Can we please keep it quiet in here!

[erick4x4]

Ok, as I have seen so many threads about a variety of topics with noise, pops, clicks, crackle, etc. I thought it would be nice to have a thread dedicated to all things noise.

This is much for me as well as all the other learners out there, maybe a place to start if you are having excessive noise, or even just fine tuning as I have been lately. I got so excited when I started I have 20-30 pedals, in my music room, and noticed that pedals I built early (when I was much sloppier, and less experienced) are consistently noisier, and I began to realize what some of the little things, and build quality did to the noise.

Here is what I have learned (and could be wrong on), please everyone jump in.

1) Always use a shielded enclosure to keep the noise down.
2) Star Grounding seems to help (though not always, might be a myth?)
3) While great for testing, sockets, trimpots, and switches can often add noise, and you may consider taking them out when you are done.
4) Using too bright of an LED (or not a big enough resistor anyway), can add popping when you switch.
5) Putting a Really Big Capacitor on the power rail can help eliminate hum if it isn't on your circuit already.
6) Some pedals (like a rangemaster, may never be quiet enough with a power supply, and you really may be best with a battery)
7) Using Shielded wire on gain pots, and even volume pots might help on very high gain circuits.
Using the least amount of wire in your enclosure rather than overly long strands, can help with noise.
9) Pedals with inductors can be sensitive to proximity and angle with power supplies.

Well that's about the extent of my very limited knowledge, please submit your ideas. I would love some thoughts on layout and proximity of parts in a pedal, I know this can matter, but I don't know how!

[David]

That "really big capacitor" should actually be a 100R resistor, followed by the "really big capacitor" across the power supply leads.  I've done this on several pedals.  No noise.  Nada.

[John Lyons]

1) Always use a shielded enclosure to keep the noise down.
Almost always true. Good rule to follow...

2) Star Grounding seems to help (though not always, might be a myth?)
It's a good rule to follow and it can make a big difference. No myth! Some circuits are more sensitive to it.
In tube amps this is a bigger rule to follow since the voltages are higher and there are higher currents. For pedals it's not as important.

3) While great for testing, sockets, trimpots, and switches can often add noise, and you may consider taking them out when you are done.
Usually they are fine to keep in the circuit. I have had some that were noisy in critical high gain builds. To many sockets is a problem waiting to happen though. When you get something you like there isn't too much reason to keep building with tons of sockets.

4) Using too bright of an LED (or not a big enough resistor anyway), can add popping when you switch.
Using an LED can cause poping but I'm not sure about the brightness or resistor value thing...

5) Putting a Really Big Capacitor on the power rail can help eliminate hum if it isn't on your circuit already.
Using a 100 ohm (47 -100ish) resistor in line from the DC jack or battery and a large cap can make a pedal quieter as David noted above.
Usually DC supplies have a 1000uf cap across the power supply but a local cap of 100uf and series resistor to power is a good idea to keep hum and PS noise out.

6) Some pedals (like a rangemaster, may never be quiet enough with a power supply, and you really may be best with a battery)
Sometimes yes. Using germaniums devices is a little tricky and batteries do better sometimes.

7) Using Shielded wire on gain pots, and even volume pots might help on very high gain circuits.
This always helps but sometimes it's not noticeable. The higher the gain/impedance the more this will help.
A decent linear layout from one side to the other on the board along with short wires as much as possible with work quieter and with less chance of oscillation. Most distortion and overdrive circuits benefit here.

8)Cool Using the least amount of wire in your enclosure rather than overly long strands, can help with noise.
Yes, always

9) Pedals with inductors can be sensitive to proximity and angle with power supplies.
Yes, that's true, same as guitars pick ups. Putting your pedals right next to or on the same circuit as SCR dimmer lamps, Wall warts, electrical lines or florescent lights (in the basement or first floor) is a good way to pick up noise as well.

There are a lot of "good practice" things to keep in mind. Partially why boss etc pedals seem so overly full of parts. They are trying to coverall the basses with the conditions people might use their pedals in. It's better to over design something and hope that the bad situation doesn't come up. Practicality and frugality is why we skimp on these until problems come up.

John



John

[frankclarke]

Components are a major source of noise too, large value resistors, germanium trannies and so on. And some individual components are "bad" and noisy.

