[Question] Why do not use FET to active coil

[vantusaonho]

Hi everybody!
I have a question as: why do not we use a FET to active a coil (Relay coil)?
As normal, we use a transitor to do that but I wonder why don't use a FET
Thanks

[borberk]

What kind of FET you have in mind?

[metal]

Why do you want to use FET any way?

[Sideshow Bob]

I guess it many reasons. Like cost, and also availability. BJTs are cheap and common, traditionaly FETs are more uncommon. FETs are voltage controlled and BJTs are current controlled. Many FETs need around 10 volt to be turned fully on. That may be unpractical in a 5 volt system. But almost any common 5 volt logic. Will be able to source the base current needed to switch on a common transistor. In order to control most basic relays

[CocaCola]

Whatever you choose don't forget to add the protective 'flyback' diode across the coil so you don't kill the FET or transistor...

[solutions]

Because bipolar guys live on myths and are bent on using a current controlled device on a voltage output. We had a thread on this - the guy spent MONTHS trying to make a bipolar work.

You CAN get logic threshold FETS now, typically a couple of volts to turn on, not 10V, which is fine for a coil. The 10V devices are more in the dozens of amps stuff, which is a pretty HUGE coil :-)

As was said, you should clamp the spike you get from the relay coil with a diode, but even then, some FETs don't need one because their intrinsic body diode is robust enough to do the job (we had another thread on that one as well). That diode can make the difference in the economics argument between FETs and bipolar.

IMO, running anything of modest power off of a logic gate BEGS for a FET and use of a bipolar is a silly waste of board real estate and a couple of $0.003 resistors (emitter degeneration, base current limiter)...

All assumes a DC relay coil, of course.

[metal]

This page will answer your question: http://www.learnabout-electronics.org/fet_03.php Think about it, you need to utilize a voltage divider in order to make Vg lower than Vs such as using a voltage divider on the Vs, or use -ve voltage on Vg when Vs = 0, this is not considered feasible solution looking at BJTs which don't require such conditions.

You CAN get logic threshold FETS now, typically a couple of volts to turn on, not 10V, which is fine for a coil. The 10V devices are more in the dozens of amps stuff, which is a pretty HUGE coil :-)

Solutions, designers use MOSFETs for amplifier output and FETs for input, so the FETs don't even see the large coil. Also, FETs and MOSFETs are usually used as direct switching devices, even for AC, but require another configuration for AC, of course.

[Sideshow Bob]

@vantusaonho
It could be interesting to know why you are asking. The answers you have got so far. Is then looking at them not very good at all. The answer from Solutions is quite bombastic, but very much based on half-truths. And my own answer is kind of in the same alley. But the bootom line. It is nothing wrong in using a FET to activate a relay. For the hobbyist a "logic level" fet may be hard to get depending on which part in the world you live. And hence it would be easy and cheaper to use a transistor.    

[solutions]

^^^ Hey - you started it 'Bob  ;-)

Metal - you are correct, BUT what you describe is a depletion mode FET. An enhancement mode device has a positive Vt, and will pinch off at about a volt or so and has full turnon at around 2-3V. FET is a transistor type, MOSFET is an implementation - sorry if I confused things by omitting "MOS".

example: http://www.irf.com/whats-new/nr060209.html

[metal]

Solutions, we are talking about JFET, not HEXFET MOSFET. I can remember an example now, 2SK170 is a FET transistor. All of us know that we can use MOSFETs to control a relay and the N-Channel MOSFETs need +ve voltage, can't you remember the well known BS170, the depletion version.

[Sideshow Bob]

^^^ Hey - you started it 'Bob  ;-)

Metal - you are correct, BUT what you describe is a depletion mode FET. An enhancement mode device has a positive Vt, and will pinch off at about a volt or so and has full turnon at around 2-3V. FET is a transistor type, MOSFET is an implementation - sorry if I confused things by omitting "MOS".

example: http://www.irf.com/whats-new/nr060209.html

That is correct. But those are almost without exception meant for power application. And to use such device to switch most common relays will be grossly overkill.  And also quite expensive. Not to mention board real estate. For those power FETs. A search at my component supplier gave me however only one component that could be interesting in such a setting. The MMFTN170. You also say that you do not any kind resistor at the gate, compared to the BJT base resistor. Well that is sort of true if you only want to switch a relay now and then. But if you move somewhat up in switching frequency. You will find that driving a FET correct. Need careful design. Just Google mosfet gate driver design

[vantusaonho]

@vantusaonho
It could be interesting to know why you are asking.   

