Why do all FET pedal I build are VERY noisy ?

[ricothetroll]

Hi,

The first one I built was ROG's English channel. Though I love its sound, mine is very noisy (broadband noise, shhhhhhh... with some dirty transients in it). I tried to change the j201, noise is still the same. Screened in/out wires : no change.
http://www.diystompboxes.com/smfforum/index.php?topic=78484.msg697268#msg697268 (reply #14)

Then I tried to put a jFET buffer (Buff'n Blend) in front of a FV-1. Also damn noisy ! I finally switched the FET circuit for TL072s, that is perfectly quiet (I have now more concerns with the noise of the FV-1 itself   )
http://www.diystompboxes.com/smfforum/index.php?topic=90776.msg773115#msg773115

Recently, when I proceeded my last order a Musikding, I ordered their kit version of ZV... Super Hard... :
http://www.musikding.de/product_info.php/info/p2639_The-HardOne---Booster-kit.html
It works good but... guess what ? Also very noisy ! A friend of mine has the original, he said he has no noise problem at all.

I tried all those circuits with different PSU, no change.

Is that some kind of curse ?  

Best regards.

Eric

[darron]

i've found FETs pretty good for noise! something's up....

[Galego]

It's because you touch yourself at night. 



I don't know, have you looked at how you do your PCB layouts, how you do your wiring, stuff like that?

[anchovie]

I tried all those circuits with different PSU, no change.

What kind of PSUs - cheap switch-mode, perhaps? Have you tried a battery too?

FET stages won't reject power supply noise - there's usually at most a resistor between V+ and the drain. If you don't get the noise with a battery then you should try a PSU designed specifically to provide clean juice to pedals.

[ayayay!]

It's probably because you bias everything to the suggested voltage, which often is 4.5v.  I think it's best to bias by ear.  We're talking 9 volts here guys, it's not gonna hurt much...  FETs are wildly different from batch to batch, or just from one another.   That's why there's so many FET tester threads.  

I think my drain on Q1 (MPF102) in my Dr. Boogey is 1.6V!  The Q2 & Q3 (J201) are about 3.2V.  (I remember all of this because I wrote the drain voltages inside the backplate.)  No one has told me yet that the pedal is noisy, but I'm often told it sounds great.  Just sayin...

[R.G.]

A professional detective would ask himself "What's common about all these builds?"

[DDD]

As per my personal experience, FETs produce less noise than resistors in the gate bias (or gate grounding) circuit. Frankly, I was surprised with this fact.
Try metal film resistors in the gate circuit and (or) try lower resistance gate bias circuit.

[ricothetroll]

Hi,
Thanx a lot for your answers !

It's because you touch yourself at night.  icon_lol

Damned ! I knew this was bad for electronics !

What kind of PSUs - cheap switch-mode, perhaps? Have you tried a battery too?

I have a generic PSU and a Boss PSA. Tried both, same result. I'll try those with my (homebuilt) lab PSU, based on LM317/337. Or add more filtering on the supply of the effects.

However yesterday I gave a try to the BSIAB2. It works pretty good, and the noise is acceptable for such a high gain pedal ! Now I'm even more confused... Why does this one work better ? I didn't even use screened cables.

It's probably because you bias everything to the suggested voltage, which often is 4.5v.

Bias didn't change anything to the noise level.

A professional detective would ask himself "What's common about all these builds?"

Right !

As per my personal experience, FETs produce less noise than resistors in the gate bias (or gate grounding) circuit. Frankly, I was surprised with this fact.
Try metal film resistors in the gate circuit and (or) try lower resistance gate bias circuit.

Here's a clue : my English Channel was built with carbon resistors instead of the metal films I generally use. I'll try changing them for metal films and see what I get.

But still : how can that booster be noisy ?
http://diy.musikding.de/images/stories/hardone/hardoneschalt.pdf
Bad specimen of BS170 ?

Best regards.

Eric

[ORK]

Wide bandwidth of the BS170. Can be helped with a little drain-gate cap of 22p to 220p. You seem to use a very bright rig?

