Shop Learn
Darlington Transistor vs. MOSFET — Parallax Forums

Darlington Transistor vs. MOSFET

ArchiverArchiver Posts: 46,084
edited 2003-02-23 00:45 in Microcontrollers
The other day I set up an automotive fuel injector driven with an IRL520
MOSFET to capture some scope traces showing how it handled the voltage
spikes. My trusty Tek 2465B scope did a nice job, and I expect to post the
photos and article sometime this weekend on my website.

With the recent discussion of the TIP120 Darlington, I thought it would be
interesting to compare it with the IRL520 MOSFET. I couldn't in good
conscience use the TIP120 with a fuel injector because it is only rated at
60 volts -- considerably less than the raw injector spikes. So for this
project I replaced the injector with a 10 ohm power resistor (the actual
resistance at operating temperature was 9.8 ohms).

One thing to keep in mind about Darlingtons is that, by their very nature,
there is a substantial voltage drop between the collector and emitter. This
causes more heat dissipation in the device, and results in less power to the
load. In some cases it is acceptable, in others it isn't. It may result in
the need for a heat sink, making it more expensive overall than a different
transistor which wouldn't need a heat sink under the same conditions.

The TIP120's characteristics can be seen by looking at the data sheet,
located here:

http://www.onsemi.com/productSummary/0,4317,TIP120,00.html

You might want to download the PDF data sheet and take a look at the graphs
on page 5. Note that the graphs for the NPN TIP120 are on the left side of
the page -- the graphs on the right are for the complimentary PNP devices...

Notice on Figure 8 (left side graph) the DC current gain at various
collector currents is shown. Each curve is for a different operating
temperature. At room temperature, the TIP120 gain is over 3,000 for
collector current from 1 to 4 amps.

The next graph in Figure 9 is even more revealing, showing how much base
current you need for full saturation (minimum collector-emitter voltage
drop) at different collector currents. Take a look at the curve for Ic
(collector current) of 2.0 amps. Notice that as you look across the
horizontal axis, the collector voltage (drop) decreases as the base current
increases. A base current of about 5 mA or so gets the collector voltage
about as low as possible. So a 1k ohm base resistor (or a little lower)
would be perfect at 2.0 amps. But notice that the collector voltage is
still 1.0 volts when the transistor is fully saturated (on). That means
that it will be turning 2 Watts into heat (power = voltage * current), which
heats the transistor and steals the power from your load. Notice that the
6.0 amp curve needs about 15 mA for full saturation, and there is now a 1.6
volt drop in the device - resulting in 9.6 watts turning to heat in the
transistor instead of powering your load. Better buy a good heat sink!

Here is a comparison between a TIP120 and an inexpensive "logic-level"
IRL520N MOSFET. Both can be easily driven by a Stamp pin. Each device was
operated as a "saturated switch" on the ground side of the load. In other
words, the +13.85 VDC power was applied to one side of the load (a 9.8 ohm
power resistor), and the transistors switched the current between the other
side of the load and ground. Ambient temperature was 71 degrees Fahrenheit.
Temperature measurements were taken with a Fluke 80T-IR infrared temperature
probe, and displayed on a Fluke 87 meter.

Using the TIP120, I connected a 20k ohm pot between the 5V control signal
and the base. Varying the resistance between 20k ohms and 500 ohms, there
was almost no variation in the voltage drop across the transistor (collector
voltage), indicating that it was saturating nicely with only a fraction of a
milliamp of base current. I left it toward the lower end of the resistance
range so the transistor was fully on. After letting the temperatures
stabilize for about 15 minutes, here are the readings...

