Linear Actuators
 

I have noticed a lot of discussion and I personally have had a number of discussions on other options for pneumatics. Often times it seems to me it would be beneficial to avoid the extra weight with a compressor, air tanks, and other peripherals on a robot when something could be done with a linear actuator.

The largest thing that comes up is that they are slow, but they are accurate. I find them to be an interesting component that FIRST should allow in next years game as an option for teams who want to go without pneumatics, or for other types of systems.

I know there are converters that could turn a regular motor into a linear actuator, and there are other work around solutions, but do you think FIRST should allow a dedicated linear actuator to be used during the 2015 season? Would your team find it useful? What would you use it for?

Re: Linear Actuators
 

The term "Linear Actuator" is a little vague... there are many types of linear actuators. All the term means is that it's something that creates motion in a straight line. It's something teams do quite frequently with cams or winches. Most commonly available linear actuators are really just motors connected to a lead screw... they would require some disassembly and replacement of the motor, but could be made to work.

I would guess, however, that you're thinking more along the lines of an electric solenoid actuator, which have been legal for a couple of years.

If you had a specific product to link to, you could probably get better discussion over what you want to discuss.

Re: Linear Actuators
 

Sorry for not being specific. I had been talking about electric lead screw acutators. Here's a quick example of the kind that I am thinking of.
http://www.bosch-ibusiness.com/boaae...%20Spindle/357
When talking about changing angles on our robot from 2013 this came up as an option. It would be able to change the position of the heavy shooter while being very precise. Also, it could have pulled shooter springs on a 2014 robot to varying tensions and differant shots. Do you think it could come in handy in similar ways in the future of First?

Re: Linear Actuators
 

All that is is a motor attached to a gear box. The form of the motor for that particular product doesn't easily match anything that's legal, but I'm sure with a little searching you could find one that utilizes a 500-series motor and swap out the motor for an RS-550. Doing so would make it perfectly legal, similar to all of the teams over the years that have taken a shop vac, swapped out the motor, and used it as a suction device on their robots. The only thing that makes it illegal is the model of motor that's attached to it.

Re: Linear Actuators
 

I think it woudl be useful. I've seen some teams try to use them before I told them it wasn't a legal motor. I think the easiest thing (for now) would be a piece of hardware that mounts to a standard CIM or banebot motor. Provide built in limit switches to attach to the control system/motor controller and you're all set.

Re: Linear Actuators
 

They've been legal for a long time, and they're not hard to build or use. They are, however, slow as death.

We sped ours up from this initial test, and it was still far too slow, and we'd have been better off going with pneumatics.... The fine adjustment ability was really cool (and led to the first successful autonomous scoring our team had ever done--and it was killer consistent), but massive overkill for the game.

https://www.youtube.com/watch?v=gTLMkUmhU2Y

Re: Linear Actuators
 

I think an allowed COTS Linear actuator (or one we get in the KOP) would be very good for teams as they could use them to expand designs a little. (Expand designs they wouldn't have thought of otherwise.)

Of course there can still be custom options that are better for achieving a particular design goal, but something like those motors would not be a bad thing.

More than likely if a game called for fine angle adjustment like 2013 did I would consider a linear actuator in a design like that. The fine angle adjustment of my shooter would allow me more versatility in shooting positions around the pyramid. Granted it was proven that such angular adjustment was not needed to be successful, my former team attempted to use a custom 2 bar linkage for angle adjustment and it didn't quite work, or package as well as I wish it would have.

Re: Linear Actuators
 

We've gotten multiple-start screws and nuts in the kit for years and years.

Re: Linear Actuators
 

You might be remembering some kit in the past. I know that my old team had some lying around, but I'm not going to look up the kits that I got when I was on 1102. Since my rookie season mentoring 4901 has ended I can soundly say, since I took part in receiving the kit this year, that the 2014 KOP did not have one.