[JDoyle]

It should be pointed out that MOSFETs and JFETs are different beasts entirely. In terms of noise, a MOSFET will always be noisier than both a JFET and a BJT. In general, a JFET is quieter than a BJT when fed from a high impedance source ('high impedance' is almost NEVER defined, but a good guess is anything higher than about 5k can be thought of as 'high impedance'), and a BJT is quieter than a JFET when fed from a low impedance source (again, 'low impedance' < 5k or there abouts, roughly; very, very roughly). And to repeat - a MOSFET will be noisier than both no matter what it is fed from.

Also - take a look at the circuit (with the BS170) you posted. I wouldn't call it a highly refined design. All of the noise inherent in the LED indicator portion from LED1, R5, D2, & D3 (reverse biased diode junctions are always a noisy choice) is NOT filtered out by C3, as it should be, but is instead injected directly into the top of the BS170's load resistor, R4. Also, R1 and R3 are entirely too big in my opinion and unnecessarily increase the noise. Same goes for R2, which is at least twice as big as it needs to be.

In my opinion, there are much better and productive ways to hook 14 parts together than is shown in that circuit.

While I can appreciate the desire for minimizing loading and whatever so-called 'tone-sucking' may occur, I think that the practice has been taken to such an extreme as to have created a problem in place of the one it corrected.

By creating circuits with one megaohm (or more!) input impedances, one is also creating a circuit which adds one megaohm worth of resistor noise in order to save an extra 0.1% of the original signal. And at the same time making the circuit that much more susceptible to stability problems, RF interference, etc.

High impedance is not a wholly good thing, too much can wreak havoc.

Personally, even if interfacing directly with my guitar, I rarely use input impedances greater than 100k; the noise you gain ain't worth the signal you save, especially considering that you almost assuredly will NOT hear a sub-1% change in level but certainly WILL hear the extra noise contribution of a 1 Meg resistor.

Anyway, sorry if I went a little off topic...

Regards,

Jay Doyle

[ricothetroll]

Hi,
I tried to supply my pedals with my lab PSU and... they are now totally silent ! As I tried two different supplies I though this wouldn't be the problem. Actually it was ! I'm feel really sorry for wasting your time...
Thanx a lot again for your answers. Jay : that's not off topic, it's actually a good clue, and a nice thing to remind when designing circuits !
Best regards.
Eric

[amptramp]

But still : how can that booster be noisy ?
http://diy.musikding.de/images/stories/hardone/hardoneschalt.pdf
Eric

Not too sure you want to plug in an external power supply with the polarity shown.  If you do, that may give you noise problems later on.  Note also that MOSFET's are very susceptible to electrostatic discharge punching through the gate oxide.  It may not destroy the device, but it may increase noise and you may never know if it has happened unless the device came to you in conductive foam and was never removed until it was used.

[ricothetroll]

Not too sure you want to plug in an external power supply with the polarity shown.

Right ! That's a huge mistake.... Nevertheless I did it the right way on my pedal 

Best regards.

Eric

[Earthscum]

Anyway, sorry if I went a little off topic...

Regards,

Jay Doyle

Actually, it was all pertinent and relevant info, for sure!

I've been using 2SK117GR's for awhile now. In the datasheet, it shows that (with these, anyways), anything above 1M starts getting noisy, as well as anything under 100k. That's actually a pretty darned good sweet spot. I've been digging on about 220k-330k for Bass guit, and 330k-560k for guit-fiddle. This is paralleled along with the anti-pop resistor preceding the cap (usually 1M gets rid of the pop for me).

The -GR's needs lower Drain Impedance, more along the lines of MPF102's (>5-6k). With higher drain resistor values, they can become quite noisy. J201's, on the other hand, are fairly quiet with the higher (10k-33k) values.

[jammybstard]

Hi,
I tried to supply my pedals with my lab PSU and... they are now totally silent ! As I tried two different supplies I though this wouldn't be the problem. Actually it was ! I'm feel really sorry for wasting your time...

I've had this as well recently; are your supplies Switchmode PSUs?

Professionally I work with video signals and at home I play with audio and the last couple of years with both I'm starting to see so many cheap SMPSs causing noise problems.

I'm starting to hoard transformers as I worry they'll be hard to get hold of one day!