Beam Director

I need to steer a large diameter beam (50mm in diameter) with remote (computer controlled) input. The speed is REALLY SLOW. The beam is pulsed at up to a maximum of 4 PPS and the pulse is essentially instantaneous and so nearly 250msec are available for the mirror(s) to slew and settle.

I originally thought a pair of big mirrors on two of the Pangolin Vrad motors would work, but these mirrors have to be really huge and the controller will also need an amplifier to drive the motors. What about adapting the moving head mount on a Gobo? With the light removed and a single mirror around 150mm square the X and Y can be handled by a single mirror and the control for these already exists as part of their standard operation. How rapidly can these be aimed? Would there be a best choice for my application? If they loose a signal or power do they go nuts or can they remain steady or move to a "safe" mode?

what about something like this?

http://www.zaber.com/products/produc...M&tab=Features

Quote:

---

Originally Posted by planters

I need to steer a large diameter beam (50mm in diameter) with remote (computer controlled) input. The speed is REALLY SLOW. The beam is pulsed at up to a maximum of 4 PPS and the pulse is essentially instantaneous and so nearly 250msec are available for the mirror(s) to slew and settle.

I originally thought a pair of big mirrors on two of the Pangolin Vrad motors would work, but these mirrors have to be really huge and the controller will also need an amplifier to drive the motors. What about adapting the moving head mount on a Gobo? With the light removed and a single mirror around 150mm square the X and Y can be handled by a single mirror and the control for these already exists as part of their standard operation. How rapidly can these be aimed? Would there be a best choice for my application? If they loose a signal or power do they go nuts or can they remain steady or move to a "safe" mode?

---

I don't think this will do it. If the beam is striking the mirror at a 45 degree angle, the optic will need to be a minimum of 3 inches in diameter and the beam may need to be directed such that the foreshortening will require as much as 5 inches. Also, the extreme resolution implies that the max speed is slow. I am hoping to get say 100 degrees/ second or better between pulses. The range at 20 degrees optical is not too bad.

is this for lab or light show?
Stepper motors?
would u b interested in a older big galvo w/ paddle mirror? there is some older stuff like for fiber fed heads for YAG

Quote:

---

is this for lab or light show?

---

Neither, really. I am trying to steer a pulsed laser to track and then target moving objects (inanimate!). I don't know how big the paddle mirrors get, but other than big industrial galvos, I suspect nearly 80mm minor axis would be very rare. The only reason I don't think this project is crazy hard is the very slow speed for repositioning. Stepper motors are a good idea if micro-stepping resolution is available. The precision jitter error of these micro-steps is acceptable. The reason the moving head is interesting me is that the Y mirror in a mirror pair becomes tremendous. For a clear aperture without vignetting the Y has to be at least 3 times as long as the 50mm beam is wide!

a small turret would work too i guess then. or a mirror on a yoke. how accurate r u trying to get? with OpenCV or Openframeworks, an arduino and microstepping drivers you could pull it off affordably.
https://www.youtube.com/results?search_query=sentry+gun

have you looked into hobby servos? Very easy to control.

I have some 3" diameter 45' directors on ball bearing bases designed to be turned. I'll take a pic when I get home. I have a LOT of them.

Steve

I got a neat suggestion from an interesting source.

http://www.hobbyking.com/hobbyking/s...ontroller.html

For those with drone experience this might not be the best choice, but in general these gimbals seem to be a really good solution. They slew to 2000 degrees/ second, are made to move weights similar to a good quality (thick) 150mm square mirror and they are inexpensive. They need to interface with a computer and the dongle suggests this one does.

http://www.lasershs.com/robolase_bounce_mirror.htm

I remembered these...

Hey Planters

The gimbal motors might be a good option as you can get them in several sizes and can be quite cheap, however you may need to take into consideration the inertial mass or the rotor plus mirror

These guys have a selection
https://aerialpixels.com/product-cat...gimbal-motors/

Not sure the best method for driving them for this situation though

On a side note for gimbal controllers, the one on my quad will shudder slightly when the quad is sitting on a bench, so assume the PID loop error or something

Planters, I have three complete sets of 6350 large mirror galvos with amps. The amps are class I servo loops with adjustable slew rate limiters. This means they love to drive big mirrors. Have everything. Amps, PSUs and interconnects. AMPS have the 12 bit DAC chip with parallel input and by adding a jumper they switch to single ended analog. These things LOVE big mirrors. You could easily go over Cambridge's speced Arpeture size. My sets are from a layup machine for fiberglass, have spent their entire lives in clean enclosures. They guided a human in laying up airplane wings. I bought them used for a Canadian guidence system project that was unfunded half way thru.
Would just like to recover my money at this point.

