"signal conditioning" for Analog Dimming Power Supplies

[corry]

I'll do another post with an intro, and some details, but I wanted to keep things short...

The first supply I found for dimming with a continuous analog signal rather than PWM dimming was at https://www.merghart.com/p/29/laser-...nd-ttl-control. I wanted to understand the entire circuit, so I simulated the input "conditioning". The author states it is to bring the signal to a 0-1.25V range. My ltspice simulation showed that this compressed the about 80% of the signal range into about 2% of the output signal range. I found this quite curious, and decided it wasn't what I wanted, and a simple voltage divider would do. I implemented a similar, but different circuit, and I think my results look fantastic. However, there is this lingering "what if"....what if the published circuit was better?? So, being that I am weak when it comes to analog, I figured I would ask some experts. Maybe the author is even a member

[Renfrosnart]

I'll do another post with an intro, and some details, but I wanted to keep things short...

The first supply I found for dimming with a continuous analog signal rather than PWM dimming was at https://www.merghart.com/p/29/laser-...nd-ttl-control. I wanted to understand the entire circuit, so I simulated the input "conditioning". The author states it is to bring the signal to a 0-1.25V range. My ltspice simulation showed that this compressed the about 80% of the signal range into about 2% of the output signal range. I found this quite curious, and decided it wasn't what I wanted, and a simple voltage divider would do. I implemented a similar, but different circuit, and I think my results look fantastic. However, there is this lingering "what if"....what if the published circuit was better?? So, being that I am weak when it comes to analog, I figured I would ask some experts. Maybe the author is even a member

I know this is months old, but there are questions about your posted results.
When you say "compressed the about 80% of the signal range into about 2% of the output signal range" are you referring to the current measured through R4 as the ourput? The circuit is very basic, a simple current regulator with an analog input that offsets the current one way or another. There is no provision for stable LD bias since the op amp + input is at 0V when there is no input signal, thereofre nothing to compare the LD current with. The diodes just clip the input signal. TTL should be fine. For undistorted analog signals, it needs more work. DC coupling is present from the input to the op amp, so of that input signal has some DC on it, things will be different. It's explained in the article about these concepts.

[corry]

I probably should have updated. I was actually talking about current through the laser itself. I think I still have the ltspice schematics, and I’ll see if I can post pictures…in the future. It’s entirely possible I connected something wrong in ltspice. But that circuit isn’t exactly complicated. Still entirely possible

In reality, what you’re looking for is a voltage controlled current source. I found many more schematics with that search term, and implemented one…not very well…on perfboard . It worked, and I was able to draw cool stuff with the laser.

For the hardware, what I’d like to do is combine the SMPS I used to get the voltage just above the max voltage the laser would be able to handle, so the mosfet wouldn’t have to dissipate much power, the opamp, the mosfet, and the pots for all three channels into a single circuit board and let it all out open source.

The problem that my perfboard implementation has is a lack of decoupling caps and filter caps. So the laser signal is a bit noisy, but only noticeable at low power. I knew I was cheating by not including them and the plan was never to run with perf board, so I figured the noise problem would be fixed when I sent plans off to oshpark. Last opportunity for that I ended up needing to spend more time than I imagined on my lawnmower pcb (I’ll describe that project below)

Other problems are things like safety. I’m ignoring the ilda interlock. That needs to be fixed. I’d also like to add a photo sensor near the output to sense light scattered off the beam, similar to how particulate matter sensors work, to give a basic idea as to whether the beam is actually on or not. It’s not as good as having power sensors in the laser diode, and I’d like to make this board I’m working on support that as well, but time is limited, so are funds

In reality the laser is my lowest priority project. I have a much more interesting research project that takes most of my spare time and money. Won’t say much on that beyond that it’s mechanical in nature so I spend a fair bit of time with cnc machines and a fair bit of money buying carbide! (A steerable tool holder for precision reaming being the latest $1800 purchase). Second side project is an autonomous lawn mower. Yes, there are commercial ones, but not for 1 acre+….and most aren’t great at the 1/2 acre level either, or they require some pretty big modifications to an existing mower. Mine is bolt on (and epoxy on some magnets for limit switches), uses RTKDGPS for navigation including a dual antenna system for heading, and should have enough power in the control motor to actually perform the “zero turn” that a human driving it could do! It’s power hungry though, so I’m not solving that aspect. Eventually, I’d like to make it all electric, but I think that will require ditching most of the original frame to hold enough batteries to mow the entire lawn on a single charge.

So, sadly, the laser isn’t top priority. It’s sitting in a conspicuous location so I don’t forget about it though! . Again, I’ll see if I can find pictures of what I had. I may have overwrote files in haste of design not valuing that which didn’t work!

[corry]

So I didn't still have the ltspice file from before, not exactly shocking.

Months later with a fresh brain, I'm having another look. I think what happened was I had disconnected the voltage divider. I managed to recreate the waveform...kind of, by adjusting the voltage divider to allow the circuit to be clamped by the diodes. Its ugly...

The problem I am having with recreating things is it seems like every time I want it to use the clamping, the sim hangs up. Probably because its simulating the death of the diodes? Not sure. Anyhow, here's the garbage circuit with a typo in the name...but I couldn't be bothered to fix it

[corry]

For anyone wondering, here's the circuit I ended up using that worked pretty well! (Other than the noise) (needs caps!)

[corry]

And here is the same circuit with diodes added...

So, am I to understand what those are supposed to do is clamp the incoming voltage? What is it clamped to?
When I input a 6V signal instead of a 5V signal, LTSpice gets stuck....sim runs super slow. I don't mind proper protection, but dislike added components for nothing Even without the diodes, 6V signals kill the sim, so LTSpice doesn't seem to be seeing any clamping action.