I asked that thing because a stepper motor driver can be made by FET of Transitor. In stepper motor, there are 2 or more coils, so that control stepper motor is to control its coil.
I think you are right when said here

I guess it many reasons. Like cost, and also availability. BJTs are cheap and common, traditionaly FETs are more uncommon. FETs are voltage controlled and BJTs are current controlled. Many FETs need around 10 volt to be turned fully on. That may be unpractical in a 5 volt system. But almost any common 5 volt logic. Will be able to source the base current needed to switch on a common transistor. In order to control most basic relays

[GunMage]

I asked that thing because a stepper motor driver can be made by FET of Transitor. In stepper motor, there are 2 or more coils, so that control stepper motor is to control its coil.
I think you are right when said here

Why not use a stepper motor driver ?? We are working on a stepper motor driver board now and are using a dual full bridge. Thing works great. It is around a $7.00 chip though.
http://www.st.com/internet/com/TECHNICAL_RESOURCES/TECHNICAL_LITERATURE/DATASHEET/CD00002293.pdf

[Gallymimu]

Allegro makes very nice stepper drivers for about half the cost (A4988 for instance).

I use FETs for driving relay coils all the time.  There isn't a problem.  Most MOSFETs today can be driven with 5V or even 3.3V logic though they may not come to as low of an Rdson unless they are specified for logic level.  FETs are so good these days I can't imagine a relay could that would need a very low Rdson.

[Ichan]

How complicated you all guys.. 

Question:

why do not we use a FET to active a coil (Relay coil)?

Answer:
we do not use fet to drive relays because it costs several times of the bjt counterparts, while bjt is enough for that purpose.

-ichan

[solutions]

LoL - $0.11 in qty 100...a major fortune for them thar MOSFETs

http://www.digikey.com/product-detail/en/2N7002WT1G/2N7002WT1GOSCT-ND/1967031

[GunMage]

Allegro makes very nice stepper drivers for about half the cost (A4988 for instance).

I use FETs for driving relay coils all the time.  There isn't a problem.  Most MOSFETs today can be driven with 5V or even 3.3V logic though they may not come to as low of an Rdson unless they are specified for logic level.  FETs are so good these days I can't imagine a relay could that would need a very low Rdson.

I have used the Allegro stepper drivers. We typically have issue with them in our application. Hence the ST part. However as a general stepper driver they are good chips.

[Ichan]

LoL - $0.11 in qty 100...a major fortune for them thar MOSFETs

Ha, I didn't know that - so the price of BC547 is just on par with 2N7000... well at Digikey.

But in my daily works bjt for this purpose is as cheap as dirt, more then $0.04 won't fit the budget.

-ichan

[solutions]

I know you guys didn't know that, which is why I've been trying to pry your eyes open.

By the time you f with biasing and deal with the insertion costs of the two (at least) extra resistors, the cost is the same or less for a FET, Mr Cheap
 

Don't forget, assembly cost dominates on "low cost" components. And we all know Digikey prices are HIGH, which is why I used them as my example.

Above all, FETs don't have thermal runaway problems if you're dealing with modest to high power levels. For huge power, we have the IGBT, which is a FET/Bipolar hybrid...again, you can drive some of those with logic levels directly.

I've given up on bipolars almost entirely now. Dinosaurs.

[bigtoy]

I've used FETs to drive relay coils. Agreed it can cost a few pennies more than a BJT, but sometimes that high-impedance gate is just what the good doctor ordered. Oh, and I should edit, for some low-voltage applications, sometimes it's hard to accept 0.6 - 1.0V drop across a BJT, so a low-impedance FET can be helpful in that way too.

[Gallymimu]

I have used the Allegro stepper drivers. We typically have issue with them in our application. Hence the ST part. However as a general stepper driver they are good chips.

Can you share what issues you had with the Allegro?  I've got a design using them right now but it hasn't be put into production yet.

[GunMage]

Can you share what issues you had with the Allegro?  I've got a design using them right now but it hasn't be put into production yet.

My issue is the enviornment I subject my components to. My PCB's are typically submerged in oil and see very high pressures (thousands of PSI).
You are not going to have anything to worry about unless you put your circuits into extremely hostile enviornments like we do.

**EDIT - Spelling. We need spell check **

[Ramnish]

In electronics IMO, keep the circuit as well as components as simple as possible. Never try and do an overkill. For driving a relay use of a simple transistor is the way to go unless you need strict logic level control, but for that too the final switching should be achieved through a simple switching transistor. Moreover for relay switching, the switching speed is not a critical factor. This is just my way of thinking.

