Maximizing Scanner Speed

(Please let me know if this is the wrong post location)

Hello all,

I am working on a university project using laser scanners. Right now I have a set of SpaceLas 30KPPS galvos. I have been reading a lot on how the KPPS rating is misleading in terms of physical speed of the mirrors.

My application requires maximizing speed. Essentially I am irradiating a rectangular area (covering every square inch) and want to scan over it as fast as possible. The scanning angle is not set in stone; ideally it would be as large as possible. Since the area is a simple rectangle, precision is not important. Do you have any ideas on what technology I should use? I have also read a little about LIDAR and MOEMs.

For smaller (1-10 degree) angles, is there a way to tune my current set of galvos to maximize speed? I don't have any experience in tuning these things.

Am I correct in assuming that for larger angles galvos are NOT the right tech?

I am interested in hearing about all ideas that are around $10K and less.

Thank you!

Hi ScoobyDoo,

Welcome to the forum! I'm the mechanical engineer for ScannerMAX ( www.scannermax.com ) , a company that designs and manufactures optical scanners. I just graduated from the University of Central Florida about three years ago.

For a starting point, I would recommend checking out the following book.

https://amzn.com/069274777X

The hard copy is around $49... but the kindle edition is only 99 CENTS and has all the same content.

This was written by the president of our company, Mr. William Benner. I did the majority of the illustrations throughout the book. It is a very easy to read guidebook on galvos and applications, what can and cant be done, tunings and applications, etc. It has a concentration on galvos but touches on other technolgies such as MEMS, Resonant scanners, etc

The key to scanning fast has several parts. The scanners, of course, are important. You want the smallest amount of inertia possible to accomplish your goals. So if your beam is 2mm diameter.... no reason to use a mirror set designed for 10mm beams. The smallest mirror which will accomplish your goals will help get where you need to go. Currently we offer a system called the "Saturn 1B" which is the fastest in the world for galvo technology at the moment. It will move a 3mm beam at 90K (7.5kHz) at around 8 degrees. 120K (10kHz) is possible at smaller (not as usuable) angles, but good for some applications. This mirror set will also scan a full 60 degree cone (albeit not at a full 90K/7.5 kHz).

If you want to scan any kind of wide, a MEMS device will not work for you.

I recently developed (and delivered this week) a custom designed mirror set which is capable of a 72 degree by 92 degree scan angle. Really anything can be accomplished, our scanners can do up to 110 degrees reliably. But there are trade offs. Wide angle, sacrifices speed, adds inertia, etc. Longer the mirror is to scan that wide, more likely you may hit resonance problems, etc.

Many of our tunings are done with an oscilloscope and others with the "ILDA test pattern". Both are useful tools depending what you are trying to accomplish. Most scanner amps I am familiar with can be tuned, servo gain, high frequency and low frequency damping will be your friends. Search "Buffo Scanner Tuning Guide", one of the members created it, it is very helpful. If you know exactly what angle you want to tune for, it is easy to optimize a tuning with an oscilloscope.

Our high end systems are tuned digitally... no pots to adjust. Several tunings can be stored on the digital servo driver and switched with the click of a mouse, for different applications. Much easier than the traditional way.

Feel free to email me (address below) with more details about your project (beam diameter, wavelength, realistic desired speed and angle (can be in kpps or kHz), waveform, etc, ) and I can see if we have anything 'off the shelf' to get you started. If not, we can discuss custom solutions, or options from our competitors which might be better for you if we can't meet your needs.

https://www.youtube.com/watch?v=DMLE...ature=youtu.be

Here's a video of the mirror set I designed for a client to scan a 2mm beam, 72 by 92 degrees. Theres also an interesting bit toward the end of the video showing how the amp is controlled via software.

Please note that there is a small angle step time and a large angle step time for any given Galvo-Amp-Mirror combination. It is not as simple as just "bandwidth".

