# Solenoids, Transistors and MOSFETs

Im looking at these 2 projects for controlling solenoids:

This one uses a TIP120 transistor and powers both the board and the solenoid from two 9V batteries.

This one uses a MOSFET and powers the valve from a 12V lead acid battery.

They are both transistors but Im interested in this heat dissipation talk.

I’ve borrowed the diagram from the first project and attached it here. From what I understand, the mosfet is just a transistor with a heat sink attached to it? So is this the reason why the second project uses a mosfet instead of a regular transistor like tip120?

1) 9V battery won't power a solenoid well/at all/for long, unless its a tiny tiny solenoid.

2) TIP120 is a darlington pair, strictly speaking, ie two BJTs in one package.

3)

From what I understand, the mosfet is just a transistor with a heat sink attached to it?

No, just plain wrong. A MOSFET is a MOSFET, its not a BJT. Either can have heatsinks or not.

The second project probably uses a MOSFET because they are much much better than BJTs and noone uses BJTs for power since 20 or 30 years ago... Typically you wouldn't have any need of a heatsink for driving a few amps through a solenoid.

With a darlington you'd need a heatsink, with a BJT you'd need something to provide enough current into its base (for switching a BJT has a gain of 5 to 20 only, darlingtons about 1000, MOSFETs are voltage driven, not current driven).

In that MOSFET circuit it is vital a logic-level MOSFET is used. You can calculate the power dissipation as I-squared-R where R is the on-resistance of the MOSFET. For instance a 5 milliohm MOSFET with 2A flowing dissipates 20mW when on. A darlington about 2 to 3W when carrying 2A since darlingtons always waste a volt or two.

If you do use a MOSFET, its still a good idea use a gate resistor (200-330 Ohm) and a pulldown resistor from gate to ground (10-20k Ohm). The second instructible shows neither.

The gate resistor to limit the current at the arduino pin as the mosfet gate charges /discharges. The pulldown is so the MOSFET doesn't turn on briefly when powering up the arduino.

I know one 9v battery wont run the 12V solenoid, I think that is why they used 2. Even so, Im sure two 9v batteries might run it for just a little while.

So back to the mosfet and bjt.

I understand one difference is that bjt's are current driven whereas mosfets are voltage driven.

MOSFET - Voltage driven, thus doesnt waste current at the Gate, which means less overheating? - Doesnt require a resistor in series with the gate, why? - Does require a pulldown resistor so the 'gate doesnt float on reboot'. Why? - Better for high power applications because switching is faster. Again, why?

BJTs - Are better for low power applications like LEDs?

I guess I want the underlying explanation of why mosfets are faster and waste less current and why that makes them better for high power applications?

Marciokoko: I know one 9v battery wont run the 12V solenoid, I think that is why they used 2. Even so, Im sure two 9v batteries might run it for just a little while.

I doubt it, you have been told, you will learn the hard way I suspect.

So back to the mosfet and bjt.

I understand one difference is that bjt's are current driven whereas mosfets are voltage driven.

MOSFET - Voltage driven, thus doesnt waste current at the Gate, which means less overheating?

The main reasons are that the Vsat can be a low lower, and switching losses are much less.

• Doesnt require a resistor in series with the gate, why?

Not that simple - the gate is a large value capacitor, effectively, so if the chip driving it can't handle large current pulses a resistor may be needed to prevent overload. If driving several MOSFET gates in parallel resistors may be needed to prevent/damp an oscillation mode. Sometimes resistors are added to slow down the switching somewhat to reduce EMI generation.

• Does require a pulldown resistor so the 'gate doesnt float on reboot'. Why?

Because the gate is a capacitor, it can retain its charge from last time and turn on unexpectedly, or be half-on-half-off and burn up.

• Better for high power applications because switching is faster. Again, why?

Because its a majority carrier device, unlike a BJT which uses minority carriers and is dependent on carrier-recombination to turn off quickly (but fast recombination means a lower gain device)

BJTs - Are better for low power applications like LEDs?