- Ramnish

[haXudon]

i've always preferred using FET for my designs than BJTs for the simple fact that it's easier and i don't need to mess around with drive currents.

just a head's up when you want to use an FET or any transistor to drive a coil or motor, make sure you implement a snubber or clamping network. this allows for faster switching and better control of the peak reverse voltage. also a diode across the drain and source is a good idea for driving inductive loads.

[LithiumOverdosE]

I learned many years ago (the hard way) that using any kind of voltage controlled switches (and devices in general) in noisy environments is a very bad idea and should be avoided if not absolutely necessary. Their comparatively high input impedance makes them ideal candidates for all sorts of problems with EMI/RFI signals, ranging from brief unexpected turn on and turn off to brief excursions into linear region. Usually one can indeed solve that by all sorts of filtering, EMI/RFI shielding, carefully optimised PCB design, negative gate bias voltages etc. However it all adds to the circuit complexity and cost.

Or one could simply use current controlled switches like BJTs and significantly reduce complexity and cost of the circuit while making it less prone to unpredictable behaviour in noisy environments. Even higher gain Darlington's are less prone to problems compared to voltage controlled switches.

So the choice of the switching element really depends on the particular application and circuit in which semiconductor switch is to be used.    

[zac]

I learned many years ago (the hard way) that using any kind of voltage controlled switches (and devices in general) in noisy environments is a very bad idea and should be avoided if not absolutely necessary. Their comparatively high input impedance makes them ideal candidates for all sorts of problems with EMI/RFI signals, ranging from brief unexpected turn on and turn off to brief excursions into linear region. Usually one can indeed solve that by all sorts of filtering, EMI/RFI shielding, carefully optimised PCB design, negative gate bias voltages etc. However it all adds to the circuit complexity and cost.

Or one could simply use current controlled switches like BJTs and significantly reduce complexity and cost of the circuit while making it less prone to unpredictable behaviour in noisy environments. Even higher gain Darlington's are less prone to problems compared to voltage controlled switches.

So the choice of the switching element really depends on the particular application and circuit in which semiconductor switch is to be used.    

I haven't had false triggering or linear effects with mosfets driving relays, but couldn't one just add a pulldown resistor (such as a 10K between gate and source) to prevent that?  I would think this could only be a problem if the drive source is high impedance. 

[LithiumOverdosE]

Ever tried working with microcontrollers and MOSFETs that they drive in extremely electrically noisy environment?  

G-S pull-down is a standard thing whenever I use MOSFETs as switches. I don't even think about it anymore.

However, one can do all the filtering and signal conditioning but what's the point of forcing the use of MOSFET if BJT can solve all of that in such cases? KISS

[zac]

Ever tried working with microcontrollers and MOSFETs that they drive in extremely electrically noisy environment? 

G-S pull-down is a standard thing whenever I use MOSFETs as switches. I don't even think about it anymore.

However, one can do all the filtering and signal conditioning but what's the point of forcing the use of MOSFET if BJT can solve all of that in such cases? KISS

I worked on stuff in automotive environments that were pretty noisy.  One problem with bipolar transistors is the power consumption is much higher than with MOSFETS (which is almost zero).  This is mainly a problem with battery powered equipment.  My driver designs used mostly bipolar transistors 20 years ago, but I mostly use MOSFETs now for things like relay drivers.  One problem in the past was the MOSFETS needed higher than 3.3V or 5V to saturate so a separate bias supply would be needed (or needed to use a MOSFET driver chip with an internal charge pump), but this limitation has been mostly overcome with the logic level MOSFETs that have been available since around 2000.   

[LithiumOverdosE]

I worked with some RF pulsed power stuff. Really hard to control induced spikes and MOSFETs did fire on their own a few times. I did solve it with carefully placed ground planes, shielding and transient suppression but in the end I simply felt more comfortable with BJTs. 

I do agree with advantages of FET technologies in low power circuits but it only goes to illustrate the point I tried to make. It all depends on application and there is no universal rule. I use BJTs because I very rarely had anything to do with battery operated devices and higher power consumption was never a concern. I almost exclusively use MOSFETs and IGBTs for power switching and I don't recall when I last used BJT in that role. 

[bigtoy]

I've used both BJT & MOSFET. They both work. Circuits are similar but not identical. Often the reason for choosing one vs the other is based upon what other components are already on the board - it's nice to limit the bill of materials when possible.