Source: http://www.usfirst.org/roboticsprogr...c/kit-of-parts

http://www.usfirst.org/sites/default...e_2014.1.7.pdf

http://www.usfirst.org/sites/default...e_2014.1.4.pdf

Re: Linear Actuators
 

Here's a differant question: Do we like the user interface of this new completed assembly better than having to put the motor, gear box and lead screw together seperatly?

I think that the motors in the included link are well packaged and could be easier and quicker to include in a robot design than to put the whole system together as a custom set up. This could open the door for newer teams to use a linear acuator that wouldn't have considared it before.

Re: Linear Actuators
 

They moved to FIRST Choice, I think we got 4 or 5 of them this season.

Re: Linear Actuators
 

Ah. That would be it.

http://firstchoicebyandymark.com/en/fc14-095

Still, the point remains. I think some teams would benefit from a complete module as opposed to a lead screw kit.

Re: Linear Actuators
 

I think more teams would benefit from a lead screw and nut and having to figure out just how they're going to make their own module!

Re: Linear Actuators
 

FYI, you are allowed to use electric solenoids <= 1" stroke

Re: Linear Actuators
 

I think those teams that would most benefit from having a module provided are those who don't have the resources to create it from the individual parts. Lets also keep in mind that a single module won't fit all uses - you have questions of speed, force, and distance to take into consideration... those variables are going to be very specific to the application at hand, and could easily drive those teams with abundant resources to create their own.

So why don't we see more of them? For one, there are design tradeoffs with going with this approach over other options. Which option is best entirely depends on the weight the team gives to the tradeoffs. But perhaps more importantly, there is probably a lack of inspiration towards this mechanism. We are all familiar with springs and motors, wheels and pulleys, grabbers and elevators. But you don't really see or interact with lead screws very often, and it's not something that's already prevalent in FRC... all that combines to make it something that doesn't really come to mind when designing a robot. I know the only times my team has talked about it was many years ago when we prototyped a scissors lift, driven by a lead screw, and when we got a box of them donated (by a wife who wanted them out of her garage, I think... I'm not sure if she even told her husband she was donating his gearboxes!). It's just something that hasn't come up in discussion, and something we don't have much experience with. For that reason, having one in hand in the KoP may provide the inspiration for a team to go in a new direction.

Re: Linear Actuators
 

When we used one to lift our robot in 2010 (our best year ever), people acted like it was magic and we were rather amazed at the number of people amazed by the mechanism. Many teams just don't even consider them, and a lack of familiarity may well be part of it.

Another part of it, though, is that pneumatics do a very similar job, don't stall and release magic smoke, and come in sufficient variety that if you're already using them (and thus have the compressor, etc already on-board), they're a better choice.

Lead screws are used in industry primarily for fine control, while FRC robots have to complete gross-control tasks. In 2010 it made sense for us because we had no other pneumatics on the robot--why add another system just for one mechanism? But if you're already using them, why use a slow mechanism specialized for fine control to execute tasks that don't require fine control?

Re: Linear Actuators
 

I've been eyeing some of these models for some time:

http://www.servocity.com/html/12v_linear_actuators.html

I would love to have a drop-in electronic shifting option for gearboxes, and some of the lighter-duty short-throw actuators there have enough force to shift under load fairly quickly. They are pricey, however, and I haven't had time to find out what type of motor and shaft they are. I called servo city and they were unable to give me any more info than external dimensions, so someone would have to buy one to find out if any of our motors would mount in the same location.

Re: Linear Actuators
 

Yes, but the power limit is so low that they have pretty limited utility.

Re: Linear Actuators
 

I can only think of one time in FRC I used a electric solenoid and that was 2012 when we designed our basketball elevator to no be able to allow the basketballs into our shooter until the solenoids were engaged then applying a clamping force on the outer wall of our elevator.

Let's just say that design worked....but the solenoids we had to use were absolutely terrible and ended up overheating and all sorts of things after continuous use during practice. That was because of the power limit among other things.