Collector voltage (drop from collector to emitter) = .774 V
Voltage across load resistor = 13.0 V
Current though load and transistor = 1.33 A
Power delivered to load resistor = 17.3 W
Power turned to heat in the transistor = 1.03 W
TIP120 Temperature = 191 deg F
Temperature rise over ambient = 120 deg F

I then dropped in the IRL520N MOSFET (no wiring changes needed) and got the
following results under identical conditions:

Drain voltage (drop from drain to source) = .224 V (equivalent to collector
voltage, above)
Voltage across load resistor = 13.5 V
Current though load and transistor = 1.38 A
Power delivered to load resistor = 18.6 W
Power turned to heat in the transistor = 0.31 W
IRL520N Temperature = 111 deg F
Temperature rise over ambient = 40 deg F

There are some pretty significant differences.... power lost in the
transistor (turned to heat) went from 1.03 W in the TP120 to 0.31 W in the
MOSFET. As a result the MOSFET ran a lot cooler -- a temperature rise of 40
deg F over ambient, vs. a 120 deg F rise in the TIP120.

The load received more power with the IRL520 - about 8% more - since less of
it was wasted as heat in the transistor. Even more dramatic differences can
be obtained by substituting the lower resistance IRL530N or IRL540N.

I will post a link when I get the scope trace photos on my website with the
MOSFET driving the fuel injector. I ran it for several hours and the MOSFET
handled the voltage spikes from the injector nicely without any other spike
suppression components. Its temperature was only a few degrees above
ambient (although the injector got pretty hot, since the fuel flow normally
helps cool it).

Hope you found this interesting...

Randy
www.glitchbuster.com

Comments

  • ArchiverArchiver Posts: 46,084
    edited 2003-02-22 15:29
    Randy,

    Good stuff, well researched and written. Thanks for sharing, I found it very
    interesting, and I will probably never make a FIS. But then, never say
    never!

    Jonathan

    www.madlabs.info

    Original Message
    From: "Randy Jones" <randyjones@w...>
    To: <basicstamps@yahoogroups.com>
    Sent: Friday, February 21, 2003 10:24 PM
    Subject: [noparse][[/noparse]basicstamps] Darlington Transistor vs. MOSFET


