# darlinton pairs

#### Fenris

Joined Oct 21, 2007
288
hi all

im planning on using a darlington pair to drive 20W/12V filement bulbs.
the transistors will be a BC548 type and the TIP41A.
Hfe of BC548 125
Hfe of TIP41A 15-30

the bulbs will draw 3.6A the TIP41 can handle 6A, both transistors are
well under their Max Voltage.

the base supply for the BC548 is 22.5mA
the multiple of the gain is 125*25(mid way on the TIP) =Hfe 3125
so 22.5mA * 3125 = 70312.5 or 70A.

the circuit is dynamic and the base supply will swing from 1uA to 22.5mA

Am i correct in my understanding that the 70A is the theoretical capacity
of the TIP if it could handle such a load and all i need to know is that as
im only drawing 3.6A and the TIP can handle 6A im ok.
All that matters is that the small current has been amplified to allow the
larger transistor to work efficiantly within its real world ratings.
so if as an extreme example the combined Hfe had been 500A
it doesnt matter so long as i stick to the 6A max. or do you have to
pick a pair that using their combined Hfe * Base current is =to or less than
the current handling capacity of the load bearing Transistor.

i have read some of the Darlington threads on here and also surfed the net for answers, i
am just trying to see if ive got a good enough understanding to implement the BC548
and the TIP41A in my circuit.

Regards

Fenris

#### GS3

Joined Sep 21, 2007
408
The transistor cannot carry more current than the lightbulb will allow. You'd need a battery for that.

Be aware that the initial surge, while the filament is cold, can damage a transistor if the circuit is not carefully designed.

The BC548 has a max collector current of 100 mA so I believe you need to select another transistor like a BD139.

#### beenthere

Joined Apr 20, 2004
15,819
Are you just using the darlingtons as a current switch? If so, have you considered using a FET like an IRF510? Just put 10 volts on the gate through a small (100 ohm) resistor and it's on. No base current.

I just used the IRF510 as an example. Any TO-220 packaged N channel FET wil probably work as well. Probably need a small heat sink to be safe.

#### Fenris

Joined Oct 21, 2007
288
hi

thanks for looking. theres a pic attached that may help explain
what im doing. basically a sound to light unit 12V battery originally
had 4 LEDS in series that flashed in time and varied in intensity to
an audible input. i have tried an IRF540 which worked up until
it fried. i have been told that this is because the gate input
was ramping rather than switching on and off this increased its
resistance causing it to over heat, yes it had a heatsink. as you
see in the diagram i intend to drive 2 20W/12v bulbs.

as the Hfe of the BC548 is 125
then Hfe * base current
125*22.5mA = 2.8A is this the point where i check the collector current
and as you say it is only 100ma. so 2.8A is to much for it.

the BD139 has a collector current of
continuos DC 1.5A
pulsed 3A

its Hfe is up to 40 so 40*22.5 = 900mW so safely under the max of
its collector current.
then on to the pair together Hfe1*Hfe2 > 40*25=1000 so
22.5mA @ base of the BD139 * 1000= 2.25A (@the collector of the TIP41A?) in theory.
have i understood this. i seem to be finding once i understand one
piece of the puzzle it then turns out theres another bit to consider lol

Regards

Fenris

#### Attachments

• 42.9 KB Views: 74

#### GS3

Joined Sep 21, 2007
408
There's more to this than meets the eye.

You are adding a big load to a circuit designed for a small load. Not complicated but you have to do it right. If you want to make sure the switch goes on/off avoiding the linear part then you would need to provide a Schmidt trigger or, at least, amplify the signal enough so that the amplifier is saturated most of the time. Other than that the transistors will be in the linear region and dissipate a lot of heat.

You also have to limit the base current of BC548 or whatever replaces it. The way you have it set up it is not limited and will be damaged. You need to insert a resistor where you took out the LEDS.

With flashing lightbulbs you really have to overdimension this because the initial current surge is huge compared to the stable comsumption. If it is the transistor doing the limiting then you are generating heat and need to dissipate it. If the transistor is not the limiting factor then you must be certain the transistor can handle the surge or it will go up in smoke. You may want to consider using a relay because relays are much more forgiving of overloads.

To really be able to assist we would need to know the voltage of the signal at the collector of the first transistor.

Also, why 12 V? There are such circuits powered from the mains which use Triacs to switch lights at mains voltage and it is much simpler.

