2
\$\begingroup\$

I'm new to electronics and I have a problem. I want to use a LED to detect light but the voltage it provides is very small (30 mV) and I don't think it can open a transistor. How can I amplify this voltage? (I don't want to use op amp) And, I have only NPN transistors. Can you help me?

\$\endgroup\$
  • \$\begingroup\$ Do you know how to select and fix a bias point for the transistor? \$\endgroup\$ – LvW Feb 28 '15 at 11:41
  • 1
    \$\begingroup\$ I'm new to this and i don't think i ever heard about bias (sory form my mistakes,this is not my original language) \$\endgroup\$ – sergiu reznicencu Feb 28 '15 at 11:44
  • \$\begingroup\$ You probably need two transistors, arranged so that the base-emitter voltages cancel out. Read up on "long tailed pair". \$\endgroup\$ – Brian Drummond Feb 28 '15 at 11:47
  • 1
    \$\begingroup\$ @Dwayne: I didn't do the downvote, but I can see how the "I don't want to use op amp" statement may be deserving of one, not to mention lower case "i" in several places and "Can you help me?", making the correct answer to the question simply "Yes". Using a opamp is the obvious simple way to do this. Since the OP gave no justification, this smells like a religious issue, probably because the OP doesn't know how to use opamps. Not knowing opamps is OK, but wanting to avoid learning about them is not, at least not here. The more I think about it, the more I'm inclined to -1 too. \$\endgroup\$ – Olin Lathrop Feb 28 '15 at 16:03
  • 2
    \$\begingroup\$ First of all: THIS IS NOT MY ORIGINAL LANGUAGE.And if I used several times "I" that means that I know why, am I right? So those "lower cases" are only mistakes, right? Everything is logic. And in my country it sounds terrible to say only what you WANT WITHOUT the permision of that person,am I right? This is about education so it's your "teritory". And I'm 13 years old so I CAN'T buy anything from the internet! \$\endgroup\$ – sergiu reznicencu Feb 28 '15 at 17:34
2
\$\begingroup\$

You can do what you want if you add some amplification to the output signal from the LED. A single op-amp can do this or you can use a pair of transistors configured as a differential amplifier (the "long-tailed pair" that Brian Drummond mentions above.

I'm going to show you an op-amp version because it is easy. I like using the LM358 or LM324 op-amps for these simple circuits - they are readily available and the input common-mode voltage range includes ground. That makes them easy to use with a single supply voltage.

schematic

simulate this circuit – Schematic created using CircuitLab

Upon rereading your question, I see that you asked if you could do this using only NPN transistors. Probably yes but it's not either easy or obvious. The problem is that the signal provided by the LED is both low voltage and very high impedance - some of the references that I checked show the LED being loaded by a very high value resistor (30M or 100M).

\$\endgroup\$
1
\$\begingroup\$

Yes, you can do this with just NPN transistors. 60mV is small, but probably enough. Resistor values will depend on your voltage source, but try this:

schematic

simulate this circuit – Schematic created using CircuitLab

In normal conditions, R1 and R2 set Base voltage about 0.57V - not quite enough to conduct. Adding 30mV gets you 0.6V. With no Emitter resistor, you should get a decent current through R3.

You can also look at "super alpha" designs for very high gain. One example is here

\$\endgroup\$
  • 1
    \$\begingroup\$ @ Alan Campbell: if you have a chance, why don't you try breadboarding this? After you have replaced the LED and transistor because they burned up, maybe you can come up with something different. Please note: I'm not trying to be sarcastic or mean. It's just that what you have drawn can't possibly work and it's not fair to the original poster to suggest your circuit. Please also note that I'm not down-voting your answer - I'd genuinely like you to have a look at it and improve it. \$\endgroup\$ – Dwayne Reid Mar 1 '15 at 1:13
  • 1
    \$\begingroup\$ You're right - I wish I had a breadboard. I'd also like to re-design the base bias [as it stands, the diode has it turned on hard]. \$\endgroup\$ – Alan Campbell Mar 1 '15 at 9:01
1
\$\begingroup\$

I've come up with something that might work. You will need to breadboard it and see.

schematic

simulate this circuit – Schematic created using CircuitLab

Q1 & Q2 need to be identical transistors, preferably from the same batch. They should also be touching each other if possible so that they are at exactly the same temperature.

Cover D1 and adjust R1 so that the output LED (D2) goes out. Then shine light at D1 and see if D2 lights up.

No guarantees that this works but it should. I'm not anywhere near my lab bench, so I can't try it myself.

\$\endgroup\$
  • 1
    \$\begingroup\$ Q3 is PNP. OP requests NPN only. \$\endgroup\$ – Alan Campbell Mar 1 '15 at 9:01
0
\$\begingroup\$

If you really want to use an LED (not good but cheap) you should probably use it as a current device and make a one op-amp current-to-voltage converter. Very simple and the output would be linear. Use a TL061 for its low cost and its low bias current which might be important depending on the amount of current the LED puts out. Do not use a TI brand TLO61 which has given me problems in years past.

\$\endgroup\$
  • \$\begingroup\$ OP said he doesn't want to use an op amp... \$\endgroup\$ – Null Mar 1 '15 at 1:20
  • \$\begingroup\$ bad choice on his part \$\endgroup\$ – user68868 Mar 1 '15 at 2:15

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.