No, but they may be cheaper in through-hole package.

I guess I want the underlying explanation of why mosfets are faster and waste less current and why that makes them better for high power applications?

They are more efficient, much faster at high power, don't suffer from secondary breakdown, need less drive power, however they are substantially less rugged - overvoltage on the gate will damage/destroy the device very easily. Typically you use a MOSFET driver chip to provide protection from this as well as allowing the fastest switching.

Ok I’ve got access to a 12V lead acid battery.

So I dont have the FQP-50N06L N-mofset but I have:

1. K2996
2. IRF510
3. 2N7000
4. TIP120 (BJT)

Could someone guide me through what I need and why I need that? I understand the Gate-Source and Drain-Source voltages are important.

I have 2 valve options:

1. Current: 2.5A/12VDC

2. Im looking for the other specs buts its a small solar water heater valve. Its 12VDC but I dont know the amperage

3. How could I find the amperage on my other solenoid?

4. What specs should I look for in the K2996, IRF510 & 2N7000 mosfets and why?

Thanks

Im thinking my Vds needs to be the voltage that I need to drive my load, which if its my vdc solenoid then its 12V min.

FQP-50N06L=Vds of 60V K2996 = Vds of 600V IRF510 = Vds of 100V 2N7000= Vds of 60V

The Vgs is the voltage of the signal from the arduino to the gate, so since it needs to be a logic level mosfet:

FQP-50N06L=Vgs of 20V K2996 = Vgs of 2-4V? IRF510 = Vgs of 2-4V? 2N7000= Vgs of 20V

Thats weird because the FQP and the 2N7000 dont seem to be logic level mosfets due to their high Vgs? Why was the FQP suggested in that tutorial?

BTW, the amperage on my other 12VDC solenoid might be 350mA

Clarification about gate voltages in datasheets:

There is the Vgs maximum rating - the absolute maximum voltage range on the gate, with respect to the source, before the device fails. You don't even go near this. For standard MOSFETs its about +/-30V, for logic level +/-16V is more typical.

There is the Vthr (threshold voltage) - this is all about the point of turn-off of the device, nothing to do with being on, all about ensuring it is off. Ignore this, but expect it to be about 30 to 40% of the working gate on-voltage.

The key thing in the datasheet is the on-resistance rating, which is quoted alongside the Vgs that guarantees it. Something like "Rds(on) = 0.01 ohms @ Vgs=4.5V"

That's the thing you need - it means 4.5V gate drive will give an on-resistance of 0.01 ohms.

It doesn't mean you can use the thing with 4V of gate drive, it means you want a minimum of 4.5V and you can rely on it working. Most 5V logic-level MOSFETs quote the resistance at 4.5V of drive. Most others use 10V (typically you'd drive these at 12V, not 10)

Smaller surface mount MOSFETs can have lower drive voltages for use with 3.3V logic for instance, but that's not normal for TO220 sized packages.

Ok so here are the values for all options:

Ok so if I supply 10V at the gate Ill get these Amperage values because the resistance will drop to those values.

They are all 10V values. So I need to determine which is best suited for my needs with the solenoid.

QUESTIONS

#1
2N7000:
Seems inadequate at least for the 12VDC/2.5A solenoid because it can optimally supply 500mA. This might be best suited for a smaller solenoid like the 350mA I might have.

#2
FQP:
The FQP seems to be able to provide quite a large amperage. If the valve is rated at 2.5A, does it mean it cant draw any more than that? Because what happens if it wants to draw 5A and the mosfet lets it, it might overheat and burn, no?

#3
K2996 and IRF510:
These have large Vds values but that just means that they can handle larger voltages so they might be able to drive larger loads, right?

#4
But what about the Vgs of 20V vs 2-4V?

The datasheets for the irf510 and the 2sk996 say a Vgs of 20 and 30 respectively. 2-4 is the gate threshold voltage range.

The mosfets actual current capacity in reality will be limited by the heat to be dissipated.

