0
\$\begingroup\$

I would like to have a precision current source ~1 μA loaded by a resistor ~100k and a voltage source. Also I would like to switch on and off the current. The problem is that when the transistor switch turns off, since the current is very small, it has a trouble dissipating the charge collected in the channel in the case of a MOSFET or in the base in the case of a BJT. As the result switching off time is very slow. Is there a quick solution?

Upd. I need to switch off the current, not just shunt it to the ground. MOSFET switch BJT switch Switching off chart

\$\endgroup\$
  • \$\begingroup\$ Your question is vague. You need to clarify what you mean by a "small" current. Also, what is your definition of "fast" and "slow" switching? \$\endgroup\$ – Joe Hass Apr 1 '14 at 17:56
  • \$\begingroup\$ The title says ~1 μA. The chart shows the switching off time is ~1 ms. It would be good if it were 3 orders of magnitude less. \$\endgroup\$ – facetus Apr 1 '14 at 19:57
2
\$\begingroup\$

You could use a PNP differential pair that always stays active. Drive the bases to either shunt the current to the load, or to ground. That should provide very fast switching of the current, although the current is always flowing somewhere. This is similar to how ECL logic circuits worked, and why they are able to get high switching speed.

There is some good information on differential pairs here: http://www.st-andrews.ac.uk/~www_pa/Scots_Guide/audio/part1/page3.html

\$\endgroup\$
  • \$\begingroup\$ Thanks for an idea. Shunt is not an appropriate solution for my task, since I want to switch the load between a current source and a current sink back and force. Btw, with a shunt a single BJT would work as well: it would be connected to the ground and a simple capacitance parallel to the base resistor would solve the switching off time problem. \$\endgroup\$ – facetus Apr 1 '14 at 20:04
  • \$\begingroup\$ Well if you turn the single BJT on (saturated) you will have much slower off time than if you use a diff pair and keep the transistors from saturating. You'll have the minority carrier recombination time slowing you down: en.wikipedia.org/wiki/Emitter-coupled_logic You could use a couple of diff pairs to steer current to the load so that it can source and sink. \$\endgroup\$ – John D Apr 1 '14 at 22:32
  • \$\begingroup\$ You could use a couple of diff pairs to steer current to the load so that it can source and sink. Hmm... can you make a sketch how to do it? \$\endgroup\$ – facetus Apr 2 '14 at 1:12
  • \$\begingroup\$ No? ECL is based on steering the current of one direction between two shoulders, I cannot use it to steer the current between source and sink. \$\endgroup\$ – facetus Apr 3 '14 at 17:44
  • \$\begingroup\$ @noxmetus Glad you found a solution. Sorry I didn't have time to post a sketch, I had a PNP and an NPN diff pair in mind, each would switch between the load (NPN to sink, PNP to source) and ground. Simpler maybe would be to set up a static sink current source , then use a diff pair to switch in the source current + static sink current. \$\endgroup\$ – John D Apr 3 '14 at 21:57
0
\$\begingroup\$

I found a solution. The key is to negate the influence of the voltage source V2 instead of trying to shut the current with a switch. enter image description here

\$\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.