Re: Linear Actuators
 

Depends on how imaginative you are. I think it's POSSIBLE to shift with them (though not necessarily advisable). They can heat up quite quickly.

Re: Linear Actuators
 

Anyone mess with vex?

http://www.vexrobotics.com/276-1926.html

Is this kit legal?

Re: Linear Actuators
 

I didn't see anything that would necessarily make it illegal (in 2014 at least--ask again next January).

However, given that it's a VEX item, and not a VEXPro item, I'd be very wary of using it with anything other than the VEX motors (which at least in 2014 were legal in FRC). Given that the VEX motors are far less powerful than, say, a CIM or 775, do your engineering on the power needed and available before committing.

On the other hand--the price ain't bad, and the VEX motors have decent power and as I recall were unlimited quantity items last year. Might have a winner if you can adapt it to shifting, or might not. Anybody want to try it during the offseason?

Re: Linear Actuators
 

These are intriguing, Tom.

The motors on the big ones (e.g., SDA6-263) look a lot like CIMs, so maybe they can be adapted to get FRC legal. The price is right at our present $400 single component BOM limit, and they weigh about 10 lb. -- but the prospect of lifting an entire alliance 6 inches in under 4 seconds is ... intriguing.;)

Re: Linear Actuators
 

Lead screws are pretty cool, and they don't have to be slow. An 8 start 1/2" lead screw travels 1 inch per turn.

We used some of these (http://www.mcmaster.com/#99030a400/=sr50n0) on our climber in 2013, and they were decently fast. To travel 30 inches carrying a 150 lb robot took under 5 seconds. If we weren't carrying the weight, we could have done it faster.

Re: Linear Actuators
 

Though our climber never came together by regional time in 2013, we were going to use rack/pinion to extend and retract our arms.

Mcmaster has a selection of racks in a good range of lengths, widths and pitches in both low load Nylon and heavy duty steel.

Re: Linear Actuators
 

I saw this link on you tube:
https://www.youtube.com/watch?v=GjfsPTJx_DY
It's not necessarily this HS-55 motor that intrigued me, but the attaching piece to the servo. While it's not truly a linear actuator it could serve that purpose with a little math. The thing I find most attractive about using servo's is the ease of setpoint operation use. The question then remains what servo's of what high torque are allowed.

I'd sure love to use something like this SDX-901 if it wasn't so expensive
https://www.youtube.com/watch?v=YDFMhyA4rHw

the HS-55 is only 18 oz-in (0.127 nm, 0.09 lb-ft)
while this SDX-901 is a whopping 402.7 oz-in (2.84 nm, 2.09 lb-ft)

Re: Linear Actuators
 

Our ball elevator used an 8-start, 5/8" diameter lead screw driven by a window motor. It was plenty fast for our needs. Over its 128 matches the thing survived an intense beating. Some of the stuff it was attached to... not so much. :ahh:

Re: Linear Actuators
 

So I have a question for everyone who has used the lead screw technique. How did you do a set point operation... like if you want it to go a certain distance? Did you:

1. Use an encoder?
2. Use a potentiometer?
3. Use a limit switch?
4. String potentiometer?
4. Something else?


If you used an encoder how do you ensure it remains calibrated (e.g. limit switch). I'd like to use this lead screw technique, but I haven't found reliable way to sense distance. (I have not tested the string potentiometer either).

Re: Linear Actuators
 

We used encoders with limit switches on the ends for continuous calibration purposes. It worked quite well, and was very PID-friendly because you could be off by 50+ counts and still in the "right" position.

Re: Linear Actuators
 

Obviously, it all depends on the design/setup. When you have an application with a limited operating range, I strongly recommend potentiometers. We haven't done it with a lead screw before, but we have done it with a winch, which is similar in that it turns rotary motion into linear motion. There are many design considerations to take into account first, though... like how many rotations are required. When we used it on our winch in 2013, the winch itself was designed for 7 rotations, and we had a 10-turn potentiometer available (It's always good to have a little extra rotation available, just in case).

