Linear Actuator Control

[joethepro]

Question:

Since a small footprint is wanted and the need of high torque, what about using a cordless drill motor (400in/lbs) hooked up to an encoder? This would give me a very small footprint (could probably go even larger on the motor) and very high torque... correct?

If the batteries of the drill motor will last one ten foot run I could change the batteries after each run, which is really no big deal.

 

[khalil]

hi just trying to help .........
why you dont try a hydraulic motor ?

 

[rsdoran]

Joe another site you may want to look at:
http://www.robotics.com/robots.html#motors

I hope to have time to perform a test but I think I know how you can make a "poor man's controller" using something like the drill you described but forget the batteries, you can buy or build a DC power supply with a small footprint. I am getting ahead of myself, let me try to run the test and see what happens.

 

[joethepro]

As ndzied1 said, your accuracy requirement points to leadscrews. Just make sure you don't discount friction. That could be a respectable portion of you total motor load and may require you to jump up a motor size to account for the friction. On the flip side leadscrews are very hard to backdrive due to the friction. So leadscrew systems should tend to reject outside forces and hold position a bit better than a ballscrew setup (OK, that's theoretical).

Hi kamenges,

I agree with you and ndziedl. I spoke with a CA company and they told me I can get "anti backlash screws" to eleminate the backlash in the system. I asked the company what was the friction wear factor of the lead screw for a 300lbs load, in other words, how far did the table have to travel in order to get 0.0625" of backlash in the lead screw. They looked it up in their speck book and said it would have to travel over one million inches (under perfect condision & without using the the 'anti backlash screw'), which is perfect.

I sent QuickSilver Controls an email and will follow that up with a phone call..... thanks for the link!!

hi just trying to help ......... why you dont try a hydraulic motor ?

Hello khalil,

The reason is because the footprint of hydraulic would be too large and too much weight. Thanks for the thought, it really is appreciated.

 

[joethepro]

I hope to have time to perform a test but I think I know how you can make a "poor man's controller" using something like the drill you described

Hi rsdoran,

Let me know if your test works! If it does I'll fly down there and KISS YOU!! lol

forget the batteries, you can buy or build a DC power supply with a small footprint.

Great! Good idea!

"Tell me and I forget, show me and I remember, involve me and I understand."

Perfect quote! Where did you get this quote? Is it yours or did someone else say it?

 

[rsdoran]

Tell you the truth I am not sure if I actually created it or its something I read/heard in the past. I didnt copy it directly from a book or webpage. I had another saying but it was used kind of in relationship to a topic so I decided to make another one and this is what I came up with. It describes me to a T, I cant always remember what I am told, read and retain (but may not be able to apply), and if I am involved with or do then I understand and can do again.

 

[joethepro]

Update:

I emailed the SilverMax company, this is their reply:

Joe,
I didn't see any solid torque or power requirements nailed down in your thread, but I believe you are looking around 1hp. SilverMax most powerful model is only fractional horsepower (3250 oz-in @stall), so by your rough estimates SilverMax as a solution would not be applicable without significant gearing. If your slide manufacturer can help you determine a solid torque requirement, sizing a SilverMax is simple.

The SilverMax mentioned above (34HC-4) lists for just under $2000 an axis. Please see our distributor info section on our website to find a distributor in your area. They would be happy to help you select and price all the components for your system.

Their motors have only 3,250 oz/in, if you divide that by 12 that's 270.83 lbs/ft which might do it (cordless drill has 400 lbs/ft of torque). The only problem is the 2k for each motor, I need 4 motors... ouch! That's not counting the slides and other hardware... double ouch! I think I'll stay with the drill idea.

rsdoran,
did you have time to do the test?.... let me know how it goes

Thanks,
Joe

 

[Gerry]

Units: the torque units are a product, not a quotient. i.e. lb-ft not lb/ft
3250 oz-in divide by 16 oz/lb divide by 12 in/ft = 16.9 lb-ft

Also horsepower is lb-ft/sec. Your HP requirement depends on how fast you want to go. I doubt that you have a 1 HP cordless drill.

I don't think you need anything near 400 lb-ft of torque. I did a quick exercise with A-B's Motionbook software using 0.5" pitch 1" diameter screw, 0.5 coeff. of friction, 90% efficiency, 1 sec. move, 5 sec. dwell. According to that, torque requirement was under 2 N-m (283 oz-in).
It also indicated that the most influential factors on motor size are lead screw pitch (speed) and diameter (inertia).

With (almost) any motion control system, the major torque requirement is for acceleration and quite often, the motor itself becomes just as significant as the load.