#### Fenris

Joined Oct 21, 2007
288
hi all

thanks for looking, ive done a bit of digging and i think if i have done the sums right a 1K8 resistor on the base of the BC548 will protect it. or did you mean above the 47K resistor and base of the circuit forming a voltage divider? the collector voltage on the BC548's collector will i think be 3.33V ive worked this out from P/I............ 12V/3.6A=3.33V. the TIP41A can handle 6A continuous the lamps will draw 3.6A as for the initial surge i dont know. this is how it is done in the off the shelf units although i know in one case the Power transistor is mounted outside the case with a heatsink.
id be really gratefull if you could steer me on this last bit, it seems im so close.

i have just fixed a friends relay powered light flasher basically a similar
circuit. preamps a mic and uses this to actuate the relay. works well to. but i wanted to do something myself whilst learning this hobby again. ive downloaded tons of reference material for transistors and darlingtons relearned my V=I*R plus all the variations in between

its done with a battery because im a member of the project dalek forum and im building this to go in a full size model for the flashing earlights
you can buy complete systems including the voice box but they are hekkish expensive £100+. for a tenner i have built a voice modulator that does a good voice and for less than that im getting this sound to light on the go. if i can then get it all working together i want to put the plans up on our forum for all us guys building on a budget. i've learned a lot but theres still these blindspots coming up. there is a schmitt trigger circuit on the web, 555 chip feeding a TIP41A to power 40W of bulb. but its behaviour is not.........cant believe im going to say this..............Dalek like. which is what im aiming at. i hope you can help id be very gratefull. pic of original circuit attached for your reference.

Regards

Fenris

#### Attachments

• 41.2 KB Views: 56

#### GS3

Joined Sep 21, 2007
408
OK, some rough 'n tumble calculations referred to the attached schematic. Bear in mind that transistor gain has a wide margin of tolerance and you cannot count on typical values but have to design for worse case scenario or measure actual gain.

TR1 collector current: 4 A
TR1 base current = 4/20 = 200 mA
TR2 base current = 200/100 = 0.5 mA
We provide a 10K resistor to limit TR2 base current to 1 mA
TR3 base current = 0.5 mA / 100 = 5 uA
A resistor of 100 K provides way more and now we just need to make sure TR4 is going from saturation to cut-off rapidly so what we should see at the collector of TR4 is a rectangular signal with vertical rise and fall and no slopes.

#### Attachments

• 1.5 KB Views: 39

#### Fenris

Joined Oct 21, 2007
288
hi there

i have entered your schematic changes to my own drawings
this weekend i will get the BC548B and the TIP41A and give it a whirl.
fitted ready to fire it up. maths was always my weak point, i wouldnt have got this far without all your help guys. thanks and i will let you know how it goes.

Regards

Fenris

#### GS3

Joined Sep 21, 2007
408
this weekend i will get the BC548B and the TIP41A and give it a whirl.
As I said, BC548 will not do. Make sure that any transistors you use can withstand the current, voltage and power requirements.

#### Fenris

Joined Oct 21, 2007
288
sorry, yes im getting a BD139 as recommended earlier in this
post, by your good self. thanks again.

Regards

Fenris

#### Fenris

Joined Oct 21, 2007
288
hi all
welli was goingto use the TIP41 family in this circuit.
Maplins are out of stock and say it is now discontinued
as far as manufacturing goes. As the intention was to
get a circuit up for my fellow builders to be able to do
and using off the shelf componants this poses a problem.
Is the TIP41 no longer being made? non of the componant
data sites say this. if it is then if i have followed the data
sheets correctly, then i should be able to just swap the
TIP for a 2N3055..........shouldnt i? i have got the BD139
installed but havent bought the other yet because i dont
want to waste money on the wrong choice. one day the
lights will flash again!

Regards

Fenris

#### GS3

Joined Sep 21, 2007
408
2N3055 should work just like any other general purpose transistor which can withstand Ic about 10 Amps.

#### Fenris

Joined Oct 21, 2007
288
hi there

seems that the TIP41's are still being made, so thanks Maplins for the
bum steer. ive ordered both the BD139 and TIP41C from Rapid online.
ive saved the data sheets for both so that should availability become
a problem i can use them to check other Transistors for compatability.

Thanks again will let you know how it goes.

regards

Fenris

#### Fenris

Joined Oct 21, 2007
288
hi

i have got the DB139 and the TIP41, they are installed and the changes you suggested have been made as regards the resistors. its not working as expected yet. when switched on the the lights come on full brightness no flashing to noise. so i pulled the 1M resistor from the base/collector of TR4.
now they flicker weakly and go solid weakly in reaction to noise, any ideas?

i have tested it now without the 1M resistor, as mentioned, on a 9V battery and a pair of torch bulbs 12V 2.2W each, with this configuration it reacts as it should. unfortunately i need to run on 12V with the pairs of 20W bulbs but would compromise to a pair of 10W bulbs they are halogen G4 bulbs.

further to this i have changed the resistor that is in series on the +VE line to 100R-120R this has brought about the proper reaction in the bulb (12V/20W on its own) what think i need to do now is increase the sensitivity of the darlington pair ( i think what i mean is the current handling) as i need to talk louder to get the bulb to become brighter) also as it will handle only one bulb at its current level of reavtive brightness. the 1M resistor has been left out still because it only causes the bulb to come on permanently. so it seems im very close (and a lot further away from despair) since i started this reply. any thoughts greatly recieved, previous attachment of circuit pics still stand bar the changes i have made to get this far. incidently the 2 transistors
that make up the darlington pair are running very cool as well which is good but suggests they are running well below what they can.