In some applications, a potentiometer may not be appropriate. For example, it may require too many rotations to make an encoder practicable. In that case, an encoder with limit switches is great... especially when you can have limit switches at both extremes, and you travel to both extremes on a regular basis.

The only real downside to an encoder is that you have to keep track of where you are in the code - with a potentiometer, the value returned tells you where you are with nothing else needed.

Re: Linear Actuators
 

Lead screws are super accurate for distance. The McMaster ones are rated per turn to within 0.009" per foot of travel. We used two end of travel limit switches and an encoder, coupled to a gearbox shaft with surgical tubing.

Re: Linear Actuators
 

Our setup used a 10-turn potentiometer, belt driven off the lead screw, which was in turn direct driven from the window motor.

Re: Linear Actuators
 

In 2013 we were a tower shooter. We used a "linear actuator" to move our shooter up and down for fin movements. We got our off of a car seat. (They are used to move seats forward and backwards in cars) We took the old motor off and then made a plastic converter on the lathe to the versa planetary gearbox we were using with a bag cim. Overall it took only a few hours to fully assemble.

Re: Linear Actuators
 

That is very cool! I googled around to see what they look like and they look like they are self contained when retracted (e.g. like a pneumatic cylinder). If you have any pictures of this setup... please share... thanks!

Re: Linear Actuators
 

For a while in 2013 our launcher was articulated using a lead screw (at first , we later replaced it with a piston) and we used a few magnetic reed switches (they sense the proximity of a magnet) to sense when the screw was close to either end, as well as a set-point in the middle.

If you just need to move to certain set-points, and are ok not knowing exactly where you are when between them, using switches that trigger at the important points can be both reliable and easy to set up.

Re: Linear Actuators
 

This is the only picture I could dig up of the part. I am sorry that I could not find a better one. I can ask for a better one from someone else on the team.

http://www.idleloop.com/frctracker/p.../2013/3398.jpg

Re: Linear Actuators
 

We prefer string potentiometers for our linear actuators. Quite simply, they are the easiest sensor to implement, and they are accurate enough for nearly every FRC application.

Re: Linear Actuators
 

A friend pointed me to this link last night: http://www.mcmaster.com/#ball-screws/=sy3xrc ... while the parts look great the prices are terrible... it makes me wonder if it would be easier to make your own using parts from say home depot or something. <I'm going to search around a bit more as well>.

Re: Linear Actuators
 

Ball screws are pretty overkill for FRC. McMaster's fast travel lead screws are more affordable, and are very smooth running. The brass nuts have very low friction.

http://www.mcmaster.com/#acme-precis...screws/=sy9ith

Re: Linear Actuators
 

This is what we used. Ball screws are just too expensive!

That said, you of course still want grease on those brass nuts.

Re: Linear Actuators
 

Tom, We bought some similar to the ones you show in the link, but without the pot feedback. We modified the shaft of a mini CIM to accept the drive gear of the original motor. Works like a charm but still dead slow at 0.5"/sec. Would be ok for small adjustments that require precision and strength but if you wanted a 6" stroke for example, you'd wait 12 seconds :ahh: - an eternity during a match. Still a very compact unit; easy to use.

Re: Linear Actuators
 

I also have been considering to use this model... Thanks for the feedback what would be great is to know how to make something similar to this buying separate parts that are affordable. From that one could buy a different lead screw like an 8 or 10 start. I'm having a hard time finding information about how to mount the motor onto the screw, and how this is self contained on the other end. I'd almost want to take one of these apart to see how they are doing it. (this is still research in progress)... Any insight to this would greatly be appreciated.

