opamp feedback loop caps??

[JRay]

Hi

I am planning on using a simple non-inverting op-amp stage for  a 25 dB gain preamplifier for my acoustic guitar pick up. I am aiming to get as high fidelity and low noise as is possible from a single 9v battery power source, as the preamp will be strap or belt mounted.  I am undecided on what feedback loop topology to go for.  Could someone help shed light on what the pros and cons of isolating the feedback loop from DC by using a capacitor as in U2 below, opposed to the config of U1, direct connection?

My thoughts were to use a cap to roll-off around 25Hz or so as this should keep the handling noise from the guitar out of the mix.  However in listening tests, the guitar base strings sound better, more solid and punchy without the cap (4.7uF and 10uF BP caps tried with the feedback resistor values as below).  But will not using a cap cause trouble with instability or with too much LF getting amped through PA systems etc?  I am using a 10K/100uF + 10K/100uf to establish 1/2Vcc bias for the opamp, 47n input isolating cap into a 1 Meg input impedance and an OPA2134 chip



Would very much appreciate your wisdom here!

Cheers

[CynicalMan]

If you think it sounds better with the really low frequencies, just double the cap to 22uF to shift the rolloff below your hearing limit. It's better to have it in, otherwise the op amp will amplify any DC coming in, which will reduce headroom in extreme cases.

[nexekho]

I see, that makes sense... I quite often see a cap used in parallel with the resistor to a voltage used to bias, what is that for?  Would it not pass the noise we don't want and force the DC we do want to go via the resistor?

[CynicalMan]

The point of that is to filter out noise on the power line, but also to make the reference voltage act as AC ground. With the cap there, the impedance to ground is very low, and AC signal can't pass through it. Without the cap, signal gets passed through the line, which can cause all sorts of annoying glitches. Non-inverting amplifiers have lower gain than they should, filters don't filter as they should, and the feedback can cause oscillation in bigger circuits. So it's less important that it filters noise, and more important that it filters signal out of the power line.

[JRay]

Thanks for your reply Cynical

I have a 22uF Bi-Polar kicking about somewhere, I will give it a try.  I assume there is no DC polarizing voltage in the feedback loop and so normal electro's would be no good. If I am right this should shift the roll off point down to around 10Hz with a 680R resister leg. I couldn't really tell any difference in the LF handling noise with the 10uF cap (which is highest value I tried) compared to when the 680 ohm resistor was connected to Vcc/2 but sitting with headphones on is a vastly different scenario to say playing a festival stage, and here I would expect LF response too low may cause issues. It is just the bass strings sounded more detailed when the cap was not used. So phones back on tonight for a re-try.

I do not want to shift the feedback resistors up a decade in order to use smaller caps as I wish to keep noise down to a minimum and I already have a 33K series resistor on op-amp input pin which hopefully will offer some protection to the op-amp from the outside world. I could lower this value a bit I s'pose

Cheers for your help!

Ray

[familyortiz]

Ray,
A few other thoughts... you are using a 47nF to ac couple the input. Any capacitive coupling will block DC but I'd up that value for better low end response to about 1uF or better.
I also like using low end rolloff for acoustic instruments. To keep your design simple, do you have the ability to do your rolloff at your mixer? Our Mackie has a button for each channel, to do this.
Otherwise, could you put a simple, switchable high pass filter at the output of your gain stage?
Regards,  Mike

[JRay]

Hi Mike

Thanks for your reply. I am using 47nF as I have a 1KV polystyrene capacitor of this value and the sound using this cap as an input into the 1 Meg input Z is night and day with the other caps I have, and I have tried several polyprops, large and small  polyesters ranging from 0.01uF through to about 1uf. The clarity is much much better with this polystyrene cap and that is without loosing any bass. I do not have any LF roll off available on my mixer and so this is why I originally thought to roll off around 60-70Hz as I thought this would reduce any sub-sonics or handling noise with the guitar. However the sound just seem to be missing something with this roll off.

I tried a 22uF BP cap in the feedback loop and the bass started to sound a bit woolly. The best sound using phones is with the 10uF BP cap as this seems to give the best balance of bass depth combined with unaltered clarity and detail, the bass strings retain their bell like quality. So I reckon this is where I will pitch it for the belt preamp and then maybe arrange for some further EQ later in the signal chain if needed on a live stage, as sound through phones and sounds through PA gear are two entirely different animals!

Thanks for all your replies, appreciated!

Ray