Replacing Scanner Mirrors

[buffo]

Your picture is a good start of what a jig should look like, but remember that you also need to prevent the mirror from sagging.

Look at the right-side picture showing the end-on view; you would need something in the bottom of that hole to support the end of the mirror so it doesn't sag downward.

Of course, this is only if you're looking for perfection. As I said, I eye-balled mine, and they were fine.

Adam

[mixedgas]

About all you need, pictured in Paint drawings below. Mirror sets on sliding pin. Gravity or rubber band holds the Galvo. Easily made on a milling machine, estimate 1o-15 minutes to make, unless you get fancy and do the V=Block version. Unless you have 90K scanners, using a magnifying glass and ruler would easily let you center the mirror.

I'd love to see the mirror gluing kit Cambridge used to sell, anyone have one? Daystar Jon can you shoot a pic of yours?

Steve

[planters]

I appreciate all this advise and it is encouraging. I think I could fabricate the jig in 30 minutes not allowing for clean up, but maybe it won't be necessary. I like the idea of the slot because it would make access to the glue up very easy to do and to see. Something this makes me think about is the sound that comes from a scanner motor as it speeds up. Could this be due to vibrations in less than perfectly balanced scanners or is the actual scanner moving the mirror at a frequency equal to these sounds? I would estimate the frequency range from 500 to 4000 Hz. This is not at all the same as KPPS. Steve's point re the 30Hz square wave makes that clear.

[planters]

I also need to add I am looking at the new 506 series motors and here the mirror can't sag. The shaft design prevents this and so only side to side and square to the shaft laterally is required.

[mixedgas]

Eric,

One thing to do if your afraid and have the proper gear:

I start tuning with the amps on current limited bench power supplies, with the positive and negative rails set to ~200 mA limits at +/- 24 Volts. It gives me peace of mind when starting on unknown tunings or amps that have been adjusted by the incompetent. The positive rail will draw a current a bit higher then the negative rail, this current is the position sensor LED current plus about 10 mA. Position sensor LEDs, depending on the Galvo specs, are usually set between 10 and 60 mA.

Circuit Specialists sells nice 0-3 Amp 0-24V Chinese bench supplies with adjustable current limiting and LCD displays for about 99$ each. I buy them when they go on sale for about 59$ each. I would not use them "RAW" to test laser diodes, they have some surges as you set the current, but for testing amps and diode drivers they are the Cat's Meow.

Quiescent current on the scanner amps can be as much as 100 mA per supply rail on each axis, when not scanning. This will rise to around 250 mA for Cambridge products scanning at small angles and climb from that point. This depends on the model of Galvo. As you start to scan, the current does not rise much until your waveform gets complex or the angle goes wide. The PEAK currents can be 3-4-5 amps, which is why I suggested external capacitors be added to power supplies in projectors.


Steve

[planters]

OK Steve, let's say I do this. This sounds good. I take a pair of what will become poorly tuned amps and current limit them and begin the tuning process. When do I transition to higher and higher current? Will the limited current affect the tuning in some clear way at some point where I can then say set the current to 1.0A or higher? Or, can I just place them back onto the unlimited supplies that were selected to drive them in the projector?

[mixedgas]

OK Steve, let's say I do this. This sounds good. I take a pair of what will become poorly tuned amps and current limit them and begin the tuning process. When do I transition to higher and higher current? Will the limited current affect the tuning in some clear way at some point where I can then say set the current to 1.0A or higher? Or, can I just place them back onto the unlimited supplies that were selected to drive them in the projector?

Lovely thing about having average reading current displays and current setting knobs. You will see the small signal behavior and the current is pretty much average, if you watch the current, you will see as image complexity rises, the current goes up in proportion. If you move the scan rate up, the current climbs. If you increase the angle, the current climbs. If things go really ape or the scanners oscillate, current climbs through the roof because the high peak currents become a very large average current. So you might be humming along at .5 amp and suddenly see 1.5A or see one rail climb way high if something is badly wrong. If things are running smoothly, both rails show about the same amount of average current. It is a beautiful thing.

When things are bad the current will easily be two times the value when the situation is good.

If you have the limit set too low, the galvo simply refuses to increase to the demanded angle. Yes, you could, in theory, box yourself into a corner where a frequency pole arises because of the current limiting circuit, but I have yet to see that in a few days of experimenting with new tuning. In fact I tried to induce it with a scrap galvo, however I could not cause a oscillation with the limiting alone. One thing I never do is set the limit to below 100 mA. I set the current limit by shorting the PSU leads together, with the knob at the low end of the scale.

When your done, there is no adverse effect of going to the standard projector power supplies. But when your truly experimenting, current limiting is a good thing.


I also find it instructive to place a 10X scope probe on the .1 Ohm current sensing resistor over by the output connector on the DT style amps. That shows what is more or less DI/DT, you see lots of current draw across the resistor at the start, end, and damping of movement, so the waveform is big at the edges of a jump, but quickly declines to near zero when the galvo is at rest or coasting. The scope probe on the sense resistor does have a habit of inducing a small amount of noise into the image.

I use the limiting mainly for troubleshooting old gear. However if I ever again teach a new user to tune scanners, the current limit will be in use.

An example: If the cable is miswired on the position sensor, the galvo will slam to one side and the power amplifier runs the full 24 supply into the galvo, minus the C-E drop in the output transistor. On a 6200 I have with one bad internal photodiode out of four, the amp will run 3.5 Amps through the galvo coil on one rail, unless I have the limit set. The Same thing happens if you reverse the A and B sensor signals, the galvo slams to one side and rapidly heats up, to dangerous levels.

It is a useful technique, especially if you have no idea where to set the potentiometers on a "mystery" Ebay amp.

Steve

[kecked]

Just a thought. Color the shaft black before you add the mirror back. I wish they were all black to start. How many times while aligning the beam have you gotten "splashed" by the beam from the reflection.

[planters]

Very interesting that you mention this. I have been testing a large quantity of scanner mirrors and I have learned a few things. Chamfer is bad. Clear epoxy is bad. And, a reflective shaft is bad. Chips and dirt are less important than I thought. The first three cause a distortion and beam spread around the far field spot that is very noticeable. The last two only reduce the brightness a little with the lost light spreading so far as to be invisible.

[planters]

OK done.

This was not hard. I used a small quantity of methylene chloride (paint stripper gel) to soften the epoxy. This took approximately 2 hours to diffuse into the slot and the gel prevents any running into the bearings (its thixotropic don't you know). This did not have any effect on the coating and is much better than acetone for removing epoxy. I cleaned the shaft with acetone and abraded the surface with 1000 grit jewelers cloth then simply cleaned off the shaft as well as old optic prior to its storage.

I used a tooth pick and 5 minute epoxy to wet the shaft notch with a very small convex bead then inserted the new mirror (correctly-both times, it does happen) and then inserted the motor into the jig and rotated the motor body until the outer edges of the mirror contacted the walls of the slot and the shaft could be seen rotating relative to the body. I then very slightly backed off to prevent the mirror from sitting tipped in the slot, but preserving the balance and alignment.

The glue is curing now. One improvement I would make in the future is to support the motor body in a rectangular slot that is open from above which can be clamped by a simple external tooling clamp. This would be easier to machine to fit the motor with minimum clearance than a precision bore and would be easier to drop the motor/mirror/wet epoxy assembly into without contacting the sides. But nevertheless, I would personally not attempt this without a jig.

Hopefully, the epoxy will hold once cured and I face some re-tuning later today.