1
\$\begingroup\$

Are there any low end limits to how low of a constant current you can make with a transistor ?

For a simple circuit like the following, is there a lower limit to the achievable constant current ? Would 100uA-1mA be asking too much ?

I'm not that concerned with (yet) about the temperature effects of the current (that might be a seperate question).

I've sim'd it, and everything appears to be in order, but are there any practical limits that a simulation will exclude ?

schematic

simulate this circuit – Schematic created using CircuitLab

\$\endgroup\$
  • \$\begingroup\$ Transistor beta falls off as current decreases, so at very low currents it would be impossible to set the current accurately using a set resistor as the base current would be on the order of the emitter current. That being said, according to the data sheet the 2N3904 has a minimum beta of 60 at 100uA, so I don't think that will be a problem in this circuit. \$\endgroup\$ – Bitrex Apr 12 '15 at 2:43
  • \$\begingroup\$ Yeah, what @Bitrex said. But I think 100uA puts you in a good zone for a wide variety of transistors. Once some time ago when I was worried about this, I found a transistor that maintained good beta at lower bias currents. If you are interested I can look it up for you, but I think you are fine at 100 uA for many transistors. \$\endgroup\$ – mkeith Apr 12 '15 at 3:48
  • \$\begingroup\$ @mkeith ya that would be great. Take your time. Every day I learn something new. Love it. \$\endgroup\$ – efox29 Apr 12 '15 at 4:02
2
\$\begingroup\$

The 2N5088 is my favorite for low collector current use. It's gain is specified down to 100uA - at 300! Given that gain rolloff is gradual, you can go considerably lower at the cost of less gain. Here is the data sheet:

http://www.onsemi.com/pub_link/Collateral/2N5088-D.PDF

\$\endgroup\$

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.