N-channel MOSFET as voltage regulated resistor

[newInRobotics]

I know that it is possible to use MOSFET as regulated resistor if following conditions are met:

• V_GS > V_threshold
• V_GS < V_DS

Say I have the circuit as attached below. LED has these specs:

• V_forward = 3.45V
• I_forward = 20mA

To have LED at it's brightest I need to hook up 330 Ohm resistor. However I want to be able to regulate its brightness by varying resistance.

Is it possible to fine tune R_DS of a MOSFET so that:

• At 9V from trimmer, R_DS = 330 Ohm
• At 0V from trimmer, R_DS = 1 kOhm
• Or vice-versa

Is there a formula to calculate R_DS?

[Soeren]

Hi,

I know that it is possible to use MOSFET as regulated resistor [...]

To have LED at it's brightest I need to hook up 330 Ohm resistor. However I want to be able to regulate its brightness by varying resistance.

Is it possible to fine tune R_DS of a MOSFET so that:

• At 9V from trimmer, R_DS = 330 Ohm
• At 0V from trimmer, R_DS = 1 kOhm
• Or vice-versa

Don't over-engineer stuff 

A 4k7 potentiometer (used as a variable resistor) plus a resistor of 270R, in series with the LED will give you a range of:
  (9-3.45) / (4700+270) = 1.1mA
to
  (9-3.45) / (270) = 20.6mA
(Look ma... No FETs)

However... White LEDs (if that's what you have) is better dimmed with PWM, as they have a substantial color shift when dimmed by reducing the current.

[newInRobotics]

Don't over-engineer stuff 

The circuit I have provided is just an example application to keep it simple. The real application will require something that changes resistance with a change of voltage or current, hence I thought about MOSFET, as they change R_DS with a change of V_GS. My questions still remain the same:

• Is it possible to fine tune R_DS of a MOSFET as I need only a fraction of available range of resistance?
• Is there a formula to calculate R_DS?

[vinniewryan]

In the datasheet for your MOSFET, there should be a chart for "SATURATION CHARACTERISTICS". This would be the only section I can find which shows the relationship of V to rDS. My only other suggestion would be to connect the FET to an ohm meter and vary the voltage to your circuits requirements to map out the change in rDS manually.

[Soeren]

Hi,

My questions still remain the same:

• Is it possible to fine tune R_DS of a MOSFET as I need only a fraction of available range of resistance?
• Is there a formula to calculate R_DS?

Not sure exactly what you mean with "fine tune"?  You cannot change any transistor in any controlled way.

There's no formula, as no two devices will be the same (even if from the same production batch).

To get exactly the range that you need (and at all temperatures) with a FET (any FET) will take a feedback arrangement.

Not knowing what you're building, but a digital potentiometer might be a better solution.

[newInRobotics]

You cannot change any transistor in any controlled way.

Aha, that's the answer I was looking for.

Not knowing what you're building, but a digital potentiometer might be a better solution.

I'm still trying to reinvent the bike for the sake of learning   All recent threads are related to PWM generator. The latest idea is based on 555 timer (as shown in the attachment).

By changing R1 i can alter duration of high side of pulse and by changing R2 - low side of pulse. Having R1 = R2 I can have perfect 50% duty cycle. I figured that if I reduce one resistor by same value as I increase the other I can vary duty cycle without changing frequency (hence all the questions about well controlled resistance of MOSFET).

Digital pot would suit well for this application, however as app requires duty cycle to be changed by analog voltage I would need ADC and I₂C (a microcontroller in essence). Having uC would defeat the purpose of the device as uC itself can generate PWM.

My latest thought regarding voltage controlled pot is to have LDR in conjunction with LED (and trimmer to fine tune resistance of LRD).

[Soeren]

Hi,

I'm still trying to reinvent the bike for the sake of learning   All recent threads are related to PWM generator. The latest idea is based on 555 timer (as shown in the attachment).

By changing R1 i can alter duration of high side of pulse and by changing R2 - low side of pulse. Having R1 = R2 I can have perfect 50% duty cycle. I figured that if I reduce one resistor by same value as I increase the other I can vary duty cycle without changing frequency (hence all the questions about well controlled resistance of MOSFET).

There is a deep secret to the 555, passed from father to son in a closed society called the brotherhood of the timer chips... If you promise never to reveal your source, cross your heart and hope to live, I'll point you towards it. *coughPin5cough*   

Seriously, you want to give the datasheet a long hard look and start experimenting with the control input. For starters just connect a potentiometer (of max. 1kOhm) from pin 8 to pin 1 with the wiper going to pin 5. That way you can get a feel for the range available (won't work at the extremes of the pot).

My latest thought regarding voltage controlled pot is to have LDR in conjunction with LED (and trimmer to fine tune resistance of LRD).

Yes, it's ancient technology, it's called a Vactrol (Google will probably have some illo's for you)
Without feedback, it would be hard to make two vactrols track perfectly and wanting to do one of them inverted ain't gonna make it any easier.

[newInRobotics]

*coughPin5cough*

Isn't pin 5 used to alter frequency of oscillation? How would it help me to change duty cycle without changing frequency?

[Soeren]

Hi,

How would it help me to change duty cycle without changing frequency?

As I said, study the datasheet... And whether you see it or not, hook up a test circuit and experiment. That's sometimes the best way to learn.

But don't give up on the datasheet too soon, it's a really good thing to know the 555 intimately, to know when it will be the best solution and when it absolutely won't.

[Admin]

Look up a potentiastat, a device that electronically changes resistance to give you a desired output voltage. There are various schematics out there.

I'm actually working on a similar project, but I haven't quite figured out the easiest way to do 16 of them with only one mcu, yet . . .