I believe I've made some horrendous mistakes in my mA vs. mW understanding/explanations on previous logs!
This probably explains why many of my burnt-diodes were burnt!
Prompted by @johnowhitaker's explanation at: https://hackaday.io/project/560-mini-laser-cutter/log/28767-getting-there/discussion-42620
Here's the bit that I think might've been horrendously-misunderstood/mis-explained throughout previous log-entries, so please read the corresponding tidbits in those entries with a hefty amount of skepticism!
P = V * I. Duh.
I've burnt-out a LOT of diodes in this project, and having only a couple remaining drives with lasers allegedly-capable of burning materials, I need to figure this out...
So e.g., as I recall, I think I configured my original current-source to drive my 16x diode with 200mA, per something I read online... That diode, on that circuit, has yet to burn out, but *several* later diodes on my new circuit *have*. Including 16x (and a 20x?) diodes. The new circuit is nearly identical to the old circuit, except for one plausibly-glaringly-related thing: I actually calibrated its current-limiting resistance (calculated: 5ohms) until the output actually reached 250mA (for a 20x diode) when the potentiometer was at its minimum.
According to a chart in these logs, copied from elsewhere, according to one particular diode's specs: "x16 speed recording – 250mW".
And, I haven't been measuring the voltage! Let's just say that the forward-voltage of the diode is somewhere around 2V...
P = V I = 2 * 200mA = 400mW.
I think, I was making the mistake that 200mA would equate to something
*less than* 200mW, not *MORE*. I guess I had something like P=I/V in my head for this case, or maybe I was taking the mW-rating to be *optical* instead of *electrical*? Something like that. Regardless, somehow I figured 200mA would be a safe value for a 200mW diode, 250mA safe for a 250mW diode, etc. CLEARLY THAT'S WRONG. P=VI. Duh.
I threw in the calculated resistor-values (and a potentiometer, in series) and let it go... and had no trouble.
As I recall, I started with two ten-ohm resistors in parallel... 5ohms, right? Then I measured the output-current which was something quite a bit less than 250mA. They were 5%ers, so I found a resistor to put in parallel that should've been about right to fix the tolerance-issue... That *still* wasn't getting 250mA, so I found *yet another* resistor to put in parallel...
Finally I got 250mA, which I figured should be well-within-spec for a 250mW diode. (But of course P=VI. Duh.)
Now, here's another thing...
The circuit contains a potentiometer in series with the (now "calibrated") current-limit-resistors. I think I used a 100ohm potentiometer... Well, it's quite likely that those 100ohm potentiometers have a bit of resistance in them, even when adjusted all the way to the "0" end...
Then there's the equation for the current-output of the LM317 current-source circuit:
I ~= 1.25V / R
Right, so R = 5ohms = 250mA...
But R = 6.25ohms = 200mA
So all it would take is 1.25ohms of error to *completely* change the output-current... 1.25ohms on a potentiometer-wiper ain't much, right...?
So who knows.
It's also plausible that the calibration I did, by placing resistors in parallel with the (5ohm, calculated) limiting-resistor actually resulted in a much *lower* resistance. Maybe my potentiometer wiper wasn't adjusted *all the way* to the left when I did that calibration. Or maybe the wiper had some dirt on it, so now with my "calibrated" limiting-resistance values the output can spike during adjustment... I dunno.
Lots of speculation. But the key is it wouldn't take much in terms of tolerance-error in this case to bring a calculated-but-unsafe current to a physically-safe value...
And thus, I've a 16x diode that's been running on a half-assed and mistakenly-calculated circuit for quite some time, and a graveyard of diodes from a carefully-calibrated circuit whose output was being watched almost constantly.
I guess tolerance-error worked in my favor on that first one.