    > The other day I set up an automotive fuel injector driven with an IRL520
    > MOSFET to capture some scope traces showing how it handled the voltage
    > spikes. My trusty Tek 2465B scope did a nice job, and I expect to post
    the
    > photos and article sometime this weekend on my website.
    >
    > With the recent discussion of the TIP120 Darlington, I thought it would be
    > interesting to compare it with the IRL520 MOSFET. I couldn't in good
    > conscience use the TIP120 with a fuel injector because it is only rated at
    > 60 volts -- considerably less than the raw injector spikes. So for this
    > project I replaced the injector with a 10 ohm power resistor (the actual
    > resistance at operating temperature was 9.8 ohms).
    >
    > One thing to keep in mind about Darlingtons is that, by their very nature,
    > there is a substantial voltage drop between the collector and emitter.
    This
    > causes more heat dissipation in the device, and results in less power to
    the
    > load. In some cases it is acceptable, in others it isn't. It may result
    in
    > the need for a heat sink, making it more expensive overall than a
    different
    > transistor which wouldn't need a heat sink under the same conditions.
    >
    > The TIP120's characteristics can be seen by looking at the data sheet,
    > located here:
    >
    > http://www.onsemi.com/productSummary/0,4317,TIP120,00.html
    >
    > You might want to download the PDF data sheet and take a look at the
    graphs
    > on page 5. Note that the graphs for the NPN TIP120 are on the left side
    of
    > the page -- the graphs on the right are for the complimentary PNP
    devices...
    >
    > Notice on Figure 8 (left side graph) the DC current gain at various
    > collector currents is shown. Each curve is for a different operating
    > temperature. At room temperature, the TIP120 gain is over 3,000 for
    > collector current from 1 to 4 amps.
    >
    > The next graph in Figure 9 is even more revealing, showing how much base
    > current you need for full saturation (minimum collector-emitter voltage
    > drop) at different collector currents. Take a look at the curve for Ic
    > (collector current) of 2.0 amps. Notice that as you look across the
    > horizontal axis, the collector voltage (drop) decreases as the base
    current
    > increases. A base current of about 5 mA or so gets the collector voltage
    > about as low as possible. So a 1k ohm base resistor (or a little lower)
    > would be perfect at 2.0 amps. But notice that the collector voltage is
    > still 1.0 volts when the transistor is fully saturated (on). That means
    > that it will be turning 2 Watts into heat (power = voltage * current),
    which
    > heats the transistor and steals the power from your load. Notice that the
    > 6.0 amp curve needs about 15 mA for full saturation, and there is now a
    1.6
    > volt drop in the device - resulting in 9.6 watts turning to heat in the
    > transistor instead of powering your load. Better buy a good heat sink!
    >
    > Here is a comparison between a TIP120 and an inexpensive "logic-level"
    > IRL520N MOSFET. Both can be easily driven by a Stamp pin. Each device
    was
    > operated as a "saturated switch" on the ground side of the load. In other
    > words, the +13.85 VDC power was applied to one side of the load (a 9.8 ohm
    > power resistor), and the transistors switched the current between the
    other
    > side of the load and ground. Ambient temperature was 71 degrees
    Fahrenheit.
    > Temperature measurements were taken with a Fluke 80T-IR infrared
    temperature
    > probe, and displayed on a Fluke 87 meter.
    >
    > Using the TIP120, I connected a 20k ohm pot between the 5V control signal
    > and the base. Varying the resistance between 20k ohms and 500 ohms, there
    > was almost no variation in the voltage drop across the transistor
    (collector
    > voltage), indicating that it was saturating nicely with only a fraction of
    a
    > milliamp of base current. I left it toward the lower end of the
    resistance
    > range so the transistor was fully on. After letting the temperatures
    > stabilize for about 15 minutes, here are the readings...
    >
    > Collector voltage (drop from collector to emitter) = .774 V
    > Voltage across load resistor = 13.0 V
    > Current though load and transistor = 1.33 A
    > Power delivered to load resistor = 17.3 W
    > Power turned to heat in the transistor = 1.03 W
    > TIP120 Temperature = 191 deg F
    > Temperature rise over ambient = 120 deg F
    >
    > I then dropped in the IRL520N MOSFET (no wiring changes needed) and got
    the
    > following results under identical conditions:
    >
    > Drain voltage (drop from drain to source) = .224 V (equivalent to
    collector
    > voltage, above)
    > Voltage across load resistor = 13.5 V
    > Current though load and transistor = 1.38 A
    > Power delivered to load resistor = 18.6 W
    > Power turned to heat in the transistor = 0.31 W
    > IRL520N Temperature = 111 deg F
    > Temperature rise over ambient = 40 deg F
    >
    > There are some pretty significant differences.... power lost in the
    > transistor (turned to heat) went from 1.03 W in the TP120 to 0.31 W in the
    > MOSFET. As a result the MOSFET ran a lot cooler -- a temperature rise of
    40
    > deg F over ambient, vs. a 120 deg F rise in the TIP120.
    >
    > The load received more power with the IRL520 - about 8% more - since less
    of
    > it was wasted as heat in the transistor. Even more dramatic differences
    can
    > be obtained by substituting the lower resistance IRL530N or IRL540N.
    >
    > I will post a link when I get the scope trace photos on my website with
    the
    > MOSFET driving the fuel injector. I ran it for several hours and the
    MOSFET
    > handled the voltage spikes from the injector nicely without any other
    spike
    > suppression components. Its temperature was only a few degrees above
    > ambient (although the injector got pretty hot, since the fuel flow
    normally
    > helps cool it).
    >
    > Hope you found this interesting...
    >
    > Randy
    > www.glitchbuster.com
    >
    >
    >
    >
    >
    >
    > To UNSUBSCRIBE, just send mail to:
    > basicstamps-unsubscribe@yahoogroups.com
    > from the same email address that you subscribed. Text in the Subject and
    Body of the message will be ignored.
    >
    >
    > Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/
    >
    >
    >
  • ArchiverArchiver Posts: 46,084
    edited 2003-02-22 17:24
    Sounds really good. Where can I get these MOSFETs?