Lets look at the irf510. At 2 amps the mosfet will dissipating 2x2x0.6. 2.4 watts of power. A heatsink of less than 20c/W would be needed to stay under 70 degrees, Also the rdsOn will rise as the heat does too making matters worse.

A larger Vds just allows a higher operating voltage. Their current is still limited the same way. Less current is required at higher voltages for the same power. Choose your mosfet for your application. A 100v mosfet or 600v is not the right choice for a 12v application. A 30v low rdsOn mosfet would likely perform much better and handle a higher current.

FQP:
The FQP seems to be able to provide quite a large amperage. If the valve is rated at 2.5A, does it mean it cant draw any more than that? Because what happens if it wants to draw 5A and the mosfet lets it, it might overheat and burn, no?

Yes, the valve draws 2.5A at its rated voltage, it can’t magically draw more (unless you short out its
terminals, in which case the fuse on your 12V supply will fail (you do know a fuse is necessary for high
current batteries like lead acid and LiPo?)

The FQP30N06L is rated at 0.035 ohms at 5V drive, completely fine, that’s the only device of the four
that is suitable. The IRF510 is not logic level, 2N7000 hasn’t anything like a plausible specification and
the 2SK2996 is a 600V part, which is why the on-resistance is so huge.

Make sure the free-wheel diode is rated for 2.5A pulses.

Ok I can see IRF510 only has a 10V Vgs for RDSon, so its not logic level and K2996 is also only 10V.

But 2N7000 has logic level RDSon at 4.5 and 5V.

FQP is the best, I know, but I dont have one of those

And yes I have a diode and fuses.

Ok so the IRF510 will not turn on all the way at 5V. Since that is what I have right now, (I just ordered an IRZL44 and FQP), does it mean the valve might open slower?

What might I expect from operating the valve with that mosfet?

Marciokoko: Ok so the IRF510 will not turn on all the way at 5V. Since that is what I have right now, (I just ordered an IRZL44 and FQP), does it mean the valve might open slower?

What might I expect from operating the valve with that mosfet?

The irf510 will not handle 2.5 amps at that gate voltage, It will get very hot and fail. Even at 10v it would need a large heat sink.

The irlz44 is a logic level fet will be dissipating less than quarter a watt. no problems.

Thanks alka. I actually just posted this about wanting to find a good tutorial that shows how mosfet and bjt's operate (http://forum.arduino.cc/index.php?topic=398004.0).

Thanks for being patient. When you say the irf510 can't handle 2.5A at the gate, why would it have to? The valve is rated at 2.5A but that would be a voltage between drain and source, not at the gate.

Marciokoko:
Thanks for being patient. When you say the irf510 can’t handle 2.5A at the gate, why would it have to? The valve is rated at 2.5A but that would be a voltage between drain and source, not at the gate.

The gate is not part of the current path. 5 volts gate drive will not allow the mosfet to turn on enough to allow that current flow between drain and source.

I should have been a little more clear with my wording.

OK that I understand because Rdson is listed at 10v. So how much would flow at 5v? Its not just V=IR, right, because it's not a linear relationship.

Take a look at this from the irf510 datasheet, Looks like about 1 amp max if I am reading this correctly.

Heat might limit you to well below 1 amp.

If the Rds(on) is not listed at 5V, you don't use 5V - no behaviour is guaranteed by the datasheet. Note that that graph is "typical behaviour", which means actual devices differ a lot, and the behaviour also changes with age due to ion migration.

The electric field across the gate oxide is a large proportion of the breakdown voltage for the oxide, it is under high stress and the properties will change with age. The manufacturer gives you values that are tested as being able to work even after a long working life. All of the graphs and values that are "typical" related to gate voltages will have variability between devices as well as with age.

Ok sorry, that was a silly question. I had already seen those graphs when I asked it. Ok so I would get 1A out of it, I guess the question would be, how well would the valve work with 1A. I guess Ill have to try it.

Thanks guys!