 

[Steve Bailey]

Joe,

The critical parameter in the analysis of your torque requirements for this system is what acceleration rate you need to achieve. I did a quick and dirty analysis assuming the two ball screws (or lead screws) are going to be mechanically coupled. I used a 1.5 inch diameter screw at 0.25 inch pitch. I used a screw efficiency of 90%, a coefficient of friction between the table and it's support of 0.2. Knowing nothing about what type of mechanical coupling you might use, I calculated inertia based on two screws, each 15 feet long. I used a traverse speed of 8 inches per second and an acceleration time of 0.5 seconds to get from zero to full speed. That gets your 12 inch move done in around 2 seconds.

From that, I come up with a frictional torque (required to maintain the load in motion) of 1.8 Lb-In, and an inertia torque (required to accelerate the load) of 53.7 Lb-In. The traverse speed of 8 IPS gives a motor speed of 1920 RPM. 55.5 Lb-In of torque at 1920 RPM equals 1.7 HP.

If you can stand a 0.75 accel time, your inertia torque requirement drops to 36 Lb-In.

You could consider using a 1.5 or 2 HP AC motor and a VFD and a micro PLC as the motion controller. Many micro PLCs can handle encoder inputs and have analog outputs. Use the encoder to keep track of table postion. Use the analog output as the speed referenc for the VFD. Use the difference between the table's current position and its desired position to calculate the speed reference. As you approach the target position, the speed will decrease so you can stop within an acceptable tolerance.

Use a three-phase, 230 VAC motor. For horsepower ratings under 3 HP, many brands of VFD permit a single-phase feeder.

The VFD approach won't give you quite the performance of a servo, but it could be acceptable.

 

[joethepro]

Gerry & Steve,

Thank you for taking the time out of your busy day to work on this problem. Though I understand just a fraction of what both of you are talking about I will not allow my ignorance to keep this project from progressing.

The prototype I built I use a standard four foot long threaded rod that I bought from Home Depot. I hooked that up to a table I built from wood and used ball bearings for the rollers. I then made a wooden frame connecting the two tables and put 200 lbs to 300 lbs on the frame. The drill was then hooked directly to the threaded rod. It moved the weight easily. The setting on the drill was set on position 1, which is the lower setting and more torque (400 lb-in max)..... (I have a correction from the previous torque numbers. I stated before that the cordless drill has "400 lb-ft" of torque, this is not true. I double checked the drill's specks and the max rating for that particular drill is rated at 400 lb "IN" not "FT"..... here is a link to the drill I used.... http://www.dewalt.com/us/products/tool_detail.asp?productID=3188)

I don't know how much difference 400 lb-in is to 400 lb-ft, but my gut tells me its big..... I'm sorry for the mistake guys.

Also, a reminder. The 200 lbs to 300 lbs I used in the test was spread out between "2" tables, so the drill was accutally moving 100 lbs to 150 lbs. In other words I hooked up 2 drills one for each slide. Both drills where exactly the same (see above link). I don't know if you calculated that in or not but if you did then just forget I said it.

Thanks guys!
Joe

 

[kamenges]

Hi Joe-

You're right; there is a big difference between 400 lb-ft and 400 lb-in. If I remember right, the Vortec V-8 that goes into GM trucks puts out just under 300 lb-ft. When I saw the 400 my mind kind of skipped over the torque units.
But 400 lb-in is still a snoot-full of torque. It would take a STUD to hold that drill if it got cranked up. (OK, I want one).

The thing to remember with the drill is that you don't get that 400 lb-in at the 1800 RPM max speed of the drill. You get it at something under 450 RPM. Also, you said you used threaded rod in your test and said the drills moved the load without a problem. I suspect you used something in the neighborhood of 3/4" threaded rod. I think that gives you about 10 threads per inch. This is about 2.5 times tighter thread pitch than Steve used in his calculations. This means your motor would need to turn 2.5 times faster than Steve's numbers indicate (or 4800 RPM) to get the performance numbers he went with.

You stated earlier in this thread that the drill almost seemed too fast. I assume that was set at the 450 RPM/400 lb-in setting. If so you weren't going anywhere near 12"/sec. Even if you ran the drills at the full 1800 RPM max speed, assuming a 10 threads/inch threaded rod, you would peak out at 3"/sec.

Also, hand drill motors are not intended to run full tilt all day long. They cook. If you intend to do one move cycle (10 feet) and then wait a few minutes for the drill to cool you might be alright. But if you plan to do back-to-back sequences the drill will melt down pretty quick.

Another downfall of using the drill as-is is that it has no built-in deceleration capability. The leadscrew may provide enough friction to handle the decel, but it may be tough to consistently stop where you want without the motor stopping the whole works.

I guess the moral of my story is I would stick with industrial motors and controls. I think you will ultimately be happier with the result. If cost is an issue, try eBay. Seriously. you can come across some pretty good stuff out there for little money.

