N-Channel MOSFET as a high side switch (no PWM)

[T_guttata]

The key to the question is to be able to have a voltage a least 5 to 10V above Vcc.
There are many solutions

1.  Use a charge pump (if this works for you. Some charge pump requires active clock which may not work for you)
2.  Use a capacitor switch voltage doubler (like ICL7660 mentioned earlier or any other voltage doubler.  This uses an internal clock)
3.  Use another voltage source which is higher than Vcc.

If this voltage can be created either by the 3 ways or any other ways, the problem is solved to use NMOS as a high side switch

There are many ways to create the required voltage, but obviously it should be at reasonnable complexity and cost.
I could use a SMPS, but it will cost probably $5-10 and will require >20 parts...

A photovoltaic coupler + N-Channel MOSFET (High Side) would look like attached, right?

Which voltage difference is created by the photovoltaic coupler? Is that called "open circuit voltage" in the data sheet?

[Zero999]

The key to the question is to be able to have a voltage a least 5 to 10V above Vcc.
There are many solutions

1.  Use a charge pump (if this works for you. Some charge pump requires active clock which may not work for you)
2.  Use a capacitor switch voltage doubler (like ICL7660 mentioned earlier or any other voltage doubler.  This uses an internal clock)
3.  Use another voltage source which is higher than Vcc.

If this voltage can be created either by the 3 ways or any other ways, the problem is solved to use NMOS as a high side switch

There are many ways to create the required voltage, but obviously it should be at reasonnable complexity and cost.
I could use a SMPS, but it will cost probably $5-10 and will require >20 parts...

A photovoltaic coupler + N-Channel MOSFET (High Side) would look like attached, right?

Which voltage difference is created by the photovoltaic coupler? Is that called "open circuit voltage" in the data sheet?

Yes, it's the open circuit voltage, since the MOSFET's gate draws no current, when charged.

Your circuit contains one error: a series resistor is required to limit the current through the LED.

[Benta]

Use the VO1263 instead if you can source it.
It's probably the best on the market right now, costs the same as the 1271 and is dual, meaning you can double the output current.
Expect switching times in the 1...10 ms range, depending on your MOSFET.

[CMTan]

Yes, that would work too.

[Zero999]

Use the VO1263 instead if you can source it.
It's probably the best on the market right now, costs the same as the 1271 and is dual, meaning you can double the output current.
Expect switching times in the 1...10 ms range, depending on your MOSFET.

The VO1263 is twice the price of the VOM1271 on Mouser and Digi-Key.
https://www.mouser.co.uk/ProductDetail/Vishay-Semiconductors/VOM1271T?qs=9vrc9xEJKyxSd3E39M%252BFqQ%3D%3D
https://www.mouser.co.uk/c/optoelectronics/optocouplers-photocouplers/photodiode-output-optocouplers/?series=VO1263

https://www.digikey.co.uk/en/products/filter/optoisolators-transistor-photovoltaic-output/903?s=N4IgTCBcDaIGoHkCMYBsBmABCAugXyA
https://www.digikey.co.uk/en/products/detail/vishay-semiconductor-opto-division/VOM1271T/3588700?s=N4IgTCBcDaIGoHkCyBGMB2FIC6BfIA

It's a decent enough suggestion. The LEDs can be connected in series, so no extra drive current is required. The downside is it lacks the built-in fast turn-off circuit of the VOM1271. Also beware that the open circuit voltage can go as high as 16.5V, which is too high for some logic level MOSFETs, although this is better  for driving some  non-logic level devices. A resistor can be put across the output to limit the voltage, if it's an issue. This will speed up the turn off time and might make an additional turn off circuit unnecessary.
https://www.vishay.com/docs/84639/vo1263aa.pdf
https://www.vishay.com/docs/83469/vom1271.pdf

[T_guttata]

Unfortunately, none of those parts is currently available.
What's about ASSR-V622-302E, https://docs.broadcom.com/doc/AV02-0259EN ?

