Experimental Brightness Attenuator / Scan fail

[Badpip]

(...) and F’n idiotic without first rate hardware and safeguards...

That is the whole point in this circuit

/Thomas

[mixedgas]

I see

I still think a PNP/PFET would be better - should the power to the scanfail disappear, the transistor will turn on and attenuate the modulation output.

/Thomas

Touche.. Nice Idea, I did not think of that. But I would have a output enable relay across all power bus lines in the system anyways. If the scanner rails, the control system rails, and the ILDA connector is not IN, do not even power up the lasers.

The goal of the project, once upon a time, was a scan fail so cheap, that we could insist that it be included on Galvo amps, and not need a tricky log log squaring circuit to calculate vector velocity like some older 1970s analog scan fails. Absolute Velocity, ie SQR ((Xvelo^2) +( Yvelo^2)) is the transfer function used on scan fails that had been accepted at one time in the US.

There is another thread some place on this, and a LONG one at that.

Steve

[logsquared]

just came across this thread. Here is a super simple idea maybe worth some thought/ discussion. Sorry for the crappy schematic i used my eagle cad layout software to draw it. Anyway, It's basically a window comparator that checks the voltage of the galvo feedback against the input. If its Vdiode above or below the output goes low. Obvioulsy needs a gain adjustment on one of the inputs to balance the two levels. Could replace the diodes with voltage sources or pots to adjust the thresholds.

[DMills]

A scan set is (pretty close, at least in small signal terms) to a second order lowpass so that wont really work.

Regards, Dan.

[logsquared]

A scan set is (pretty close, at least in small signal terms) to a second order lowpass so that wont really work.

Regards, Dan.

Shouldn't the phase response be fairly flat below the scanner knee frequency? I did consider possibly needing a lowpass filter on "Vin" to compensate for the lag of the galvos. My original thinking was.. the "window" created by the diodes might be enough to compensate for the lag of the mechanical system. If the lag of the galvos is several degrees or more this approach certainly won't work. Its been awhile since I tuned a galvo on a scope... can't really remember what the phase relationship was like??

Had another "awkward" idea as well... Maybe integrate the driver error signal then compare that to set threshold. This way if the scanner is not where its supposed to be for longer than a set amount of time (calculated by integrating the error voltage) we can assume the galvo is not responding and blank the lasers, close shutter etc.

Just typing some random ideas, as I really haven't put too much real thought into it.

[DMills]

Differentiate the feedback to produce x' and y', square them with a pair of AD633s, gives you |velocity|^2, threshold this and if this value is below threshold and Y is below the horizon line then blank the output (Probably 3 or 4 chips for the core none larger then 14 pins).

It also needs some glue logic to ensure the power rails are valid and that both pots are returning valid voltages and such, but it is not rocket science.

It is just a pity that the AD633 is so expensive.

Regards, Dan.

[Qumefox]

Sorry to dredge up an old thread but i've been thinking of trying to prototype a scan-fail.

How expensive is 'expensive' here? AD633's seem to run $6-$10 depending on quantity. That seems fairly minimal compared to cost of the rest of a projector.

[taggalucci]

I could do with at least 4 scanfails if this project progresses.