    Thanks

    Bill



    --- In basicstamps@yahoogroups.com, "Jonathan Peakall"
    <jpeakall@m...> wrote:
    > Randy,
    >
    > Good stuff, well researched and written. Thanks for sharing, I
    found it very
    > interesting, and I will probably never make a FIS. But then, never
    say
    > never!
    >
    > Jonathan
    >
    > www.madlabs.info
    >
    >
    Original Message
    > From: "Randy Jones" <randyjones@w...>
    > To: <basicstamps@yahoogroups.com>
    > Sent: Friday, February 21, 2003 10:24 PM
    > Subject: [noparse][[/noparse]basicstamps] Darlington Transistor vs. MOSFET
    >
    >
    > > The other day I set up an automotive fuel injector driven with
    an IRL520
    > > MOSFET to capture some scope traces showing how it handled the
    voltage
    > > spikes. My trusty Tek 2465B scope did a nice job, and I expect
    to post
    > the
    > > photos and article sometime this weekend on my website.
    > >
    > > With the recent discussion of the TIP120 Darlington, I thought
    it would be
    > > interesting to compare it with the IRL520 MOSFET. I couldn't in
    good
    > > conscience use the TIP120 with a fuel injector because it is
    only rated at
    > > 60 volts -- considerably less than the raw injector spikes. So
    for this
    > > project I replaced the injector with a 10 ohm power resistor
    (the actual
    > > resistance at operating temperature was 9.8 ohms).
    > >
    > > One thing to keep in mind about Darlingtons is that, by their
    very nature,
    > > there is a substantial voltage drop between the collector and
    emitter.
    > This
    > > causes more heat dissipation in the device, and results in less
    power to
    > the
    > > load. In some cases it is acceptable, in others it isn't. It
    may result
    > in
    > > the need for a heat sink, making it more expensive overall than a
    > different
    > > transistor which wouldn't need a heat sink under the same
    conditions.
    > >
    > > The TIP120's characteristics can be seen by looking at the data
    sheet,
    > > located here:
    > >
    > > http://www.onsemi.com/productSummary/0,4317,TIP120,00.html
    > >
    > > You might want to download the PDF data sheet and take a look at
    the
    > graphs
    > > on page 5. Note that the graphs for the NPN TIP120 are on the
    left side
    > of
    > > the page -- the graphs on the right are for the complimentary PNP
    > devices...
    > >
    > > Notice on Figure 8 (left side graph) the DC current gain at
    various
    > > collector currents is shown. Each curve is for a different
    operating
    > > temperature. At room temperature, the TIP120 gain is over 3,000
    for
    > > collector current from 1 to 4 amps.
    > >
    > > The next graph in Figure 9 is even more revealing, showing how
    much base
    > > current you need for full saturation (minimum collector-emitter
    voltage
    > > drop) at different collector currents. Take a look at the curve
    for Ic
    > > (collector current) of 2.0 amps. Notice that as you look across
    the
    > > horizontal axis, the collector voltage (drop) decreases as the
    base
    > current
    > > increases. A base current of about 5 mA or so gets the
    collector voltage
    > > about as low as possible. So a 1k ohm base resistor (or a
    little lower)
    > > would be perfect at 2.0 amps. But notice that the collector
    voltage is
    > > still 1.0 volts when the transistor is fully saturated (on).
    That means
    > > that it will be turning 2 Watts into heat (power = voltage *
    current),
    > which
    > > heats the transistor and steals the power from your load.
    Notice that the
    > > 6.0 amp curve needs about 15 mA for full saturation, and there
    is now a
    > 1.6
    > > volt drop in the device - resulting in 9.6 watts turning to heat
    in the
    > > transistor instead of powering your load. Better buy a good
    heat sink!
    > >
    > > Here is a comparison between a TIP120 and an inexpensive "logic-
    level"
    > > IRL520N MOSFET. Both can be easily driven by a Stamp pin. Each
