disgusted, back from Selem

[LASER-AD]

Hey steve, any further Info on the "interfacing" Board yet?

[dave]

This is a chunk of a email on this subject from Bill..

If I were you, I would not go so far as to make the "Ready for anything" filter.

Really, that page has been there for a long time. I never finished it to my satisfaction, so I never linked it into the rest of the site. Having tested that general circuit on many lasers now, I can tell you that the most that you need is the simple dual-slope filter curcuit (standard filter with an extra diode to quicken the leading or falling edge). We use slew rate limiters in our industrial servos, so I figure it was worth a try with lasers as well. It works, but it is not really necessary. What would be better is a non-linear circuit on the output of this filter, before the laser. I will try to come up with a quickie demo of this and put it on that page.

Bill

Like you David, I still haven't gotten round to building these up for our rgb's,. but should do soon..

Race?

[dsli_jon]

...SUGGESTIONS WELCOME

Hi Steve-

Just finished perusing this post, don't have a lot of time to think this thru right now (on a deadline...) but we have used this excellent board http://www.lasershs.com/TomCat_diffe...n_LD_board.htm in the past for similar issues, and at first glance, it seems (??) like it might be of help to you guys for your meets... just a friendly fyi... ciao..

- j

[yaddatrance]

We used to run with tomcats during the latter half of the QM32 era,
From what I recall of the circuit, it uses a simple (and effective) inverting
and unity gain opamp circuit to generate the ILDA differential outputs from
the QM32 source input, but I don't recall having the option to adjust
either the slew or linearity of the color lines...

On our own systems we use a quad 8-bit ADC and 12-bit DAC with a fast
uC doing a "per color" slew rate and color level correction. Since scansystems
also have slew, we usually tune the color so that each color lands cleanly on
a standard "point lookahead"... The reason to use a 12-bit DAC for color
output is that on most analog diode lasers, the range between the lasing
threshold to full brightness is only 1 or 2 volts, so you would end up with very
coarse color control using a just an 8-bit DAC.

The really cool thing about using digital signal processing techniques is that
you do not have to slow everything down to the slowest laser to get
everything matched up... The devil is in the details, but you can "average"
the rapid color changes using a modified kalhman filter for the slower laser
modules and get a perceivable color "accuracy" boost (in comparison to
just feeding a slow laser fast data and letting it deal with it itself)... This is
really important for us because most of our shows and effects are realtime
generated using procedural elements so we do not often have the luxury of
hand-tuning the colors for each frame...

[The_Doctor]

The reason to use a 12-bit DAC for color
output is that on most analog diode lasers, the range between the lasing
threshold to full brightness is only 1 or 2 volts, so you would end up with very coarse color control using a just an 8-bit DAC.

Could use a scale and offset adjust based on a single analog gain stage. That way you can map your digital range to anything in reach. That would be standard practise in audio. (Gain, anyway, offset is usually only done to set zero by DC blocking caps). Still worth having the 12 bits though, to make things easier to control remotely. Last time I tried the LambdaPro laser analog mod it wasn't even linear, so the top end wouldn't see any change over several values, but the bottom few would leap in sharp jumps. If you used a crude analog curve compensation, as well as that offset/scale adjust, I bet 8 bits would be plenty.

[mixedgas]

while I would love a DSP solution, right now the analog devices SSM2241 is doing a great job and at 4$ a chip, its not bad, bandwidth is 100+ kilohertz.

Steve

[LASER-AD]

when can i buy one completed from you??

[The_Doctor]

while I would love a DSP solution, right now the analog devices SSM2241 is doing a great job and at 4$ a chip, its not bad, bandwidth is 100+ kilohertz.

Steve

Exactly. DSP can't sensibly be the answer. If you need to output via a DSP based correction for scale, offset, and perhaps curve as well, it makes sense to get it right at the outset, and that means greater demands on the main system outputs. If the maker of those isn't going to do that, then the weakest link in the DSP chain is right there. Throwing more DSP at it can't ever improve the signal quality, not even at huge expense. Why even try, when a resistor network and a decent op-amp can do it easily, cheaply, and with no significant loss?

This isn't the first time recently I've seen a perfectly good simple op-amp gain stage design utterly ignored in favour of the new, the shiny, the digital. Time to stick to common sense, I think. And if anyone thinks I'm wrong, that digital coding on PIC's or anything else is somehow easier, better, or more practical for solving basic voltage scale correction issues, they'd do a lot worse than take the issue up with Joerg, a consultant who posts a lot in sci.electronic.design and other newsgroups. Come to that, what about Pangolin? I remember there was some circuit that used op-amps for correction there. If Bill Benner thought DSP would be better, he'd do it that way. The only thing DSP could really score on is a lookup table to adjust for seriously dirty nonlinearity, but if the laser's mod input is that bad it's hardly worth the effort of trying to adapt to it.

[Lampy]

Lets see some schematics then chaps! I posted something a while back but I don't think it works properly. Perhaps I should try and remove it incase anyone takes the time to build it.

[The_Doctor]

Here's one, as simple as I could make it, but it cures the two worst problems.



Another thing it doesn't do is cope with differential inputs, this one's for single ended, but it will reach the ground rail cleanly. Supply with more than 5V if you need full 5V output. Or, stay with 5V supply and omit the zener, the clip will limit the output.

Adjustment is easy: set controls to minimum, adjust offset up from zero till you reach threshold emission, then raise scale till full input signal gets you the full laser output.