Asking For Ground Loops?

[kurtlives]

Is this PSU design just asking for ground loops?
http://www.generalguitargadgets.com/pdf/ggg_reg_power_supply.pdf

Don't commercial PSU units that include daisy chains have issues with ground loops?

Thanks...Chris

[Cliff Schecht]

No, those are not ground loops. The entire system is "star-grounded", which helps to prevent power currents from leaking into your signal grounds (although they all do become common at some point).

[kurtlives]

O really nice!

So good or bad PSU would you say?

What is the advantage of having a separate x-former ands rectification system for each DC output jack?

[sjaltenb]

I believe the only way to completely avoid ground loops is to have a seperate transformer (or seperate winding) and power supply for each DC output.

For my 14 output power supply, I am using a 12 winding transformer, which outupts 12vAC to 12 completely isolated DC power supplies like you show. The ground for each power supply (which is passed along to the FX board) will all be star grounded to the enclosure independently with 18ga solid core wire.

So, the ground basically goes:

Star-to each PSU seperately- and that ground is passed to the FX board, so in the end, the FX board is directly connected to the star ground. = No ground loops.

I BELIEVE this is right, however, I have not yet tested it. This is how R.G.'s Spyder is setup.

[kurtlives]

Ok would I have any "issues" if I hooked up three of those PSUs....attach all the ground together...hook all the transformers primaries together?

[zyxwyvu]

Is this PSU design just asking for ground loops?
http://www.generalguitargadgets.com/pdf/ggg_reg_power_supply.pdf

Don't commercial PSU units that include daisy chains have issues with ground loops?

Thanks...Chris

Yes, and yes, though daisy chains will usually have less of an issue with ground loops than multi-output supplies, because their loops are smaller.

No, those are not ground loops. The entire system is "star-grounded", which helps to prevent power currents from leaking into your signal grounds (although they all do become common at some point).

While the power supply itself is star grounded, once you connect it to more than one pedal, the star grounding is compromised. RG has explained this quite well here:

http://www.geofex.com/Article_Folders/Spyder/Spyder.htm

Basically, a loop is formed through the power cable ground to effect 1 -> a 1/4" cable between effects -> a power cable ground to effect 2 -> power supply again.

What is the advantage of having a separate x-former ands rectification system for each DC output jack?

Having a separate transformer (or secondary) and rectifier/regulator sections cuts the ground loops at the power supply end. All the grounds are still connected, through the grounds in the 1/4" cables, and through each pedals input, output, and power jacks.

I believe the only way to completely avoid ground loops is to have a seperate transformer (or seperate winding) and power supply for each DC output.

For my 14 output power supply, I am using a 12 winding transformer, which outupts 12vAC to 12 completely isolated DC power supplies like you show. The ground for each power supply (which is passed along to the FX board) will all be star grounded to the enclosure independently with 18ga solid core wire.

So, the ground basically goes:

Star-to each PSU seperately- and that ground is passed to the FX board, so in the end, the FX board is directly connected to the star ground. = No ground loops.

I BELIEVE this is right, however, I have not yet tested it. This is how R.G.'s Spyder is setup.

This is almost correct. The idea with using multiple secondaries is to have no connection between grounds at the power supply. If you connect them all inside the power supply, even if it is a star ground, you will have ground loops. RG's spyder has all of its outputs completely isolated.

[sjaltenb]

This is almost correct. The idea with using multiple secondaries is to have no connection between grounds at the power supply. If you connect them all inside the power supply, even if it is a star ground, you will have ground loops. RG's spyder has all of its outputs completely isolated.

I am now confused. If the transformer has seperate windings, like the Weber does in R.G.'s spyder, or like a Voodoo Lab PP2,  these windings feed the PSU, and each PSU has its OWN wire to star ground, and each PSU only powers 1 effect, how can a ground loop form?

I can see how if one PSU powers multiple effects a ground loop would form (daisy chaining), but not if it only powers 1 effect. 

thanks!!!

[zyxwyvu]

I am now confused. If the transformer has seperate windings, like the Weber does in R.G.'s spyder, or like a Voodoo Lab PP2,  these windings feed the PSU, and each PSU has its OWN wire to star ground, and each PSU only powers 1 effect, how can a ground loop form?

I can see how if one PSU powers multiple effects a ground loop would form (daisy chaining), but not if it only powers 1 effect. 

thanks!!!