[]
You can tune the existing scanners for a very high speed at a very small angle. You do this by monitoring the position sensor and galvo current on scope while driving with a square wave generator, then a specially designed ramp waveform, and re-tuning. However it involves understanding PID theory. You basically have Low Frequency Damping, High Frequency Damping, Integration, Servo Gain, to tune. Most beginners are better off ordering a set of Scannermax, tuned for your use, then doing it yourself. This is because the controls all interact.
[]
However it is possible to retune what you have, however a beginner is likely to fry the analog amplifier the first time he tries this.
[]
For very fast speed, you need modified waveforms that are somewhat out of the scope of a laser show forum. You also need to ignore the ILDA test pattern, which is optimized for vector imaging, not raster scanning. You can then optimize your optics train for expanding angle vs. spot size vs. linearity. The Laser Media test pattern and Grid test patterns are actually somewhat applicable for marking, as is the "Circle in the Square" portion of the ILDA pattern.
[]
You will note that Mr. Smith is giving you both KHz (Sine Wave) and ILDA numbers. This is part, but NOT ALL, of the picture when scanning. Scan speed is angle dependent. Small jumps or ramps can be amazingly fast, but the Galvo will slow down in a non-linear manor as angle increases. A good first order impression of a galvo is" "A DC Coupled Low Pass Filter with interesting Non-Linearities and Resonances"
[]
There are peer reviewed papers and corporate white papers out there on how to optimize scan speed for bidirectional rastor imaging, which seems is what you want to do.
[]
Part but NOT ALL of the Square Wave initial tuning procedure is hosted here for me by a friend of Lasers. Ignore setting the scale factor, that has been done at the factory for your Existing Galvos. This gives you an idea of how an "industrial" tuning procedure starts. It is by no means complete, and the part numbers in this document DO NOT DIRECTLY correspond to your existing scanner amp, even though it is copy of what is mentioned here. This is an excerpt of three different documents, and we've had it up for about ten years, so it is outdated. Part of the procedure for an older laser show technology is included, so ignore that.

http://www.skywise711.com/lasers/scanner/scanner.html
{}
As you develop this application you will invariably develop your own test pattern(s) as well.
{}

RANT MODE:
{}
The ILDA test pattern is not intentionally misleading, it is a old test devised by a group of very serious, highly qualified, Engineers and Technological Artists to allow information interchange in a specific application. The people who call it misleading have never learned enough about Galvo Mechanics and Electrical Engineering to understand what they are talking about in the first place. You need a bit of Calculus and time domain Electrical Engineering to understand the equation of state for a Galvo, and most of the complainers don't have that. So it is easy to "Dog" on the ILDA test pattern, but NONE of the complainers have EVER posted a successful alternative.
{}
END RANT MODE
[]
If you have a bit of time, Bill Benner's book (as mentioned) is very much worth reading for a Galvo Scanning beginner. It will let you visualize the tradeoffs between mirror sizes and scan angles. As mirror sizes increase, a given galvo slows down.
[]
Either the velocity or position sensor feedback test point is brought out on most "Clone" scanner amps, and you can learn much by hooking an oscilloscope there. Then inject a small square wave and look at the overshoot, undershoot, and ringing. Keep the square wave at low amplitude, as most of the clones save money by removing the protection circuitry that prevents you from overheating the galvo. Note as you expand the angle, the time required to compete a given "jump" grows larger. If your really good with electronics, you can then find the 0.1 Ohm current sensing resistor on the board and see how the amplifier waveforms change using the oscilloscope. Look at the settling time on the leading edge as well.
{}

Steve

:cat:
@ "RANT MODE:"

Sure beats dumping the stress onto your better half. >:)
Cheers :)

Quote:

---

Originally Posted by scoobydoo

(Please let me know if this is the wrong post location)

Essentially I am irradiating a rectangular area (covering every square inch) and want to scan over it as fast as possible. The scanning angle is not set in stone; ideally it would be as large as possible. Since the area is a simple rectangle, precision is not important.

---

Some simple points I think others have not addressed:

1. "I am irradiating a rectangular area (covering every square inch)" - It's impossible (from what I've ever seen) to scan a solid rectangle. The nearest you can get is using raster scanning which basically consists of alternate lines (so small areas are missed in between the lines). Raster scanning is very very scanner intensive and impossible to scan large and wide. A rectangle scanned ordinarily by laser is an outline only.

This is raster:

http://pangolin.com/images/LD2000/MonaLisa_in_laser.jpg




2. "Precision is not important". Hmm. If your scanners cannot handle the speed or angle required of them then you will see bowed lines in the rectangle not straight, so precision at the required speed IS important.

If you need to irradiate a volume I'd suggest optics to form the image. Now it's infinite speed as it covers the whole area evenly.
A second approach since your pattern is fixed is to dump the xy scanners and go to a resonant type system. It would help a lot to know the end use.

Ahh here is a second idea that eliminates a scanner. Use a line optic and scan the line in the other axis. Now there is no back and forth just up and down as a sweep. You can now use a spinning mirror wheel.

If you are going to use a laser for what ever reason, use cylinder lens to make a rectangle and then a telescope beam expander to make the size rectangle you want to illuminate. Infinite speed.

Find a light source that is the wavelength you want and build a mask to shape the image with a pair of lens in a spatial filter like arrangement. Basically a gobo like that for a stage.

Lots of ways to do this sans scanners.

That said if you do do scanners. Scannermax scanners rock.

Thank you all for the very informative replies. It will take me some time to process them all but I'll respond to a few now.

