mosfet drivers high vs low side

pbateman

New Member
I recently had a post about mosfets getting really hot when using pwm to control a 12v dc motor @ 10 amps. The post is here:
http://www.picaxeforum.co.uk/showthread.php?t=8266

Anyhow, I purchased some new mosfets which turn on at 5V so that I could drive them directly from the picaxe and not worry about them getting hot. Unfortunately, even the new mosfets are getting hot, so I can only assume that the picaxe is unable to supply enough current to turn them on quickly enough.


So, I have decided to go with a mosfet driver to avoid these limitations. I am unsure as to the meaning of a high side driver or a low side driver. Can anyone explain the difference to me?

Currently, my npn mosfet is hooked up as follows. Source to ground, drain to motor, and gate to picaxe pwmout. I assume that I will need a high side non-inverting mosfet driver. Is this correct?


Thanks!!
 

BCJKiwi

Senior Member
Just going over ground many others have covered before as I have followed these discussions closely as I am working on the same thing i.e. a PWM'd supply rail for supply to many LEDS for dimming purposes.
My apologies to the others who have posted on this before for re-stating their contributions.

My understanding of what has gone before is as follows;
Don't know which MOSFET you purchased but a true logic level MOSFET will turn on fully at well under 5V and the current is well inside the PICAXE output range.
I have just ordered IRL520 MOSFETS
These are basically the same as IRF520 MOSFETS
The "L" MOSFET has a gate threshold of 1-2 V and is hard on at <=4V.
So if you have the true Logic MOSFET it may be the PWM that's the problem.
HOWEVER the "F" MOSFET has gate threshold of 2-4 V but does not really come on fully until 10V
The MOSFET generates heat as it transitions from high resistance (off) to low resistance (on). So it will produce heat if it's in the transitional state too much of the time.
This "transitional state" can come from two areas:-
1. Not enough voltage so it never gets out of the transitional state (not a problem if you have the true logic level MOSFET),
2. Too many switch on/off cycles.
So the first thing to check is that the PWM is running as slow as you can get it. Use the calculator first - check that you have the right chip and clock speed selected and just keep reducing the frequency until you get to the minimum. Under clocking the PWM is another option if you still have a problem.

It will also produce heat if it's undersized for the current load of course but I presume you have that well covered. If you don't have a heat sink it should be significantly over-rated.
 

Rickharris

Senior Member
High side low side refers to where the FET is - If it is on the ground side of the load - i.e. between the load and 0 volts it is low side.

If it is between the load and the positive voltage it is high side

(assumes an N channel FET)
 

pbateman

New Member
Rick - thanks for the explanation. Why should the driver care if the mosfet is high or low side? Unless high side drivers create a much higher gate drive voltage than low side drivers?


BC - thanks for your thoughts. THe mosfet that I chose is the irf3706 you can get the data sheet here:



if you are interested.

This mosfet has a max on resistance of 10.5 mOhms at a gate voltage of 4.5 V. If I run the load with the gate switched on and no pwm, it does not get hot. The problem is the transitional state. If the picaxe had enough oomph (current) to turn the mosfet on quickly enough, then there would be minimal heating, but the picaxe cannot supply enough current so it takes too long to build up the gate voltage. Turning down the pwm frequency helps a little, but this is not an acceptable solution because it is audible. As a side note, I am not sure why so many people prefer to run their pwm in the audible range. Anyhow, a proper driver should solve this problem readily.

Any other thoughts?
 
Last edited:

moxhamj

New Member
As pointed out by BJCkiwi if there are two device numbers and one is 10V and one isn't it can get very confusing. I did a lot of researching 10 years ago, found one with a 5V drive and a low Ron and stuck with it. BUK555 if my favourite mosfet but I'm sure there are others.

Mind you, W=I^2xR so at 10A even with a Ron of 0.04 ohms that is 4 watts. And 0.04 ohms isn't much higher than the resistance of the lead going into the device. I don't know what the Ron is for your mosfet but if it is 0.04 ohms or higher it is going to get hot. Next step is a heatsink and heatsinks are rated as degrees per watt so if you know the Ron you can work out the watts and then if you are happy with a 20 degree temp rise you can work out which heatsink to get.

