I've read many PL threads about using ILDA test patterns to determine or calibrate a laser's performance.
I am NOT asking that same question.
My question is how would a competent lab empirically and objectively measure the specifications of an RGB laser projector? It may sound like a subtle difference, but I imagine the answer would include a setup to enable measurement of the laser output power, wavelengths and maybe a high-speed camera or device capable of measuring the difference between the demand position present in the control signal and the actual position of the optics, not to mention the divergence of the combined beam. Likewise galvo strain, while audible to the naked ear is presumably also objectively measurable with the right instruments.
Top manufacturers must have such a setup in order to make defensible claims about their product's specifications. Developers of laser control software must also have their heads around this in order to make simulations and to do signal processing to "optimise" the output. I've read lots of discussion on PL about what various software does.
Also, related, are there independent testing or certification bodies that do this kind of testing? My googling has so far failed!
My goal is to learn about the engineering, physics and mathematics of show laser performance and how it is measured. I expect this is a deeper dive than is typically required for putting on a great show so while it's not everyone's cup of tea, I am hoping some of you here have the experience to answer this.
I've read many PL threads about using ILDA test patterns to determine or calibrate a laser's performance.
I am NOT asking that same question.
My question is how would a competent lab empirically and objectively measure the specifications of an RGB laser projector? It may sound like a subtle difference, but I imagine the answer would include a setup to enable measurement of the laser output power, wavelengths and maybe a high-speed camera or device capable of measuring the difference between the demand position present in the control signal and the actual position of the optics, not to mention the divergence of the combined beam. Likewise galvo strain, while audible to the naked ear is presumably also objectively measurable with the right instruments.
Top manufacturers must have such a setup in order to make defensible claims about their product's specifications. Developers of laser control software must also have their heads around this in order to make simulations and to do signal processing to "optimise" the output. I've read lots of discussion on PL about what various software does.
Also, related, are there independent testing or certification bodies that do this kind of testing? My googling has so far failed!
My goal is to learn about the engineering, physics and mathematics of show laser performance and how it is measured. I expect this is a deeper dive than is typically required for putting on a great show so while it's not everyone's cup of tea, I am hoping some of you here have the experience to answer this.
It's certainly possible. In fact, there are several ISO and CIE standards for measuring many of the quantities you mention. The light show sector doesn't really lend itself to knowing or publishing such data though, due to the additional cost. Sadly there has been a race to the bottom for many years, with folks just fixated on output power, and different crude methods of quoting divergence.
I've measured the performance of a lot of audience scanning lasers over the years, and only found one manufacturer to stand out by a country mile in terms of consistency between projector output. With the majority, it seems to be output power that is main driver.
It's certainly possible. In fact, there are several ISO and CIE standards for measuring many of the quantities you mention. The light show sector doesn't really lend itself to knowing or publishing such data though, due to the additional cost. Sadly there has been a race to the bottom for many years, with folks just fixated on output power, and different crude methods of quoting divergence.
I understand what you're saying, and the result seems to be that the minority of quality manufacturers are treated differently in terms of their published specs. ILDA and manufacturers do publish specs, however, and I suppose what you're saying is that customers are unable to verify these claims and perhaps they don't really care.
Originally Posted by JStewart
I've measured the performance of a lot of audience scanning lasers over the years, and only found one manufacturer to stand out by a country mile in terms of consistency between projector output. With the majority, it seems to be output power that is main driver.
At first I was confused by how output power was so exclusively discussed. To me, output power often seems to be more of a curse than a benefit. I have been more focused on scanner performance. However, then I realised that the faster you scan, the more power you need to spread across that scan path, otherwise brightness suffers as the path length grows.
I've been just getting into lasers and am writing software as a part of this. It's becoming quite an education.
Creaky Old Award Winning Bastard Technologist Infinitus Excellentia Ion Laser Dominatus
Join Date
May 2007
Location
A lab with some dripping water on the floor.
Posts
10,039
The problem is a Galvo is not a simple device. A small or large jump response time, as used in the marking industry, does not come anywhere near describing what happens in laser show vector graphics. While everybody says "Oh the ILDA Test Pattern doesn't do this..." , most have little idea of what would be involved in presenting a huge 3D vector set of data representing Galvo performance. Performance changes with jump angle, waveform frequency components, and past history of motion. Mirror inertia varies widely with manufacturers, etc.