First clarifying my requirements:

We have no set speed requirements (as fast as possible, given physical and financial constraints). My application involves exploring the effects of scanned lasers on plant growth. This is where the rectangular and "lack" of precision criteria come from. Precision is not important in the sense that I am not attempting to form an image of any sort. I simply would like to bathe a certain area (starting with a single lettuce cultivar or similar and then increasing in surface area from there) in a uniform intensity of highly coherent light.

It makes sense that small angles produce faster speeds, along with the ideas that physical parameter such as size and mass affect the overall speed of the system.

Currently we have no laser size requirements, so I cannot really comment on the required beam diameter.

Raster scanning seems more like what I want to do. I am assuming that the spacing in between lines would be so small that the light intensity is effectively uniform across the whole area.


I wish I could provide more concrete criteria for you all. Such is the nature of my research... we play with technology and see what happens. I have some mechanical and electrical experience but most of my work is centered around biological systems.

To Ryan Smith, I will check out those resources and will no doubt be in contact later. Thank you again.

If the requirement is as fast as possible and a coherent light source I again go back to several non scanned solution.
You first need to tell us the wavelength of interest. If it is Blue you have up to 9W of light in one diode in as small a beam as you like which will incinerate the plant. This means you need to define an upward bound of the light input and afford the opportunity to spread the beam using optics and skip scanning entirely. There is an upward boundary for photosynthesis limited by the ability of the plant to supply the required ingredients and remove the waste products and even more important is heat. Even if the plant can keep up, the rate of reaction increases rapidly with rising heat. So you are introducing variables you are not controlling. That means that your results will be garbage.

You MUST insure that your keep the temperature at the same as your control. You can't do that just by measuring the leaf temperature because as the beam hot spot moves you are very rapidly heating and then cooling much faster than you can see with a thermometer or maybe even a photometer.

It seems to me the first thing you need to do is determine the flux of photons you want as a control in say lumens/cm2 and then assure you have the same flux on the test area. It will be extremely important to diffuse the coherent light and NOT to scan a concentrated beam on the leaf by spreading the beam to produce a uniform flux like the control.

Now once you do that then you can start asking questions like can I essentially q switch photosynthesis by rapidly scanning an area in such a way the leaf has time to recover from the excessive radiant input, thermal load, and waste removal issues.

It seems to me you could make a much more robust light source by using a xenon lamp and a monochromator. You then need only expand a single axis with a cylinder lens to make the square shape you desire.

I must be missing something you intend to do because this experiment does not make a lot of sense. Are you planning to take samples of tissue and look at the chromophores and chloplasts to see morphology changes or do analysis for CO2 uptake changes? Maybe do electrode implants to watch redux changes?

Last, and I think this is most important, how in the world are you thinking you can see growth changes in a plant on such a small scale of the entrie plant and isolate the background input from other sources of light while not killing the plant in the process. You surely will not see changes immediately except maybe smoke from 9W scanned on a leaf. You have to keep the plant alive. Ok one more point, no two plants are exactly alike so I suggest you get a clone which is available.

I'm not trying to piss you of or make you feel bad. I'm just saying that if you are going to take the time to do this you should know that there are a lot smarter people than me that will peer review you into a closet.

Quote:

---

Originally Posted by kecked

If the requirement is as fast as possible and a coherent light source I again go back to several non scanned solution.
You first need to tell us the wavelength of interest. If it is Blue you have up to 9W of light in one diode in as small a beam as you like which will incinerate the plant. This means you need to define an upward bound of the light input and afford the opportunity to spread the beam using optics and skip scanning entirely. There is an upward boundary for photosynthesis limited by the ability of the plant to supply the required ingredients and remove the waste products and even more important is heat. Even if the plant can keep up, the rate of reaction increases rapidly with rising heat. So you are introducing variables you are not controlling. That means that your results will be garbage.

You MUST insure that your keep the temperature at the same as your control. You can't do that just by measuring the leaf temperature because as the beam hot spot moves you are very rapidly heating and then cooling much faster than you can see with a thermometer or maybe even a photometer.

It seems to me the first thing you need to do is determine the flux of photons you want as a control in say lumens/cm2 and then assure you have the same flux on the test area. It will be extremely important to diffuse the coherent light and NOT to scan a concentrated beam on the leaf by spreading the beam to produce a uniform flux like the control.

Now once you do that then you can start asking questions like can I essentially q switch photosynthesis by rapidly scanning an area in such a way the leaf has time to recover from the excessive radiant input, thermal load, and waste removal issues.

It seems to me you could make a much more robust light source by using a xenon lamp and a monochromator. You then need only expand a single axis with a cylinder lens to make the square shape you desire.

I must be missing something you intend to do because this experiment does not make a lot of sense. Are you planning to take samples of tissue and look at the chromophores and chloplasts to see morphology changes or do analysis for CO2 uptake changes? Maybe do electrode implants to watch redux changes?

