[acannell]

okay i actually am not sure this would work..but i think it might..and if it did it could be really useful for anyone who wants to know precise spindle output power and torque, but their machine doesnt have any provision for it.

this should measure torque accurately even if no rpm signal is available. add the rpm signal and you get power too.

so i was looking at my mill and i thought of this. i am sure im not inventing something new, but i havent seen anyone make one of these, so its new to me.

the way my spindle is setup, its a motor, which drives a pulley, which goes to another pulley with a belt. the second pulley is directly connected to the spindle shaft.

i noticed the motor is mounted with two bolts. one appears to be spring loaded so that its not clamped so it wont move, but its held tightly with the two disc springs. the other bolt does clamp down tightly with no springs. so if you loosen the "clamp" bolt, you can pivot on the spring loaded bolt, and in this way, adjust belt tension.

so it occured to me that if you left the "clamp" bolt loose, the motor housing would want to rotate in equal torque and opposite direction to the applied torque!

my particular motor has a little 3" arm screwed into its housing on the same side as the clamp bolt, to grab when you are adjusting belt tension. i think if i turned the spindle on and held onto that arm, id have to put equal torque on it as the spindle was applying torque. so if i replaced me with a luggage scale from that arm to somewhere, i could measure force and compute torque!

now after a little more thinking, i think that since the pivot point is the other side of the housing and not the axis of the motor, it wouldnt actually work. so to make this actually work, id have to make some kind of plate which allowed the motor to freely rotate about its axis, but maintain belt tension (distance between pulleys).

ill have to think about exactly how this plate would work, but i think its a simple idea and shouldnt be hard to work out.

just because you are freeing up the rotation of the motor housing doesnt mean it will actually move. you can use a strain gage or load cell or luggage scale (which uses a strain gage) and it will all be "solid"..if you use a spring of some kind, it will move though.

so does this make sense?

the reason i call it universal is that pretty much any motor could have its output torque measured using this method. and since most of us can machine a simple plate, i would think alot of people could make their own adapter and rig up a very accurate spindle power/torque meter!

so now ill try to think of how the plate would need to be designed.

-must allow motor axis to spindle axis distance to be adjusted and locked in place (no freedom) for belt tension
-must allow motor housing axis rotation freedom (for torque measurement). no actual rotation is really needed, just freedom of rotation. (i suppose even with a strain gage there is some actual rotation, but its on the order of tenths of thousands)

heres my mill, which im sure is similar to many other mills. the tallest black box is the motor. the pulleys and belts are behind the grate with the zillion little holes in it. the spindle is also pictured

[keebler303]

You're on the right track.

Many of the commercial ones are just as you describe except with a water cooled eddy current brake as a load. The torque arm and load cell technique is widely used.
Attachment 207046

[acannell]

You're on the right track.

Many of the commercial ones are just as you describe except with a water cooled eddy current brake as a load. The torque arm and load cell technique is widely used.
Attachment 207046

thats good to hear.

i actually think this idea is not pie and the sky and could be fabbed up in a day or so..

but i think some crowdsourcing of the design here would help..ive never designed something with ball bearings and i have a feeling i wont make a very optimal design even if it works..

so heres what ive come up with:

the bill of materials will be:
"motor mount flange" (custom machined metal block)
two flanged spindle style (radial/angular) ball bearings, with ID much bigger than motor shaft..off the shelf. from what i see on ebay this will be about $15 to $45 total for two
"flange to bearing tube" (custom machined)
"main block" (custom machined)
small hydraulic cylinder/piston (custom machined) and off the shelf pressure sensor
OR
strain gage load cell (custom design)

so..drawings would be nice here but i dont have them made up yet..the way this works is:

the ball bearings get mounted in the standard back to back fashion in the "main block"..just like a ballscrew on a mill. their flanges prevent them from moving towards each other. otherwise they can fall out.

the motor gets mounted on the "motor mount flange"..which is just a square flange threaded screw holes for the motor, and a hole in the middle to pass the shaft through. it also has a threaded nose which is large OD enough to meet up with the top bearing inner race.

the preload tube is a tube with a threaded end. it gets installed from the bottom, through the bottom ball bearing, over the shaft, and threads into the "motor mount flange". the end of the tube on the bottom has a flange big enough to press on the bottom bearing innear race,so by tightening it, you are pinching the ball bearings together, and also locking everything together (motor mounted on flange, locking to ball bearings in main block).

so now its one big piece, and the motor housing can spin freely. the shaft is passed through to wherever it used to go.

now just mount the "main block" to the original mounting point of your motor.

so now we need a measuring device. something that will measure the torque expressed when the motor housing tries to turn.

i was thinking strain gage..which for this app might not actually be unreasonable..

but easier would probably be a small hydraulic cylinder, connected so its tangent to the radius of the motor mount flange. so when the flange tries to turn, it pressurizes the fluid, and then you just need a standard off the shelf pressure sensor or gauge. do a little arithmetic and you have torque!

add a couple safety devices like a stop so if the cylinder breaks the motor housing doesnt spin out of control.

you could use the electrical signal from a pressure sensor to do all kinds of neat stuff, like have a "torque fault" if there is too much (or too little?) torque, calculate power by combining it with an rpm signal, monitor tool wear or tapping torque to improve stuff.the list goes on and on!

and this is real torque delivered to the spindle (or at least a pulley or gear away from the spindle)..not electrical quantities assumed to equal torque

[keebler303]

You have to be careful how you setup the system. Drag in the bearings has to be accounted for as it takes away from the actual torque. I would think a hydraulic cylinder would have substantial seal friction which would not make it viable for pressure measurement.