Keith

 

[joethepro]

You're right; there is a big difference between 400 lb-ft and 400 lb-in. If I remember right....... When I saw the 400 my mind kind of skipped over the torque units.

I thought so, not knowing anything about this kind of stuff... just had a gut feeling.

I suspect you used something in the neighborhood of 3/4" threaded rod. I think that gives you about 10 threads per inch. This is about 2.5 times tighter thread pitch than Steve used in his calculations. This means your motor would need to turn 2.5 times faster than Steve's numbers indicate (or 4800 RPM) to get the performance numbers he went with.

It was either 1/2" or 3/4" rod, can't remember. One thing I'm finding out is I never dreamed it would be this difficult making this thing. The whole problem is stopping the drive (motors) machanism at the precise spot (every 12"). What if I made a wheel that when rolled the circumference would be 12 inches? Then put some kind of knotch in the wheel and hook up a stopper so that when the wheel turns one full turn (12") the knotch lines up with the stopper and the stopper falls into the knotch thus stopping the wheel. I could hook up the stopper to a solenoid, small servo or a timed level and after so many seconds the stopper disengages from the knotch so that the wheel can turn again........ just thinking out loud.

Thanks for everything!

 

[Steve Bailey]

Forget about using a screw of any kind to move your table over a ten foot span at the kind of speeds you're talking about. I looked at a table of critical speeds for horizontal screws in a Nook Industries catalog. At a ten foot span, the critical speed for a 1.5 inch diameter ballscrew is between 500 and 800 RPM, depending on the end supports.

 

[kamenges]

What about....

A thought popped into my normally scattered brain last night. It seems to me a significant portion of the total energy in the system is used to accel and decel the drivetrain. When I ran some quick numbers on what it would take to move the load 12" in 1 sec using a triangular velocity profile I came up with well under 0.25 HP.

What if you tried a rack and pinion setup. Put two pinion gears on a common shaft. Have the pinion gears engage racks running the length of travel on each side. This will maintain alignment as well as give you something to drive against. And because the shaft velocity (and therefore acceleration) is much lower than a leadscrew you won't be fighting as much drivetrain inertia.

The two downfalls are backlash and possible loss of resolution. But is positioning accuracy can be less than about +/-0.03" or so I think a rack and pinion might do the trick. The cost of a rack and pinion set versus a leadscrew should come out in your favor also. You also avoid the problem of leadscrew whipping.

This doesn't completely solve your motor/drive issue other than it may significantly reduce your horsepower requirement.

Just a thought.
Keith

 

[joethepro]

Man this is a GREAT site with GREAT people!! I really want to thank all the people for helping out, you guy are the BEST!!

For the last 48 hours I've gotten about 2 hours of sleep. And since last October (when I started working on this thing) I've gotten about 4 hours (on average) of sleep trying to figure out how to do this. I have been working on this project for almost a year now. I get an idea then I redesign, idea, redesign and so on. The other day I counted up about 200 various designs.

Though I can not tell anyone what this thing is, I can say there is nothing like it in the world. The plan is to eventually patent this thing, which of course is a whole different ballgame. There are many other parts to this system that I have not elaborated on, so all the advice I'm receiving from this forum has to be weighed with the other mechanical parts as a whole. As for Matthias offer, I have to tell you Mattias that I agree with kamenges it is extremely kind of you and I really thank you for it, but I really don't want to show this invention to anyone at this point.... can I take a rain-check on the offer? Once I get a working prototype and a patent pending application I'll give you a shout. Send me your email address, ok?

UPDATE:
All these great idea's from all these electrical guru's out there makes me feel like Elmer Fudd. I really don't know what you guys are talking about but again I will not allow my ignorance in this area stand in the way of progress. So that being said... After working all night searching the web I came up with an idea that is more in my ballpark, but before I get into that I would like to say something to kamenges in the "Linear Actuator Control" post.... kamenges, that is what I was thinking too. Maybe I should sideline the actuator idea and hook up a drive line directly to the cross member that connects the two tables. Use gears as wheels and set those wheels onto a linear geared track (rack and pinion), that would give me more room for a motor and would eliminate all the problems with the lead screw actuator. But again the stopping problem that remains..... Which brings me to an idea I had last night....

If I took the cross member (that connected the two tables) and put wheels on them, then glued a small magnet to the outside of one of the wheels and hooked up a reed switch, every time the magnet crossed the reed switch the switch would activate sending a signal to another switch which would shut off the motors. To solve the drift problem I could hook up a what rogerhollingsworth mentioned, a "dynamic break resistor" or even an electrical brake might work. This would keep the system very simple, easy to design and build.

Thanks!
Joe