[optikon]

I worked with some RF pulsed power stuff. Really hard to control induced spikes and MOSFETs did fire on their own a few times. I did solve it with carefully placed ground planes, shielding and transient suppression but in the end I simply felt more comfortable with BJTs. 

I do agree with advantages of FET technologies in low power circuits but it only goes to illustrate the point I tried to make. It all depends on application and there is no universal rule. I use BJTs because I very rarely had anything to do with battery operated devices and higher power consumption was never a concern. I almost exclusively use MOSFETs and IGBTs for power switching and I don't recall when I last used BJT in that role. 

Put a pull down resistor on the G-S of those FETs, thats pretty tough to get RF to reach threshold V to turn on. I'm guessing when you had problems, you didnt have a low impedance gate on the fets.

[LithiumOverdosE]

Put a pull down resistor on the G-S of those FETs, thats pretty tough to get RF to reach threshold V to turn on.

I already stated that I do that as a standard procedure.

G-S pull-down is a standard thing whenever I use MOSFETs as switches. I don't even think about it anymore.

[zab]

1 mosfet and igbt nodoubt create problem when testing with input control components removed. Ie gat circuitry open while others pins alive. may burn some time.so pull down resistor save you from such problems

[Vineyards]

Relay driving is a copy-paste unit really. I spend most of my energy for the core functions of a circuit. What do you expect to gain beyond what you effortlessly achieve using small signal transistors? FETs are actually like black magic. It is hard to match one to another and their specs may signicantly change with temperature. Bipolar transistors meanwhile are pretty bulletproof.

[Gallymimu]

We used MOSFETs and IGBTs in about the noisiest environment I can think of (vacuum deposition plasma arcing).  We used pull downs snubbers and filter caps, especially on signals coming into and out of the control boards.  Never had any problems with mosfet false triggering, usually a 1k pull down was fine.

We had more problems with CPLDs having spurious behaviors than issues with our switches.

We used the MOSFETs for driving relays, low power PWM for fans, LEDs etc.  We did use BJTs though for things like class B amp stages on opamp outputs and a few other places where linear region control of a mosfet is a huge PITA

[LithiumOverdosE]

This one was electronics for automatic control of a large Tesla transformer (not SSTC). For the reasons I still cannot fathom a decision was made (not by me) to locate electronics enclosure right between spark gaps and heavy duty primary. I suspect that extreme induced transients from primary and close vicinity of spark gaps made large enough spikes to trigger small MOSFETs. I didn't bother to investigate further due to dead line. I simply modified circuit for BJT and problem was solved. After that I simply use BJT in all critical places where it is convenient to use them just as a precaution.

[Gallymimu]

This one was electronics for automatic control of a large Tesla transformer (not SSTC). For the reasons I still cannot fathom a decision was made (not by me) to locate electronics enclosure right between spark gaps and heavy duty primary. I suspect that extreme induced transients from primary and close vicinity of spark gaps made large enough spikes to trigger small MOSFETs. I didn't bother to investigate further due to dead line. I simply modified circuit for BJT and problem was solved. After that I simply use BJT in all critical places where it is convenient to use them just as a precaution.

HA, wow that's one heck of a good EMI test for sure!

[optikon]

This one was electronics for automatic control of a large Tesla transformer (not SSTC). For the reasons I still cannot fathom a decision was made (not by me) to locate electronics enclosure right between spark gaps and heavy duty primary. I suspect that extreme induced transients from primary and close vicinity of spark gaps made large enough spikes to trigger small MOSFETs. I didn't bother to investigate further due to dead line. I simply modified circuit for BJT and problem was solved. After that I simply use BJT in all critical places where it is convenient to use them just as a precaution.

It may have been capacitive coupling to the spark gaps. With enough dV/dt even *fempto* Farads might couple enough charge the gates to switch them on momentarily. I assume a crazy scheme like this was at least enclosed in a grounded electrostatic shield.

As far as low level analog "noise" /offset  goes, BJT's are more susceptible to RF rectification than are JFet or MOSFET's. However, as far as a switching application goes, BJT are more immune due to the lower input impedance (they require uA to turn on).

Guess it depends on the app!

[EHHS1979]

Both MOSFETS and transistors are used to activate relay coils.  It is a preference thing, can be application based.  When driving relays with either MOSFETS or transistors, place a diode across either the relay coil or the MOSFET to protect the semiconductor device from voltage/current induced when de-energizing the coil.

I use devices like a DRDC3105 to drive relays, small, surface mount, SOT23 or SOT26 package, has the diode built in, to drive 3.3 or 5 VDC relays.