[Gus]

3
sockets and trimpots will have a better chance of being a problem if the effect is used on the road.

6
A good power supply should not be an issue.

run signal wires at 90 degrees if possible for min coupling (if that is what you want TW )

[brett]

RE: 6
Could it be that in some circuits the internal resistance of a battery (higher than a regulated supply) lowers the gain and gives an appearance of less noise?

[mac]

9. True.
Having 2 tv stations 100mts away from my building, the only way to avoid noise when playing is to lie on my bed and pointing to the north! Uncomfortably numb...

mac

[Processaurus]

Components are a major source of noise too, large value resistors...

Some designs use higher impedances than are neccessary throughout the circuit, I've tried the same gain (same ratio of resistor values) opamp stage in a hi gain distortion pedal, with higher and lower value sets of resistors, and noticed less hiss and better stability with the lower value resistors giving lower impedances.  Lower impedances also lessen potential crosstalk from other parts of the circuit.

Opamp inputs are very sensitive, feedback resistors should be as close as possible in the layout to the opamp, and if the input needs to be wired to a pot, shielded cable is desirable. 

[petemoore]

  Gain stage = amplification.
  Amplification = Amplification of any available input noise.
  Inputs often amplify signals which are then re-amplified, a great place for a little noise to get in and become a lotta noise.
  This of course includes all wires between the pickup magnets and the effect box.
  special case:
  Compressors are noistorious for helping any noise to become present in the mix..*when input signal drops, if there's any 'dirt on your noise floor', it's going to be 'looked at through a magnifying glass', because it will *turn up the amplification gain then, what it does.
    

[R.G.]

There are many full-blown postgraduate level EE textbooks on the subject of electronic noise. You can in fact spend a whole EE career on noise reduction.

It helps to establish the taxonomy. What you're calling noise is made up of at least:
1. power line hum
2. DC power supply ripple
3. thermal and active-device hiss
4. RF pickup
5. RF oscillation
6. bad contacts in the build
7. design and operational problems, including wiring problems

Taking them one at a time:
1. Power line hum
We live in a sea of emitted AC power line hum. If you leave an opening for it to get in, it's there and waiting.

Power line hum gets in by the paths of (a) radiation and (b) conduction. Radiation is easy to fix - seal up your circuit inside a grounded metal enclosure. This is the slam-dunk solution for radiated hum. You ...may... be able to get by with less if you do everything else right in terms of wiring, grounding, frequency response, etc. etc. It's far simpler to use a metal enclosure. For conduction, there are two paths - the shielded signal wire and any wall-plugged DC power supply. Usually it's the difference between the DC power supply and the amplifier's "ground" voltage that makes for hum. Sometimes it's ground loops. Ground loops you fix with attention to the difference between signal and DC power grounding (See "The Spyder Power Supply" at GEO). Never, ever solve ground loop problems by lifting the AC power ground on your amp. Better to have hum than be dead, which is one possibility. There are whole books written on the subject of ground loops.

2. DC power supply ripple
If you're going to feed your effects DC from an external source, it had better be free of ripple from the rectification of the AC power line. This is only an issue because the majority of wall-wart power supplies that claim to put out 9Vdc have little or no filtering in them unless they were made explicitly for powering guitar effects. Uninformed guitarists plug in anything that says 9Vdc on it and expect it to work. Wall warts that are intended for guitar are well made, regulated and quiet. Otherwise, unless your pedal has internal filtering it can be expected to hum. The "hum" is actually at twice the power line frequency, one octave up from the bass hum that power line ground noise is. Some people can hear the difference, some can't. It is good practice to put the 100 ohm resistor/100uF to 1000uF cap inside each pedal to preclude the problem.

3. thermal and active-device hiss
Again, there are whole textbooks on this one. Every resistor produces noise proportional to its resistance and its temperature in degrees Kelvin (i.e. with respect to absolute zero; room temp is 273 degrees Kelvin). The higher the resistance, the bigger the noise. The higher the temperature, the bigger the noise. For low noise, use low value resistors. No problem - except that guitar pickups have to have resistances in the 1M region to not lose a lot of treble. So keep the resistance low everywhere except where the guitar pickups drive it. Many classical effects are poorly designed in this regard, and need redesigned to lower impedance levels to get less hiss. Some resistors produce excess noise, above and beyond the unavoidable thermal noise; mostly, these are carbon comp resistors. This is the reason I chuckle every time some grinning new "effects designer" advertises using carbon comps in a pedal. It's the same as saying "I don't know anything about carbon comp resistors!" The cure in cases where all you can do is replace the same value resistor is to use metal film, which have only thermal noise.