    device
    > was
    > > operated as a "saturated switch" on the ground side of the
    load. In other
    > > words, the +13.85 VDC power was applied to one side of the load
    (a 9.8 ohm
    > > power resistor), and the transistors switched the current
    between the
    > other
    > > side of the load and ground. Ambient temperature was 71 degrees
    > Fahrenheit.
    > > Temperature measurements were taken with a Fluke 80T-IR infrared
    > temperature
    > > probe, and displayed on a Fluke 87 meter.
    > >
    > > Using the TIP120, I connected a 20k ohm pot between the 5V
    control signal
    > > and the base. Varying the resistance between 20k ohms and 500
    ohms, there
    > > was almost no variation in the voltage drop across the transistor
    > (collector
    > > voltage), indicating that it was saturating nicely with only a
    fraction of
    > a
    > > milliamp of base current. I left it toward the lower end of the
    > resistance
    > > range so the transistor was fully on. After letting the
    temperatures
    > > stabilize for about 15 minutes, here are the readings...
    > >
    > > Collector voltage (drop from collector to emitter) = .774 V
    > > Voltage across load resistor = 13.0 V
    > > Current though load and transistor = 1.33 A
    > > Power delivered to load resistor = 17.3 W
    > > Power turned to heat in the transistor = 1.03 W
    > > TIP120 Temperature = 191 deg F
    > > Temperature rise over ambient = 120 deg F
    > >
    > > I then dropped in the IRL520N MOSFET (no wiring changes needed)
    and got
    > the
    > > following results under identical conditions:
    > >
    > > Drain voltage (drop from drain to source) = .224 V (equivalent to
    > collector
    > > voltage, above)
    > > Voltage across load resistor = 13.5 V
    > > Current though load and transistor = 1.38 A
    > > Power delivered to load resistor = 18.6 W
    > > Power turned to heat in the transistor = 0.31 W
    > > IRL520N Temperature = 111 deg F
    > > Temperature rise over ambient = 40 deg F
    > >
    > > There are some pretty significant differences.... power lost in
    the
    > > transistor (turned to heat) went from 1.03 W in the TP120 to
    0.31 W in the
    > > MOSFET. As a result the MOSFET ran a lot cooler -- a
    temperature rise of
    > 40
    > > deg F over ambient, vs. a 120 deg F rise in the TIP120.
    > >
    > > The load received more power with the IRL520 - about 8% more -
    since less
    > of
    > > it was wasted as heat in the transistor. Even more dramatic
    differences
    > can
    > > be obtained by substituting the lower resistance IRL530N or
    IRL540N.
    > >
    > > I will post a link when I get the scope trace photos on my
    website with
    > the
    > > MOSFET driving the fuel injector. I ran it for several hours
    and the
    > MOSFET
    > > handled the voltage spikes from the injector nicely without any
    other
    > spike
    > > suppression components. Its temperature was only a few degrees
    above
    > > ambient (although the injector got pretty hot, since the fuel
    flow
    > normally
    > > helps cool it).
    > >
    > > Hope you found this interesting...
    > >
    > > Randy
    > > www.glitchbuster.com
    > >
    > >
    > >
    > >
    > >
    > >
    > > To UNSUBSCRIBE, just send mail to:
    > > basicstamps-unsubscribe@yahoogroups.com
    > > from the same email address that you subscribed. Text in the
    Subject and
    > Body of the message will be ignored.
    > >
    > >
    > > Your use of Yahoo! Groups is subject to
    http://docs.yahoo.com/info/terms/
    > >
    > >
    > >
  • ArchiverArchiver Posts: 46,084
    edited 2003-02-22 17:52
    In a message dated 2/21/2003 10:37:31 PM Pacific Standard Time,
    randyjones@w... writes:

    > The next graph in Figure 9 is even more revealing, showing how much base
    > current you need for full saturation (minimum collector-emitter voltage
    > drop) at different collector currents. Take a look at the curve for Ic
    > (collector current) of 2.0 amps.