Regards

Fenris

#### Fenris

Joined Oct 21, 2007
288
hi all

well today i have got the circuit working as it should. 2*12V/20W in parallel
working at an excellent brightness and very reactive. i think with a fresh eye this morning helped, i spotted a resistor out of line. the main change has been the 1M resistor has been left out because it only causes the bulbs to
come on fully and none reactive, the resistor thats in series on the +ve line
has been changed. the series resistor is still being experimented with as depending what Wattage bulbs i put in it needs to be changed to get the circuit to work, i think this is more related to the series resistor effecting the
variable resistors control of the circuit, as the working range varies with the
series resistors value. the only worry i have now is the heat from the TIP41, its hot!! it has a snap on thin metal heatsink. now obviously its going to get hot, should i consider a larger heatsink perhaps the exstruded ali type or do i need to be worrying about the temp the tip is running at and how do i know? short of it dying!

regards

Fenris

#### Fenris

Joined Oct 21, 2007
288
hi all

well things are going well. thanks to all.
however i have just noticed you can get Darlington
transistors. my current setup is a BD139 to a TIP41C,
based on the info of the previous posts what is the
equivelent darlington unit.......i have seen some rated
at 8A and upwards. but im still whoefully lacking in
understanding to know what i can or cant select

Cheers

Fenris

#### nanovate

Joined May 7, 2007
666
The TIP141 is a Darlington.

You will need a heatsink. What is the voltage across the transistor V_ce? If this is more than about 0.4V (assuming 3.3A) then you will be heating the poor transistor junction up past the boiling point -- 100 degC.

You might be better off with a MOSFET like a IRLZ44N.

#### Fenris

Joined Oct 21, 2007
288
hi

yep it runs hot but i have a heatsink that is rated @ 5oC/W. the circuit is 12VDC i have tried a mosfet 'n' type. the IRF540. it fried, due to,
im told, the gate was ramping rather than being on or off.
can the IRLZ44N deal with this effect better? ok so the TIP41
is a darlington, hmmmm. i dont mind trying the mosfet again, but id also like to try a darlington again as this hasnt died (yet), so given that im using a
BD139 as the lead for the TIP41 can i just pop a bigger replacement for the TIP in place such as a 2N6284? which can handle more Amps so will be running below its max.

Regards

Fenris

#### Fenris

Joined Oct 21, 2007
288
hi all
ive been getting to grips with multisim to help get some grasp of
whats happening in my real world circuit at the load end.
the models output matches all the tech you have given me so that helps.
now im modeling the effect of using the IRLZ44N mosfet as mentioned above.
using the oscillascope function i know the gate is passing 10.9V when the lights are on
and the quiesent voltage when they are off is between 56uV and 116uV the trace on the
osc' is a square wave, there is a slight incline showing the gate is taking about 1ms to go from the quiesent voltage to peak voltage.the grid is set to 10V/Div and the timebase is 10mSec/Div. on a purely visual basis this is good isnt it as the Mosfet is switching on off rather than ramping (i believe the scope would show a much shallower climb) so am i on a winner here, have i understood the visual reference enough? ( i know its more complicated than this) i can just get a decent heatsink and im away? comments please, they are gratefully recieved.

Regards

Fenris

Joined Feb 10, 2008
7
Dear Fenris,
I am not an expert in electronics but I am planning to do a rather similar circuit that controls a battery back up current in a 12 V DC usp unit that I am planning to do so may be we could help out each other since you had some experience now with your kind of circuit,

now I will give you my end of information,
you will have to know wat wattage dissipation for each transistor that you use along with the permitted degrees that the component may indure the rule says as you might know total C/w that means how many c per each wat the heat will go up and it consist of three main things component thermal risistance which is located on its data sheet as to case c/w and the heat sink c/w so the larger the heat sink the better and ambient degrees so the total rise in tempreature would be total thermal risistance X wattage + ambient temprature (the environment temp .surrounding the component)
understand me ?
so if you use 0.5 C/w heat sink then (0.5+2.5 (that of the component) ) X total wattage that needs dissipation)+ ambient temprature (let us say 50 celsus degrees) so let us say you want to dissipate 7 w (3X7)+50
21+50=71 c the degree of the component, if you want to know what is the thermal risistance of a heat sink if you don't have it written somewhere then it is the number 50/the square root of total square cm space of the heat sink
so if you have a heat sink that has a total space of 2500cm then 50/50=1 so its thermal risistance is 1C/W I have applied this to a heat sink in jemeco and i was just about right only 0.3C error because the space is not so accurate for me, any body has any corrections or comment please help us out here