world's simplest variable DC power supply EVER? / General Science and Electronics / Forums

Forums

4hv.org :: Forums :: General Science and Electronics

« Previous topic | Next topic »

world's simplest variable DC power supply EVER?

1 2 next

Move Thread

LAN_403

haxor5354

Sun Mar 20 2011, 12:40AM

Registered Member #2063 Joined: Sat Apr 04 2009, 03:16PM
Location: Toronto
Posts: 352

I was just messing around with LM317 and mosfets trying to make a variable supply by amplifying the output of the LM317 by using a mosfet.
after spending a few hours with no progress, I just had the craziest idea ever. WHY NO JUST AMPLIFY A VARIABLE RESISTOR?!!
I tested it out with a 10 turn 4.7K pot and a IRFP260N mosfet heatsinked and fan cooled, and it actually works!!
works great as a 12V halogen light dimmer.
no op-amps,LM317,capacitor,chokes,ICs.... just a pot and a mosfet, can't get any simpler than that!!
any circuit design flaws?

Inducktion

Sun Mar 20 2011, 12:43AM

Registered Member #3637 Joined: Fri Jan 21 2011, 11:07PM
Location: Buffalo, NY
Posts: 1068

haxor5354 wrote ...

I was just messing around with LM317 and mosfets trying to make a variable supply by amplifying the output of the LM317 by using a mosfet.
after spending a few hours with no progress, I just had the craziest idea ever. WHY NO JUST AMPLIFY A VARIABLE RESISTOR?!!
I tested it out with a 10 turn 4.7K pot and a IRFP260N mosfet heatsinked and fan cooled, and it actually works!!
works great as a 12V halogen light dimmer.
no op-amps,LM317,capacitor,chokes,ICs.... just a pot and a mosfet, can't get any simpler than that!!
any circuit design flaws?

...I tried the same thing once. The mosfet just got hot and the voltage didn't really change that much coming out. I think it's mainly because you're operating the mosfet in it's linear region, and they get REALLY hot there because they're not really on, or off. a Transistor might work better for this, if i'm right, or an IGBT.

haxor5354

Sun Mar 20 2011, 12:55AM

Registered Member #2063 Joined: Sat Apr 04 2009, 03:16PM
Location: Toronto
Posts: 352

mine stayed cool when fan cooled.
I remember a few years back I made a variable power supply with a 2n3055 to amplify a Lm317 and it get really hot even when fan cooled.
transistor are current controlled and mosfets are voltage controlled right?
PWM would work great for simple loads such as motors and lights but in my case I want pure DC not pulsing DC

Inducktion

Sun Mar 20 2011, 12:56AM

Registered Member #3637 Joined: Fri Jan 21 2011, 11:07PM
Location: Buffalo, NY
Posts: 1068

haxor5354 wrote ...

mine stayed cool when fan cooled.
I remember a few years back I made a variable power supply with a 2n3055 to amplify a Lm317 and it get really hot even when fan cooled.
transistor are current controlled and mosfets are voltage controlled right?
PWM would work great for simple loads such as motors and lights but in my case I want pure DC not pulsing DC

PWM is what you want, yes. You can make it pure DC just by adding a parallel cap to filter it out, no rectification needed.

Mattski

Sun Mar 20 2011, 02:11AM

Registered Member #1792 Joined: Fri Oct 31 2008, 08:12PM
Location: University of California
Posts: 527

Haxor, in a circuit such as you have posted here there will inevitably be a large power loss turning into heat in the pass transistor since you are attempting to reduce a voltage while passing a current, so P_heat = (V_source - V_load) * I_load. So the higher this power dissipation is the bigger heatsink/fan you need.

Also this circuit has little regulation so if the load draws less current then the voltage will rise. If the load draws more current then the voltage drops. There is some feedback because as the source voltage falls the transistor is turned on more strongly partially compensating for the change in load, but not fully. The regulation might be good enough for some circuits. If you implement some type of closed loop feedback to automatically control the gate voltage then you would have a linear regulator like a 7805, or lm317. If your load is fairly constant and you are constantly monitoring it then it might be

wrote ...
PWM is what you want, yes. You can make it pure DC just by adding a parallel cap to filter it out, no rectification needed.

With just a capacitor the voltage would still jump up to the source voltage when driving it with a PWM to turn it fully on and off. Adding an inductor and diode as in a buck converter would do the trick though, and can be made to have fairly small ripple.

wrote ...
transistor are current controlled and mosfets are voltage controlled right?

BJT's and MOSFETs (both are transistors) are both voltage controlled at the physics level. But it is frequently convenient to treat BJT's as current controlled in circuit analysis because the collector current vs. base-emitter voltage has an exponential relationship, which can be inconvenient.

James

Sun Mar 20 2011, 02:13AM

Registered Member #3610 Joined: Thu Jan 13 2011, 03:29AM
Location: Seattle, WA
Posts: 506

Inducktion wrote ...

PWM is what you want, yes. You can make it pure DC just by adding a parallel cap to filter it out, no rectification needed.

Not quite. You need an inductor and a diode, then you have a buck converter. If you just put a capacitor across a PWM output, you will have enormous peak currents and both the mosfet and capacitor will see huge stresses. The only thing preventing the capacitor from charging to the full peak voltage on even the shortest pulse width is the series resistance of it, the mosfet and the wiring.

Sure you can use a mosfet as a variable resistor to drop the line voltage, it's just like using a giant variable resistor, which was done for decades. For theatrical stage lighting it was common to use vats of salt water with an electrode that could be raised and lowered to control the lights. Giant variable resistors were commonly used in theaters and such where budget didn't allow for variacs on everything up until the 1960s or so when thyristor dimmers appeared. Any sort of resistor is a hugely inefficient way to control any significant amount of power though as you're burning up all the excess voltage as heat.

haxor5354

Mon Mar 21 2011, 10:24PM

Registered Member #2063 Joined: Sat Apr 04 2009, 03:16PM
Location: Toronto
Posts: 352

so i just made a buck converter with a 555PWM generator, but it seems like the filter cap is storing residual voltage when the duty cycle is turn down. what value filter cap should i use?

Arcstarter

Mon Mar 21 2011, 11:48PM

Registered Member #1225 Joined: Sat Jan 12 2008, 01:24AM
Location: Beaumont, Texas, USA
Posts: 2253

haxor5354 wrote ...

so i just made a buck converter with a 555PWM generator, but it seems like the filter cap is storing residual voltage when the duty cycle is turn down. what value filter cap should i use?

The filter cap is fine, you need a load on the output if you want a good reading.

haxor5354

Mon Mar 21 2011, 11:54PM

Registered Member #2063 Joined: Sat Apr 04 2009, 03:16PM
Location: Toronto
Posts: 352

so, a bleeder resisotr?

Mattski

Tue Mar 22 2011, 12:08AM

Registered Member #1792 Joined: Fri Oct 31 2008, 08:12PM
Location: University of California
Posts: 527

A buck converter without any load will approach the power supply voltage even with extremely small duty cycles, so the capacitor would keep charging up until it is near the input voltage. Without feedback the output voltage depends on the load current, so really you want to put a resistor which will draw the same current as the load you intend to power.

A 1,000Ohm will draw a bit of current and a standard quarter watt resistor should will run reasonably cool up to 12V output, at P=V^2/R=144mW. For larger loading you want a smaller value resistor, but make sure that your resistor is rated for enough power dissipation.

Edit: James, we need to stop posting at the same time

1 2 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob