Diode Driver And Cooler Driver

[mixedgas]

Professional diode driver and TE cooler circuits from dead YAG project,


In ZIP

[The_Doctor]

Nice. Instructive on how to safely control a DPSS. Might be nice to have the TEC cooler drive based on an oscillator though, if the speed is high you can filter the output for the TEC drive but avoid a large heatsink and wasted power.

I noticed no optoisolator there to stop mod input nasties from reaching the diode. I don't know if anyone does that. I always thought it would be a cool idea but if it was that cool I'd see it done, and I don't. I still think it's cool, at least it would be if it can be either very linear, or be controlled in an op-amp's feedback loop.

[lucerne]

@mixedgas
Thanks for that. Always nice to see differents diode drivers and projects
But I've seen, its an old project. Old 308-Op-Amps and -5V blanking-input (and maybe not very fast?)
Doesnt matter, I've got my own ones

@The_Doctor
What you mean exactly? Interlock? Or something else... Or do you mean an optocoupler-input do shut down the driver?
My ones have a optocoupler-output, so you could put the driver into an interlock. (Also overheat-protection of the driver+laserdiodes, etc)

At the moment, I'm testing new drivers, which drive the diode highside, so I dont need to isolate it. Just put the cathode (or mounting-case) on ground

[The_Doctor]

Just a way to stop ugly signal errors or even things like mains plugged into the mod input (don't laugh, at least one person on this forum has done it).

I can't get the idea out of my mind, that an optical link is going to make it much harder to destroy a diode, and that no amount of other limiting tricks are going to give better protection. The tricky bit is getting a good fast accurate proportional modulation though it because any feedback would also have to be isolated too, so it would have to be very linear to start with.

[drlava]

wow it's all analog. I wrote a PID control loop in a microcontroller for TEC control a while back to take care of that, but it's nice to see how things were done in the 'old days'

[mixedgas]

there are newer magnetic and capacitive isolator chips that have 10-20 mHz bandwidth and the same or better isolation then a opto. For a opto, you have to put one opto in the feedback path to compensate for the nonlinarities of the other one. I have a PDF file on how to do that up to ~1 mhz, but I doubt I could legally link it here as the university I work for paid for access to that journal.

Phil Hobb's book on designing optics has a wonderful chapter on isolation and diode drivers, including one section that shows what happens when the wiper in a potentiometer in the classic LM317 current source goes off the resistive surface for a few milliseconds and how to add another resistor to prevent that particular kind of "poof" from occuring. Phil is a research scientist who hangs out on Alt Lasers, hes worked with galvos, but for lidar, not light shows.

Steve

[mixedgas]

[QUOTE=lucerne;33010]@mixedgas
Thanks for that. Always nice to see differents diode drivers and projects
But I've seen, its an old project. Old 308-Op-Amps and -5V blanking-input (and maybe not very fast?)
Doesnt matter, I've got my own ones

adding a unity gain inverter is not that hard :-)

That was posted for somebody who likes old analog, also because what parts he has acess to are old analog. I couldnt link it in a PM so I posted it here.It is old school, but he can get started from it. I have better stuff, but I can't convert to pdf and can't find my scanner cable, and HP wants more for the ^%$# cable then a new scanner and I'm broke right now.

That protection relay next to the switch that selects light/current mode needs to go across the diode
as well. But thats for another day.

Steve

[The_Doctor]

Phil Hobb's book on designing optics has a wonderful chapter on isolation and diode drivers, including one section that shows what happens when the wiper in a potentiometer in the classic LM317 current source goes off the resistive surface for a few milliseconds and how to add another resistor to prevent that particular kind of "poof" from occuring.

In the voltage reg config that can happen, but in the current reg config, it's impossible for a pot to jump into a lower resistance by wiper failure, so the current can only fall, as the resistance rises.

What can happen is you can overshoot a set point with a pot (especially the single turn types) and drive excessive current that way, so excessive that you can burn the pot and the load, leaving the regulator managing to survive on its short-circuit protection if you're lucky. That's what the other resistor is for, in series with the pot to set a safe upper limit.