Wiring advice

[stefcuypers]

Hallo,

I'm building a 19 inch pedalfx wall. Below is an Hammond DD box that I will use.
Pots, rotary switch and jack are mounted on the side of the box.
Rotary switch selects one of the 4 effects effect. (In this example it are 4 boosters).
Input and output are shielded wire. ( the strange black wire coming out of the red or blue wire is the shielding going to ground)
Every effect has it's own power supply filtering.

I really want to build it the most silent way possible. I prefer "overkill", just to be sure.
Are there any improvements over these way of wiring? Or any mistakes?
Advice would be fantastic.



Thanks,

Stef

[MikeH]

I wouldn't tie all of the outputs together like that; it will most likely load your signal down a lot.  I'd use a '2P4T' or a '2W4T' (same thing) rotary switch to fully disconnect each circuit from the signal path. 

[potul]

Yes, hooking together all outputs is a bad idea. Better use a 3P4T and do true bypass, and you can add LEDs as well.

[stringsthings]

as MikeH posted, it would probably be best to use the same rotary switch to select the output .... ( i.e. the switch position selects the input and the output from the same booster ) ....

or you could install a simple mixer at the end of the signal chain ....

[stefcuypers]

Thanks a lot for the reply's! Will try to find out to install an 3P4T.

Are there any improvements over the grounding and 9v+ part?

Thanks!
Stef

[MikeH]

I think your power section is fine- just be sure to avoid ground loops.  Technically you have one between you input and output jack shields, but the only way to avoid that is to not run a ground wire to the sleeve of your output jack.  But that can become a problem if the jack gets loose.  I run a wire to mine, it's a trade-off, but I think it's more important to ground the jack.

[stefcuypers]

Hello,
Thanks all for the reply's.
I read last days a lot about wiring and star grounding.
I reworked the wiring. Is this an improvement? I put a ground wire to the chassis for the input ground.
So there isn't an ground loop any more. All grounding wires are going to the ouput jack, because there is the "real" grounding someone said.

If there are more improvements possible, please post them.
Thanks a lot.

[Pablo1234]

I personally Like using the Ground plane technique over star grownding when hooking up lots of devices. Takes allot more wire and can turn ugly at times but grown loops are virtually eliminated. If I am doing mixed signal Creating a ground Plane is by far superior to star grownding.
Basically creating a grownding plane is the same as star, but every device ground would have a ground wire to every other device ground and each ground wire should be equal in length.

Input jack - output jack, power supply jack, fx 1, fx 2, fx 3, fx 4
output jack - power supply, fx 1, fx 2, fx 3, fx 4
fx 1 - fx 2, fx 3, fx 4, power supply
fx 2 - fx 3, fx 4, power supply
fx 3 - fx 4, power supply
fx 4 - power supply
find the longest ground wire and make all others = example 6" x 25 wires = 75' of wire just for ground, but each and all ground loops will be diminished to a point where they can't even exist.

I know it sounds crazy but its a sure way to use mixed signals and lots of LFO's along with Analog signals.

[R.G.]

I personally Like using the Ground plane technique over star grownding when hooking up lots of devices. Takes allot more wire and can turn ugly at times but grown loops are virtually eliminated. If I am doing mixed signal Creating a ground Plane is by far superior to star grownding.

There are bits of reality in there. I like ground planes too, but they're not a panacea. For one thing, they can't be used well when you have more than one circuit board. 

There is a continuum of grounding. At low frequencies (audio qualifies!) star grounding is a good way to keep ground return current from causing interference between sections. As frequency rises into the radio range, this gradually shifts to a ground plane being a better solution because of the inductive effects of isolated wires. So yes, I can fully believe that you do like ground planes with mixed signal circuits, and it is in fact a superior solution. The ground plane keeps the fast edges of the digital side from getting into the audio side by providing fast, low inductance local ground loops for decoupling. Ground planes let the ground currents flow wherever they want to, and for high frequencies this is generally a big advantage in keeping cross contamination of signal down.

However, star grounding for low frequencies forces the ground current to flow where YOU want it to, not where it wants to. This is a big advantage in keeping ground currents from causing I*R drops which other stages with low or no ground noise rejection from amplifying the ground noise.  Star grounding is not perfect by any means, nor is it the only way to get quiet audio ground distribution. It's just the only way that is known ahead of time to be able to do it correctly without experimentation.