Addit - crossed posts - if the Ron is 0.01 ohms and it isn't hot when not being pulsed then you may well need to drive the gate harder. A picaxe can supply 20mA. If you run the picaxe into a BC547 signal transistor via a 1k and pull the collector high with a resistor and connect the mosfet gate to the collector you can pull it high and low with more current. Eg the BC547 is good for 100mA. V=5V V=IR R=V/I = 5/0.1 = 50 ohms. If that still doesn't work then you can go for a B337 and drive up to 800mA but I suspect 100mA or even 50mA may be enough.
 

Attachments

Last edited:

BCJKiwi

Senior Member
Well I'd be concerned as that device has it's ratings specified at 10V.

The Logic level devices have their ratings specified at 4.5V.
If you look at the IRL520N or IRL540N data sheets you will see the very first line of the feature set is "Logic-Level Gate Drive". Other manufacturers clearly state that Logic-Level gate drive devices are designed for direct interface to PIC micro controllers.

I just did all the research on this for myself and came to the conclusion that to be sure, I would only source an IRL type device as it is specifically a Logic Level device. I must say I found these rating charts very confusing!

The current required for the IRF3707 is in the region of Dr Acula's analysis at 17 mOhm / 0.017 Ohm

@Dr Acula
I see your favoured BUK555 is specified as a Logic Level Device. How would you expect it to behave in the application being discussed? Would you expect to need to use the additional drive transistor with a BUK555?

In my app I'm looking to supply around 1A to an array of Leds and will PWM the IRL520 to provide dimming to all the LEDs. So I will be using the lowest frequency PWM I can.
 
Last edited:

lbenson

Senior Member
BCJ--I'm also having trouble understanding/reading these specifications. For the IRF520 and IRL520 (logic level), do the accompanying charts show the distinction you're making--with the IRF520 VGS going to 8 volts (VDS 25V, 16A) and the IRL520 at the same VDS at only about 4.5 volts? For the untrained (myself), it can be hard to know what to look for in these data sheets. So whether the mosfet is hard on depends on what current you're trying to drive as well as on your gate voltage?
 

Attachments

moxhamj

New Member
Re: "Well I'd be concerned as that device has it's ratings specified at 10V."

The BUK555 or the other mosfets? The BUK555 is a "logic level mosfet" and the turnon volts are listed as 1-2V.

Going off on a tangent, if you use the higher current driver schematic above then the transistor doesn't care what volts it is switching so you can put 10V on the high side of the 100R resistor and it will perform the function of both increasing drive current and also voltage level translation. So you may not need to buy any new mosfets.

I've certainly switched the BUK555 at 1Khz directly off a picaxe. Not sure about how it would go at above audio frequencies (>20Khz). If audio whine is a problem then maybe go the other way, ie below 20Hz which is also not audible. One could get technical and calculate the input capacitance and current and do some sums. Try the driver circuit - you must have a small signal transistor lying around - any transistor would work.

Also if you do end up using low frequencies then the higher drive currents are not really needed - if you still need the level translation though to get 10V then replace the 1k with a 2k7 and the 100R with a 1k.
 
Last edited:

BCJKiwi

Senior Member
@ Ibenson
Yes I think those two graphs pretty well sum up the story.
You can see the 'L' device turns on fully at 5 V (20A) whereas the 'F' device needs 8V to achieve the same level of turn on. At 5V it is still in the switch-on transition and not fully on. This might not matter if the load is light and there are not a lot of switching cycles but for PWM it's a problem.
You won't get 8V from a PICAXE without an additional driver circuit of some sort such as the one proposed by the Dr, or a special driver chip. you would also need a higher voltage supply.

There are a number of other devices that do the same job such as the Dr's favourite the BUK555 (this is a Philips Logic level device). The main issue I had when trying to purchase in NZ was to find a TO220 type device at a reasonable price. Most new Logic level devices are surface mount.

Shafto's approach is to replace the MOSFET with a Voltage regulator with a separate on/off pin but personally I don't like that approach as I don't know what sort output you will get at turn on unless you put in all the extra filters and I don't believe the device is intended for that use. However he insists it is used that way by many without issue. PWM'ing the on/off pin of a Voltage regulator just doesn't feel right to me.
 
