Beginner: why a MOSFET?

[nyoo]

If I may, could I please revisit a beginner's question from 12.Mar.2008?

Please find attached sketches of two circuits, before and after. The "after" picture encorporates changes suggested by on1aag, which:
1. changes the "after" circuit to stop the pump at temperatures above something set by P1 (140 degrees F, in my case);
2. smooths hysteresis; and
3. uses an IRF540 (or similar) MOSFET to switch on/off a 12VDC 500mA pump, where a small-signal 2N2222 transistor had been used previously to switch on/off a 500mA relay.

As a novice, I took all this as wisdom from my elders; that is, as a cookbook recipe. Now, making a second pass but still as a novice, I'm trying to understand why the MOSFET was chosen.

There seem any number of "power" transistors, like the TIP31A, that could carry the required current. Some textbook suggested that the transistor loses voltage. Was the MOSFET chosen because it would maintain the 12VDC from drain to source? Or is there some obvious difference between relays and pumps that makes the MOSFET more attractive? All things being equal, transistors seem cheaper, less prone to ESD damage, and easier to acquire.

From scratching my head over the textbooks, I'd like to ask one other question too, if I could, please. In the "before" picture, the books tell me the size of the R5 resistor is chosen relative to the minimum hfe (or beta, or DC current gain) of the transistor. But on the datasheets, the hfe never shows for the voltages I'm using, 5V or 12V. They'll show hfe for 100mA and 1V, they show for 1A and 4V, but never for 500mA at 12V. How do I do that conversion?

Thanks very much.

 

[Hero999]

Not only does a MOSFET loose less voltage, it also doesn't require any gate current to turn on fully. Generally transistors require a base current of 1/10 of the load current to turn on fully.

MOSFETs are normally only used at higher power levels which is why you don't see that many small ones.

In the first example a MOSFET isn't really advantagous as the relay is only small. In the second example a MOSFET is more worthwhile as the load current is much greater and the op-amp (which can only supply 20mA at most) will struggle to fully saturate a transistor, unless it's a darlington.

 

[walshlg]

thanks fromt he lurkers club!

 

[DigiTan]

From scratching my head over the textbooks, I'd like to ask one other question too, if I could, please. In the "before" picture, the books tell me the size of the R5 resistor is chosen relative to the minimum hfe (or beta, or DC current gain) of the transistor. But on the datasheets, the hfe never shows for the voltages I'm using, 5V or 12V. They'll show hfe for 100mA and 1V, they show for 1A and 4V, but never for 500mA at 12V. How do I do that conversion?

Thanks very much.

Since you want all 12 Volts on the relay, you actually want the Vce to be as small as possible. This happens when the transistor is conducting as well as possible (or at least close). For a 2N2222, the smallest Vce you'll get is between 0.3 and 0.2 Volts. Since your supply voltage is 12 Volts, the voltage on the relay will be 11.7 to 11.8 volts after the power settles.

On Semiconductor's datasheet is a little more detailed than most. There's a hfe vs. Ic graph on page 3 that you can use as a reference. At room temperature and 500mA current going through the transistor and relay, it says you're looking at an hfe for about 50.

But say that graph wasn't there. If that were the case, you have to take the data points they give you and do a non-linear regression to find the hfe values in between. There's also (very expensive) software that can reverse-engineer this for you. But yeah, it's definitely something the book don't tell you.

 

[crutschow]

All things being equal, transistors seem cheaper, less prone to ESD damage, and easier to acquire.

A small clarification. They are both transistors. One is a Bipolar Junction Transistor (BJT) and the other is a Field Effect Transistor (FET).

MOSFETs (Metal Oxide Semiconductor FET) are the most common FETs and have a high impedance gate that looks like a small capacitor.

Less common are JFETs (for Junction FET) which have a gate that looks like a reverse biased diode.

CMOS circuits (for complementary MOS) uses both n-channel and p-channel MOSFETs in the same circuit to achieve high speed and minimize power consumption.

