Transistor Leakage and Gain @ 5V? - Transistor Tester

[rullywowr]

I am trying to build a microcontroller based version of RG Keen's transistor tester as seen here http://www.geofex.com/Article_Folders/ffselect.htm  My main question is if I use +5v instead of +9v, will the leakage and total gain readings be off?  Does leakage & gain increase with voltage?

The main challenge at hand is to ensure that no more than +5v gets into the analog input port of my MCU which would subsequently toast it.  I have been thinking about using a 4.7v Zener + Resistor combination as a clamping diode however I want to ensure the readings are still accurate.

The idea would be to chuck a germanium transistor in the tester, and hit the button.  The MCU would display the leakage on the screen and when it stabilized it would switch voltage to the base and do all the calculations for you.  I think this would be a cool project and handy to have around the bench.

Any input would be great!  Thanks.

[RobA]

I don't have any input on the 9V versus 5V levels and leakage for the transistor question. But, to protect the MCU, there are a couple of ways to go about limiting current and voltage into the ADC pins on an MCU that I've seen. The one I decided to use has a two clamping diodes, one from ground to the signal line and one from the signal line to +V (5V in this case). It uses two current limiting resistors. The first goes between the diodes and the input pin and limits current into the MCU pin. The second goes before the diode clamp (between the input signal and the diode clamp point) and limits the current draw through the diode clamp which in turn protects the voltage regulator from failing and causing too much voltage from hitting the power pins on the MCU. I've used 10k resistors and 1N4148 diodes. In some cases, you may need to drop down the resistor values and using BAT46 or similar diodes may help in some situations. The max current limit into an AVR is something like 1mA.

There are a couple of issues with the zener clamp, one of which is that the zener knee is soft and it has some capacitance that can alter the signal.

Higher input impedances on the ADC pins of MCU's slow them down and can make the readings less accurate. If you run into problems with this, you may want to buffer the input signal of the with a good rail-to-rail op-amp buffer. The TS922IN is pretty ideal in these cases as it has rail-to-rail on the input and output and has really good drive characteristics. It's cheap too.

Here's a discussion about that talks about it http://www.avrfreaks.net/index.php?name=PNphpBB2&file=printview&t=69346. The diagram by "microcarl" is the one I'm talking about. I've used this method for reading input CV from pots and it works very well.