Discussion in 'The Projects Forum' started by André Ferrato, Jan 6, 2016.

1. ### André Ferrato Thread Starter Member

Apr 5, 2015
204
1
Hello, i having some troubles with a few BJT i am using, it may be a mislearned thing or two that i picked while reading about BJT, so i came to you guys, the circuit is like the follow that i downloaded from the web:

But instead of one LED, it has 15 LEDS, and the Vcc is 5V, but the circuit is basically the same. I'll guide through my process of calculating the dc bias resistors: First i pick the transistor, for the first time i chose BC337, i took a look at it's datasheet and followed the same process i did last time when i was passing 100mA through it's collector and emitter from a last project. I first choose the typ value of LED forward voltage, 3.2v, so after the LED, 1.8v remains to be used, then i say that collector voltage will be 1v, so it's a 0.8 drop to there, applying ohm's law:

0.8 = R * 0.02
R = 40

Then after that i check the gate current that i need to pass 300mA(15 LEDS), using this datasheet:
http://pdf.datasheetcatalog.com/datasheet2/3/06i96k727i3oloo2gis1ozg0jlyy.pdf
The gate current is about 2.1mA when Vce is 1.0v, so i make the gate voltage about 0.7v, so that the drop from base to emitter can happen and still maintain a Vce of 1.0v. Using ohm's law: 4.3 = R * 0.0021 ; R = 2047

And the emitter is tied at 0v, so it doesn't need resistors.

Then i plug these values into PROTEUS 8.3 and simulate it, and got a perfect result with the BC337AP model of bc337, so i went to test it and for my surprise the maximum collector current that i managed to achieve with this values was 180mA, giving 12mA for each LED. That was really a surprise, because i used this method 3 times, in three different projects and it worked everytime, it maaaay be a really nice coincidence of values, but i want to know where is my mistake.

I thought at first that was my transistor, so i went and bought BC638, TIP31C and a mosfet IRF822. I tested with TIP31C using the same method and i got the same result. One thing that is also surprising, is that in PROTEUS there is a BC337 model that is different from the BC337AP and this model is working more closely with my results, but it's still different. Can anyone point my errors ? Thanks in advance.

2. ### dl324 Distinguished Member

Mar 30, 2015
3,116
602
Post a schematic of your circuit.

3. ### André Ferrato Thread Starter Member

Apr 5, 2015
204
1
The image with 15 LED is too big, the circuit is described above and the values are also.

4. ### MikeML AAC Fanatic!

Oct 2, 2009
5,450
1,066
When a single LED has a Vf of ~2V, how do you expect to light 15 in series starting with only 3.3V? The supply would have to be a minimum of ~35V

5. ### André Ferrato Thread Starter Member

Apr 5, 2015
204
1
I forgot to mention that they are in parallel between collector and Vcc, i thought it was pretty obvious, sorry. And like i described, that figure it's just representative, my Vcc is 5v and all the other characteristics are different.

6. ### MikeML AAC Fanatic!

Oct 2, 2009
5,450
1,066
So do you have a resistor in series with each LED? you should!
So what do you expect the transistor to do? Why have it there at all?

7. ### André Ferrato Thread Starter Member

Apr 5, 2015
204
1
Yes, of course i have, this thread is about what is going on with my transistor calculations, they are wired in parallel, they all have resistors, as i described... This is not about the circuit itself, it's about what kinda o mistake i'm making while calculating the collector current and the base current to the BC337.. This will be used to make the hight current 300mA go around a 555 timer. The calculation i showed is for one LED, this is an array of 15 LEDS

8. ### MikeML AAC Fanatic!

Oct 2, 2009
5,450
1,066
If you expect the collector current to be 300mA, then the base current should be 300mA/10 = 30mA to get the transistor saturated so that it has a min Vce.

Ain't gonna happen if the base resistor R2 is 2.2KΩ. Also, you can't run a 555 on 3.3V. It is quite marginal at 5V. Even if you run it at 5V, when sourcing 30mA, the 555 output pin will be lucky to get to ~2V, so the voltage across R2 is 2V-Vbe = ~1.3V. R = E/I = 1.3/0.03 = 43Ω.