Point of the matter, the ILDA test pattern provides a nice set of constraints to measure the behavior of an extremely complex beast that can be done by engineers, technicians, and artists, quickly. But the test pattern was not really intended for absolute measurement. Why that is the situation comes up later in this post.
The circle in the square roughly measures the -3dB response frequency for small jumps at a given angle.
Fact of the matter, this has been debated to death, for decades, and not once has anyone ever came up with a better idea for our industry other then the various test patterns. Quantifying Galvo performance is difficult.
How do you read a complex table of jump angle vs frequency vs inertia vs coil resistance vs ma/Per degree vs Back EMF constant vs position sensor constants, vs resonances vs shaft diameter blah blah blah ? The answer is you don't bother with it, as most users simply do not care, and Galvo technology is nearing it's technological limits. How do you test an analog scanner amp with nine variables adjusted by twenty turn potentiometers from model to model vs manufacturer to manufacturer when they often use differing PID loop topologies?
The point of the matter is your facing an infinite data set to display and interpret, let alone compare. The majority of users just don't care to have to try to interpret that, and most do not have the required technical education.
Just because the standards lack "precision engineering documentation" does not represent a major failure in this industry.
While a Galvo manufacturer might have a Dynamic Signal Analyzer at the plant, the great majority of us simply do not have the budget. I have a used one, but I'm a minority and my day job means I need a large variety of instruments, even in the home lab.
I understand where your coming from, but there is very little desire in our industry to bother with this topic. History has shown that the ILDA test pattern and some other patterns used in tuning have provided a reasonable standard for information interchange vis a vis the Galvo's complex mechanical response. That is its job, to insure artwork from XYZ laser corporation can look reasonably good on any properly tuned projector in the world. If you spend some time with actual galvo tuning, scanner amp design, and know your laser show history, you'll find it is actually a pretty good solution to a complex problem.
Laser modulation and beam quality is a separate issue. Beams from overlapping diodes in arrays do not nicely fit the mathematical models and instruments used in laboratory or industrial lasers in terms of mode structure and divergence. Unless your using a OPSL or Ion Laser source, the data can be quite meaningless. Often times no two diode arrays from the same manufacturer are even close to identical. Again, as diode array beam patterns vary wildly with distance and instantaneous intensity, you can end up having to quantify a nearly infinite set of data for each wavelength. I own both CCD based and mechanical scanned beam analyzers. I have found it very difficult to express what I measure with the scientific instruments on show lasers, whereas lab lasers often present simple beams with a good Gaussian fit. Testing at just full power is not a full solution, these sources vary widely. Often the outer portion of the beam is distorted and the inner portion contains most of the power and is quite good quality. How do you define a standard with that condition?
The current ILDA standards let you test a small subsample of the complex data at one easily constrained set of test conditions.
how would a competent lab empirically and objectively measure the specifications of an RGB laser projector? <snip> I imagine the answer would include a setup to enable measurement of the laser output power, wavelengths and maybe a high-speed camera or device capable of measuring the difference between the demand position present in the control signal and the actual position of the optics, not to mention the divergence of the combined beam. Likewise galvo strain, while audible to the naked ear is presumably also objectively measurable with the right instruments.
Everything you've mentioned can be quantified. But as Steve mentioned above, most end users don't need this level of precision, thus there's never been a market need to develop a better testing standard, which is why we don't have one.
Something else to keep in mind is that the performance of the laser source itself is much easier to quantify than the performance of the scanning system. There are off-the-shelf instruments that can give you a very accurate picture of the beam profile with detailed measurements of the optical power gradient from center to edge. Same goes for analyzing fluctuations in output wavelength. You'll often see this information included in the datasheets for very high end laser sources.
But the scanning system is vastly more complex, with multiple performance variables, so any testing standard will have to start with a number of assumptions regarding the use case. (That, or else the testing protocol, and the resulting data, will end up being prohibitively exhaustive.)
The suite of test patterns (including the official ILDA test pattern) all represent a very specific use case (including a defined scan speed and scan angle) while being easy to implement and - most importantly - easy to evaluate. You can glance at the pattern on the wall and instantly get a very good idea as to the performance of the system, including what controls to adjust to make improvements. That's why they've been "good enough" for the laser projector industry for a very long time.
Yes, one could define a new test standard that only looked at the input signals and the position feedback signals, but this doesn't say anything about what those input signals should be, or what the pass/fail criteria should be. And there's the problem: galvos respond differently across a wide range of frequencies. Do we test from 0 Hz all the way to 5KHz? At what amplitude? And how much distortion is acceptable? The answers to these questions change depending on your use case.
