Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Using a transistor as a (relatively) high speed switch

  1. Nov 14, 2013 #1
    So I'd like to pulse a current to a load from a transistor. I can use a µ-controller to trigger a transistor. The µ only has a minimum pulse length of about 100ns or greater. But I would like to have the current flow through the transistor for a much shorter pulse length than 100ns. Maybe a pulse length of 5ns or even 10ns would do.

    1) What is the best way to do this type of fast switching? Should I have the current always flowing through the transistor and then turn it off for the short pulse length, or have the current off normally and have the transistor control when it flows (for 5ns or 10ns)?

    2) What options do I have regarding specific model numbers?
  2. jcsd
  3. Nov 14, 2013 #2


    User Avatar

    Staff: Mentor

    What is the application? Getting that kind of large-signal switching speed out of discrete transistors is non-trivial. Can you just use a fast logic gate instead to source/sink the current?
  4. Nov 14, 2013 #3
    I need a low cost method of switching, and the transistor is the lowest cost device I think. A logic gate would be a lot more expensive right?

    I am charging a capacitor and then using the switch to dissipate the charge from the capacitor into another, unknown, resistance and capacitance. And measuring the voltage before and after.

    The 5ns pulse would be ideal, but a 100ns or less pulse would also work I suppose.
  5. Nov 14, 2013 #4


    User Avatar

    Staff: Mentor

    It sounds like you need an analog switch function. Why so fast?
  6. Nov 14, 2013 #5

    I am really looking for a part, in production, that can respond to the input pulse consistently. I've seen some datasheets (for example, this one) list the turn on time as 9ns typical to 12ns maximum and a turn off time as 13ns typical to 18ns maximum. I would want the difference between these typical and maximum values to be as little as possible.

    Also, what is the "storage time" listed in that datasheet?

    The µ-controller that will be sending the control pulse to the switch (i.e. ~ this transistor) will have a 100ns minimum pulse length currently, but that may change. Repeatable because there will be 8 of these in parallel, each doing multiple measurements every 100µs or less.

    I just need a low cost option for repeatable switching using a transistor instead of an actual, packaged switch.

    I think what I'm seeking is actually a Voltage Controlled Resistor. This is commonly made with a JFET, right?
    Last edited: Nov 14, 2013
  7. Nov 14, 2013 #6
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook