Thank you for the effort. I'll dig in to that link little later. The reason I'm going for PWM is actually efficiency but from a bit another point of view: Easiest variabe for my setup to control heat (and humidity) is adjusting light. There's some swings from day to night, but also day to day. I feel that shooting full 100% (or even 110% with pot) is just waste and counterproductive, if Tamb is beyond 30C (for summer) or humidity exceptionally low (winter months coming). I have a vision that I could achieve best possible overall conditions, even while cutting a bit off of light level, with minimal equipment and effort. At this point you surely undertstand why I don't do it with pot.
Actually I did it alreay with pots. Got few working, but then got a batch of DOA pots or my soldering skills sucked bad that night, but I ran out of pots while trying to do the last driver and didn't want to buy more while I could go for my next goal: automation. (Best pot did do dimming, BUT it had some leak and shorted for a ms once now and then. Didn't even notice it at first, but started wondering why just one of the fixtures flickered and others won't. After desoldering all joints, the short was still there => it's inside the pot, I assume.
This starts to sound like I'm really bad at electronics or I have very bad luck. Probably bit of both. Last night I tried Timothypauls' circuit with just mosfet between DIM+ and DIM- (connected also to common ground). Results were veery confusing. At the beginning everything else worked as expected, but PWM didn't have any effect on power output. As testing other options I found combination that gave me control ... in region of 5-10%. Measured DIM-wires' voltage and it was about 1.3V at that moment (regardless of swithcing PWM-bytevalue between 200...100). If I remove connections to Fet, I get full 10V between DIMs. So my Mosfet seems to leak even while off and PWMing or pulldown resistor on gate doesn't do much difference. Maybe I've just killed it with some static discharge, so have to try again later with more attention.
I don't know if my "kill-a-watt" isn't telling me the whole truth on low powers (higher watts are atleast close to truth), but I have had many 5% (of total, with fans) power consumption states while testing these dimmers. For example yesterday's test, I had 160w total consuming fixture with HLG-120 1050mAh and got watts like 7-15 (with PWM 100...200). I'm just hobbyist and don't understand why, but just saying it happened.
Note to myself: Do tests again with new parts.
edit. digged in to link. Short conclusion: my dimmer is bad. Doublecheck last note.