Possible way to get more output from a transistor than normally?

[realflow100]

I was thinking of a new way to run a transistor. using PWM and an upstepper for a higher maximum output by upstepping the voltage to a seperate voltage from the power supply voltage. by a circuit that uses PWM for the output. like a class-D amplifier.

but it'd be more efficient because there would be a massively better maximum output.
The transistors could be MOSFET or darlingtons

Since the voltage is seperated from the supply voltage (through an upstepper design of some kind)

It'd cause the output to be greater than if only biased with a non-upstepped voltage from the input combined with the supply voltage.

like using an AA battery instead of using a resistor between the power supplies positive and the transistors base.

so it'd upstep the voltage to a seperate voltage supply which would turn the transistor on from a lesser voltage than the supply voltage. but be able to turn on the transistor much much more.

Is that a normal thing in class D amplifiers? or is it done in a similar but different way?

the reason I was thinking PWM + upstepped (not to a higher voltage but maybe seperate voltage)
was because it would be more efficient and need less energy to actually fully turn on the transistor.. with less voltage drop. practically full rail voltage possibility.. especially with mosfets?

 

[jstott]

Congratulations, you've just re-invented the class-D amplifier.

-JS

 

[nigelwright7557]

The irs2092 class d IC creates a voltage higher than the b+ rail to drive mosfet's right up to the rail.

I think the IR2113 works the same.

 

[realflow100]

would that work with BJT transistors too? is that what some amplifiers do?

 

[rayma]

would that work with BJT transistors too? is that what some amplifiers do?

It's a very old idea, having an auxiliary LV supply to allow operation to the HV rail.

 

[wg_ski]

You don't really need a PWM supply to generate that extra voltage. As long as the average current demand over a cycle is miniscule, a bootstrap circuit works just fine. Most 'high side' MOSFET driver chips use some sort of bootstrap. Even the ones for non-audio applications.

 

[realflow100]

very interesting though...
I find using a transistor.. Hooking up a 1.5v AA battery to the transistors base somehow "adds" to the voltage output of the transistor to whatever its powering.

I tested it with a small motor and listened to the difference between transistor itself and a resistor.. straight wire to the motor. and using an AA battery with the transistor.
quite facinating.

and what is and how would you use a bootstrap?
to me it sounds like something to do with a computer bios lol.

I was thinking of the idea for both transistors and mosfets.

 

[rayma]

what is and how would you use a bootstrap? to me it sounds like something to do with a computer bios lol.

In the case of a high side gate drive, it's just a diode from the positive terminal LV supply to the capacitor's positive terminal,
and the capacitor's negative terminal connects to the output. This bootstrap capacitor is recharged during
the low output state, when the diode conducts. The cap is sized to hold up the gate voltage during the longest
time interval required. The negative LV supply terminal must be referred to the negative HV bus.

 

[realflow100]

hmm.. what if the transistor needed to be continuously on though? like in a different situation?

 

[rayma]

hmm.. what if the transistor needed to be continuously on though? like in a different situation?

Then you need a true floating supply, on an isolated winding, referred to the positive HV bus.

 

[wg_ski]

Define "continuously". A bootstrap cap will hold a MOSFET on for a long time. Maybe not for days without a refresh of charge, but for certainly seconds or minutes - perhaps hours if the cap is low enough leakage. The gate itself won't draw squat when it's not actually switching.

 

[realflow100]

I was thinking more for a BJT... it'd draw enough current unless the capacitor is enormous..