sorted... Why have Darlingtons?

I understand what they do: gain the gain, so to speak.

But presumably there are technical reasons that a transistor of “Darlingtonian gains” cannot be made made to provide that gain in one pass? If so I’d be interested to hear why that is so.

Words of one syllable, please… 8)

Old transistors didn't have a very high beta so it was often necessary to get a high enough gain.

With newer transistors you don't need them so much, although they're still used when you need a massive gain.

http://en.wikipedia.org/wiki/Darlington_transistor

Darlingtons are in my list of "don't use" devices and techniques, see http://miscsolutions.wordpress.com/2011/10/16/five-things-i-never-use-in-arduino-projects/. A single transistor (especially a high gain low saturation voltage transistor such as ZTX851) or a mosfet is almost always a better solution.

Ah ok, so they're basically superseded.

Thanks.

Historically, back in silicon dinosaur days, when the 2N3055 power transistor and the 2N2222A General purpose transistor were king... the power transistors seemed to be able to gain their current capacity at the expense of Hfe (gain). So, to create a high current solution, the darlington configuration was conceived, where a lower current/higher gain transistor would drive the base of a power transistor that had less gain... with the resulting configuration accomplishing the overall goal of hi-gain, high speed transitions, and high power.

The configuration was common enough that it eventually became a "device" you could buy, like the still popular TIP120. There is really no compelling reason "not" to use them where it makes sense to and that is why you can still purchase them. Using one "just because you have one" maybe not the best idea... but they have their uses... I use them to drive solenoids.

dc42: Darlingtons are in my list of "don't use" devices and techniques, see http://miscsolutions.wordpress.com/2011/10/16/five-things-i-never-use-in-arduino-projects/. A single transistor (especially a high gain low saturation voltage transistor such as ZTX851) or a mosfet is almost always a better solution.

Some good info there. I've had many of the same experiences. On the site, you also mention "When might I use external pullup/pulldown resistors? Only when the internal 20k pullup resistor isn’t suitable, for example, if the switches are connected with long wires that pick up noise".

You call this "Five things I never use in Arduino projects" [my emphasis],

but I might add ... external pulldowns are useful for building voltage dividers to interface voltages outside the normal 0..Vcc range. This scheme is also especially useful on A/D converter input channels.

the power transistors seemed to be able to gain their current capacity at the expense of Hfe (gain).

This is also interesting. Older bipolar transistors might have high current gain, hFE, in the 100-200 range for low collector currents, eg a few ma, but hFE would drop to 10 or so at higher currents, eg > 100 mA. Anyone know how they fabricate newer bipolars to get around this problem?

oric_dan(333): You call this "Five things I never use in Arduino projects" [my emphasis],

but I might add ... external pulldowns are useful for building voltage dividers to interface voltages outside the normal 0..Vcc range. This scheme is also especially useful on A/D converter input channels.

What I actually said I don't use is "Pullups or pulldown resistors on digital input pins used for reading switches". I guess the unspecified assumption is that I have control over how I connect both ends of the switch.

dc42: What I actually said I don't use is "Pullups or pulldown resistors on digital input pins used for reading switches". I guess the unspecified assumption is that I have control over how I connect both ends of the switch.

Capacitors and debouncing on rotary encoders? I haven't seen that thread yet, but LOL anyway...

dc42:

oric_dan(333): You call this "Five things I never use in Arduino projects" [my emphasis],

but I might add ... external pulldowns are useful for building voltage dividers to interface voltages outside the normal 0..Vcc range. This scheme is also especially useful on A/D converter input channels.

What I actually said I don't use is "Pullups or pulldown resistors on digital input pins used for reading switches". I guess the unspecified assumption is that I have control over how I connect both ends of the switch.

I realize that, but was thinking of expanding to more general cases.

Some Darlington arrays are pretty useful tho - ULN2803 can be quite a handy part when you need to pull lots of lines low. I use TPIC6B595 more often tho, like having a 74HC595 but with much improved current sinking ability. Can't drive high tho, need something else as the current source.

Modern devices use ion-implantation to get far greater control on the base region dimensions and doping levels and gradients. Older technology was limited to thermal diffusion of dopant atoms. Precise base thickness and accurate doping profiles are essential to high beta and low emitter resistance I believe.

Curiosity gets me to look at ZTX851:

Collector-Emitter Saturation VCE(sat) 10 50 mV IC=0.1A, IB=5mA* Voltage 50 100 mV IC=1A, IB=50mA* 100 150 mV IC=2A, IB=50mA* 200 250 mV IC=5A, IB=200mA*

Well, quite impressive 250mV at 5A, only there is no way arduino could deliver 200 mA base current.You can't use such transistor even for 1A load, I don't know any micro with 50mA outputs driving capability. TIP120, for comparison, needs only 3 mA for 3A collector current.

Other things, ztx851 according to "safe operating area" can supply 5A during 1 sec. TIP120 will run 5A forever.

MOSFET's are nice alternative, same time even logic level transistor would require special high capacitance driver IC for most PWM application.

It seems to me, that old TIP120 is ONLY the best friend for microprocessor application, with one nuisance - base resistor

These #s don't look good tho: Vce Saturation (Max) @ Ib, Ic: 4V @ 20mA, 5A

IC = 3A, IB = 12mA, Vcesat = 2V

P=IV = 6W, heatsink recommended I think!

P=IV = 6W, heatsink recommended I think!

Yes, I agree with that. Same time it's applicable to any BJT, Darlingtons just have saturation little bit higher than others. So the trade is TIP120 + heatsink or MOSFET + TC426.

So for say a TIP31 which can handle up to 2amps at the base..

Would you say to power ANY TIP31 transistor (until now i used a 200ohm base resistor) would you say it's ok to use a 2222a with a tip31 in a darlington array? (i only have 2 n mosfets)

Magician:

P=IV = 6W, heatsink recommended I think!

Yes, I agree with that. Same time it's applicable to any BJT, Darlingtons just have saturation little bit higher than others. So the trade is TIP120 + heatsink or MOSFET + TC426.

No, the trade for medium current (a small number of amps) is TIP120 + heatsink or logic level mosfet + series resistor. You only need a mosfet driver if you are switching much larger currents than the TIP120 can handle, or are using fast PWM (which the TIP120 wouldn't handle well anyway).

You only need a mosfet driver if you are switching much larger currents than the TIP120 can handle, or are using fast PWM

Not quite. Logic level P-channel MOSFET has high Gate Threshold Voltage (example, SFE FQP27P06) and again you need a driver IC (charge pump voltage multiplier) for 3.3V uCPU. Darlington TIP127 has no trouble with voltage as low, as 2V at the base.

Magician: Darlington TIP127 has no trouble with voltage as low, as 2V at the base.

Actually 2.5V max @ Ic=3A according to the datasheet. If I was only switching 3A, I would probably use one of these http://www.farnell.com/datasheets/1648222.pdf which only needs 1.8V.

Actually 2.5V max @ Ic=3A according to the datasheet. If I was only switching 3A, I would probably use one of these http://www.farnell.com/datasheets/1648222.pdf which only needs 1.8V.

I was referring to P-channel , when I need high side switch in low voltage (< 3.3V) design.

Magician:
I was referring to P-channel , when I need high side switch in low voltage (< 3.3V) design.

What, switch 3.3V @ ~3A high side using a PNP darlington? Even if you did have a high current 3.3V device, the Vce(sat) of the darlington would render it pretty much useless at that supply voltage.