all the work. Is this the case? The handwheel torque resisting device will need to provide, at most, 40 inch pounds of resistive torque, and this will require little current, or more accurately, few watts. Less than15 watts on a continuous basis.The servo motor that the handwheel is controlling, the one that is actually doing the work will oftentimes be consuming 1800 watts. Would this work with your scheme? As mentioned in my original post each handwheel will be connected to an encoder. These encoders will be providing the information needed to control the position and speed of the main servo motors. I am not considering something like a SloSyn system. If I turn the handwheel such that an axis should move .0001" inch I expect that it will. I can, and have, done this in the past. Used a hand wheel to turn an encoder to provide a signal to a servo amp connected to a servo motor connected to a ballscrew connected to a slide in order to move it in .00005" increments. What I haven't done is to add drag to the handwheel that is proportional to the load on the servo. Even if the drag on the handwheel isn't exactly proportional the positioning will and must always be correct, the encoder to circuitry to servo motor cannot be affected by whatever is used to add drag to the handwheel. Cheers, Eric

Can you explain a

Well, a very small resistance would be put in series with the main motor. This resistor would be of much lower value than the DC resistance of the handwheel motor. The handwheel motor would then be put across the resistor. Some portion of the main motor current would flow through the handwheel motor, producing a torque. Possibly, you'd also place an adjustable resistor in series with the handwheel motor to set the level of torque feel.

I really don't know if this will work, as when turning the handwheel fast that motor will generate a voltage that will be dumped in the resistor. But, since the servo system with the encoders will make the main motor synch to the handwheel, these back EMF's should be able to be equalized.

This would be a cheat that would allow you to do this without a bunch of electronics, like an extra, small servo amp. As I say, I'm not sure it would work.

Jon

Perhaps the mechanism from a force feedback gamer steering wheel would work?

--jsw

Greetings Jon, If it works it is not a cheat. The handwheel motor would need to be wired backward with respect to the rotation of the main servo so that it will oppose rotaton of the handwheel. I will investigate your idea. Simple and easy to implement. If it works. Cheers, Eric

I think that some gearing could be part of it, but not a lot. Probably no more than 10:1 gearing. If you don't use gearing then you'll need to find a big pancake motor with low or no cogging -- such things exist for aerospace applications, but they're expensive new, and it'd take a monstrous stroke of luck to find one surplus.

Yup.

Yup. The trick is to get a motor with a significantly lower torque constant than your servo motors. It takes some careful motor choosing.

It may not be possible without custom-wound motors (which would be the

1940 way to do it) -- if that's the case, then sensing the current to the servo motors and echoing it, suitably scaled, to the feedback motors would be the way to go.

I think you need to stop thinking in terms of "drag" on the handwheel and start thinking in terms of "torque". It's mostly the same thing, but not entirely quite.

handhwheel speeds are usually slow. gear or belt drive up to a DC motor and short it out with some type of bridge recitfier (for use in both directions) constant current setup. Even a cheapo hacked LM317 might work if you can adjust it with whatever your controls are. Hell, maybe just PWM it if you're using a computer. A single triac might pull this off.

If you want it to feel nice, use coreless motors as they don't cog as much. Yaskawa "Minertia" type servo motors are good too as they don't cog and feel REAL smooth when run backwards.

How about using DC to brake an AC servo? Higher the voltage the more resistance to motion.

How would that work? How are AC servos made? I'll look online. Eric