[Benta]

Unfortunately, none of those parts is currently available.
What's about ASSR-V622-302E, https://docs.broadcom.com/doc/AV02-0259EN ?

Would work, but you'll need a logic-level MOSFET.

[Zero999]

The APV1121SX is available from Digi-Key. It will also require a logic level MOSFET.
https://www.digikey.co.uk/en/products/detail/panasonic-electric-works/APV1121SX/2804730

[T_guttata]

Nice :-) There are nice logic level MOSFEts from Nexperia, for example: BUK9Y6R5-40H

I might order some to test the switching behaviour.

For which current do I have to set the limiting resistor at the input? In the datasheet it says: typical LED operate current 0.85mA, maximum 3mA. Absolute maximum rating LED forward current: 50mA.

[CMTan]

Hi T_guttata,
I am wondering why to take so much trouble to use a photovoltaic coupler just to drive the NMOS high side switch.
If you really need a NMOS part for the high side switch it will be much easier to use a smart high side switch that takes care of the problem we have discussed about.

Please take a look at ITS4140N.
This is Smart High-Side Power Switch that allows direct logic interface and also have several protection mechanism built-in.  No need to worry about the gate voltage for the NMOS at all.
 

[Zero999]

Hi T_guttata,
I am wondering why to take so much trouble to use a photovoltaic coupler just to drive the NMOS high side switch.
If you really need a NMOS part for the high side switch it will be much easier to use a smart high side switch that takes care of the problem we have discussed about.

Please take a look at ITS4140N.
This is Smart High-Side Power Switch that allows direct logic interface and also have several protection mechanism built-in.  No need to worry about the gate voltage for the NMOS at all.
 

That's a decent suggestion. I doubt the ITS4140N is suitable. It has an on resistance of 1Ω, whereas the MOSFET has an R_ON of 6.5mΩ. A big problem is availability.

Nice :-) There are nice logic level MOSFEts from Nexperia, for example: BUK9Y6R5-40H

I might order some to test the switching behaviour.

For which current do I have to set the limiting resistor at the input? In the datasheet it says: typical LED operate current 0.85mA, maximum 3mA. Absolute maximum rating LED forward current: 50mA.

Which opto-coupler are you talking about?

The higher the LED current, the higher the output current and the faster it will turn the MOSFET on. The output voltage is weakly affected by the LED current, whice has to increase exponentially, for the open circuit voltage to increase linearly.

The turn off time is independent of the LED current. If there's a turn off circuit, it will be faster, that a basic pull-down resistor. I had a play in LTSpice. The opto-coupler is modelled as a behavioural current source,  with the current set to 1/714 of the emitter current, mating 14µA out, with 10ma in given on the APV1121SX datasheet and there are a load of diodes in series to limit the open circuit voltage.

Q1 and D1 form the fast turn-off circuit. If they're removed, the MOSFET takes more than one cycle to turn off.

[T_guttata]

Hi T_guttata,
I am wondering why to take so much trouble to use a photovoltaic coupler just to drive the NMOS high side switch.
If you really need a NMOS part for the high side switch it will be much easier to use a smart high side switch that takes care of the problem we have discussed about.

Please take a look at ITS4140N.
This is Smart High-Side Power Switch that allows direct logic interface and also have several protection mechanism built-in.  No need to worry about the gate voltage for the NMOS at all.

Thanks for the idea. Unfortunately it will not work.

The main point is to use N-channel MOSFETs instead of P-Channel because Rdson is lower and they can candle much higher current. So if you choose a N-channel with 5 mOhm Rdson because 25 mOhm for the P-Type counter part is too high, the Rdson of 1 Ohm (ITS4140N) is obviously not an option.

Hi T_guttata,
I am wondering why to take so much trouble to use a photovoltaic coupler just to drive the NMOS high side switch.
If you really need a NMOS part for the high side switch it will be much easier to use a smart high side switch that takes care of the problem we have discussed about.

Please take a look at ITS4140N.
This is Smart High-Side Power Switch that allows direct logic interface and also have several protection mechanism built-in.  No need to worry about the gate voltage for the NMOS at all.
 