Basically creating a grownding plane is the same as star, but every device ground would have a ground wire to every other device ground

I think I disagree a little with that. The whole point of star grounding is that there is NOT a cross-wire between every ground point. So it is impossible for the ground currents to flow from place to place. The wire paths force them to flow back to the power supply first, and this forces the I*R drop on the ground wires to NOT be presented to other circuits as AC ground offsets.

You are correct in that a low frequency visualization of a ground plane is connecting every ground point to every other ground point with a wire; however that is not what star wiring is about.

I know it sounds crazy but its a sure way to use mixed signals and lots of LFO's along with Analog signals.

It is, but it's a lot more work than just star grounding the various sections any time RF signals or fast logic edges are not present. And star grounding is also a sure way to use lots of LFOs along with analog signals.

I reworked the wiring. Is this an improvement? I put a ground wire to the chassis for the input ground.
So there isn't an ground loop any more. All grounding wires are going to the ouput jack, because there is the "real" grounding someone said.

Here's the problem. The chassis is metal. If you use metal-bushing jacks for input and output, and also "ground" the power supply to the chassis, you are no longer ground loop free.

The point of star grounding is that there are NO secondary loops of "ground" conductors between points that have current return to ground. What you've drawn is not star grounding, although it's a common misconception of how to do it. However, the real grounding requirements on low power audio circuits with lowish gain are so lax that you can do almost any grounding scheme and often get away with it. But it is getting away with an error, not doing it well.

Here's how to star ground your box.
1. Make exactly 1.00000 ground wire connections to the metal chassis. Ideally, this would be a single wire so you could in concept use an ohmmeter and read continuity from your power supply ground terminal to the metal chassis, then disconnect that one wire and read open circuit. That makes it impossible for currents to use the chassis to circulate through the metal chassis to get into your signal circuits. There is no circuit for the circulating currents to circulate in.
2. It is OK to use either the input jack or the output jack as that one ground point for the chassis. But it's important to make that a solid connection with something like a star washer because over time the nut fixing the jack may come a little loose and get corrosion under it. When that happens, you can get hum.
3. Most people use metal bushing input and output jacks and so unknowingly create a current loop between them. Mostly with low power circuits you get away with this. But it is getting away with it, not good design. On the other hand, if you get away with it, that's fine.
4.Run ONE ground wire from each circuit section to the power supply ground star point. Ideally this is the point which should run a wire to the chassis, too. But that requires isolated-jack bushings for the input and output jacks.
5. Make sure the input and output jacks have ONE ground wire each. There are alternatives on where this goes. In the ideal case, this wire goes back to the circuit which is reading from or driving the jack. That requires insolated-bushing jacks. You could take it back to the power supply star ground. Better is taking the input jack to a single buffer circuit, and then distributing that buffered signal to the various circuits. Now the input jack ground can be isolated at the chassis, go directly to the buffer (which has its star ground wire) and the buffered signal is distributed as a low-impedance version of itself. This is much quieter and less prone to interference. Something similar would be ideal for the output jack. However, the output signals from effects are usually low impedance and not susceptible to the same problems as high impedance input signals. A single wire from an isolated output jack back to the star point works well. And this can be the chassis ground connection too if you don't like isolated jacks.
6. If you do isolate the input jack bushing, connect a 0.001uF ceramic cap from the jack ground to the chassis right at the input jack. This shunts RF into the chassis there and keeps it out of the audio circuits. At RF, ground planes do matter, and this forces conducted RF into the chassis, making it effectively a Faraday shield for the audio circuits, which is why you have a metal chassis in the first place.

As a side note, someone will pop up here and complain about buffers and fuzz face pedals. Yeah, yeah. It ain't perfect because the FF is an oddity. There are ways to deal with it. I'm trying to explain star grounding to a newbie, and he doesn't need the complications til he gets the basics down.

As I said earlier, there are many ways to get away with imperfect grounding because of the forgiving nature of low power, low frequency, low gain audio circuits. About the only part of that you're likely to violate may be low gain, so you can still get into grounding problems if you stack up several high gain distortion monster pedals in there; possibly also if you violate low frequency by using a defective circuit which oscillates at RF.

I'm sure I've been confusing in places here. Questions?

[Pablo1234]

Right on R. G.

I did not mean to imply that plane grounding and star grownding are the same, their really quite diffrent. I meant to say that you avoid ground loop problems with it as you do with star grounding.

