Beam shutter circuit

[myself]

Good morning all! First time here. If there's a schematic repository somewhere, kindly nudge me in that direction, eh?

I just picked up a cheap Chinese galvo kit as a toy, and because there's no such thing as too much safety, a rotary solenoid shutter to go with it. I'm planning to do all my setup with a 5mw pointer, then drop in the RGB diode module later. The control board in the kit doesn't seem to include a shutter control (which is fine; I want this to be separate), so I'm planning to make a circuit that basically looks for the AC component of the X/Y drive signals, and holds the shutter open as long as there's motion. If the signals stop for any reason (static beam), one of the signals looks like DC, drop the shutter in front of the beam.

That seems straightforward enough, and an ideal "first op-amp project" for a digital nerd who rarely plays in the analog domain. Still, it'd be nice to use a proven schematic if there's one out there.

Then it hit me: I could conceivably drive this off the position sensor feedback coming off each galvo, which would also let me detect a seized galvo, mechanical damage, etc. Trouble is I bet those signals are really sensitive to some goober tapping their wiring... is this a terrible idea? Has it been done? Again, looking to avoid reinventing the wheel if possible.

Third-and-a-halfth question: I was looking at the ILDA DB25 pinout and noticed that the interlock is specified as just a short on the source, to be sensed somehow by the laser. How is this usually implemented? Seems to me that the most robust way would be to make a small oscillator and capacitively couple an AC signal into one of the interlock pins, then sense it coming back on the other. If it goes away for any reason, or if one of the pins is shorted to ground, or shorted to power, or floating, then there's no return and the laser doesn't enable. Is that how it's done?

[laserist]

Laserium had a little beam kill circuit way back when - I'm on the road at the moment, but I'll try to remember to see if I have a copy at home. It started out as a circuit that injected noise when the signal dropped off - god was that a bita! (But we did shows live and something that got in our way was not a good thing.) You don't want a single missing signal to kill your laser - going to a line is a valid effect.

[mixedgas]

You probably want to change your technique to one of prevention, rather then one of a scan fail safeguard. CDRH and Worldwide rules favor some form of physical beam masking rather then scan fail safeguards, except when deliberately audience scanning where permitted by law. BTW, the beam staying moving if there is a system failure improves safety to a point, however the math for that is complex and outside the scope of this post.
~
Scan Fail circuits have been discussed time and time again, and they rarely solve the problem. Your show equipment should be designed and operated in such a way that you don't need one to begin with, especially under US rules. In other words, beams are three meters up from the floor and two meters horizontal from any audience accessible point, and that you have some form of masking or other hardware to prevent that kind of direct access... You do need a safety shutter, but you should design your show and equipment so you NEVER are solely dependent on it. Build a big red switch and call it a day, then design your projector to implement masking using sheet metal or other means.
~
You also need to ensure whatever your projecting on can take a static beam at full power anyways.
~
No light emission is a valid artistic effect, and animations often have blanked portions...
~
30K scan control signals update position about every .000033 seconds, so mechanical shutters can't move that fast.
~
Modern galvos have a SCAN ANGLE DEPENDENT bandwidth of about 2500-3500 Hz depending on model etc... The motion speed and instantaneous bandwidth available is nonlinear and highly dependent on the commanded leap angle between points, as well. Its not an easy equation from a microcontroller to solve, and the equation is known for some galvos.
Going a long jump is far slower then moving three resolution "points" away.
~
Most older systems worked on velocity, as the square root of the sum of the squares of the galvo feedback signals being greater then a certain value. This has the disadvantage of often blanking the system when a scanned effect shrinks to zero. This happens VERY often during a show, making a simple scan circuit fail a non-starter.
As the vector sum of the squares can be done with some simple op-amp log amps, it used to be a common method, ignoring doing the square root step and running the output thru an RC filter.
~
Downside, with modern control software, the galvos are usually still scanning while blanked, so your shutter still pops open. This is done so images do not "POP ON" and so during certain effects, the image maintains equal brightness, ie, the image does not get progressively dimmer if your "writing out" George Washington's signature during a Forth of July show etc...
~
Scan fail boards are available commercially from Europe, and it is probably far cheaper to buy one then re-invent the wheel, you have enough problems with making a home built projector for the first time.
~
Most commercial software has a TTL output that opens at the start of each song and closes at the end of each song, as well as an "Escape" key to blank the show in an emergency. For most regulatory bodies, unless your doing unmanned operation, a keyswitch, an emergency shutter, and optional logic to use the control system "beam on" TTL line to drive your shutter is enough. In the US you may need pairs of series interlock switches for the projector lid, and these may be defeatable in certain circumstances for maintenance.
~
US rules used to/may still require a scan fail safeguard to actuate within 200 nanoseconds of the failure condition forming, so its nearly impossible for a sole inventor/designer/programmer to meet that with modern scanners.. The rules provide mechanical masking as the alternative to a "compliant" scan fail. If your going to do it, you might as well look at something like Pangolin's PASS system and gain an artistic effect from it, however it is very difficult to integrate PASS for a beginner...
~
Most people just run current limited 5VDC from the projector thru the interlock pins back to the projector, the idea being to use a few tens of mA maximum, this way you have a "sealing current" (telephone term) to break thru crud and oxide on the connector pins, yet do not risk damaging your controller PCB if the cable is miswired or hits other contacts during connection.
The ILDA spec has a limit on voltage and current for this task.. Usually an optocoupler or certain small relays work just fine as the detector in the projector. Also if paralleling multiple projectors, the interlock problem can become interesting. But you do need that " DC sealing current" to break thru the oxide, I've seen people have issues with just using a pull down or cmos signal etc...
~
The position sense test tap points on nearly all scan amps are an un-isolated, un-buffered high impedance signal right off the position sensor summing opamp, normally designed for driving a 10X scope probe and nothing else. You need to keep any cable short and shielded and immediately buffer the signal. The velocity test point if brought out is the high end of a 0.1 ohm current sensing resistor from the low end of the galvo coil to output stage ground. Loading it in any way can result in your galvos oscillating or distorted images. If the galvo fails the signals can look just like a valid show scan.
~
One of the problems of using the scanner amp test points is during a jump, the high peak currents in the scanner coils (six amps some times) can lift the test point off small signal ground.
You really should read Pangolin's projector wiring guide before trying this...
~
I don't have time today, but I can link you a scan amp schematic late this evening.


Steve