What's new
What's new

Synchronous Motion Linear Motor Questions

Johnny SolidWorks

Hot Rolled
Joined
Apr 2, 2013
Location
Rochester
Hey gang - I'm looking for anyone who has experience using multiple linear motors doing synchronous motions (i.e. 'electronically geared' motions, like a CNC.)

I've used SMAC 'Moving Coil' linear actuators in the past, and I'm using two LinMot motors right now in a non-synchronous manner, but I'd love to hear if anyone has ever used 3+ linear motors in synchronous motions. What brand, type, etc.

I think I remember hearing that a couple of EDM manufacturers are using linear motors, but those would really only be 2-axis systems, right?

Is anyone out there even doing anything like this that anyone is aware of? I've been asking the almighty Google, but I'm not coming up with anything.

Thanks All!
 
Linear motors were all the rage in CNC machines about 15 years ago. All of the big names were playing with them. Then reality set in.

A linear motor needs commutation sensors and linear scales to work. Then, you basically have a giant magnet laying exposed, so they need extraordinary covers and seals to keep shit out. Then you have to deal with the heat from all of those electromagnets distorting everything.

Unless you need something lightening fast with no exposure to contamination, you'd be a lot better served with a servo motor and a ball screw, just like 99.9% of the machine tool industry.
 
Linear motors were all the rage in CNC machines about 15 years ago. All of the big names were playing with them. Then reality set in.

A linear motor needs commutation sensors and linear scales to work. Then, you basically have a giant magnet laying exposed, so they need extraordinary covers and seals to keep shit out. Then you have to deal with the heat from all of those electromagnets distorting everything.

Unless you need something lightening fast with no exposure to contamination, you'd be a lot better served with a servo motor and a ball screw, just like 99.9% of the machine tool industry.

I actually didn't know that linear motors were a "thing" in CNC - interesting. Maybe I can find one to check out. Any big name/common machines that tried it? One of my major concerns is controlling the system, as well as being able to program it. All without reinventing the wheel, of course.

I don't follow the machines still requiring linear scales though - every linear motor I've ever used has built-in positional feedback that is as accurate or more accurate than an external scale.

And high speed, low contamination is exactly the application I'm looking at :-D
 
Last edited:
I have seen some 5 axis PAR water jets using liner motors on at least two of the linear axis. 3 motors on 2 axis, was a gantry style machine.

Not directly related to your question but I have seen a system with linear electric motors with resonant frequencies mapped and then programmed to play music through the machine tool.
 
All my pick and place machines are linear motors. That's 3-12 axes synchronized and moving 500 pounds around scary fast.

They're linear switched reluctance motors (no permanent magnets) and they do have external scales. I didn't build them though.
 
My little Citizen R04 (4mm max bar dia swiss lathe) has 5 Fanuc linear motors with Heidenhain scales on each one.

I would think an applications engineer at Fanuc would be able to put a system together for you.
 
We have integrated upwards of 200 linear motors of all sizes . . . Largest at Janicki Industries on a 5-axis mill with 80+ feet of travel, ~65,000 lbs of payload capable of 1/4G accelerations and capable of rapids approaching 4m/second although cable energy chains don't deal well with these speeds.

X-axis gantry uses two paralleled motors at each end of the gantry 4 motors total (If I remember correctly developing peak thrust of 16,000+ lbs) . . . Y-axis is a single motor as is Z. Wrist axes are each dual rotary motors (4 total) operating in contra-torque mode for zero backlash.

We use mostly Bosch linear motors and favor the AMO (recently acquired by Heidenhain) absolute encoders and use either Siemens S120 drives with Simotion or Sinumerik controls or Control-Techniques drives with any number of motion controllers either through EtherCat or an analog torque command. This combo eliminates the need for Hall effect devices required for commutation by most systems.

For really high speed positioning, check out a paired linear motor parallel kinematic arrangement by Bell-Everman - the Kaos stage. We have used this arrangement with rotary motors and timing belts with good results (accels to 5G's) but the linear motor solution is an order of magnitude higher performance with respect to accuracy and repeatability.

We have deployed in CNC applications and large number of coordinated motion applications (sub-micron positioning with air bearing stages in the semiconductor world on up to harsh environment glass container manufacturing industries.)

Here is a video we made about 10 years ago by a marketing guy that used to work for us . . . this machine is still running today.


This machine was originally designed in 1999 and is running Control-Techniques UniDrives, Anorad Linear motors, Lika Linear Encoders, and Delta Tau motion controller. Encoder leads in this case carry quadrature incremental RS422 level signals and cables for motors and encoders are 300 feet long. We are using the Anorad Halls that are then converted to differential signals using a US Digital single ended to differential encoder adapter. The signals are brought into the UniDrive as typical commutation U, U/, V, V/, W, W/ signals and the UniDrive thinks it is running a regular brushless rotary servo motor.
 
Last edited:
We have integrated upwards of 200 linear motors of all sizes . . . Largest at Janicki Industries on a 5-axis mill with 80+ feet of travel, ~65,000 lbs of payload capable of 1/4G accelerations and capable of rapids approaching 4m/second although cable energy chains don't deal well with these speeds.