Last edited:

pbateman

New Member
Wow, been pretty active since I last had a look.

First off, I had a typo in my previous post my mosfet is an irf3706 not irf3707. I had a link to the datasheet, but it is no longer in my post. The datasheet can be found https://ec.irf.com/v6/en/US/adirect/ir?cmd=catSearchFrame&domSendTo=byID&domProductQueryName=irf3706:

As you can see it has Rds specs for 10, 4.5, and 2.8 V. Even @ 2.8V, the Rds is only 22mOhms, the two logic level devices that were mentioned irl520 and irl540 have Rds values of 180 and 44 mOhms @ a Vgs of 10 V respectively. That is, even when these two mosfets are driven at 10V on the gate they have a much higher on resistance than the 3706 when it is driven at only 2.8V.

The irf3706 and irl540 have similar delay time and capacitance values (specified at 4.5V at the gate) while the irl520 has lower delay times and capacitance (both will allow for cooler running at low gate currents), but the irl520 has too low of a current rating (10 A max) for my application.

Dr A: I am having trouble following a couple of things you said. How did you drive a mosfet at 1khz with a picaxe? The pwmout command will only go as low as 3.9khz. Also, anything between the lowest picaxe frequency 3.9 khz and 20khz is definitely audible. For things like leds, you can't hear the whine, but motors and coils scream like banshees, you can even hear the whine coming from and incandescent bulb.

I really appreciate everyones thoughts on these matters. Hopefully our discussion will help others when they get ready to do the same things that we are doing.

It is my firm belief that you cannot drive a mosfet, logic level or otherwise, at non audible pwm levels (>20khz) with high currents (~>10A) directly from a picaxe without the mosfet getting hot.
 

Michael 2727

Senior Member
High Side, Low Side - What Rick said -
High side low side refers to where the FET is - If it is on the ground side of the load - i.e. between the load and 0 volts it is low side.

If it is between the load and the positive voltage it is high side

(assumes an N channel FET)
You can have the load on either side, but if you use
an N-ch MOSFET Low Side, e.g. load between the MOSFET
and the NEG/0V it messes with the VGS voltage/resistances
and can be less efficient.

MOSFETs can be tricky beasts in some circumstances but
they are "usually" easier to use than BJT transistors.
5V-10V on the gate and they turn ON, doesn't get much simpler.
( with virtually no current (minimal) draw from the supply to GATE )

When a MOSFET starts to turn ON it acts like a variable
resistor until the full Gate (saturation) voltage is applied, e.g. 10V Vg.
This is assuming you want to use the full current capability
of the device, e.g. 22A, and the RSD-on stated, e.g 0.014 Ohms.

However this can change if you only require 1A.
The MOSFET may start to switch ON at 3.2V and you will
get near the RSD on with only 4.7V Gate voltage applied.
If you then try to draw more than the 1A the resistance
RSD on will shoot up and the mosfet will quickly get hot.

In many cases excessive heating points to the MOSFET
not being turned ON hard enough, too little Gate VOLTAGE.
(never mind the current MOSFET Gates don't use any, just the voltage)

Another thing that bugs me is the insistance of the "know it nothings"
when they refuse to use a Gate resistor, ALWAYS USE A GATE RESISTOR,
the resistance is not needed but they are used as a FAILSAFE device
in the case of a complete MOSFET failure where the full supply voltage could
backfire up the gate input line to the rest of your circuit.

Gate discharge resistor-
In some cases a Gate can be held ON with no input present. This is due to
internal capacitance. As good practice a resistor is placed from the Gate
to the Source (Neg/0V, N-ch). This resistance can be very high, from 47K
Ohms to 1 Meg, it will ensure that MOSFET always turns OFF when the
input is Low. (or removed )

Another thing that can catch you out, is when using a Gate resistor in
conjunction with a Gate pulldown (as above) it does form a Voltage Divider.

Using a 5V Gate input supply via a 4K7 Ohm resistor and a 47K Gate
pulldown, you will now only get 4.55V into the Gate. Be aware of this.