Top manufacturers must have such a setup in order to make defensible claims about their product's specifications.
Yes, they do. For example, Pangolin's ScannerMax division has links to detailed performance data for their Saturn scanners on their website. But you'll notice that every test case has defined frequency ranges and amplitudes. Thus, each test represents a use case. Also note that most people would be hard-pressed to identify which ones are relevant to laser projectors. And while ScannerMax provides much more information than most other scanner manufacturers, they can't test for everything.
ScannerMax has also published data about rotor inertia, coil resistance, heat transfer, mirror stiffness, torque, etc, even though most of this information is of limited utility for laser projector manufacturers. The reason they do this, of course, is because ScannerMax is selling scanners to lots of companies that have nothing to do with laser light show projectors. However, for the laser projector industry, the general consensus is that the existing test patterns provide sufficient proof of performance for the intended use case.
Regarding the official ILDA test pattern specifically, you may enjoy these articles that discuss the pattern in greater detail:
My goal is to learn about the engineering, physics and mathematics of show laser performance and how it is measured.
This is getting out of my knowledge zone, but if you really want to go down that rabbit hole, I suggest you start by looking up the patents for the original Cambridge Technology 6800 scanners. Then look at the patent for the 6215 scanners. Finally, look up Bill Benner's patents for the ScannerMax Saturn scanners, which pretty much reflect the current state of the art.
After you've read up on the history and evolution of scanner technology (assuming you can follow the patent language - I couldn't!), you might want to reach out to Bill Benner directly (via the ScannerMax website) and see if he would be willing to discuss things with you. If Bill is unavailable, another good person to reach out to is Aron Bacs. (He's on LinkedIn.)
Finally, if you don't mind drinking from a fire hose (), you can always pump Steve Roberts (above) for more information. Steve has been doing the sort of lab testing you've been asking about for a *long* time, mainly because he's been working for university researchers who want to squeeze every ounce of performance out of the equipment. He basically lives and breathes this sort of stuff... He can also recommend test equipment (and parameters) to you, should you ever want to do some of this advanced testing for yourself.
Best of luck!
Adam
PS: I'll leave you with this interesting factoid about the performance of the Saturn 1 scanning system from a demonstration by Bill Benner at SELEM several years ago: The maximum rotor torque is equal to 40,000 Dyne-Centimeters per volt applied across the coil. So at 48 volts (which is the maximum for the +/-24V differential PSU), that's 1.92 x 10 ^6 Dn-cm, or about .14 Ft-Lbs. Which is just crazy for something that small.
Then realize a galvo test bench
starts with a audio signal generator with at least 10 bit waveform resolution and dc to at least 22. KHz sine, square , and triangle waveforms. An analog scope with a crt is preferred to avoid digitizing resolution issues in most cases.
Toss in a true RMS voltmeter. A good one.
A true test bench is going to have some wickedly complex gear to verify position sensor performance optically.
Most laser show users are best advised to leave position sensor linearity and position sensor gain controls to the factory. With the exception of a few US manufacturers, getting position sensor calibration constants is like trying to obtain the nuclear launch code of the day.. That adjustment
sets the slew rate of the electrical signals in the galvo PID loop. So one has to ask oneself. how do I determine if the factory technician, or which there may be twenty in the room in Asia, performed well that day.
If I'm going to do a really good evaluation of a given galvo pair, will I assume that setting is perfect? Tough question, especially if your reviewing the hardware for online consumption.
I can flood you with data back to 1985 or so. Including materials from company founders, academic publications, and guidelines from scanner amp engineers, and the only textbook on galvo design ever written. How far do you want to go d own the rabbit hole?
I'll post a few manuals with good tuning tutorials in a day or two. Take a look and let me know if you really, really, want to pursue this.
Steve
Last edited by mixedgas; 06-25-2023 at 20:08.
Qui habet Christos, habet Vitam!
I should have rented the space under my name for advertising.
When I still could have...
You've provided a lot of excellent information which I'll take some time to digest.
The challenge inherent in measuring and specifying scanner performance is clearly due to the inherently nonlinear physics of galvos (inertia, impulse, emf, impedence, inductance, dac resolution, etc.) I'm not trivialising this by oversimplifying. I do have a basic understanding of the construction and operation of such electromechanical devices.