Neat thing about a slew rate limiter, these galvos will try to scan the ILDA test pattern, and do so quite well. The limiter keeps them from scanning the portion of the image that is too fast to follow. Which means even with a huge Y mirror, these things will actually scan a decent graphics show.

Mill your own large mirror mount and be very happy.

Attached are some docs.

Steve

Steve,
Wow! This will take me a little while to wade through, but your description makes it clear these would work. Short of bending the shaft, these will simply refuse to move too fast if I load too much mass or demand too much speed. I like that. Let me scan through this and I'll send you a message. Thanks.

I want to bump this thread with a request.

I would like to try out my idea of a modified moving head mount as a beam director. I am looking for suggestions for a manufacturer or better yet, for offers from anyone that has a nice, solid, fast and functioning yoke style mount. I do not need the head or light source, the optics or any of the electronics that drive the head and it seems like a waste (and expense) to buy one of these things and scrap this. I can fabricate the mirror and its mount and the large size of this 2 axis mount means that I can remote this beam director from the laser which is much too heavy to move to say an elevated location and not worry that the beam has to be dead on center otherwise it will spill off the sides of the mirror.

Would this work

http://www.lasershop.de/second-hand/...irror-352.html

If 1.8 or 0.9 degree steppers with micro-stepping is good enough angular resolution for your application, two mirrors would most likely be less mass than a single mirror on a yoke or a yoke holding a fiber launch. that is, unless you want a very wide scan angle, then a yoke could do 360 on one of the axis. they real key would be accelerating and decelerating the mirror at a rate that matches the motor and moment of inertia of the mirror assembly to avoid missing micro-steps without encoders/servo loop.

a thing not on this CAD model would be hard stops for dumb homing upon initialization.

Attachment 46550Attachment 46549

you could do this with surplus parts for pretty cheap.

loopee,
That would work if the speed is acceptable. I don't know what the degree/second is. The mirror size is good so even if I have to replace it with a mirror having a higher damage threshold and it is a thick mirror, this should be able to move it.

lightlinked,
This should work as well. The torque of the steppers and the speed should be more than enough. The angles I want to cover are on the same order of a typical laser projector and the resolution should be a few mrad. The beam will cross free space to the tripod mounted beam director and there be reflected to the target. The motors would have to operate with micro-stepping or be geared because even 400 steps/revolution is 1.8 degrees with the 2:1 optical reflection. The hassle is I need everything. I need motors, drivers, a controller and mirrors and then I have to integrate it. This isn't impossible, but it is a fair project.

Anaheim Automation makes my preferred microstepping driver, cost is reasonable, and they are a stock part.

Watch out for motor/load resonances and the need to ramp the driver speed up and down. Steppers do not start/stop on a dime. They have a nasty nonlinear load/speed curve, and the thus torque falls off at high speed. (Which I am now fighting on a motion stage development project)

you might be much better off with a gear train between your motor and mirrors, with a brushless motor.

The Cat's Meow of fairly easy to use motion control chips are made by LSI:

http://www.lsicsi.com/products.htm

Steve

The two mirror option is more ideal if the laser aperture is close to the assembly since it is ideal to enter it at a precise angle. If you wanna hit this device from far away, a single mirror on a yoke becomes a little more practical. However, it will be slower on the axis that holds the second axis and mirror due to having to move the mirror motor, yoke frame and mirror. however, it is still quite quick.

are we to understand that this application has nothing to do with art/entertainment? ie are you trying to avoid using lightshow parts? ie. you want something more industrial or R&D oriented?

how are you going to "computer control" this? what are you doing for software?

This application is not LIMITED to art and entertainment, but that's where I'm starting. I think heavier duty components whatever the source are more appropriate. Unlike a typical scanner application where the mirrors are light and fast this will be a "slew,stop, settle,pulse...repeat" at a rate from less than once/second to 4Hz. One way around the precision entry challenge is to fix both the laser and the scan head and then fire the laser at a fixed target and a known scanner orientation. This then establishes a parking position from which X and Y slews are measured.

I am planning to use open CV:
http://docs.opencv.org/trunk/modules.../tracking.html

Steve may have a good solution with the scanners because they are digital and hopefully can interface with a PIC or Arduino and I know that there are algorithms in open CV that will output a signal based on frame position to operate a camera scan motor. They could as easily output a 12V signal to the laser to fire after a sufficiently long period for the scanners to reach a target location.

is this going to be a high power laser? ie class IV? is this going to be around people?