9. ### André Ferrato Thread Starter Member

Apr 5, 2015
204
1
The datasheet of 555 says when is sinking 50mA the drop is only about 0.4v, with 30mA i should expect lower right ? What do you mean get transistor satured so that has a min Vce ? I think that is here my blank space of knowledge about BJT, could you explain to me a little bit ?

Another question, please if you don't mind, look at the graphic that says "Figure 4. Saturation Region" at the bc337 datasheet:

BC 337 DATASHEET

I thought, if i choose a Base Current and a Vce, that would give me my Collector Current , or i am wrong ?

And the calcs i showed, why they do not match?

EDIT: I wrote sinking, but my device will be sourcing, there is a difference ? I mean difference in the dropout.

Last edited: Jan 7, 2016
10. ### MikeML AAC Fanatic!

Oct 2, 2009
5,450
1,066
When using a BJT as a switch (not a linear amplifer), the base current is set to 1/10th of the collector current. Assume that the Hfe is 10; not what is shown on the data sheet.

This is why I stopped using BJTs and replaced them with FETs in about 1990.

11. ### André Ferrato Thread Starter Member

Apr 5, 2015
204
1
Hm, i think i should read more about Transistors, i have an IRF822 here, it's N-channel enhancement MOSFET with this datasheet:

IRF822

Can i run this "thing" i am trying to do with it ?
The problem with replacing BJT with Fets is that where i live, "Brazil", the big electronic shops of my state are in "São Paulo", there the Fets are cheap, but here on my city, let's say that for a BJT i would pay 15cents, for a Fet i would pay 1.15 dolars, converting that to my money and wanting to do a project that uses 4-5 fets, it's a problem. To work around, i buy from china, wait 2-3 months, and get 25 fets for a price of one )))))))))) happy

12. ### MikeML AAC Fanatic!

Oct 2, 2009
5,450
1,066
From the TI data sheet:

A NE555 operated at 5V is only spec'ed to pull-up to 2.75V when sourcing 100mA. I guessed 3V (but wrote 2) in my previous post, however, I got the R2 calc correct.

13. ### André Ferrato Thread Starter Member

Apr 5, 2015
204
1
I see thanks, and about my other doubts ?

14. ### MikeML AAC Fanatic!

Oct 2, 2009
5,450
1,066
The 822 has a rather high Rdson = 4Ω because it is designed for really high voltage. There are NFets with an Rds of a few mΩ with a Vgs of ~4V.

If you want to use BJTs, look at Darlingtons, (or homemade Darlingtons).

Here is a homemade modified Darlington. Note the current through R2, the current into the base of Q1, and finally, the current through D1 as a function of the simulated output voltage from the 555.

R1/D1 simulate your bunch of LEDs/resistors...

Last edited: Jan 7, 2016

Apr 5, 2015
204
1
16. ### MikeML AAC Fanatic!

Oct 2, 2009
5,450
1,066
Huge overkill, but they would work.

Using the 10K pull-up to the gate makes it work with the 5V-powered '555. Note that now, the 555 is sinking current (and not much of it) instead of sourcing it.

Last edited: Jan 7, 2016
17. ### André Ferrato Thread Starter Member

Apr 5, 2015
204
1
I thought about that, but the pwm wont reach the mosfet will it? The diode should block it i mean

Last edited: Jan 7, 2016

Apr 5, 2015
204
1
bump

19. ### ErnieM AAC Fanatic!

Apr 24, 2011
7,320
1,571
The resistor is the source of drive voltage fo the fet. The diode is there to let the 555 drive the gate low, to about .7 volts.

That should work (hey it sims good, right?) but I don't see the need fovthecresisto o the diode: just drive th gate direct from the 555.

20. ### MikeML AAC Fanatic!

Oct 2, 2009
5,450
1,066
@ErnieM: The Vgs threshold of the IRFZ44N is 4V. Will the output pin of a 555 operated from a Vcc=5V pull-up high enough to turn on the gate of the IRFZ44N without some additional help?