However, it is clear that a single test pattern can, at best only function as a point sample in what we might think of as a high-dimensional space of nonlinear functions of time, orientation, angular velocity, rotational acceleration, etc. Reducing all this with linear extrapolation makes a terrible fit for the true performance profile. I don't expect this to be interesting to most busy laser show professionals who have better things to do than calculus since their intuition supplies all the answers they need.
My ultimate goal is an effective geometry to signal stream optimiser, enabled by more comprehensive scanner performance measurement and then compensating for it. I'm attempting to achieve this without specialist equipment or sophisticated diagnostic skills although it's still at a very early stage. I'd like to hope that I will have something of value to share, but I'm too inexperienced to know yet whether this is sublime insight or sublime foolishness.
Creaky Old Award Winning Bastard Technologist Infinitus Excellentia Ion Laser Dominatus
Join Date
May 2007
Location
A lab with some dripping water on the floor.
Posts
10,039
For the class 0 amplifier used in Laser Shows, it will be somewhat foolish to tack on outside feedback, as the inherent error is high but not not particularly reproducible. Generally you do this in a DSP based scanner amp. Trying to "tag on" a feedback correction to the existing amplifier is non trivial. Its been done for repetitive waveforms, and a correction table of pre-calculated waveforms derived, but the researchers stuck to sawtooth and triangle waveforms. After all, industry uses two Galvos that are speed matched or tuned for extreme accuracy. . Laser show uses two Galvos that are matched to a known artistic derived transfer function standard. One factory manual refers to four different available tuning styles besides ILDA.
I've dumped ILDA artwork into a class 1 driven Galvo pair professionally tuned with equal mirror speeds and optimum settling time. The results are not pretty. I suspect many Ebay surplus galvo amps have been discarded or returned when purchases by persons not aware of the differences in available tunings.
The attached files have data for moving iron in most cases. moving coil (modern) will be quite different.
Again, how you make this relatable to the average laserist is difficult. Anyone can Bode plot a galvo's response. It doesn't reveal much for artists.
Anyone can digitize the V and P testpoints on the amp. Other then to minimize ringing , it doesn't tell you much about the image aesthetics. It does let you fix the corners in images when tuning. It may let you see when oscillations or overshoot or bow are breaking out. Besides, Galvo induced visual distortion is part of the inherent art form of laser scanning. The resonances, ringing , inertia, lag, all contribute to the "flowing" images with smooth illumination. Comparing the difference between DAC out and POS out from the scanner amp may help in rough tuning, easy enough to do on a standard two channel scope with sum / diff math. But for the final tune, you will be implementing corrections to match the artistic pattern. I generally tune for sensor calibration, speed / low ringing / high accuracy / balance using a square wave or quadrature square for XY, ( Pseudo Dirac Spike!) and then detune to match the ILDA pattern. The detuning can be substantial. But whenever possible I will use an Oscilloscope for diagnostics.
One of your issues will be obtaining enough actual galvo pairs to build a data base. Optimization or performance ratings derived from one sole pair, are not a victory.
The Gold Standard for implementing the test frames as designed was the Pangolin QM32 board in point mode, not vector. More modern cards have software that may, may not make their own pre-optimizations to the test pattern.
At every display industry conference I've been to, two things happen. 1. "Hey can you retune these for me?" 2. "Arguments about how to read subtle differences in the test pattern". Generally by showtime, retuning and matching of the projectors shared for the conference has happened. Usually at least three people go away understanding how to tune for the first time. (Yes Buffo, I mean YOU!, your always helping with tuning during setup, and so is WRB.)
We have not even got to tangent distortion in the image yet...
One of the things I've been known to watch is the average power consumption on the amplifier power rails. It can be rather telling as tuning improves for gross tuning only. (Note, for many amplifiers the AGC circuit, voltage regulators, reference voltage source, and position sensor add 20-80 mA to one rail)
Sorry for the "firehose" rant of concepts.
Check your PMs. Openly stating due to my unpaid beta test relationship with a current project, I'm not posting "Everything" openly.
Steve
Last edited by mixedgas; 06-28-2023 at 18:27.
Qui habet Christos, habet Vitam!
I should have rented the space under my name for advertising.
When I still could have...
Ideal laser specification testing
Reply With Quote
Originally Posted by christo
buffo
), you can always pump Steve Roberts (above) for more information. Steve has been doing the sort of lab testing you've been asking about for a *long* time, mainly because he's been working for university researchers who want to squeeze every ounce of performance out of the equipment. He basically lives and breathes this sort of stuff... He can also recommend test equipment (and parameters) to you, should you ever want to do some of this advanced testing for yourself. 