Bimba makes some low pressure hydraulic cylinders, perhaps they would have low enough friction to be negligible compared to the motor torque.
Bimba Original Line® "HL" Option - Bimba Manufacturing

[acannell]

You have to be careful how you setup the system. Drag in the bearings has to be accounted for as it takes away from the actual torque. I would think a hydraulic cylinder would have substantial seal friction which would not make it viable for pressure measurement.

Bimba makes some low pressure hydraulic cylinders, perhaps they would have low enough friction to be negligible compared to the motor torque.
Bimba Original Line® "HL" Option - Bimba Manufacturing

remember there wont really be any movement. the torque will be transmitted through the bearings and the hydraulic piston or strain gage with only material strain, ball bearing play, and the (in)compressibility of the fluid accounting for movement. so dynamic losses wont be an issue.

i think the breakaway torque of the hydraulic cylinder wont be an issue either. i made a small hydraulic ram in order to measure drawbar force, and i was able to measure only a couple pounds of force on it and have it return to zero without sticking, even though the load wasnt centered by anything. and it wasnt even lubricated, it just had water in it. it wasnt properly machined either im sure lol. this will be how i make the piston for this setup, if i go that way. it doesnt seem like the most elegant way though..something about it bugs me. there must be a better way? i think what bothers me is that, even though there should be almost no movement, using a piston requires its axis to be tangent to the radius of the motor housing rotation in order to be accurate. (right?) i suppose with almost no movement this shouldnt be hard to maintain..but it just seems icky..

as far as static breakaway torque, i.e. the bearings sticking, i think thats a little unusual to worry about..brand new ball bearings are going to have a teeny resistance to rotating compared to the forces that will be measured.

but the devils in the details and im all talk until i make a prototype. i wish DaOne or Hood would join in like they did on the drawbar force gauge. i really have important things to be working on and this is the last thing I should be doing lol

[acannell]

okay i couldnt resist so i spent this morning drawing this up

this setup uses a single, unflanged bearing

its meant to be as easy as possible to construct, to make a prototype quickly and see if this is a workable idea

this drawing doesnt yet include the provision for measuring torque. this is just the "free floating motor mount" structure.

the green parts get bolted together
the blue parts get bolted together

the upper green part also gets bolted to the black motor. and in doing so "pinches" the inner bearing race together, thereby locking the motor, the ball bearing, and the green parts all together.

the blue parts "pinch" the outer bearing race. so now all the parts are secured together, yet the motor _housing_ can rotate freely relative to the blue block.

the bottom blue part gets bolted to the original location the motor bolted to..i.e. the z axis casting on my machine

i was able to fit all this in the same height as my original motor mount on my mill, so thats nice since now i dont have to worry about the motor shaft stickout being different or what have you. it will just replace my belt tensioning block

the bearing is one i had laying around, a NSK 6908Z..you could probably use a zillion different bearings

the thin blue plate has holes for set screws which screw down onto the outer bearing race to secure it. this is pretty ghetto of a way to do it, but for a prototype itll work..looks like i forgot to put reliefs in the big block for that..but no big deal

i didnt draw holes that allow mounting of the thin blue plate to the big blue block..just fyi

also, to actually make this, i dont yet know my motors actual shaft diameter because i cant find accurate specs and it has some special split bushing thing that connects the shaft to the pulley..so i may have to deal with that in order to correct these models..

[Hood]

Interesting idea, sadly for me it would be quite difficult to do.
My spindle motor is coupled direct to the spindle and is located by the motors spigot and held by 4 bolts. That in itself would not be a problem as I could remove the 4 bolts which would allow the motor to rotate. The problem would be the lack of space around the spindle motor to fit any kind on stops/hydraulics/force gauge/whatever, see pic below.
Will be intersting to see your progress and results if you manage to set something up.
Hood

[acannell]

Interesting idea, sadly for me it would be quite difficult to do.
My spindle motor is coupled direct to the spindle and is located by the motors spigot and held by 4 bolts. That in itself would not be a problem as I could remove the 4 bolts which would allow the motor to rotate. The problem would be the lack of space around the spindle motor to fit any kind on stops/hydraulics/force gauge/whatever, see pic below.
Will be intersting to see your progress and results if you manage to set something up.
Hood

i think all hope is not lost. it may be that the biggest problem for your machine would be dealing with the motor being further away than it currently is...mine already has a "plate" of some kind used to adjust belt tension, so i can essentially replace that with the bearing gizmo and not have to deal with anything else being different..

i think on yours, a strain gage between the stationary surface and the motor housing would fit in the space that you have...so basically just a short section of metal with a little gauge glued to it..

in thinking about this a little more i realized a couple things:

-the motor reversing direction has to be taken into account. so if a cylinder is used the hydraulic setup needs to be a little more clever to allow for measurement in both directions.

-this is really a perfect application for a strain gage...and engineering one into this for a beginner might actually be pretty easy since the forces are constrained by the ball bearing to one direction..

-this is also a really good opportunity for the hydraulic setup, if someone wants to make it really quick and dirty, or make it very drawn out and elegant, since you could make the hydraulic piston very fancy and have a remote mount gage, and even design the cylinder size so the gauge reads directly in in/lbs or ft/lbs with no conversion needed