Active devices hiss proportional to their internal construction, temperature, and operating conditions. Again, whole textbooks on low noise semiconductors. Modern silicon devices are actually quite quiet when used at an appropriate impedance level and operating current. Germanium in particular is prone to excess noise caused by lack of surface passivation on the chip and by poor processing. There's no fix for a noisy germanium. Put in a different one or live with the noise.  Active devices can get extra noisy when electrically abused. See "When good opamps go bad" at GEO for how and why, as well as some fixes.

4. RF pickup
Many textbooks. Same cures as AC power line radiation, except that radio waves can creep in on the incoming/outgoing signal cables if the shielding is not perfect. Some RF situations - like living between high power transmitting antennas - can't be helped. The only solution is to go somewhere where there's not so much RF in the air.

5. RF oscillation
If your circuit is poorly wired, with high gain, high impedances, excess unshielded wire, it can make itself oscillate at RF. This causes either no sound (it's busy making RF) or loud, angry hissing, or harsh sounding audio, or a combination of the above. Shielded internal wires and better wire routing can fix this. You have to know what each wire does to go a good job of this that's more than easter-egging.

6. bad contacts in the build
Bad contacts introduce high and variable as well as nonlinear resistances where there should be only linear conduction. Cold solder joints are one way, poorly contacting sockets are another. Old, dirty switches, jacks, connections to the metal enclosure, etc. Contacts should CONTACT.

7. design and operational problems, including wiring problems
These are the ones you've done to yourself. Any cap that is switched simply MUST be pulled to the proper DC condition both before and after switching by a high value resistor, or you'll get a pop. Cap switching should only short out the high value resistor which holds the cap at the right DC condition. LED popping can be pernicious. It's caused by the sudden change in power supply current or sudden change in voltage across the LED when the switch closes or opens. You can fix this by either (a) making sure the current transient does not flow through wires shared by the input signal of the effect or (b) slowing down the turn on/off so it cannot radiate that sharp edge, or both. The common practice of using the input jack to switch all DC power to the effect is the ...worst... setup for making something like LED popping cause ground noise.

[mountainking]

Got a question RG. I always use shielded wire for all connections from the input jack to the input of the circuit when building high gain effects to help reduce outside interference from entering the signal path, but I keep asking myself, "what is the shielded wire doing that the shielded enclosure isn't already doing?"

[foxfire]

i'm by no means R.G. but, i believe that the shielded wire is helping to keep interference from the rest of the pedal getting into the signal. or at least that's how i think of it.

[mountainking]

i'm by no means R.G. but, i believe that the shielded wire is helping to keep interference from the rest of the pedal getting into the signal. or at least that's how i think of it.

Yeah I understand that, but I was referring more to things like rf interference. I also realize that if the cable going into the pedal isn't shielded it like having a long antenna connected to the input of the circuit. My designs are very quite(noise-wise), even when on the breadboard, considering the huge amount gain they have, but I'm just wondering if I'm doing something unnecessary in my construction. Maybe I should rephrase my question.

Does using shielded hookup wire help to prevent rf interference if the circuit is build into a shielded enclosure?

[miqbal]

Ok, as I have seen so many threads about a variety of topics with noise, pops, clicks, crackle, etc. I thought it would be nice to have a thread dedicated to all things noise.

This is much for me as well as all the other learners out there, maybe a place to start if you are having excessive noise, or even just fine tuning as I have been lately. I got so excited when I started I have 20-30 pedals, in my music room, and noticed that pedals I built early (when I was much sloppier, and less experienced) are consistently noisier, and I began to realize what some of the little things, and build quality did to the noise.

Here is what I have learned (and could be wrong on), please everyone jump in.