P.S. I was thinking of using this as a lead screw http://www.use-enco.com/CGI/INSRIT?P...&PMPXNO=942694

Re: Linear Actuators
 

You don't have to drive the screw. If you drive the screw, you'll most likely need to support both ends with bearings and thrust bearings, and you'll need another bearing solution for the nut/carriage, which wants to twist. If you do drive the nut, you could probably mill a hex on the top/bottom and put a gear/versahub/sprocket on it.

You can also drive the nut, which is what we did on our climber. We found this to work well with the geometry of our climber, as the screws were non rotating and fixed to our climber arm.

We put the nut in an 1/8 inch wall aluminum box, drove it with a sprocket, and used two of these (http://www.mcmaster.com/#60715k11/=sysaez) bearings.

Multi start lead screws are awesome and can go pretty fast. We prototyped an 8 start climber that could lift the robot one level up in under 4 seconds, but were unable to use it due to other gearbox issues.

The material selection of both the nut and the screw are important too.

Re: Linear Actuators
 

I almost understand this and see the benefit of doing it this way... how exactly does the sprocket interact with the nut though? I know the vex #25 sprocket has a nice 1/2 inch Hex ID... does a sprocket like this mate with the driven nut in what you were saying?

Re: Linear Actuators
 

http://www.mcmaster.com/#fast-travel-acme-nuts/=syswre

We used nuts identical to the ones at the top of that page. All the half inch multi start lead screws and their nuts can be purchased from amazon.com for a lot less. IIRC, they were from nook industries. To connect the nut to the sprocket, used a 14 tooth sprocket from McMaster, removed the hub and put a clearance hole for the lead screw.

To transfer the torque, we cut a small (0.1 thick?) groove about .25" deep on the diameter of the nut. We left behind part of the sprocket's hub, and it fit into this groove. There were also 2 4-40's that went into holes we tapped in the nut.

McMaster also sells flanges that can be screwed onto the end of the nuts too. You could also make your own if you have the ability to cut the threads. The 1/2" screw nuts have 15/16-16 threads on the outside.

If you aren't carrying a ton of weight, you could probably get away with using a few 8-32's to transfer the torque. You can buy "machinable" nuts that have room to add larger tapped holes to the end.

Re: Linear Actuators
 

Thanks for reply as it has inspired my to try an experiment with 3D printing just to see how far I can push the envelope in that direction using nothing more than the VEX motor, which has about 9 inch-pounds of torque on the higher setting. In this picture I want to create a lead screw which encases itself within a cylinder casing, and the motor can drive a printed nut that is fused with a gear that is compatible with Vex gears. The pitch on the screw is going to be very large like one revolution per inch, and I may be able to have multiple helix (e.g. 2 or 4) cut grooves for the screw. I'm not sure yet of the optimal geometry of the groove profile, but I'll study the acme angles... This will end up being able to push weak load (probably less than 9 pounds) that has a quick time probably 2-4 seconds for full range motion. Since the Vex motor turns less than 2 revolutions per second the friction should be minimal... although ABS material does have a high melting point. This is probably a few months out to complete, but I'd like to share this experiment to see what people think about it and any advise on how to make it actually work.

Thanks again... wish me luck!

Re: Linear Actuators
 

What internally is preventing the (what I presume to be) threaded rod from rotating inside the casing? It would have to be held radially stationary by some sort of slot it rides in which, if there, can't be seen.

Re: Linear Actuators
 

Yes... let me try to explain the proposed parts not yet posted:


Thruster Bearing | Geared nut | Thruster bearing All within the same inch where the vex gear would be. The thing I do not know yet is how far the screw can go before it needs support bearings when the rod gets retracted. I was thinking maybe 6" (halfway) of placing a bearing there.

I've been trying to learn from this video:

I know this video is not driving a geared nut, but it does give me some direction as to what has worked before.

Thanks for the feedback... any ideas about how to solve that would greatly be appreciated.