Yes, this is going to be high power and no it is not. It is inside a structure and the area is evacuated during testing.

here is a sketch of a yoke with motors close to on axis to the center of the mirror's center, which allows for the use of a round mirror instead of an elliptical one as needed when you have the tilt motor behind the mirror like with the stage lights. this also gives you more to tilt and pan, or even continuous tilt. this simple design is unbalanced though on the pan, which is acceptable since its not supposed to be spun around all fast, and the pan range is limited by the fact that wires gotta go to the tilt motor.

keep in mind this is a drawing so...

you can use any motor, geared or direct or even timing belt driven.
if you do direct drive you can get steppers with the shaft coming out the back where you can add an encoder, or a magnet and a magnetic position sensor. there are also servo combos of motor/encoder/controller for industrial applications with rs485 or rs232 or ethernet.
you could have a micro controller handle all the motion control tasks separate of the image processing computer. ie accell decel rates, keeping track of position of the mirror. then just send this mirror controller position commands.

Attachment 46551

I see what you mean. What about the link I provided in post #9 above:

http://www.hobbyking.com/hobbyking/s...ontroller.html

Not that this is the "right one", but the design is similar and presumably these have controllers that are compatible with digital output.

the brushless camera gimbals rely on an IMU for positioning. its a different approach since its designed to be moved around, the application is pretty different unless you want to move the gimbal around willy nilly.

some of the digital hobby servos are actually pretty nice. idk about repeatable positioning or accuracy though. i dont think its much better than 10 or 12 bit.

http://www.trossenrobotics.com/p/cus...bot-joint.aspx
http://www.hobbyking.com/hobbyking/s...16Sec_75g.html
http://www.hobbyking.com/hobbyking/s...hort_Arm_.html

I just remembered FLIR makes automated pan/tilt mounts.
http://www.flir.com/mcs/view/?id=53484

Lightlinked,
I liked your link to the Tossen site and at first was intrigued with:
http://www.trossenrobotics.com/robot...ision-tracking

But, your link is structurally superior and the very high torque (cost) of the included motors is unnecessary. I contacted them about a lower torque, but equivalent resolution version.

luxlid,
I will contact them, but I suspect that the cost may be very high. I also wonder if these mounts are designed for slow, stable positioning and tracking (they have that look) vs rapid slew and pointing. I'll find out.

I feel like I'm stumbling through the forest bumping into obstacles and taking some false paths. There may be several ways to get there, some better than others, but I have a sense I am going to get there.
Thanks

I believe that the Trossen solution may be the right one. However, their web site needs work. They have no reviews...on anything. Say you pick a servo. You will be linked to some compatible components such as brackets or drivers, but not all of them and often not the best options. Nevertheless, their pan and tilt turrets with the two servos daisy chained and the USB compatible controller looks sturdy, fast and configurable. Their Roborelm vision software may also be more complete than the freeware open CV, but here I will defer to the computer geeks in the house. I'll use what THEY want. I want to use an inexpensive camera, but I believe that I may need to use something like the Point Grey cameras because I'll need to shutter the camera for the brief period during the laser flash.

I also may add a diode laser into the beam line with a short pass dichroic or a small pick off that sends the diode laser on a parallel course, sort of like knife edging, so that setting up the remote turret (or scanner) is easier.

I have yet to here from FLIR, but I fear the worst.

What is the exact solution by Trossen you're referring to? Having no reviews at is also pricking.

The WidowX with the AX/MX DYNAMIXEL Servo Manager Kit and the Roborelm software package looks promising. This is without the use of any modified kitchen utensils:)

But seriously, the lack of reviews is strange. They have been around for a few years, I talked to them on the phone and the technician really knew his stuff. However, other than Hobbyking's rather lightweight components, I am unaware of any mid or upper level robotics supply sources.

something along similar lines over at the university of Tokyo. it looks like they are using big galvos
Lumipen 2: Robust Tracking for Dynamic Projection Mapping
http://www.k2.t.u-tokyo.ac.jp/mvf/Lu...n/index-e.html
https://www.youtube.com/watch?v=ZuSUHuSceYc
https://www.youtube.com/watch?v=p7IL0Gvux7U

Planters should have received a box by now with a pair of big galvos, bigger amps, and very , very, very large overkill power supplies.

Steve

lightlinked,
Thanks, that is basically what I need to do and hopefully Steve's large galvos will accommodate it. The only significant difference will be that the high energy levels and low divergence will benefit from maximum galvo size and I need to shutter the camera when I discharge the laser. The latter might may effect the tracking performance.