X-axis gantry uses two paralleled motors at each end of the gantry 4 motors total (If I remember correctly developing peak thrust of 16,000+ lbs) . . . Y-axis is a single motor as is Z. Wrist axes are each dual rotary motors (4 total) operating in contra-torque mode for zero backlash.

We use mostly Bosch linear motors and favor the AMO (recently acquired by Heidenhain) absolute encoders and use either Siemens S120 drives with Simotion or Sinumerik controls or Control-Techniques drives with any number of motion controllers either through EtherCat or an analog torque command. This combo eliminates the need for Hall effect devices required for commutation by most systems.

For really high speed positioning, check out a paired linear motor parallel kinematic arrangement by Bell-Everman - the Kaos stage. We have used this arrangement with rotary motors and timing belts with good results (accels to 5G's) but the linear motor solution is an order of magnitude higher performance with respect to accuracy and repeatability.

We have deployed in CNC applications and large number of coordinated motion applications (sub-micron positioning with air bearing stages in the semiconductor world on up to harsh environment glass container manufacturing industries.)

Here is a video we made about 10 years ago . . . this machine is still running today.


I was hoping you'd chime in Motion - thanks for the great info! And your bottle stacker is a hoot! It looks so simple unless you realize what's actually going on. (I used to work for a company that did bottle packing equipment for some of the biggest name brands out there.)

Out of curiosity, is integrating a Sinumerik controller any harder than integrating a typical PLC?
 
We use mostly Bosch linear motors and favor the AMO (recently acquired by Heidenhain) absolute encoders and use either Siemens S120 drives with Simotion or Sinumerik controls or Control-Techniques drives with any number of motion controllers either through EtherCat or an analog torque command. This combo eliminates the need for Hall effect devices required for commutation by most systems.

At a previous job, we used Fanuc linear motors with Fanuc drives, controlled by a Fanuc CNC. Keep in mind that the linear motor needs cooling. You can use air, but the maximum RMS power output is quite low. We always used a chiller, which also cooled the spindle.

Feedback was from Heidenhain LC sealed glass scales. For short travels we used absolute scales, for longer travels the scales were distance coded. During startup, the CNC/drive was able to use the incremental track to run the axis until it figured out where it was.

I know the controls guys tried the inductive sensors (I think they were AMO's, but I'm not sure), but they ended up tearing them out and putting in a Renishaw system. I don't recall if it was optical or magnetic. They explained why they made the switch, but I don't remember the answer as it was years ago.

Motion Guru, have you tried out the Bosch setup where they laser weld the AMO scale to the rail and the readhead is mounted to the car? It looked quite nice but I never got a chance to use it.
 
Hi Daniel . . . yes - we have used the Guided Rail Scales with good results and in fact we are shipping a machine today with that setup. On smaller machines, we use the AMO linear scale as one of the support bearings (these are Bosch linear bearing rails / cars). The scales are not welded . . . a simple shallow slot is machined in the top of the rail and clamps added at the ends of the rail to hold the inductive slotted strip and stainless cover in place. You pull the clamps, lift the scale out and then you have access to the socket head cap screws that mount the rail. Once the rail is in place, set the scale in the shallow slot and slide the bearing car on and then put the end clamps on.

We do RMS calcs for every system and water cool when necessary or if water cooling is not possible, we simply upsize the primary and mount to a heat dissipating surface. I am fairly certain that the Fanuc motors were made by Anorad initially as they were physically identical with identical mounting features and performance specifications, motor lead and hall sensor wire colors - when Rockwell bought Anorad, they had a rash of quality issues and we switched at that time to Bosch and haven't looked back. I believe but don't know for sure that Fanuc now manufactures their own linear motors based on the Anorad original design.

Reneshaw encoders are good quality but seem to be very fussy with regard to alignment issues - we have had very good results with AMO both incremental and absolute. With very long travel using incremental - we laser burn a grid at the same pole spacing as the magnets and we mount three prox switches to emulate Halls A, B and C . . . simple, easy to trouble shoot and not proprietary.
 
Hi Daniel . . . yes - we have used the Guided Rail Scales with good results and in fact we are shipping a machine today with that setup. On smaller machines, we use the AMO linear scale as one of the support bearings (these are Bosch linear bearing rails / cars). The scales are not welded . . . a simple shallow slot is machined in the top of the rail and clamps added at the ends of the rail to hold the inductive slotted strip and stainless cover in place. You pull the clamps, lift the scale out and then you have access to the socket head cap screws that mount the rail. Once the rail is in place, set the scale in the shallow slot and slide the bearing car on and then put the end clamps on.