    Hi Randy,

    Good information.....
    I completely understand everything you wrote, One point though. In response
    to the original post regarding the use of a TIP120, that user indicated
    he/she needed to switch a 2 amp relay.

    I am relatively sure the 2 amps that person was talking about was the switch
    side of the relay and not the coil side, and that would be the basis of
    his/her concern for the value of the base resistor. As I am sure you are
    aware, if a relay requires 2 amps to be acitaved, there is something wrong
    with that relay.

    None the less, your information on Darlington vs Mosfet is still acurate,
    applicaple and thorough!!!!!!!!!
    Good work!

    Regards,

    Ken


    [noparse][[/noparse]Non-text portions of this message have been removed]
  • ArchiverArchiver Posts: 46,084
    edited 2003-02-22 18:04
    In a message dated 2/22/2003 9:25:59 AM Pacific Standard Time,
    billak@f... writes:

    > Sounds really good. Where can I get these MOSFETs?
    >
    > Thanks
    >
    > Bill
    >

    <A HREF="www.glitchbuster.com">www.glitchbuster.com</A>


    [noparse][[/noparse]Non-text portions of this message have been removed]
  • ArchiverArchiver Posts: 46,084
    edited 2003-02-22 22:16
    At 12:52 PM 2/22/03 -0500, smartdim@a... wrote:

    >Good information.....
    >I completely understand everything you wrote, One point though. In response
    >to the original post regarding the use of a TIP120, that user indicated
    >he/she needed to switch a 2 amp relay.
    >
    >I am relatively sure the 2 amps that person was talking about was the switch
    >side of the relay and not the coil side, and that would be the basis of
    >his/her concern for the value of the base resistor. As I am sure you are
    >aware, if a relay requires 2 amps to be acitaved, there is something wrong
    >with that relay.

    You simply can't make that assumption.

    Grab a horn relay from an American-made old clunker in the junkyard. The
    kind of horn relay I am talking about is in a metal can with 3 terminals:
    Batt, Horn, Switch. Measure the coil resistance. Awfully low, isn't
    it. Now hook it up to a 12V battery and measure the coil current. Darn -
    its over 2 Amps. Even more darn - the relay is getting too hot to hold.

    Now try the same thing for the starter solenoid from the same clunker. You
    might find a coil current in excess of 10 Amps.

    The original poster did not say where the relay was from or even what kind
    of relay it was. An awful lot of stuff is built for intermittent use for
    cost reasons. One of the ways of saving cost in relays and starter
    solenoids is to reduce the size and number of turns of wire used to make
    the actuator coils.

    dwayne

    --
    Dwayne Reid <dwayner@p...>
    Trinity Electronics Systems Ltd Edmonton, AB, CANADA
    (780) 489-3199 voice (780) 487-6397 fax

    Celebrating 19 years of Engineering Innovation (1984 - 2003)
    .-. .-. .-. .-. .-. .-. .-. .-. .-. .-
    `-' `-' `-' `-' `-' `-' `-' `-' `-'
    Do NOT send unsolicited commercial email to this email address.
    This message neither grants consent to receive unsolicited
    commercial email nor is intended to solicit commercial email.
  • ArchiverArchiver Posts: 46,084
    edited 2003-02-23 00:45
    Thanks for all the info Randy and all.

    I was trying to simplify my question by using a 2 amp relay. But to be
    totally honest its a Push Tubular Solenoid with a resistance of 5.8 ohms.

    FYI I'm using it to depress a button that needs about 500 grams of force at
    1/8 inch, so this rube goldberg affair. Any advice on a better way then a
    Push Tubular Solenoid would be great as the solenoid does take a lot of
    power and heats up when cycled a dozen times or so.

    I haven't any experiece with muscel wires or electric pistons, could these
    be a option ??

    Bob Phillips
Sign In or Register to comment.