That said, the reason I like plane grounding is mostly due to me using mixed signals, but he did also ask for over kill and plane grounding will do a superb job if he keeps the wires the same length. IR drops divide in parallel so for every device you add you get one more drop, sounds like a bad thing but it brings the drop across each down exponentially.

say you have a ground loop issue that has a IR drop of 1, if you add another equivalent IR drop in parallel it becomes .5 and a 3rd would bring it down to .33 and so on. so for every device with a grown you decrease the drop by a whole lot.

Star grounding is best for audio frequencies simply because you can predict your ground paths and its a whole lot simpler.

[stefcuypers]

R.G. and Pablo1234 thanks a lot for the amazing reply's.
It took me 2 hours to understand a little about it.
Guess grownding plane is a little too complicated for me.
I tried to implement the suggestions of R.G. for the star ground wiring.
Is this the way to go?

Here's how to star ground your box.
1. Make exactly 1.00000 ground wire connections to the metal chassis. Ideally, this would be a single wire so you could in concept use an ohmmeter and read continuity from your power supply ground terminal to the metal chassis, then disconnect that one wire and read open circuit. That makes it impossible for currents to use the chassis to circulate through the metal chassis to get into your signal circuits. There is no circuit for the circulating currents to circulate in.

This one isn't clear for me. Would you mind to try to explain it a little further?

Thanks a lot for this amazing advice.

[R.G.]

Well, this took a while.

That actually looks pretty good. I'm guessing that the block in the upper right corner is the power supply inlet. If so, there is one "ground" wire from the power source to each circuit board, and only one wire from the power source to the chassis. This ensures that the power current from more than one pedal does not travel in the same wire, causing voltages that make both circuits/pedals think ground is moving around. There is one wire from the power supply to the chassis, and the input jack has a low-value ceramic cap to the chassis at the input to allow RF from the guitar-cord-antennas (!) to be passed to the metal shielding chassis. This is all good.

The grounds from the input and output jacks are attached to the single power supply ground point. This will probably work fine in this setup.

However, grounding gets equivocal at this point. Each of the black wires from the circuit boards to the power supply point carries the circuit's own return currents, and so there is a voltage, however small, from the circuit's operation caused by the circuit's own ground return current through that wire. The input and output jack ground wires have (essentially) no current in them, so they have no voltage across them. So each circuit board "sees" its signal/reference ground voltage on the PCB as different from the input and output jack grounds by the amount of voltage caused by its own return currents.

In a big system with large currents flowing through the wires, this could cause problems. Ideally, the signal/reference grounds should be separated inside each circuit board, and the signal/reference grounds returned to the jacks and/or power supply "ground" by separate wires. That would stop the issues with ground currents in the power return wires on each circuit board. But this is not a big system. The currents are small, and the wires are short. Even more importantly, the individual circuit boards were not designed with power ground separated from signal/reference ground, so the internal separation is impossible without a lot of circuit cutting.

So I'd say, go with it this way. It'll work a lot better than most other alternatives.

[stefcuypers]

Thanks a lot for the verification R.G.. Really appreciate the advice.

[GGBB]

Hope I'm not intruding here, but I don't understand how the .001uF cap has any impact since the path I've highlighted in green below appears to short it out:

[PRR]

> I don't understand how the .001uF cap has any impact since the path I've highlighted in green below appears to short it out:

Re-read this post:

6. If you do isolate the input jack bushing, connect a 0.001uF ceramic cap from the jack ground to the chassis right at the input jack. This shunts RF into the chassis there and keeps it out of the audio circuits. At RF, ground planes do matter, and this forces conducted RF into the chassis, making it effectively a Faraday shield for the audio circuits, which is why you have a metal chassis in the first place.

Audio can run around the bush with little harm.

RADIO waves (we can't get away from them) should "often" be taken to large metal chassis ASAP.

The path you high-lighted is a near-short for DC and audio but has significant inductive impedance at many MHz and would be a great antenna for high TV band signals. It is liable to spray external radio waves all over your audio circuits.

0.001uFd is 10K at the top of the audio band , 100 ohms at the top of the AM band, 2 ohms in the FM band.

[GGBB]

The path you high-lighted is a near-short for DC and audio but has significant inductive impedance at many MHz and would be a great antenna for high TV band signals. It is liable to spray external radio waves all over your audio circuits.

0.001uFd is 10K at the top of the audio band , 100 ohms at the top of the AM band, 2 ohms in the FM band.

Thank-you Paul!  I've been educated!