According to my research, mostly N-Channel is used for low side switching and P-Channel is used for high side switching. But if I want to use N-Channel for high side switching, what are the ways and what are the drawbacks of using N-channel for high side switching. What I understand is the use of some kind of charge pump or bootstrap(haven't understood this yet), so that the Vgs can be made greater than Vin. But it would be great if someone properly explains this in a easy way. Please help. Thank You.
2 Answers
You will need a voltage that is well above the power rail to supply enough gate voltage to turn the N mosfet full on.
In PWM H-bridges this is often done with a charge pump which is pumped by the PWM cycle.
If you want the N mosfet to be on in a stationary situation you will have to provide that higher voltage by some other means, maybe with a standalone (self oscillating) charge pump.
-
\$\begingroup\$ Thank You. Can you please guide me to something, where I can understand the charge pump or bootstrap circuit. \$\endgroup\$ Commented Dec 18, 2020 at 6:56
-
\$\begingroup\$ As comments on the other answqer mention, check relevant datasheets. \$\endgroup\$ Commented Dec 18, 2020 at 10:49
You can drive a high-side NMOS switch with special gate driver. It takes care of generating the high voltage necessary to fully turn on the high side NMOS. The search term is "high side gate driver."
PMOS is OK sometimes but because of the way the physics and chemistry of semiconductors works NMOS will always have a slight edge in terms of switching efficiency. High power applications typically use NMOS switches on the high side as well as the low side.
-
\$\begingroup\$ I am trying to build something of my own and not use some kind of special driver. Also on the use of PMOS, what I understand is PMOS provides more resistance to current compared to NMOS(that is the current carrying capacity). Please guide me with charge pump or bootstrap circuit as I understand that is what is needed to provide the extra Vgs needed. \$\endgroup\$ Commented Dec 18, 2020 at 7:03
-
\$\begingroup\$ Well, PMOS is much easier to drive. You should determine if PMOS resistance is low enough for your purposes. If it is, then go that route. In my opinion it is not worthwhile to try to implement a high-side driver for NMOS using discrete components. \$\endgroup\$ Commented Dec 18, 2020 at 7:06
-
\$\begingroup\$ you can make your own photovoltaic mosfet driver with several flat red LEDs glued together flat-sides and one Jumbo white LED glued shining into the red LEDs. Wire the reds in series, then to GS. When you turn on the white LED, it generates a voltage in the reds and thus raises G a few volts above S. A logic-level N-channel works best for this. Or you can just use a battery or secondary PSU. \$\endgroup\$– dandavisCommented Dec 18, 2020 at 8:42
-
1\$\begingroup\$ @mkeith I've actually done it a few times. with reds 4 in series I can get ~7.5 volts, which is weird because the LEDs spit out more than their forward voltage. I have no idea the frequency, probably depends on gate capacitance since there is not much current, but it can stand-in for a relay no problem, and seemed to handle 1khz PWM fine on an IRLZ44N (no scope, so it could be rounded AF...) \$\endgroup\$– dandavisCommented Dec 18, 2020 at 8:48
-
1\$\begingroup\$ @mkeith: I thought the same, but the white emitter gave a far higher voltage, but perhaps it was just much brighter. commercial PV fet drivers use infrared instead of red, which produces even higher voltage. I read somewhere that frequencies (or is it wavelength?) above or equal to the LED color will convert to voltage; anything excites an IR, nothing exits a UV. I tried about 20 combinations and 12mm bright white into rectangular red top-firing was the most effective for me. see top 3: youtube.com/results?search_query=photovoltaic+mosfet+drivers \$\endgroup\$– dandavisCommented Dec 19, 2020 at 3:37