help me lower the noise on my new acoustic guitar preamp project, please?

[maximebodson]

Hello,
Here's the schematic of my new "in progress" acoustic preamp. It's basically a buffer with variable gain (+3/+19dB) followed by a 3 band eq. I tuned the frequencies to suit the sunrise magnetic pickup.  (100Hz bass, 700hz lo Q mid control,  and 5,5Khz sharp presence Q)
It should be noted that this is actually only one half of the preamp actually, as another channel with different eq frequencies are chosen. the guitar that it's built for has  2 pickups (sunrise mag and K&K pure western mini).
while the sound of the preamp is really nice and the eq frequencies really on the right spots, I would like to ask you for tips on how  to lower the noise floor which is a bit high. an acoustic guitar shouldn't be noisy, right?
it seems like artificial gyrators are a bit on the noisy side..




thanks

all the best

Maxime

[dschwartz]

hey, nice preamp!
please describe the noise with more details.. it´s a hiss? hum?? brrrrr? pfrfrfr? ptptpt?
different noises have different solutions...


my blind shot is that you're getting noise from the magnetic pickup..
and yes, acoustic guitars are very quiet, but most of them have piezo pickups, yours is magnetic, and that´s why you probably have some floor hum noise.

solution? hmm, only a noise gate comes to mind..or any of those weird humbucking techniques like coils around the pickup..

[petemoore]

  Remove the source to see how well the pickup sources 'noise'.
  If the power supply is a 9v, a battery can test this as a noise source, no-noise with battery means the power supply could be improved.
  Those are the two big sources when it's working right.
  Testing the grounds is a top of list item also.

[BubbaFet]

Umm.. you have a positive feedback loop on your 1st preamp stage, which makes it more of a comparitor than an amplifier.

[dschwartz]

Umm.. you have a positive feedback loop on your 1st preamp stage, which makes it more of a comparitor than an amplifier.

wooha you're right
but i suppose he corrected it on his build because there´s no way any intelligible sound would come out of that...

[PRR]

> an acoustic guitar shouldn't be noisy, right?

Better if quiet, true. But if the pickup or player is weak, and the output has to be boosted 19dB(!), then the hiss will come up also.

"Noise" is a bad word. As Daniel is asking: Hiss? Hum? Buzz? Rumble? Radio stations?

Noisy at low gain? HIGH gain? EQ flat? 5KHz boosted full?

What opamp?

Has the pickup pole height been correctly adjusted?

[Eb7+9]

it seems like artificial gyrators are a bit on the noisy side..

that's your main source of added noise here ...
replace them by physical inductors and you're one step ahead/below

[maximebodson]

hello guys,
thanks a lot for your replies !
bubbafet, that's a nice observation ! I made a mistake in the drawing. the first opamp input is non inverting actually, to preserve high input Z.. by the way, the opamps are TL074 powered +/-9volts regulated with7809/7909 chip.
and my noise type is hiss definetly!

the purpose of 19dB availability of gain is that the next stage uses large eq dips  (the sunrise pickup need a large midrange dip to sound "natural", around 9dB). also, as it's meant to ne a 2 channel pre, the output will be using a passive blender circuit, and will loose 6 dB of gain so..
For now my setting are 9dB mid cut and 9dB high cut and around 12dB of gain.
Also I am quite a loud fingerpicker and there pole pieces are set high ..
I was also thinking of the gyrators as the source of noise. Maybe I could build some very small inductors? (the preamp should be very compact) I was initally worried of using inductors on this piece of stage equipment and the possible higher interference sensibility.

any ideas on how to improve the circuit?

[PRR]

Try an SE5532 chip; it may be more suitable.

Compare hiss before and after the EQ/gyrator stage. Let's be sure where the hiss is coming from.

[maximebodson]

Hello,
it's summer time again..  finally found time to work on the preamp.
I tested the preamp with the inputs shorted both from the output and also right before the gyrator.
the hiss dominates at high input gain (+18dB), even before the gyrator.
however, when the gain is lowered, I can surely perceive noise coming from the eq section (with eq cuts and boosts).
do you think a 2-stage preamp would be less noisy? should I add a fet before the amplifier? would it be of use to put bypass caps on the opamps (the psu is very near though)?

any ideas?

thanks a lot

maxime

[Gurner]

My outside bet would be on your supply rails - as petemoore said, try some batteries as temp replacements.

I realise you using linear regulators, but I've just had my first experience with LDO regulators - it swamped my small guitar signal with more white noise than a boom mike thrust into hurricance henry. Talk about white noise .....ugly - I have since written to the head of NXP to ask him if he's taking the p1ss.

[R.G.]

If it was mine, I'd first short out that 33K series resistor into the (corrected!) noninverting input. It is likely to be adding a lot of noise, and not much else.

Why is it there?

[maximebodson]

Hello
RG and Gurner, thanks for your insight..
I tought the 33k resistor made a lowpass filter with the 100pF cap.  (my take on one of your pages, RG!)
I'll try to short it right away.
I also measured noise on the rails, but it's fairly low (1-2mv AC).

maxime

[maximebodson]

RG you were right,  the input series resistance added noise !!   
I don't know if it's still sufficient but that's a good start, for sure!

thanks

[R.G.]

I tought the 33k resistor made a lowpass filter with the 100pF cap.  (my take on one of your pages, RG!)

Yeah. I've met the enemy and it is me. 

Any series resistance you add to the signal also adds its thermal noise. Whether this is significant or not depends on the characteristics of the thing it's amplifying.

I make 33K and 100pF be 48kHz, roughly, ignoring everything else. That's not a bad thing to do, but for noise reasons, it's probably better to do in the opamp feedback section than the input Rs and Cs. Killing RF is only really possible at the input, so a 100R to 1K and a 100pF would get the RF while lowering the thermal noise in proportion to the resistance going down. A 100pF to 200pF cap across the 47K feedback resistor will kill the opamp's response to higher-than-audio stuff down to a gain of one.

You might want to mess with that approach a bit.

[maximebodson]

I'll follow that route!

thanks RG

[maximebodson]

I tested the preamp and it works really nicely! it's now very quiet.
I need to widen the eq bands a little bit, but otherwise it meets my expectations...
I uploaded a clip if you feel like hearing it:
http://dl.free.fr/vjI7xiC6o
first bit is sunrise+KK, then sunrise only then k&k only

Maxime