That's a decent suggestion. I doubt the ITS4140N is suitable. It has an on resistance of 1Ω, whereas the MOSFET has an R_ON of 6.5mΩ. A big problem is availability.

Nice :-) There are nice logic level MOSFEts from Nexperia, for example: BUK9Y6R5-40H

I might order some to test the switching behaviour.

For which current do I have to set the limiting resistor at the input? In the datasheet it says: typical LED operate current 0.85mA, maximum 3mA. Absolute maximum rating LED forward current: 50mA.

Which opto-coupler are you talking about?

The higher the LED current, the higher the output current and the faster it will turn the MOSFET on. The output voltage is weakly affected by the LED current, whice has to increase exponentially, for the open circuit voltage to increase linearly.

The turn off time is independent of the LED current. If there's a turn off circuit, it will be faster, that a basic pull-down resistor. I had a play in LTSpice. The opto-coupler is modelled as a behavioural current source,  with the current set to 1/714 of the emitter current, mating 14µA out, with 10ma in given on the APV1121SX datasheet and there are a load of diodes in series to limit the open circuit voltage.

Q1 and D1 form the fast turn-off circuit. If they're removed, the MOSFET takes more than one cycle to turn off.

APV2121S

What does it mean "takes more than one cycle"?

[freda]

Hi T_guttata,
did you  miss my prior post ?
with it you can choose whatever low RdsON MOSFET you can afford (and get), not worry whether its a logic threshold  or not, or
be too concerned with Gate charge, the dc/dc converters small enough to put on your PCB, can drive many mosfets.
Whats more you can switch it as slow as you want as its not a boostrap circuit.
No exotic parts required.
The only nuance if your switching a hi voltage hi side, then the isolation potential of the dc/dc needs consideration
and the direct coupled pull down should instead be a opto-coupled switch.
I just rechecked your OP, and my circuit should be fine to re-configure to yours.
Please..please

Please take a look at ITS4140N.

Oh look, No Stock!!
Maybe those posters work for Vishay or Infineon Marketing department?

[T_guttata]

@ freda

I did not really understand your circuit. You would suggest to use the B0509S-1WR3 to create a voltage difference of 9V to switch the N-Channel MOSFET on the high side? What else would I need?

[CMTan]

Yes, it would work too.  Like what I said earlier, you need to create a voltage at leave 5 to 10V higher than Vcc.
Any isolated DC-DC will work.

You just need to connect the output reference ground of DC-DC to Vcc.  Since it is isolated, this connection is not a problem as long as the isolation (Vin to Vout) is < 1500V (for the one mentioned).

With this higher voltage you can drive the gate and use whatever NMOS of your choice. 

[T_guttata]

So basically this B0509S-1WR3 is just a "one component solution" for a SMPS circuit, right?

[Zero999]

opto-couplers are generally slow. My suggestion is use a isolated DC/DC converter to supply Vgs. see circuit fragment.... (Attachment Link)
(Attachment Link) (Attachment Link) " alt="" class="bbc_img" />

Of course i would only recommend this circuit if the high side voltage is not too high, say 50V or less.

The original poster didn't say how fast it needs to switch.

An isolated DC-DC converter  will work, but that circuit will have a very slow turn-off time, with that 10k resistor between the gate and source. A proper MOSFET driver is required for fast switching.

APV2121S

What does it mean "takes more than one cycle"?

Without a fast turn off circuit, the MOSFET will take a long time to turn off, 10s of ms, which is longer than the period of the square wave used in the simulation I posted. If you want to see for yourself, LTspice is a free download.

You've not state how fast it needs to be. If slow switching is acceptable, just use the opto-isolator +MOSFET+pull-down resistor. Need it to turn off a little quicker: add the BJT+diode.

If you really need fast switching, use the ADum6132, which includes much of the circuitry required, including \an isolated DC-DC converter.
https://www.analog.com/media/en/technical-documentation/data-sheets/adum6132.pdf