PWM Speed and Motors-
If you read the MOSFET Application Notes from various sites, they
usually recommend PWM of 4KHz to 20Khz for use with Motors.

I tend to use 4KHz as it seems to work well and the motor noise
isn't that bad (it will depend on your particular motor type).
Another reason you may not immediately realize is that @ 4KHz
PWM 50% ON and 50% OFF is fine. But what happens at 1% - 2%
Duty and 98% - 99% Duty the ON/OFF pulse width is VF-FAST !
You can divide /5 that if you use 20KHZ PWM.

Which brings us to BACK EMF protection.
Never rely on the MOSFET internal Doide for full BACK EMF
protection when using Inductive loads or Motors.
If you use very high PWM frequencies you will need High
Speed diodes for protection as standard ones are too slow.

There are a lot of very good data sheets and application notes
on MOSFETs and how to use them, at sites like IRF and ST, etc.
Go get some, even if they are only for future use.
 

Michael V

Senior Member
High Vs Low Mosfet frustrations

Having trouble with mosfets i searched the forum, and this thread gave me some clues. My problem is not with the logic level thing, more on the behaviour of the mosfet depending whether it was attached to the + rail or the - rail.

I am building a 12V battery powered field data logger, measuring pressure. The pressure transducer is a 9-30V excited transducer which gives a 0-5 V output. With a 10K resistor and 100nf capacitor it mates with the picaxe ADC input nicely, and the output will nicely go via SEROUT to a SParkfun Logomatic SD card data logger. Compared with the picaxe the transducer is comparatively power hungry, demanding 10 mA. I only want to measure the pressure every 30 seconds or i minute, so i came up with the idea of only turning on the Transducer a few seconds before taking the reading, then turning it off again. Picaxe can handle that, right?

Well the standard Picaxe Mosfet (or Darlington transistor) circuit has the "source" (or emitter) to ground, which is on the "low side" of the load. The load is attached to the power providing voltage, i.e 12V. This didn't suit me. It's just a switch i thought, you turn it on with a signal from the picaxe. I have seen industrial proportional solenoids ( or linear actuators) switched this way, with the solenoid connected to ground and the PWM control to +24V. Should work!

At first i tried a MPSA13 Darlington transistor with Collector attached to + 12 battery voltage. ( The battery also supplies the picaxe via a 7802 regulator plus capacitors). Even going down to a 100R resistor on the base it wouldn't give me more than about 4V to power the transducer. I only want 10 mA, but need more than 9 volts.

Try a Mosfet I thought. I replaced the transistor with a 2N27000 200mA n-channel Mosfet. Small TO-92 package. Same result, 4V. So then i followed the picaxe manual and rejigged the circuit to have the mosfet "switch" on the low side. Great, i got my 12 V back. (mental note to read the manual) But when i connected the transducer it was being very silly. This is because the ADC on the Picaxe is measuring between it's own 5V and ground, and with the mosfet in there ithe transducer is not on ground any more.

Searching the net i learned about " High and low" sides - so THATS how they run those solenoid valves. Searching the forum i found this thread, but don't have the solution to switching the mosfet on the "high" side yet.

Looking at the attached diagram, the top sketch is the standard Picaxe Mosfet circuit which is good for motors, as per the manual. However, where i want to be is the bottom diagram, where the mosfet switch is on the plus side ( high side) , and the load is on ground.

How do i do it?

Is this what they use p- channel Mosfets for? I tried the Manufacturers data sheets, but all i found was special purpose "High Side Driver" chips. Must be a simpler way with a transistor or something.

Any help appreciated.

Michael
 

Attachments

BeanieBots

Moderator
All you want to do is switch it on and off so it can be simply done just with a another NPN transistor and two resistors.
Create a circuit as if you were going to use a transistor to switch an LED on/off from a 12v supply. Replace the LED/resistor with just a resistor.
Connect your FET gate to the collector/resistor junction.
PICAXE high will turn on first transistor which will then pull the FET gate low which in turn will turn on the FET. Hope that's clear and makes sense.
 

Dippy

Moderator
Using a P chan MOSFET is ideal for a high-sided switch as (nearly) in your lower drawing.
In this sort of application it nearly behaves like a PNP tranny.