 

[RadioRon]

One of the reasons that MOSFETs are favored at higher power levels is the way they behave when they heat up and how that is better than a bipolar transistor. Since transistors are not ideal they get hot when passing a lot of current. When a bipolar transistor gets hot, its Vbe decreases (referred to as a negative temperature coefficient device) and in many biasing situations this encourages more collector current to flow which tends to add even more power dissipation and adds to heating the transistor even more. This increased heating builds on itself and can cause temperature to increase out of control, something we call thermal runaway, where the transistor breaks down. The MOSFET, on the other hand, has a positive temperature coefficient. That is, as it heats up, its drain impedance increases and this tends to decrease the current in typical circuits. So there is low risk of thermal runaway in the MOSFET unless it is configured as a constant current source. This is another reason the MOSFET is popular in power circuits.

In the circuit in question, the amount of power dissipation in the transistor is low so this MOSFET advantage is not important.

 

[Leftyretro]

One of the reasons that MOSFETs are favored at higher power levels is the way they behave when they heat up and how that is better than a bipolar transistor. Since transistors are not ideal they get hot when passing a lot of current. When a bipolar transistor gets hot, its Vbe decreases and in many biasing situations this encourages more collector current to flow which tends to add even more power dissipation and adds to heating the transistor even more. We call this a positive temperature coefficient. This positive feedback can cause temperature to increase out of control, something we call thermal runaway, where the transistor breaks down. The MOSFET, on the other hand, has a negative temperature coefficient. That is, as it heats up, its drain impedance increases and this tends to decrease the current. So there is no thermal runaway in the MOSFET. This is another reason the MOSFET is popular in power circuits.

In the circuit in question, the amount of power dissipation in the transistor is low so this MOSFET advantage is not important.

You have the cause and effect correct, but I think you have positive and negative coefficient reversed for the device types. MOSFET exhibits negative TC while bipolar have positive TC, at least that's my interpretation on it.

https://en.wikipedia.org/wiki/Positive_temperature_coefficient

Lefty

 

[RadioRon]

You have the cause and effect correct, but I think you have positive and negative coefficient reversed for the device types. MOSFET exhibits negative TC while bipolar have positive TC, at least that's my interpretation on it.

https://en.wikipedia.org/wiki/Positive_temperature_coefficient

Lefty

How embarassing. Yes, you are correct, the industry defines the temperature coefficient of a MOSFET as Positive in that the increase in temperature causes an increase in on-state drain resistance. I was thinking of the effect of that increasing resistance as a negative thermal feedback in a typical circuit. Anyways, thanks for catching that, and I have edited my post so as not to confuse future readers.

 

[picbits]

Never overlook the obvious - a 30A mosfet can be had for as little as £0.30 in single quantities while a relay + transistor will set you back a fair bit more than that.

You can also use PWM to vary the speed of the motor with the MOSFET which you wouldn't be able to do with the Relay.

Anyone else think the 1N4148 would be a little underrated to provide protection for a 500ma 12v motor though ?

 

[Leftyretro]

Never overlook the obvious - a 30A mosfet can be had for as little as £0.30 in single quantities while a relay + transistor will set you back a fair bit more than that.

You can also use PWM to vary the speed of the motor with the MOSFET which you wouldn't be able to do with the Relay.

Anyone else think the 1N4148 would be a little underrated to provide protection for a 500ma 12v motor though ?

Yea a 1N4004 is cheap and easy to get and would be a better choice.

Lefty

 

[nyoo]

Thanks for responses that are pretty clear, even to a beginner.

For the 1N4148 vs. 1N4004 discussion, I'm again poring over the datasheets and wondering what's the critical characteristic for "underrated". Is it Io, Average Rectified Forward Current, which is 150mA for the 1N4148 and 1A for the 1N4004? If so, would any diode do, whose Io was greater than 500 mA?

About the MOSFET. The pump motor says it will switch on at 8V (and run weakly), but it runs best at 12V. Did I hear that the MOSFET will allow all 12 volts will make it from the power supply through the pump?

Please could someone show me how the output from the LM741 is limited to 20 mA?

I was surprised that this circuit has no capacitors. But as I know so little, I didn't want to mess. Are they unnecessary here? If they would be useful, do I go with the rule of thumb "capacitor voltage should be twice the supply voltage" or, in this case, 25V?

 

[Leftyretro]

Thanks for responses that are pretty clear, even to a beginner.

