Hello all,
As the title somewhat states ,what would be the advantages or dis-advantages of …
Incorporating a dc mosfet led rotary dimmer , to use in a custom build ( aside from size constraints) and without the use of any PWM at all?
(No clicky, common momentary, or ramping modes driver being used )
Thanks for any replies in advance.
Isn’t that how “linear” drivers work? They use a FET like a variable resistor to control the current to the LED.
Probably in some way, but it has been my experience that all of the drivers in lights I have purchased over the years have used PWM to control the output.
For example ,take any common torch put it into a lower mode shine at a wall 20 feet or 5-10 meters away and move the beam very fast all over the wall and you will see a dashed pattern on the wall and not a continuous line. This dashed / residual flicker effect I do not want for my build.
Essentially what I need is a dimmable, 100 % direct drive , like a rheostat in the old days , but something more modern with a rotary knob and minimal heat.
If this is possible or commercially available what would be issues to contend with.
Thanks
You’ve probably tried mostly inexpensive lights then. Most better lights don’t use PWM.
Using a FET is not “direct drive”.
I guess I should also say.
Does this mean that of the many led lights I own I do not have the correct driver, in any of them to get rid of this kind of effect , or that the correct driver reduces (key word also minimizes) this PWM flicker to a degree where if you perform the test as I outlined above , I or you would not see a dashed beam trail, or the dashes would be just closer together?
Which driver or other electrical/ electronic device would you advise I use ,which has no flicker or signal gate/ PWM and IS direct drive and can pull 4-5 amps.
Thanks appreciate the comments.
Get a flashlight with linear driver and Anduril, something from Emisars.
Remember this is a potential build… Which driver or other electrical/ electronic device would you advise I use ,which has no flicker or signal gate/ switching channel PWM etc.
and IS direct drive and can pull 4-5 amps but still can be dimmed down or blast out full battery output other than a straight up rheostat (sorry no disrespect but if you are under 25 years you maybe do not know of this device)
Adjustable with a knob or slide switch* (see my other posts above )
Thanks appreciate the comments.
If such a device is available it probably was invented in the years where analog devices were still in common use but with a bit of digital thrown in as well.
Thanks much
Is what happen in these basic regulators that use D882 and pot.?
Like SammysHP said many lights don’t use PWM dimming, brands like Acebeam, Fenix, thrunite, Olight usually use constant current SMPS ( buck, boost) drivers with analog dimming, though they might still have some cheap models with PWM dimming, also at min mode, depending how low it is, there might be low frequency ripple, but usually min modes aren’t low enough for than.
Anyway a constant current current driver can be dimmed by using a potentiometer, this for example is the Noctigon linear driver (Convoy’s linear driver is similar) :
Like you proposed it uses a Mosfet, a device with varying resistance depending on the Gate-Source voltage, but the exact resistance/Vgs is unknown so instead of directly changing Vgs an Op-Amp increases or decreases it if the voltage across a sense resistor is too high or too low.
By applying a voltage to the non inverting input of the Op-Amp (pin 1) you get the same voltage across the current sense resistor. Here dimming is done by a filtered PWM signal, divided down to desired Vsense by a voltage divider (R3-R4), but if you want you can replace R3 and R4 by a potentiometer (or rather add it in between, with different R3 and R4 value in order to have a min/max voltage).
The ”PB3” wire would be directly connected to the 2.8V LDO and ON/OFF is the Op-Amp power supply (pin 5), that way the MCU is completely bypassed.
The same modification can be done on a buck or boost driver, they usually use a similar circuit except the IN+ and IN- are swapped.
Though if I were to make a driver with a dimming knob I would instead wire the potentiometer to an ADC pin of the MCU, which would in turn set the current, that way the MCU can still handle thermal throttling, low voltage protection, button UI…etc.
Thank you all for the advice,
I do believe , freeman you nailed it,
thankyou very much, exactly what I was looking for!
So freeman when I obtain the driver, I may have an additional question or two as I am unfamiliar with the task you kindly detailed ,looks easy enough, I hope you do not mind, mainly about the particular board layout ,it’s components and termination points.
Many thanks for the help, such a wealth of info from everyone,
it’s great.
I am interested in this project too. I’m not techy enough to design a driver but I can solder a pot to something if I’m told what goes where.
To me rotary controls are the pinnacle of intuitive design. Having looked at everything on the market it seems they all fall short of the ideal. Most have PWM and/or unacceptably high parisitic drain from magnetic switches.
The usual solutions to the parasitic drain seems to be having a second clicky switch which is just bad design.
The mechanical off should be incorporated into the rotary switch.
There are potentiometers with on off switch that could work but somebody smarter than me would have to figure out how to do it.
I note that I have recently acquired a Fenix LD42 which has 5 brightness levels on a mechanical rotary switch. Constant current in all modes. No blinkies… I don’t know whether trying to reverse engineer this would be useful.
Torchmun, do you mean e-switches? The typical "e-switches" (I don't think there's anything magnetic about most) in most lights may drain a battery in months. I never thought of the second switch as a guard against that. You can just partially unscrew the head if you're really worried about long term storage and readiness. I always thought of the extra switch as being because people like the extra switch, also how I've seen it advertised. Some are forward clicky, for "tactical" use, and it allows holding the light by your head while turning it on more easily, which may also be "tactical." Unfortunately, because of poor UI's many do not allow turning it off (and thus on) with the e-switch, so then even if you want to control it with the side switch, you still have to turn on it with the rear switch. That's the annoying part.
As for PWM, there are definitely buck lights on the market without PWM. That said, a knob is kind of cool in its own way, although probably requires two-handed operation.