Where you have your "what goes here" the easiest/cheapest/safest thing is to use a little NPN transistor, which when switched by your PICAXE takes the Gate pin to ground. Don't forget the base resistor. So PICAXE--Res---Base: Col-Gate : Em-Gnd. AND a (say >10Kohm) from the Gate to 12V. (You may like to put a small res 1k in between Coll-Gate).
When you do a PICAXE Basic HIGH this switches 'on' the bipolar which 'grounds' the MOSFET gate, switching it ON. When LOW the bipolar is 'off' allowing the gate to go high via resisotr and MOSFET is OFF.

But there are other methods too. And certainly for fast switching things have to be more complicated as in a Switcher/PWM circuit, but for this a simple approach is good enough and drivveling on about Gate-Source voltages will just confuse.

PS Snap!

PPS. BB's faster than hippy at typing!
 

Michael V

Senior Member
Like this?

High BB.

Do you mean like the attached sketch? Something like that would make sense, except the resistor values - not too sure. there would be a ratio thing in there, to include the transistor resistance. DO they look right?

Michael
 

Attachments

Dippy

Moderator
Almost. Change that 1K for 22K and you'll be there.
And there's no need for 47.
And some poeople wouid recommend a (say) 1K between coll and gate
 

BeanieBots

Moderator
That's exactly it. Perfect for simple on/off control.
I wonder who will be the first to use that circuit for PWM and then complain about the FET getting hot?
For PWM a proper bidirectional current capable driver must be used.
 

Dippy

Moderator
Ah, the zeners on the gate... to protect Gate from too high voltage .... a la Data Sheet 'Max G-S'. Nout to do with dumping gate charges. Oh DnT you worry me sometimes :)

Note to MichaelV. This is something completey different, so don't confuse your question with this image. This is a low sided driver, which can be simplifed hugely but I won't.

DnT : Post it on the thread YOU started as if it's like what you just posted it isn't appropriate to MichaelV's question and may induce unnecessary head-scratching.
 

Michael V

Senior Member
Thanks again

Hi guys,
Ive just wired it up, and after some initial confusion it works just fine. The confusion is that the mosfet is on when the pin is low. I guess you knew that, but i'm a beginner.

Thanks once again.

Regarding PWM control, As mentioned earlier i've seen 1 amp proportional solenoids for hydraulic circuits (technically actuators) driven with mosfet on the high side. Previous two posts are now confusing me. Could you not use this circuit for PMM control ? (bigger Mosfet of course)

And if you can tell me how to do current sensing in the same circuit to limit the power output and calibrate the solenoids (all slightly different resistance) then a $7 Picaxe and some mosfets can replace $1700 programmable controllers.

Michael
 

BeanieBots

Moderator
You,ve something wrong then!
A HIGH on the PICAXE should turn the FET ON.
The PICAXE O/P will put base current into the NPN base. It will then turn ON pulling the FET gate LOW with respect to the supply voltage. That should turn the FET on.

You can't use such a circuit for PWM because there is not enough current available from the 22k resistor to discharge the gate capacitance quickly. The result is a lot of time spent in linear mode which will result in a lot of heat dissipation. Not a problem when used just to switch on/off occaisionally but will build up very quickly when used for PWM.
 

Dippy

Moderator
"The confusion is that the mosfet is on when the pin is low. I guess you knew that, but i'm a beginner."

1. BB, maybe he meant the 'Gate' when he said "pin"? Crystal ball only working at 15% Duty Cycle today.
2. Michael, please always define what you are talking about as there are many darned 'pins' that go low or high or somewhere in the middle. Yes, ambiguity is a sign that you are a beginner.
... and one day you won't be a 'beginner' any more.
 

BeanieBots

Moderator
Good point Dippy. I'd made the ASSUMPTION that pin meant PICAXE OP pin.
Never heard gate LEG referred to as a PIN before.
I was a little confused by the statement that it was working but had an inversion issue. The only way I could see that happening was with an OC on the NPN emitter and Vload the same as PICAXE Vcc.
(positive earth configuration with 'redundant' NPN).
 

Dippy

Moderator
Well, I saw some blokes a while back admiring a filly and they said she had nice pins.

