Scanner Limits

[planters]

As an alternative to computer generated graphics, I was thinking of taking one of my low cost 30K CN clones which has worked well and with out complaint in a projector for about a year and driving it with a sine wave generator. I suspect it is essentially equivalent to a DT scanner in quality. The purpose of this is as one component to a lumia projection system. The question is how to know the reasonable frequency/throw limits or if there is a way to measure or see the warning signs when approaching these limits short of throwing a mirror. Fidelity is not too important as a feed to a thermal,water or crystalline lumia unless the fidelity can indicate the safe operating limit.

[mixedgas]

With a properly cooled galvo, and a sine wave, you'll quickly see the 3 db point by watching the feedback signal and integrator signals on the scope while increasing wave frequency. You'll also see it visually.

Other waveforms are a bit more difficult to discern.

Manually Integrate the area under the curve on the scope, and use that with Ohms law to calculate the coil dissipation. Or Google the math and just do it.

The last warning sign if you go too far on Cambs or Camb clones is when the stator heats up and the rotor starts dragging against the coil. The rotor is directly adjacent to the spiral wound coil on those. Right before it starts dragging, you have to slam more and more drive into it as the magnetics heat up.

If you force it past the obvious warning signs, If your amp does not have the coil temperature calculator circuit, you'll note the feedback going nuts and the mirror vibrating in strange ways.
Usually at this point, the fuse on the galvo body (if it has one, Older Cambs do,) burns out.


Soon after that something breaks permanently.

I don't know what a Eyemagic or other modern galvo will do.

Its not this complex, an astute observer will see when your getting limited or hitting limits.

A few months ago I pushed a dying scanner to find this out.

Steve

[dsli_jon]

As an alternative to computer generated graphics, I was thinking...driving it with a sine wave generator....as one component to a lumia projection system.

Heh...Hey Brian - should we introduce Sir Planters to the aqueous wonder-world of the Raster-Bar?



...of course, you can-also overlay with scanned-anim's / abstracts, ie: https://www.youtube.com/watch?v=TFzH7_BNeNY , but I kinda like seeing the raster-fx 'shine' on their own..

..To your-point, ya just gotta watch-out for them breaking into osc's, etc.. Perhaps put the input sigs on a pot, so you can 'open the valve', nice and slow-like..

Oh, plus everything Steve said.. re: EMS, the 4k's will just start making 'bird's nests' / freaking, till input is limited; the 7k's will, indeed, 'collapse' into a straight beam, like the CT 6200-family, when breaching limit.. I guess this is in-lieu of the old 6800-style 'fuse-approach'... Yes, fuses are a 'pita', but.. less 'dangerous' than scans collapsing into a straight beam, if you've not got a scanfail-sys, inline..

.02
j

[laserist]

Heh...Hey Brian - should we introduce Sir Planters to the aqueous wonder-world of the Raster-Bar? snip

You know me, I'd publish it all down to the resistor values. Evolution in action...

[polishedball]

Jon, your gonna cause me sleepless nights wanting all the effects. Toys etc

[buffo]

The question is how to know the reasonable frequency/throw limits

The details of the ILDA test pattern have your answer. Specifically, the center circle. Remember that it's actually made up of 12 points * in a dodecahedron that all lie completely outside the center square, but because the scanners are already ballistic, the corners are rounded off to make the circle. (This is the 3dB distortion point Steve was referring to.)

Thus, if you take your point speed (30K) and divide by the number of points to make 1 trip around the circle (12), you get a small signal bandwidth of 2500Hz.

Note that this is an absolute maximum, and you will still have 3dB of distortion at this frequency even if your scanners are perfectly tuned. But it's a good baseline.

Also, remember that this is for a small-step signal. So however wide the circle would be at a given distance from the projector will define the maximum width of the scanned pattern at 2500Hz. If you need to go wider than that, you will need to reduce the frequency even more.

Manually Integrate the area under the curve on the scope, and use that with Ohms law to calculate the coil dissipation.

Why bother? If you already have the galvos tuned, it's far easier to use the test pattern to define the maximum performance of the scanners.

If you force it past the obvious warning signs, If your amp does not have the coil temperature calculator circuit, you'll note the feedback going nuts and the mirror vibrating in strange ways.
Usually at this point, the fuse on the galvo body (if it has one, Older Cambs do,) burns out.

Soon after that something breaks permanently.

Yup. Be on the lookout for wobble in the pattern, or audible squealing from the galvos. Either one can be a sign that you are over-driving them. It also helps to monitor the heat of the scanners. (Use a laser thermometer.) If they get much over 140 F, you're probably pushing them too hard.

Adam

* Technically there are 36 points, since the beam makes 3 complete trips around the circle, but you get the idea...

[mixedgas]

Buffo is essentially correct except for one tiny detail. in one operations region of the amp tuning, there is more bandwidth then the ILDA pattern shows. About 15-20% more, but only for pure sine waves with tuning set for reduced feedback. I had a few papers on the subject, long ago. I no longer have them. If you look at the amount of the information complexity of a abstract composed entirely of sine and cosine waves, you'll see a clue. This is how the classic planetarium show imagery looked so good, often with open loop galvos. State of the art (past 5 years) mechanical bar code scanners used this regime.

I posted that, and then two er I mean three afterthoughts came up:

I do have a old GSI graph some place that shows this, however I'm on the road as usual and thus cannot scan it in, wait one week.

For discrete steps in the ballistic regime, the ILDA pattern holds.

Second afterthought: From the physics point of view, This thread effectively silences the frequent critics of the ILDA tuning scheme. It also starts to hint at why none of those critics has ever came up with an alternative tuning scheme.

Third Afterthought:


The pure sine wave closely approximates the waveform needed for maximum acceleration performance of the galvo while exciting fewer resonances in the galvo shaft.

Steve

[planters]

This has been helpful. I am going to use a two channel waveform generator that will allow me to slowly up the voltage while watching/listening to the galvos.

Jon,
Those are neat videos. What is the origin of those horizontal bright and dark bands?

[swamidog]

it's quite an interesting effect.

maybe raster scanning through a stationary lenticular optic with some color modulation and blanking effects.

This has been helpful. I am going to use a two channel waveform generator that will allow me to slowly up the voltage while watching/listening to the galvos.

Jon,
Those are neat videos. What is the origin of those horizontal bright and dark bands?

[mixedgas]

Quoting Buffo re area under the curve.

Why bother? If you already have the galvo tuned, it's far easier to use the test pattern to define the maximum performance of the scanners.

End Quote.

He can check the dissipation against the corresponding data sheet and see if he can push further with the sine wave. I'd take whatever commercial model that the clone is emulating, check its data sheet, derate 10 or 20 % and drive accordingly.

Steve