Re: Linear Actuators
 

I know it's been quite a while since I've last posted but finally I have something that works well:



Here you can see it in action:
https://www.dropbox.com/s/y49mpgqedi...84622.mp4?dl=0

and Here is where it is pushing the battery:
https://www.dropbox.com/s/h8bvnyu8ui...85045.mp4?dl=0

More media... shared here via ftp:
ftp://ftp.termstech.com/LeadScrewPics/
username sharedmedia@termstech.com
password SharedMedia!!!!

It is unfortunate that the 393 is no longer a legal motor for FRC, but we still intend to use it for another project, and intend to have it work fully with the roboRIO... there are 3 solutions we are currently exploring with that:

1. Find out what the real-world current draw of the motor is actually using and see if this is more than what the roboRIO can provide using a direct connection. I suspect it's under in most cases as it was on the Digital Side Car... but I also believe is issue is why they removed it from the list and probably won't put it back.

2. Use external battery... Vex Support has helped us find a solution that is sound. Doing this, and some simple wiring that will get the 7.2 volts we need (and correct current)... of course this would not be allowed in FRC

3. Use a custom voltage regulator with the PDP... this one seems to be the most robust assuming that we may pull more current than what the roboRIO can provide. We're looking to several possible voltage regulators and may try 2 parallel circuits of 2amp with 7 volts since these are common to find... still work-in-progress.


That said... it's back to the drawing board for FRC... we'll have to pick a motor that is on the list if we wish to have something... the problem there is the speed of the motor... so if you want to help us out... please reply with which motor you think would be ideal replacement. The ideal case is slower speed motors that have a lot of torque like the 393... from that we can try to figure out how to make it work.

This linear actuator shown here is the first revision... we intend to add the 10 turn potentiometer, and limit switches in the future revisions. Also the back half of actuator it is not shown in the picture this one spans 7 inches, but can go up to 12, by screwing on the back end of the appropriate size. It's lightweight because it is 3D printed, and very cost-effective solution. Will be nice to have something for FRC though. ;)

Re: Linear Actuators
 

This is a very cool little project, I like it a lot.

We ran the 393s off TalonSRs in previous years, you can probably just use a small speed controller like an RC one if you want to. We actually overvolted the motors to about 10 or 11 volts without issues but we were only having them turn about 100 degrees.

Re: Linear Actuators
 

Thanks for the feedback,
That is interesting to run 10-11 volts... but this will be turning up to 10 revolutions. We tested using a voltage regulator that shows how much current is being drawn (as well as set it for ~=7.2 volts)... using the higher voltage gives more torque, so it may be worth it to use... it really depends on the load requirements of the manipulator. If we can get by with 6 volts for lighter loads (checking the current as well) then a direct connection may be fine (will need to research that a bit more though).

Re: Linear Actuators
 

I can't believe that nobody has yet chimed in with the following information:

As of the 2015 season, there IS an FRC-legal COTS electric lead-screw linear actuator available via FIRST Choice!

http://firstchoicebyandymark.com/fc15-037



As an FRC alumnus & mentor working at Bosch, I actually helped to coordinate the first donation of these motors by making sure that the documentation was detailed enough for the FRC crowd. That first donation was somewhat smallish by motor count, so they went pretty quickly; given the positive reaction to this first-year offering, however, I wouldn't be surprised to see the donation increase for the 2016 season.

Re: Linear Actuators
 

1/3" per second is pretty slow and there isn't an easy way to change that gear ratio, stroke, or mounting options like there would be with a custom printed one.

Re: Linear Actuators
 

Agreed! There are many benefits to designing an actuator specifically for the application at hand, and I have nothing but respect for the innovative approaches described above. I mean seriously, James' 3D-printed actuator is crazy awesome! :yikes:

I just put two and two together, however, that the OP was actually a Bosch intern scouting out the potential demand for the self-same linear actuator that I posted above. In other words, the Bosch linear actuator was only donated as a FIRST Choice option-- and only became an allowable COTS motor-- because of the initial positive response in this thread that showed that FRC teams would indeed be interested in having that option available.

Pretty cool to see this come full circle like that.