I love this:
http://en.wikipedia.org/wiki/Filly

Check out the text under the picture: "A Filly with her Dam Mother"
- how many times have I said that!! Hahaha.

Yes, pins, legs,leads etc. It's a funny old world!
 

BeanieBots

Moderator
Yep, commented on many a "filly pin" myself. Never thought I'd witness several blokes discussing the legs on an electronic component though!
Maybe THAT's the fascination with electronics. All those legs.
 

Michael V

Senior Member
Catching up

Just catching up. I'm pretty sure the mosfet opens when the pin is high, but i'll have another look tomorrow. Late now here in SYD. The thing with the filly and the pins, i don'tknow......
 

Dippy

Moderator
Michael , when you awaken tomorrow can you please clarify:-
"I'm pretty sure the mosfet opens when the pin is high.."

"opens" = MOSFET off? Like mechanical switch/relay terminology???
"pin" = PICAXE pin??? - or "pin" = Leg hanging out of MOSFET???

If the P chan MOSFET switches off when PICAXE pin is high then you have done something wrong - as described by BB about 6 posts ago..... If so recheck your circuit and make sure you've got your MOSFET the correct way round.

Fillys? Pins? You don't know? You ought to get out more mate! ;)
 

inglewoodpete

Senior Member
Michael_V is right. The MOSFET will turn off when the PICAXE output is high.

PICAXE High turns the NPN on. Collector is low, therefore gate is low and turns the MOSFET off.

PICAXE Low turns the NPN off. Collector and gate go high so turn the MOSFET turns on.

The NPN is working as an inverter/buffer/level shifter.

This could be a problem at bootup, since the output of the PICAXE will be low, turning the MOSFET on. Make the first command at startup "High {pin}". If the initial 'on' flick is a problem, you may need an RC delay on the base of the NPN.
 

Michael V

Senior Member
Output pin low - Mosfet passes current.

Thanks pete, i was doubting my sanity for a while there.

Sorry for my dodgy terminology, i'm learning every day.

The circuit is definitely as per my sketch, with the 1K and 22K resistors in place. Got out my trusty multimeter and put in some simple on off code. Definitely when the output pin is high the mosfet is not passing current, and when the output pin is low it is passing current.

so the command " low 1" turns the power on to my transducer, which is firmly connected to ground and reading correctly. The drain of the 2N7000 n channel mosfet ( the top part on the circular mosfet symbol, the right hand side when looking at the flat side legs down) is connected to +12.

For this application, initial switching on the power to a transducer, it is not a problem that it starts high for a few milliseconds, but i can see that the "flicker" might be a problem for some. The RC delay tip is a good one.

Given this "inversion" are BB's comments about using this for PWM and the current through the resistor still correct? Thinking further, that these little mosfets are less than $1 from Jaycar, the mosfet could be used in the place of the transistor. That may help the current get away quicker, and make the switching quicker. (sorry if that terminology is wrong). If so This would be a great simple circuit for PWM proportional solenoid control ( 1 amp max, 24V) - and one that would benefit from the RC delay.

Yes, i need to get out more, too many of those 8 and 14 pin fillies and not enough of the two legged kind.

Michael
 

inglewoodpete

Senior Member
The circuit provided is sufficient to drive a 2N7000 for on-off type switching or PWM*. It can also drive a power MOSFET for simple switching. *2N7000 has a typical gate capacitance of 20pF.

However, PWM of a power MOSFET introduces some problems. Due to the relatively large capacitance of the MOSFET gate, typically 1220pf (1.22nF) for an IRF540N, the 22k resistor will turn the 'on' edge of pulses into a (relatively) slow ramp. As discussed in earlier posts, slow turn-on equals heat. Many slow turn-ons, as is the case of PWM, causes lots of heat. The NPN, having a comparatively low 'on' resistance will cause the the MOSFET to turn off quickly, which is fine.

A low impedance driver is required to operate a large MOSFET in PWM mode. The solution is to use a PNP/NPN complimentary pair with a 3rd transistor (NPN) to shift the voltage for the 'upper' transistor or the pair. The curcuit would look a bit like the output of an audio amplifier, without all the feedback circuitry.
 

hax

New Member
Maybe the brains trust here could figure out a nice circuit to do this. As I am sure many would be interested in a solution, including me.
 

