Noisy Buff'n Blend (FV-1 circuit)

[ricothetroll]

Hi,
I added Beavis Audio Research's Buff'n Blend circuit and did a new layers with it. All works pretty good, excepted the pedal is noisy ! I think it comes from the Buff'n Blend circuit, because noise is also present when I turn the blend pot fully ACW.

The differences between my schematic and Beavis's one are : I used 2N5457s instead of MPF102, FETs are biased with 2M2 voltage divider instead of 1M to keep input impedance about 1M, I used a 10u input DC blocking capacitor to keep LF intact (I know 10u is abusive - 0.015Hz cut - but can that be a problem ?).

Here are the related files :
http://www.beavisaudio.com/techpages/blocksfragments/ (scroll down about half of the page)
http://www.wuala.com/ricothetroll/public/FV-1_Schematic.png/
http://www.wuala.com/ricothetroll/public/FV-1_Board.png/

If anyone has a clue, please tell me, that's one of my first "designs", I still have a lot to learn !

Best regards.

Eric

[ricothetroll]

I forgot to say that by noise, I mean electronic "pink-like" noise...

I also noticed that there is a loss between wet signal and dry signal (wet one is weaker). Anyone encountered that also ?

Best regards.
Eric

[Quackzed]

there is a note on the schem saying to include r4 only if the wet signal is overpowering the dry signal. if your wet signal is too quiet you can (jumper) r4 and that will help the volume loss of the wet side. it may also get rid of the pink noise.
if its still noisy, my next guess would be the 10uf cap. is it old? oriented right?
after that, are you using a battery or power supply.... if power supply, then it could be ps noise, but i'd guess that the fv-1 has all the power suppy  noise issues already solved so probably not the ps...

[ricothetroll]

Hi Quackzed,

Thanx for your answer !

I must precise that the noise issue isn't catastrophic : at a "home volume" I must be very attentive to hear it, but it could be really annoying at loud rehearsal levels, and painful if followed by some overdrive...

I actually didn't install the serie resistor, I just put it in case I would need it. I'll definitely try to change the 10u cap, although I quite doubt it as it's an good quality Elna Silmic II (btw it could be a bad specimen), and normally bigger input capacitor induce less thermal noise.

I can hear the noise is modulated by FV-1's effects (I actually als hear FV-1's volume drop on the noise when blend pot is CW ) so I guess it has to come from the input stage. I also put sockets for the fets and tried to switch to some other 5457 and J201, without any noticeable change.

One thing I forgot about this schem is RF protection I usually put right after the 2M2 pull down resistor (serie 34k+100p to GND), who knows, it could be a solution... Or I could reduce the 2x 2M2 voltage divider, 300k-500k input impedance is enough anyway. Or some PSU filtering rignt near the drain.

I'll also try to breadboard a TL07x input stage (actually that's what I usually do but for whatever obscure reason I wanted to try a discrete input stage that time) that I would plug in the JFET socket (source) to see if S/N is better.

Concerning the volume loss issue, I actually should have expected it and put some boosting option. By the way I think instead of the blend I'll implement separate controls for wet and dry, to avoid a dry signal loss when using the "reverb only" settings.

To be continued...

Best regards.

Eric

[ricothetroll]

Well I finally dropped the jfet input idea. I did another layer with TL072 opamps and separate dry and wet pots (with FV-1's output boosted). If it appears to work well I will share it here.

But I'd like to understand the issue with jfet boosters anyway, so any explanation would be welcome !

Best regards.

Eric

[ricothetroll]

TL072 I/O version done ! It works very well, a lot quiter than the jfet one. Now with only the dry pot turned up I can hardly notice the difference, in terms of noise level and sound (the slight hum of my old Ampeg B25B might be masking it). By the way, I can now hear that the FV-1 is the noisiest one ! Looking at the datasheet, they claim ADC noise = -93dB max and DAC noise = -93dB max, that makes ~-89dB max total. Shouldn't be so bad, but it still can be heard. Maybe I should boost the input as much as I can before saturation to improve the S/N, but I'm afraid that with a 3.3V supply chip I won't have much headroom : strong attacks with high output pickups may come close to the datasheet's 2.6V-3V p-p input level limit. I guess I'll have to live with it !

A few mistakes remained on the board I etched (I had to hook up 2 more caps and to cut&hardwire a trace), I'll correct them and post the "good" layer in a new thread.

Best regards.

Eric

[ORK]

Did you ever try "noiseless biasing" on the jfets?

[ayayay!]

Eric, after reading this thread and the other one about your FET builds... don't get offended, but are you dead sure you've got your pinouts correct?  I know on my first JFET build, I assumed it was identical to a CBE transistor.  Reason I'm wondering is because I built my first Beavis Buff N Blend circuit about 3 months ago and it worked perfectly.  I had to add a Tillman boost to the end of it to help volume recovery a little, but it was dead silent. 

Again, don't take it the wrong way, but I've been where you're at with the frustration.  Hang in there man. 

[ricothetroll]

Hi,
I just compared again my BnB schem/layout to the 5457 datasheet. Pinout is correct.
But you're right sometimes those things happen (especially with me )
Best regards.
Eric