reducing out of band noise. how, or should I even care?

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

mikep

Well-known member
Joined
Feb 18, 2006
Messages
450
Location
Philadelphia
I built a simple Complementary feedback pair emitter follower and really like how it sounds. it measures well too, except for out of band noise. 20k bandwidth, unweighted noise is better than -100dBV. it measures about 20 dB worse with 100k bandwidth. It is not oscillating. the bandwidth of the circuit is really beyond what I want, many MHz. any idea how to slug the BW to 200k or so? inductive output filtering is the only thing Ive come up with that works. Or should I even care?

the circuit is similar to this, but I'm using a JFET for the input transistor:

(from dself's site):
cfp2.gif


mike
 
[quote author="Kit"]Cap from drain to gate?[/quote]

Actually thats the first thing I tried. it seems to work somewhat in the simulator, especially if add a series R with the C but in real life, no go. I get evil peaking.
 
[quote author="mikep"].. the bandwidth of the circuit is really beyond what I want, many MHz. any idea how to slug the BW to 200k or so? inductive output filtering is the only thing Ive come up with that works. Or should I even care?[/quote]
What's it driving? Is that band-limited? The only reason you might want to limit the noise bandwidth is if you are driving an ADC that has inadequate anti-aliasing filters. Otherwise, just bask in the transient response of your circuit.
 
One thing worries me a bit---although you say the circuit isn't oscillating, the noise, if "white", shouldn't rise that much for the difference in bandwidth (I'd expect about 7dB, not 20dB). I suspect you have something that may be stable, but only marginally so. Alternatively, you may be picking up ~RF and see that when you measure at the higher BW.

In any case, that stage often needs a feedback C of a few ten's of pF or more from output collector to base (the second stage) when used at unity gain. I would look at the transient response using a good fast square wave and adjust for low overshoot/ringing. Load the output with what you expect to drive and have the source impedance also be what you're going to use.
 

Latest posts

Back
Top