Dippy

Moderator
Hang on, hang on.... oh I'm going mad.

Right, first, are we still talking about your final hand-drawn sketch with a P Chan MOSFET and a bipolar transistor 'driving' it?

If so, PICAXE High turns the NPN on (yes). Collector is low(yes), therefore gate is low(yes) and turns the MOSFET off (No, it goes ON for a P Channel).

Just get the P (I repeat P) chan MOSFET.
Stick its Source (pin 3 usually for a TO220/TO251) onto your 12V supply, put it's Drain (middle leg pin 2) out to an LED+2k2 resistor.
Then try a resistor say 10K whatever between Gate (pin 1) and Source (pin3 / power supply). LED OFF.
Now stick the res between Gate and Gnd > LED ON.

VERY similar to a PNP behaviour.

The reason I am 100% sure is that I'm doing a circuit right now where part of it is a smple MOSFET P high sided switch to provide power to another part of the circuit - and guess what? It works.
And as my Autorouter is not clever enough I'll have to do it manually, so I'll leave you to it.

DnT posted a driver circuit for PWM driving on his thread recently I believe.

Here is a dead simple on-off P chan high sided switch. When PICAXE O/P=1 then MOSFET switches ON - so easy it rhymes:-
(This is only for slow 'power supply' switching, for fast switching you need a proper driver)
 

Attachments

Dippy

Moderator
Also, 2 wobbly photos of breadboard.
1. Gate to ground = MOSFET ON
2. Gate to Vss/V+ = MOSFET OFF

Your NPN does the swiching of Gate to Ground . So Gate goes to ground when NPN=ON i.e. PICAXE O/P=ON (High).
And now I'm puffed out...
sorry they are a bit blurred, it was a bit gloomy i.e. slow shutter.
And very grubby 15 year old breadboard.
 

Attachments

inglewoodpete

Senior Member
Steady bro'
You're right for a P channel MOSFET, but the hand drawn diagram has an N-Channel MOSFET (Arrow pointing inwards).

I agree with you: normally I'd use a P-channel MOSFET for a high-side switch, but that's something different what was being discussed (I hope).
 

Dippy

Moderator
You're quite right Pete, I wasn't watching his arrows. I had leapt before I looked and (wrongly) assumed we were talking about Ps.


Anyway, Michael, just get yourself a P chan and all will be happy. I'm exhausted now... and hallucinating badly.
 

Michael V

Senior Member
Sorry guys,
didn't mean to create friction.

I'm a beginner, and there is only n channel in the manual, and i started with that. I didn't really know what a p channel was for and i was guessing , sorry to confuse. My sketch was not that clear, but the arrow is pointing left. That little mosfet is a great compact size (and price) for switching up to 200 mA. The p Channel mosfets seem to be not common.

I can live with the initial low. Future design i will go with a P channel in your config, especially for you Dippy because you are a nice bloke.

Pete, if you can sketch up the " complimentary pair with third transistor" as Haxby says, a few of us could learn. Today i learned that the controller with eight proportional outputs for solenoid control is going to cost $2100, plus $2400 in programming. I thought about duplicating the PWM outputs in this applications with Picaxe 14ms or even 8ms, one for each channel. But the FETS would be on the high side, and after all this discussion on heating mosfets and slow switching i don't want to try it.

Thanks again guys.
 

inglewoodpete

Senior Member
Dippy, You're not completely bonkers (yet). I'm now home from work and was having a think about the whole thing on the busride home.

Michael_V's circuit will not work very well at high (ie significant) currents. A high-side N-channel MOSFET will need a gate supply that is more positive than the positive rail.

Ie for an N-channel MOSFET the gate must be 5v more +ve than the source. If the MOSFET is hard-on, the source would be milliohms to the drain and therefore the +ve rail, which would require the gate to be 5v higher than that.

The practical options are to use a P-channel MOSFET as Dippy discussed, or an N-channel MOSFET with a special supply at least 5v higher than the 'normal' +ve rail to drive the gate.

I got a bit confused between BJTs and MOSFETs. There are similarities and there are differences.
 
Top