MOSFET as switch and power monitor

Hi all!

I'm designing a circuit to power some heating bands that go around the telescope to avoid dew.
My plan is to use a MOSFET to enable one or more channels, so that I can decide which bands have to be powered.
Power is drawn from a 12V battery and the Arduino activates and deactivates the heaters, depending on the detected temperature. I won't be using PWM, I just need to turn on and off the heaters for 5-10 minutes. Each band draws about 3A.

I know that I have to use a P-channel device for this application, but which one in particular? (eg are the IRF9Z24N, the IRF9Z34N or the FQP27P06
any good?) What considerations should I do in order to pick up the best option?

I also have another question. What can I do to monitor the power usage? I tried with a shunt resistor, and it works this way, but I would like to monitor the power usage from the battery not only for the heaters, but also for the rest (mount -3A-, computer -5A- and camera - 3A). Considering that, combined, the draw would be more than 10A, I think that the shunt resistor is just a waste of energy... what can I use instead?

Thanks for helping me!

Stefano

I think that the shunt resistor is just a waste of energy

You don't want a p channel you want a N channel MOSFET Like this one COM-10213 http://dlnmh9ip6v2uc.cloudfront.net/datasheets/Components/General/FQP30N06L.pdf

And you don't use a shunt resistor that waste amps you use a small one .001 ohm I wouldn't buy one I would make it with a piece of wire

Thanks for the reply.
My understanding is that the MOSFET has to go on the positive side, to avoid leaving the electronics "hot". That's why I was considering a P channel. At home I have some IRF540, that are N-channel power MOSFETs.

Considering the current that will flow in my circuit, is it better to use a Hall effect sensing device? I found some LEM sensors (even if they are way more expensive than a shunt - but they should be safer and more power efficient) that may be a solution, but I'm not 100% sure how to use them.

What is wrong with a shunt resistor? If you used the resistor suggested by be80be, how much power waste will you have? Would the device you want to use in it's place use more or less power?

E = IR
E = 10A x .001 Ohm
E = .01v

P = IE
P = 10A x .01v
P = .1w

Power flows from negative to positive

negative charged electrons move toward the positive

So you are cutting off the Power using a N channel

And a a 1 inch piece of 18 gauge wire would read 10 Millivolts @12 volts 10 amp and and put out 100 Milliwatts of heat I'm sure it's not going to get hot you'll have to scale it with a opamp.

I have a DVM it's not a cheapo and it reads across a 3 inch piece of wire it's 14 gauge

Chrono251:
Thanks for the reply.
My understanding is that the MOSFET has to go on the positive side, to avoid leaving the electronics "hot". That's why I was considering a P channel. At home I have some IRF540, that are N-channel power MOSFETs.

'Hot" is only relevant if you've got other devices around you which are grounded and there's a danger of touching both with something conductive. A 12V battery on a hill probably isn't much risk.

Either way will work, but with a P-type you'll need to be able to switch the gate of the MOSFET between GND and +12V, which the Arduino can't do directly. You'll have to add a pullup resistor to 12V on the gate of the MOSFET ("off") and another smaller transistor between MOSFET gate and ground for the Arduino to be able to switch the MOSFET on.

@ what fungus said your power loss is more with the P channel then the N channel.
And like fungus said you can't use just a P channel MOSFET you have to add a driver.

be80be:
@ fungus And your power loss is more with the P channel then the N channel.

Sorry, I didn't mean a P-channel is the way to do it...I mostly replied because of the 'hot' thing. A floating 12V battery connection isn't going to electrocute anybody.

fungus I was just adding to what you where saying. In all reality A relay be it Solid State or mechanical. Would give you better heating.

@ what fungus said your power loss is more with the P channel then the N channel.

Only if you select the incorrect P channel mosfet.

My choice:
http://www.fairchildsemi.com/ds/FQ/FQP47P06.pdf
Rds = 0.026 Ohms @ -10v Vgs

Your choice:

Rds = 0.035 Ohms @ 10v Vgs

Both require 10v gate to operate at that Rds, so you really need a driver for both.

2 quick things. 1 copper wire has a positive temperature co-efficient that is when it gets hot it's resistance increases which means that while it "Looks" like the current is going up it is actually going down because of the increase in resistance of the copper. It's all right to use copper for connecting things as the rules that that into consideration but not to measure current with because of it's 3300 PPM positive temp co..
And 2. There are many "Logic Level" mosfets available, using a driver and a higher voltage Vgs mosfet is kind of redundant.

Bob
{Edit RKJ}

My previous post was somewhat confusing.. So
OTOH there are 2 common kinds of drivers both open collector. An NPN transistor is used to bias a P ch Mosfet on and a PNP transistor is used to bias an N ch Mosfet. Or both or either as required.
There are also hybrid IC's that are used to drive mosfets where there are no sources of voltages greater than 5 to 10 volts as some high power Mosfets require.
The 'common drivers' are trivial to implement and the 'Driver' IC's do require some special handling and this was the reason for my comment on redundancy.

Bob

The need for high-side switching is when controlling another circuit - not necessary for a light or a heater.

For current measuring a hall-effect current sensor is the way to go, there are various types, probably the simplest
if you can handle surface mount components is the ACS710/711/712 family which are available in various current
ranges. Being fully isolated you can use a hall-sensor to measure current anywhere in the circuit.

There are other through hole hall sensor modules available, there are resistive shunt modules available too.

Thanks all for the answers!

My idea for hot side switching is related to the fact that the battery won't be operating inside a building, but outside where dew may be present. I understand that nobody will die for touching a wire with a 12V battery (fuses are present anyway), but I would like to make the device as safe as possible. If I use a switch on the hot side, I have the feeling that the risk for short circuits is reduced as the only way to create one is to touch the PCB. The battery doesn't power just the heaters, there is also the telescope mount, the laptop and the camera that get power from there.
If you say that a N-channel is ok, though, I trust that and that's better for me as at home I only have some N-channel MOSFETs (IRF540).

For the schematic, I was thinking something like the one attached to this post (possibly using the FQP47P06). Will it work this way?

Using a 1 mOhm shunt resistor won't waste much power, but that'd give me a max voltage drop of 15mV at 15A. I can connect that to an OPAMP (like I did in my tests on breadboard - not using the heater but a bunch of resistors and a bigger shunt), but I guess that the readings won't be very accurate. I don't need perfect readings, but something that makes sense and doesn't change much with temperature. That's why I was asking if Hall fx devices are valid for this. I also like the idea of having the sensing circuit isolated from the other one (but an optocoupler may be used for this).

MOSFET-schem.jpg

That schematic looks good to me, however I would decrease the 10K resistor to 1K. For current measurement, I suggest a Hall current sensor such as http://uk.farnell.com/allegro-microsystems/acs712elctr-20a-t/sensor-current-20a-soic8-712/dp/1329624. Alternatively, use an axial leaded resistor of about 0.02 ohm such as http://uk.farnell.com/vishay-dale/lvr05r0200fe73/resistor-precision-0-02-ohm-1/dp/1108083 and an op amp. The trick to getting good accuracy with a low value resistor is to solder separate sense wires on to the resistor leads, between the resistor body and where you connect the power. The sense wires feed the op amp. That way, the resistance of the connections doesn't affect the accuracy.