Wanting an Nmos

[Nominal Animal]

I like to control backlighting LEDs (at less than half a watt of power) using current control, instead of PWM'ing, because of an utterly unreasonable hate towards flickering.  As my enclosures are small (palm-sized to smaller than a shoe box), the problem with linear current control is the waste heat.  Thus, I've switched my focus to current-mode buck and boost controllers; specifically those that adjust the feedback voltage threshold using a filtered PWM input, like TI TPS92201A.  (That particular .one is a buck/step-down controller limited to 5.5V or less, so is not suitable for OP.)

I wonder if such a buck or boost controller (depending on what kind of input voltage OP has available) could be a better solution for OP too?

For example, TI TPS922051D2 you could control using a potentiometer in a voltage divider configuration.  These are cheap, too, this one less than 1€ apiece in lots of ten from Mouser or Digikey, although you'll likely spend more on a good inductor to use with it.  (I haven't used this one in particular, and only bring it up as an example to start your IC search/selection from.)

[Tenderfoot]

For future reference, please post the schematic in the first place. A lot of time has been wasted guessing.
[/quote]

Re-read my original post. I asked about the availability of a component with particular characteristics. People were free to say 'Such a thing doesn't exist, you are an idiot' or possibly 'Have you tried this one'. There was no need of a Schematic. Schematics seem to lead to more arguments and off-track comments. That is why I purposefully did not show one. Now regarding  using some kind microcontroller or whatever. I have zero interest in that. Electronics for me now is a hobby I dabble in it for enjoyment, no other reason. Taking a circuit from some application note, or buying a module from China and then getting chatgpt to write the code is not what I want to do. Digital is the future that's obvious, but some people like to tinker with old cars, I like 'old' electronics. So I will stick with analogue thank you. Maybe I'm weird, but there is no law against that.

[Tenderfoot]

That looks an interesting circuit, I'll give it a try as soon as I have time 

[Zero999]

For future reference, please post the schematic in the first place. A lot of time has been wasted guessing.

Re-read my original post. I asked about the availability of a component with particular characteristics. People were free to say 'Such a thing doesn't exist, you are an idiot' or possibly 'Have you tried this one'. There was no need of a Schematic. Schematics seem to lead to more arguments and off-track comments. That is why I purposefully did not show one.

I have read your original post several times.

On the contrary, the lack of schematic resulted in off-track comments. Once you posted the schematic, it became abundantly clear, the turn on characteristics, had nothing to do with the transisor type but how it was configured: follower vs  common source.

Now regarding  using some kind microcontroller or whatever. I have zero interest in that. Electronics for me now is a hobby I dabble in it for enjoyment, no other reason. Taking a circuit from some application note, or buying a module from China and then getting chatgpt to write the code is not what I want to do. Digital is the future that's obvious, but some people like to tinker with old cars, I like 'old' electronics. So I will stick with analogue thank you. Maybe I'm weird, but there is no law against that.

Fair enough. No one is telling you what to do, merely offering alternative solutions.

I also suggest a buck converter, similar to Nominal Animal's suggestion. It would improve efficiency because power woudn't be wasted in current limiting resistors, but if it doesn't interest you, then no problem.

[temperance]

For light dimming, you need a logarithmic dimming curve. You can create this dimming curve with a PN junction "for free". The same can be achieved with MOSFET operated below the threshold voltage.

[Zero999]

For light dimming, you need a logarithmic dimming curve. You can create this dimming curve with a PN junction "for free". The same can be achieved with MOSFET operated below the threshold voltage.

The OP's circuit already acomplishes this with an RC circuit.

[Nominal Animal]

The shape of the LED current ramp isn't that important to human perception, it is the maximum slew rate — fastest rate of change — that really matters.

Seeing that the OP has around six chains of LEDs they want to turn on and off "slowly", I think I'd use a TI TPS922050D2DGN per chain, so that each chain has a 24V input, GND, and a 5V active-high enable signal.  One could use e.g. shielded phone cable, two twisted pairs within a common shield, for longer runs.

The active-high enable signal would be connected to a 2.5V regulator enable pin, say Analog ADP121, whose output would be connected to a low pass filter (say, three 33kΩ resistors in series, each followed by a 1µF capacitor to ground, for about one-second turn-on/turn-off time), and to the analog DIM pin of the TPS922050D2.

The LED control circuitry costs about 2€ per LED chain — TPS922050D2 is 0.435€ apiece in lots of ten at Digikey, 0.836€ apiece in lots of ten at Mouser; the inductor depends on the current each LED chain needs, with the calculation procedure described in the datasheet linked to above.  Because you can adjust the turn-on/off time per chain, you can do stuff like have them turn on/off almost sequentially, for an even more pleasing effect.  With a bit of care in the design of the control PCBs, you can have them made and assembled in sets of ten, leaving out just the parts that vary (like the RC low-pass filter and the sense resistor setting the maximum current).  After populating and testing the board and soldering the wires (to connectors) to it, use UV-curing nail polish to protect the PCB and components from moisture.

As you can see, one can definitely use "modern" ICs to get this kind of stuff done pretty effectively, without any microcontrollers or PWM at all.  Also, TPS92205xD2 are definitely not the only LED drivers with an analog input control pin you can find; there are many, many alternatives from several manufacturers that will work well.  You only need a microcontroller if you want to use the PWM controlled one LED drivers and a programmable ramp; I recommend the 3·(t/T)² - 2·(t/T)³ one for turn-on during 0≤t≤T, and 1 - 3·(t/T)² + 2×(t/T)³ for turn-off during 0≤t≤T in that case, which I find particularly pleasing.  Note that many of the cheap (~0.5€ in singles) microcontrollers easily programmed in the Arduino environment can achieve this, so it really would not affect the price much.