Tieing Linear Voltage Regulator Ground to Positive Voltage

[terry_dactle]

Hi,

I have a project with an adjustable voltage being controlled by a MOSFET and a esp

My MOSFET is ideal at a Vgs of 10V, max of +/- 20v and the Vadj is 6-65V

I have a gate driver but it wants a voltage source for the high side MOSFET driver.

My question is, can I use a Linear regular such as the LM7812 and tie the ground to the source 6-65V via a diode
 Would this safely and reliably give me 12v over the Vadj rail to drive my MOSFETs?

Thanks in advance.pl

[Kim Christensen]

I doubt it, but post a schematic of what you have in mind.

[tom66]

If the MOSFET is being switched more often than about 1kHz, use a standard charge pump circuit to generate the high-level gate voltage with a high-side gate driver (e.g. IR2117 datasheet).  Such a circuit only needs about 12V for the gate bias, the level shifting is performed by the charge pump and gate driver.

If the MOSFET needs to be turned on continuously, with undefined off time, then you will need to either use a P-channel MOSFET (which can be switched by pulling the gate towards ground, assuming the source is positive and the Vgsmax is not exceeded),  or you need to use a circuit to generate something like +10V above the input supply rail to saturate the N-channel MOSFET.  In most cases, unless on-resistance and solution space is critical, you will do better with a P-channel MOSFET as it is far simpler to drive.

You can't use a linear regulator to do what you want without some kind of charge pump somewhere to generate the high-side voltage (if you're using an N-FET).

[Zero999]

A rough guess is the original poster is using a P-channel high-side switch, in which case a negative voltage regulator, with its 0V terminal ties to +V, along with a suitable N-channel driver and level shifter could be used to drive it.

[tom66]

A rough guess is the original poster is using a P-channel high-side switch, in which case a negative voltage regulator, with its 0V terminal ties to +V, along with a suitable N-channel driver and level shifter could be used to drive it.

There's no need for anything so complicated though.  A zener diode and resistor will readily prevent the maximum gate voltage from being exceeded.

[Zero999]

A rough guess is the original poster is using a P-channel high-side switch, in which case a negative voltage regulator, with its 0V terminal ties to +V, along with a suitable N-channel driver and level shifter could be used to drive it.

There's no need for anything so complicated though.  A zener diode and resistor will readily prevent the maximum gate voltage from being exceeded.

That's fine at DC. The fun starts when it needs to switch at a much higher frequency.

[tom66]

That's fine at DC. The fun starts when it needs to switch at a much higher frequency.

Indeed, but I can't see how having a floating linear regulator will be much better.  Unless you're talking about a convoluted circuit where the gate drive would be switched between Vcc and Vcc-10V, but that would require a fair number of discrete parts and could still be achieved without the regulator given you don't need much average current to switch a FET.

You really want a dedicated P-channel or N-channel driver once you get into the several-kHz range, and most of those contain internal logic/limits to avoid overstressing the MOSFET.  If you are getting into the 10's to 100's of kHz, there is no reason (IMHO) to do anything other than use an N-channel FET with a standard high side driver IC, which automatically gives you Vgs protection as Vgs will never exceed Vcc (except if you violate dv/dt, size your gate resistor appropriately.)