1) Always use a shielded enclosure to keep the noise down.
2) Star Grounding seems to help (though not always, might be a myth?)
3) While great for testing, sockets, trimpots, and switches can often add noise, and you may consider taking them out when you are done.
4) Using too bright of an LED (or not a big enough resistor anyway), can add popping when you switch.
5) Putting a Really Big Capacitor on the power rail can help eliminate hum if it isn't on your circuit already.
6) Some pedals (like a rangemaster, may never be quiet enough with a power supply, and you really may be best with a battery)
7) Using Shielded wire on gain pots, and even volume pots might help on very high gain circuits.
Using the least amount of wire in your enclosure rather than overly long strands, can help with noise.
9) Pedals with inductors can be sensitive to proximity and angle with power supplies.

Well that's about the extent of my very limited knowledge, please submit your ideas. I would love some thoughts on layout and proximity of parts in a pedal, I know this can matter, but I don't know how!

Hi erick,

nosie is still a mystery for me too, to be honest.
Btw, from a newbie point of view like me: Metal Zone MT-2 (Is this a bad sample?? ) is probably one of the phenomenal hi-gain solid state circuit ever made under the sky of the earth, but I almost (or even) never  seen any shielded cable,etc. We can sometimes see a lot of jumper spreaded all over the board (probably more than fifteen jumper on an MT-2 board), so it almost impossible to implement a 'text book' star grounding, left to right signal flow, or some thing that. <<--need corrections by all diy-guru here Some says the IC used is not the best low noise type (M5218L), and they use cheap resistor too (5% tolerances), but still.... very quiet.

Sorry, I did not mean to mention that those tips are wrong, but I always wonder: Is there any secret  from the factory to make such a hi-gain pedal so quiet? I bet, if you built one exactly the same pcb size, the same cable, the same enclosure, but you buy your own components on your own local market, the result will be different.
Well, I plugged it right into my sound card, and no noise at all. It makes me  sometimes think that  it is still better 'buying' than 'diy-ing'.
Just an opinion, hope some one will open up my mind 

IQB -  Jakarta, Indonesia

[R.G.]

Does using shielded hookup wire help to prevent rf interference if the circuit is build into a shielded enclosure?

No, once you're inside a shielded enclosure, the shielded wire doesn't help with RF pickup. As mentioned, the purpose of shielding there is to keep the circuit from coupling back to the high impedance input node.

[R.G.]

We can sometimes see a lot of jumper spreaded all over the board (probably more than fifteen jumper on an MT-2 board), so it almost impossible to implement a 'text book' star grounding, left to right signal flow, or some thing that.

Actually, jumpers are used on single sided boards to prevent going to double sided PCBs. Think of them as PCB traces that didn't quite fit on the bottom. The fact that there are jumpers on a board say nothing at all about whether the board is star grounded or has good signal flow. No indication whatsoever.

Some says the IC used is not the best low noise type (M5218L), and they use cheap resistor too (5% tolerances), but still.... very quiet.

There is a lot more in those textbooks I mentioned than just which IC, how to wire, etc. A big one is to limit bandwidth. Other things being equal (which they never are!) noise voltage is proportional to the square root of the bandwidth. So limiting bandwidth limits noise as well. Noise is also impedance sensitive, and sensitive to bias currents and voltages. There is a LOT about noise that I just can't type in here. And in fact, very little that a DIYer can do about the things that take the longest to write about.

Sorry, I did not mean to mention that those tips are wrong, but I always wonder: Is there any secret  from the factory to make such a hi-gain pedal so quiet? I bet, if you built one exactly the same pcb size, the same cable, the same enclosure, but you buy your own components on your own local market, the result will be different.
Well, I plugged it right into my sound card, and no noise at all. It makes me  sometimes think that  it is still better 'buying' than 'diy-ing'.

Your mind is plenty open already. The factory can do something that not many DIYers can do: have qualified, trained engineers work on all the aspects of noise during the circuit design. They can take the time to calculate best impedances for low noise, do careful power supply decoupling, cut the bandwidth to just what's needed, etc.

I should have been really clear in my post - what I wrote and the tips here are just the barest outer surface of designing for low noise. There is much more to the total subject than I could spend the time typing in.

[mac]

Good lecture about propagation of EM waves in conductors. Or "can an EM wave penetrate an enclosure or shielded wire?".

http://farside.ph.utexas.edu/teaching/em/lectures/node102.html

mac