mosfet load switch

[andy257]

Hi All,

I am playing around with various configurations of a mosfet load switch and have stumbled across a problem i dont quite understand.

Ive attached the circuit below. Basically its a very simple P mosfet load switch. I dont understand why the mosfet does not turn on at the value of the zener. As soon as the zener conducts it should pull the gate low with respect to the source pin and the mosfet should switch hard on. However in reality it actually switches on at around 11 to 12 volts.

I am using an led as a simple load so its hard to see when it actually does switch on because there is no snap action, only a gradually increasing brightness until its fully on.


Any ideas what is causing this?

 

[Thunderchild]

the mosfet will have some conduction at any gate voltage, with the small current requirement of the led you will easily pass enough current to switch it on. the value that the datasheet reports is the gate voltage required to fully switch the mosfet on to it's minimum resistance

 

[Hero999]

It will switch at the MOSFET's threashold voltage plus the zener voltage. The turn on will be slow and the voltage drop across the MOSFET will decrease as the current rises, untill it's fully on, then the voltage across it will increase with the current.

 

[crutschow]

Ive attached the circuit below. Basically its a very simple P mosfet load switch. I dont understand why the mosfet does not turn on at the value of the zener. As soon as the zener conducts it should pull the gate low with respect to the source pin and the mosfet should switch hard on. However in reality it actually switches on at around 11 to 12 volts.

It sounds like you expect the zener voltage to somehow drop once the zener voltage is reached. The doesn't happen. The zener does not pull the gate low. It conducts with a constant voltage drop across it (it looks much like a 9V voltage source). Thus the mosfet will start to slowly turn off as the voltage is increased above 9V plus the mosfet threshold voltage.

If you want the mosfet to fully turn on a 9V then you will need to add a comparator circuit to drive the mosfet.

 

[Hero999]

You nead two transistors to make it suddenly snap on when the threshold is exceeded.

Here's an example.
https://www.electro-tech-online.com...-with-transistor-as-switch.105748/#post864468
**broken link removed**
Thish circuit is designed for 25V, you'll need to change some of the component values for it to work at 9V.

Tr1 can be replaced with a P-channel MOSFET, if you like.

 

[Thunderchild]

are you trying to use a variable input ? if so you are a lot better off maybe with an N channel mosfet and proper drive circuitry

 

[andy257]

hi all,

thanks for the responses. I made the post because i didnt understand what i was seeing on the bench.

Ive actually posted a similar thread about a month or two ago and Mike answered with a very similar circuit. I dont see the difference between his and mine only the zener he has chosen is one of theose adjustable ones. His has hysterisis but i dont really see how that makes it any different to the circuit i posted.

original link below

https://www.electro-tech-online.com/threads/high-side-switch.103808/

 

[crutschow]

Ive actually posted a similar thread about a month or two ago and Mike answered with a very similar circuit. I dont see the difference between his and mine only the zener he has chosen is one of theose adjustable ones. His has hysterisis but i dont really see how that makes it any different to the circuit i posted.

It's the hysteresis that makes the difference between a fast, snap-action switch, and the slow switching that you observed.

 

[Hero999]

R3 provides positive feedback which causes hysteresis: read the post I linked to for more information.

 

[MikeMl]

Hi All,

I am playing around with various configurations of a mosfet load switch and have stumbled across a problem i dont quite understand...

The turn-on of the PFET is reasonably abrupt.

 

[Hero999]

I'm sure it depends on the MOSFET and the load current.

That's also a pretty slow turn on, if the voltage rise/fall is very slow 100mV/min then that MOSFET is going to burn up when it passes through the saturation region. Adding hysteresis will make it so it's not possible for the output switch to be half on and half off which would be lead to meltdown.

 

[MikeMl]

hi all,

thanks for the responses. I made the post because i didnt understand what i was seeing on the bench.

Ive actually posted a similar thread about a month or two ago and Mike answered with a very similar circuit. I dont see the difference between his and mine only the zener he has chosen is one of theose adjustable ones. His has hysterisis but i dont really see how that makes it any different to the circuit i posted.

original link below

https://www.electro-tech-online.com/threads/high-side-switch.103808/

You missed the critical point: There is much more to an LM431 than a Zener. The LM431 is used as an open-loop comparator with a voltage reference (2.500V Band Gap) connected to one of its inputs. As the voltage at the tap of the voltage divider (REF input to the LM431) swings from a mV above 2.5V to a mV below 2.5V, the LM431's Cathode suddenly switches from low to high, switching off the PFET with only a couple of mV of input voltage change..

Look at the equivalent circuit of what is inside an LM431:

 

[andy257]

Hi Hero & Mike,

Starting to understand now. I will try with some hysterisis.

One thing i noticed with the bjt circuit was that it has a potential divider on its gate unlike the mosfet equivalent circuit. Why is this?

 

[Hero999]

Are you talking about R4 and R5 on my previous reply to this thread?

R4 limits the base current and isn't needed with a MOSFET because its gate is an open circuit, R5 is a pull-up to ensure the transistor turns off and serves the same function with a MOSFET.

 

[indulis]

If the circuit is wired up as shown in the original post, it won't work... the MOSFET is backwards... the body diode will always be "conducting".

It's the hysteresis that makes the difference between a fast, snap-action switch, and the slow switching that you observed.

Hysteresis relates to "WHERE" (what voltage) it switches (changes state), not how fast the slew will be.

 

[Hero999]

If the circuit is wired up as shown in the original post, it won't work... the MOSFET is backwards... the body diode will always be "conducting".

You're right, I can't believe no one else spotted that.

 

[MikeMl]

Andy used a symbol for a NFET, but labeled it PFET. In my circuit, it is a PFET with the source connected to the + input. The LM431 based circuit has no intentionally added hystersis, but the opamp has an open-loop gain of over a 1000, so the circuit switches from On to Off with a very small change in battery voltage.

 

[Hero999]

It looks like a P-MOSFET symbol to me
**broken link removed**
**broken link removed** P channel MOSFET

 

[MikeMl]

The arrow would be pointing the other way if it was a PFet symbol.

 

[indulis]

The arrow as shown is for a "P" channel. The arrow would point towards the gate if were a "N" channel.

**broken link removed**