I haven't seen that setup before. Does the slot come from the factory or do you put it in? I was thinking of their IMS system, where the scale is on the side and there's a stainless cap laser welded over it. Obviously, you're limited to the maximum rail length (is it 5m for Bosch?), but it looked stupid-simple to design in and install.

https://md.boschrexroth.com/modules/BRMV2PDFDownload-internet.dll/R999000487_2017_06_EN_IMS_media.pdf?db=brmv2&lvid=1201902&mvid=13009&clid=20&sid=4FB432DF37AA855774F2B1BBC7B90B1D.borex-tc&sch=M&id=13009,20,1201902


Reneshaw encoders are good quality but seem to be very fussy with regard to alignment issues - we have had very good results with AMO both incremental and absolute. With very long travel using incremental - we laser burn a grid at the same pole spacing as the magnets and we mount three prox switches to emulate Halls A, B and C . . . simple, easy to trouble shoot and not proprietary.

The only secondary-feedback, long-travel systems I ever worked on had the Renishaw optical scales. After the beds and rails were aligned during install, we'd fasten down the optical scale. Two readheads were mounted to a carriage. The carriage was on mini-linear rails and sprung towards the scale mounting surface. Cam followers kept the readhead at a constant distance from the scale. There were two readheads so you could go over a gap in the scale. While we installed a one-piece scale, in the event of damage you can cut out a damaged section and splice in new scale. When it goes over the gap, one readhead is on the old scale and one on the new, so the CNC never loses track of where it is.

The only short axis I ever used the Renishaw optical scales on was a pivot axis. IIRC, the scale was epoxied to a machined surface and the readhead was bolted to the other part. That only had one readhead, as the scale was only ~8" long. I remember it being pretty easy to set up.

As far as reliability, the scales were under the way covers, but everything in that factory was covered in CFRP dust and it was right by the ocean and non-air conditioned, so everything wanted to rust. We wiped the scale clean during yearly PPM, and I don't recall ever having issues with it.

Now that I'm thinking about it, I'm going to say that the reason the controls guys dropped the AMO was a noise issue. They probably could have gotten it to work, but I think it was easier to switch to a known system. Are there any tricks to the AMO system? For example, the trick our controls guys figured out for the Heidenhan LC scales was to make the scale housing and readhead mounts out of G10. This isolated the scale from the machine and apparently made everything work much better.
 
That system is the Bosch system that uses AMO components . . . the system we purchase is an AMO system that uses Bosch components . . . it is slotted and pinned from the factory and you can buy scale tape lengths to 9m for absolute with shorter rail sections set up for mounting. You can order exact lengths as well - if your design is solid, it is a good solution and if you have a little motion stage, you can add Bosch bearing cars for load carrying and then tether the encoder.

lmf-310-lmf-3150.png


and . . .

LMK_Guided.png


In one application, we have used these in a reactor vessel cleaning machine with glassified carbon dust and sulfuric acid vapor . . . very bad environment for rust. We added automated oil lube to the system (which flooded the bearings, read head, and electronics with oil) and this seemed to work well for years of operation while the larger size 45 THK bearings for the actual machine needed replaced every 2 years due to ingress of carbon powder and rust issues.

There is an interpolation box that we mount near the read head and then go DriveClick for the long distance to the drives cabinet . . . we haven't tried carrying the low level signals further 10m tops . . . more often less than 3m.

Also, depending on what drive we are using, we often change the encoder fault bit to a warning bit and keep track of how many times it drops out (absolute systems only) . . . then if it becomes apparent we have a problem, we set an alarm to tell the operator that they need to inspect the encoder. Generally, this has never been an issue and the only time we have had problems is when someone has used a prybar on top of the bearing rail to "adjust" something that didn't need adjusting.
 
A note in passing.
Linear PM motors have really large loose magnets.
Make sure you know how to handle them safely.

Definitely true.

That magnetic attraction doesn't go away once it's installed either. The magnetic pull will try to deform the structure the motor is bolted to, so take that into account during the design process.
 
Hey Motion - do the motors you guys are using having built-in e-stop functionality/programming? The vendor it was decided for me to work with has essentially told us that in the event of an e-stop, their drives will just drop power, meaning the linear motor motions will either drop (vertical) or 'coast' to a hardstop (horizontal.) This seems kind of unacceptable to me.

And in the event of a total system power drop (i.e. plant power outage) do you guys have brakes built it? Linear rail brakes maybe? The problem with our application is the linear rail brakes don't have nearly enough braking force to stop the motion, so I'm essentially looking at building a custom e-stop/power-drop brake that can stop our system and still be manually overridden for maintenance, etc.
 
I can't speak to the motors that Motion Guru uses, but on Fanuc, e-stop kills power. Fanuc drives (on both linear and rotary motors) will actually bring the axis to a faster stop under active control than they will if you break the e-stop loop.

Relying on the motor to actively stop the machine is risky. What do you do if there's a power outage?

Spring-set, pneumatic release rail brakes are very easy to integrate, but, as you've figured out, they aren't great for dynamic braking. They're much better at holding something in place once it's stopped. On some large linear motor driven vertical axes that my old company did, they used an external spring-set, pneumatic-release brakes that were designed to brake wind turbines. These were caliper brakes and instead of a round rotor, they ran on a long piece of bar stock. I don't remember the brand, but off-the-shelf brakes do exist that can be used for this.
 








 
Back
Top