Electronics > Beginners
What is the PSRR of a one-transistor amplifier?
cvanc:
This question arises out of the troubleshooting thread linked below. I honestly don't know what the PSRR of a simple common emitter stage is expected to be. I spent some time looking for a citation and so far have come up empty. What does the group mind here have to say about this? Can you cite a reference?
Schematic for the circuit that triggered this inquiry is attached, but I'm looking for the general rule on this issue. Thanks!
https://www.eevblog.com/forum/repair/can-you-double-check-my-troubleshooting-here-(old-kenwood-receiver)/
exe:
Had similar questions as well for years. This article answered them: http://sound.whsites.net/articles/followers.html .
Basically, you can use a simulator to verify the design without building it.
T3sl4co1l:
About zero.
The collector output impedance is high, so the voltage at the output is essentially Vcc - Ic * Rc. The Vcc isn't reduced at all, it has gain 1, so the part of the output due to variation in Vcc is a ratio of 1, or PSRR of fuck all dB.
Perhaps a better way to think of it is a shift in the ground reference for the signal. Suppose you follow an NPN common emitter stage with a PNP common emitter stage. The PNP's input is relative to its emitter, which is now near Vcc. The Vcc term in the NPN's output cancels out with this, and PSRR is quite good (roughly Rc/Rearly).
Early effect is the name given to collector output impedance (i.e., change in collector current for a change in voltage -- the incremental resistance). It depends on transistor type and bias current. Typical figure would be in the 100s kohms for a preamp circuit like this, so PSRR in the 20-40dB range would be expected.
Note this assumes signal frequencies. That is, where the base bias divider doesn't contribute because the coupling caps and source impedance dominate. Base bias being derived from Vcc (as in the traditional four resistor, one transistor, 2-3 capacitor, common emitter circuit) does have an effect at other conditions, of course.
You can always improve Early effect say by stacking transistors in cascode. In that case, the PSRR about doubles, again contingent on being able to ignore bias dividers of course.
Tim
cvanc:
--- Quote from: T3sl4co1l on March 24, 2019, 11:21:39 am ---About zero.
--- End quote ---
Yeah, that's more or less what I was guessing. Many thanks for the explanation.
In the specific troubleshooting case I have before me I nonetheless suspect the hum is not caused by rail ripple. See linked message - I'm barking up the right tree here, right?
Thanks again.
https://www.eevblog.com/forum/repair/can-you-double-check-my-troubleshooting-here-(old-kenwood-receiver)/msg2292795/#msg2292795
T3sl4co1l:
Right, PSRR is a design issue, not a repair issue. If it was a bad design, it would always be bad, no matter how many components or boards you replace.
For repair, you need to find which component(s) have failed, introducing that noise. Possibilities include saturated amp stages, which consequently run at low PSRR. Or if supply ripple itself is so large, it's causing stages to saturate sometimes, introducing noise that way.
In short: replace all electrolytics and try again. You can maybe measure some (ESR meter) and see if they're okay, but the best strategy is to dump them all and then see if there are other issues present.
Tim
Navigation
[0] Message Index
[#] Next page
Go to full version