For the 1N4148 vs. 1N4004 discussion, I'm again poring over the datasheets and wondering what's the critical characteristic for "underrated". Is it Io, Average Rectified Forward Current, which is 150mA for the 1N4148 and 1A for the 1N4004? If so, would any diode do, whose Io was greater than 500 mA?

About the MOSFET. The pump motor says it will switch on at 8V (and run weakly), but it runs best at 12V. Did I hear that the MOSFET will allow all 12 volts will make it from the power supply through the pump?

Please could someone show me how the output from the LM741 is limited to 20 mA?

I was surprised that this circuit has no capacitors. But as I know so little, I didn't want to mess. Are they unnecessary here? If they would be useful, do I go with the rule of thumb "capacitor voltage should be twice the supply voltage" or, in this case, 25V?

I think the key specification for a diode used in suppression of inductive kick back voltage is it's voltage rating, not necessary it's Io current rating. Inductive spikes are short transient events that don't generate sustained current levels, but PIV excursions can quickly damage a semi-conductor.

Lefty

 

[Pommie]

I think the key specification for a diode used in suppression of inductive kick back voltage is it's voltage rating, not necessary it's Io current rating. Inductive spikes are short transient events that don't generate sustained current levels, but PIV excursions can quickly damage a semi-conductor.

Lefty

Isn't the largest reverse voltage it will see 12V?

Mike.

 

[Leftyretro]

Isn't the largest reverse voltage it will see 12V?

Mike.

No, the collapsing magnetic field causes a much larger EMF then just the applied voltage. The rule of thump I was taught was to expect 4X the applied voltage, but I think it's a rate of change thing and the amount of induction, so I suspect there is a formula that can give a more accurate estimate. Anyway working with 24vdc industrial relays I was able to see usec pules of over 100v on a scope when checking once. It was interesting to notice that the protection diode caused the drop out time of the relay to increase and the sound of drop off was 'softer' then without the protection relay.


Lefty

 

[Pommie]

No, the collapsing magnetic field causes a much larger EMF then just the applied voltage. The rule of thump I was taught was to expect 4X the applied voltage, but I think it's a rate of change thing and the amount of induction, so I suspect there is a formula that can give a more accurate estimate. Anyway working with 24vdc industrial relays I was able to see usec pules of over 100v on a scope when checking once. It was interesting to notice that the protection diode caused the drop out time of the relay to increase and the sound of drop off was 'softer' then without the protection relay.


Lefty

Isn't this pulse going through the diode in a forward direction?

Mike.

 

[RadioRon]

"The rule of thump... "
Lefty

a useful simplification learned with the help of physical persuasion?

 

[crutschow]

Isn't this pulse going through the diode in a forward direction?

Yes. The diode conducts in the forward direction to provide a path for the inductive current until the inductive energy is dissipated. As such the maximum voltage the diode sees in the reverse direction is the supply voltage across the inductor.

The softer sound of the relay and longer drop out time as noted by Leftyretro are because it takes longer for the inductive energy to dissipate with the diode across the inductor. If the drop out time is too long it can be reduced by putting a small resistor or zener diode in series with the diode to dissipate the energy faster.

 

[spuffock]

If using PWM to control motor speed, a 1N4004 will take its time to turn off when the mosfet switches on. Due to its still conducting with the full rail voltage across it in reverse, it can easily dissipate enough power to destroy itself, in conditions where a 1N4148 would survive, due to its higher speed.
A better bet is the UF4004, a much faster 1A diode. Consider 1N4004 as suitable only for 50/60 Hz rectification, and UF4004 for anything faster.

 

[aljamri]

But using relay gives you the following advantages:

Physical isolation between control and load circuits
More than one load ( depends on sets of contacts )
Normally Open / Close options
Different supply for load than that for control circuit

 

[Leftyretro]

But using relay gives you the following advantages:

Physical isolation between control and load circuits
More than one load ( depends on sets of contacts )
Normally Open / Close options
Different supply for load than that for control circuit

I agree, a relay can be a very useful and effective component and not likely to be totally replaced any time soon. I recently purchased some single coil 5volt DPDT latching relays on E-bay to play with. Latching relays are a little off the beaten trail, but they can be very useful for things like auto power off circuits and can be useful as an external memory function.

Anyway I like relays, a lot

**broken link removed**

Lefty