Last, and I think this is most important, how in the world are you thinking you can see growth changes in a plant on such a small scale of the entrie plant and isolate the background input from other sources of light while not killing the plant in the process. You surely will not see changes immediately except maybe smoke from 9W scanned on a leaf. You have to keep the plant alive. Ok one more point, no two plants are exactly alike so I suggest you get a clone which is available.

I'm not trying to piss you of or make you feel bad. I'm just saying that if you are going to take the time to do this you should know that there are a lot smarter people than me that will peer review you into a closet.

---

Our wavelengths of interest are approx 450 nm, 590-610 nm, and 630-650 nm. We will be doing both morphological changes and CO2 uptake.

As far as optical ways of spreading laser light, could you offer some specific technologies for the wavelengths of interest?

We have not decided on power yet, but will most likely start with 1W diodes and work our way up from there.

Quote:

---

Originally Posted by mixedgas

Please note that there is a small angle step time and a large angle step time for any given Galvo-Amp-Mirror combination. It is not as simple as just "bandwidth".

[]
You can tune the existing scanners for a very high speed at a very small angle. You do this by monitoring the position sensor and galvo current on scope while driving with a square wave generator, then a specially designed ramp waveform, and re-tuning. However it involves understanding PID theory. You basically have Low Frequency Damping, High Frequency Damping, Integration, Servo Gain, to tune. Most beginners are better off ordering a set of Scannermax, tuned for your use, then doing it yourself. This is because the controls all interact.
[]
However it is possible to retune what you have, however a beginner is likely to fry the analog amplifier the first time he tries this.
[]
For very fast speed, you need modified waveforms that are somewhat out of the scope of a laser show forum. You also need to ignore the ILDA test pattern, which is optimized for vector imaging, not raster scanning. You can then optimize your optics train for expanding angle vs. spot size vs. linearity. The Laser Media test pattern and Grid test patterns are actually somewhat applicable for marking, as is the "Circle in the Square" portion of the ILDA pattern.
[]
You will note that Mr. Smith is giving you both KHz (Sine Wave) and ILDA numbers. This is part, but NOT ALL, of the picture when scanning. Scan speed is angle dependent. Small jumps or ramps can be amazingly fast, but the Galvo will slow down in a non-linear manor as angle increases. A good first order impression of a galvo is" "A DC Coupled Low Pass Filter with interesting Non-Linearities and Resonances"
[]
There are peer reviewed papers and corporate white papers out there on how to optimize scan speed for bidirectional rastor imaging, which seems is what you want to do.
[]
Part but NOT ALL of the Square Wave initial tuning procedure is hosted here for me by a friend of Lasers. Ignore setting the scale factor, that has been done at the factory for your Existing Galvos. This gives you an idea of how an "industrial" tuning procedure starts. It is by no means complete, and the part numbers in this document DO NOT DIRECTLY correspond to your existing scanner amp, even though it is copy of what is mentioned here. This is an excerpt of three different documents, and we've had it up for about ten years, so it is outdated. Part of the procedure for an older laser show technology is included, so ignore that.

http://www.skywise711.com/lasers/scanner/scanner.html
{}
As you develop this application you will invariably develop your own test pattern(s) as well.
{}

RANT MODE:
{}
The ILDA test pattern is not intentionally misleading, it is a old test devised by a group of very serious, highly qualified, Engineers and Technological Artists to allow information interchange in a specific application. The people who call it misleading have never learned enough about Galvo Mechanics and Electrical Engineering to understand what they are talking about in the first place. You need a bit of Calculus and time domain Electrical Engineering to understand the equation of state for a Galvo, and most of the complainers don't have that. So it is easy to "Dog" on the ILDA test pattern, but NONE of the complainers have EVER posted a successful alternative.
{}
END RANT MODE
[]
If you have a bit of time, Bill Benner's book (as mentioned) is very much worth reading for a Galvo Scanning beginner. It will let you visualize the tradeoffs between mirror sizes and scan angles. As mirror sizes increase, a given galvo slows down.
[]
Either the velocity or position sensor feedback test point is brought out on most "Clone" scanner amps, and you can learn much by hooking an oscilloscope there. Then inject a small square wave and look at the overshoot, undershoot, and ringing. Keep the square wave at low amplitude, as most of the clones save money by removing the protection circuitry that prevents you from overheating the galvo. Note as you expand the angle, the time required to compete a given "jump" grows larger. If your really good with electronics, you can then find the 0.1 Ohm current sensing resistor on the board and see how the amplifier waveforms change using the oscilloscope. Look at the settling time on the leading edge as well.
{}

Steve

---

Do you have any info on how to go about accurately determining the linear scan speed (perhaps in real time) of my galvos?

Also any further resources on tuning (especially for high speed) would be much appreciated.

Thanks,