I might be misunderstanding you. Does each secondary/rectifier/regulator section have it's OWN star ground? Or are they all connected to one star ground?

[sjaltenb]

Well, I'll try to "draw" it out in my mind...

120 mains in, with 3rd prong forming A Star Ground.

AC Primaries into a 12 secondary transformer, so, a dozen 12VAC outputs.

Each transformer secondary connects to an independent PCB which contains the rectifier, regulator and filter and output +9vDC and a ground to each FX board.

EACH of these Power supply boards is grounded independently to the Star Ground with a seperate wire. This way the ground is transferred to the effect through the independent outputs.
_________________

RIght?

[zyxwyvu]

Well, I'll try to "draw" it out in my mind...

120 mains in, with 3rd prong forming A Star Ground.

AC Primaries into a 12 secondary transformer, so, a dozen 12VAC outputs.

Each transformer secondary connects to an independent PCB which contains the rectifier, regulator and filter and output +9vDC and a ground to each FX board.

EACH of these Power supply boards is grounded independently to the Star Ground with a seperate wire. This way the ground is transferred to the effect through the independent outputs.
_________________

RIght?

You do not need to connect the power supply boards to any separate ground. They should be floating with respect to ground, at least until they are connected to effects. It should be as if you took multiple complete power supplies, and just put them in the same box, completely independent of each other.

If you haven't already, take a close look at RG's diagram at the top of this page: http://www.geofex.com/Article_Folders/Spyder/Spyder.htm.

If the ground connections are connected at all inside the power supply, then ground loops will be formed by the pedals and their cables.

[sjaltenb]

Ahhh I see.

Ok. Thanks, Sorry I hope I did not hijack the thread!

[petemoore]

  If the power supply is floating, there is no ground connection at the power supply.
  The 1 ground is made by the shielded signal cable.
  With 2 grounds, one of them might be closer to 'ground', the voltage difference between 'them grounds' = noise.

[sjaltenb]

Ok so if the ground connection is NOT being made by the signal cable, such as in a multiFX board...the power supply must be connected to ground, which you can then carry to the Effect... right?

[petemoore]

  If the power supply is floating, there is no ground connection at the power supply, no chance for ground loop to rear it's ugly head and decide to make noises.
  The 1 ground is made by the shielded signal cable.
  With 2 grounds, one of them might be closer to 'ground', the voltage difference between 'them grounds' = noise.he signal voltage is referenced to ground.
  The signal ground connection is made through the sleeves / shielded ground of the guitar cable.
 

[zyxwyvu]

Ok so if the ground connection is NOT being made by the signal cable, such as in a multiFX board...the power supply must be connected to ground, which you can then carry to the Effect... right?

If you don't have a ground connection directly between your pedals (as could be the case in a multi effect box), then you would have to connect the grounds at the power supply.

I think it is a better idea to connect the grounds between effects than in the power supply, though, to get a more stable ground across effects. In any case, I doubt ground loops will be much of a problem if your effects and power supply are so close together.

[sjaltenb]

Ahh I think I have a better understanding now!

Thanks for both ya'lls help!

[Cliff Schecht]

No, those are not ground loops. The entire system is "star-grounded", which helps to prevent power currents from leaking into your signal grounds (although they all do become common at some point).

While the power supply itself is star grounded, once you connect it to more than one pedal, the star grounding is compromised. RG has explained this quite well here:

http://www.geofex.com/Article_Folders/Spyder/Spyder.htm

Basically, a loop is formed through the power cable ground to effect 1 -> a 1/4" cable between effects -> a power cable ground to effect 2 -> power supply again.

This is something that most people need to just learn to live with. Star grounding is quite tricky in a guitar-->pedal-->amp setup, especially when you have to at some point connect all three of these devices together. The "ideal" solution is to have completely separate signal and power grounds down to one point, but it'd take some careful planning in a special setup to pull this off realistically. My noise problems aren't this bad, so I'm not going to bother.

But to really answer the initial question asked... I prefer this power supply:
http://www.generalguitargadgets.com/diagrams/ultra_clean_ps_sc.gif

Functionally this and the version you posted are very similar, just make sure to have enough load capacitance to make the part unconditionally stable (which usually isn't a problem with linears).

And to comment on the multiple transformer based option: I like this idea myself and since the input and outputs can all be wound on one core, it's a very attractive option. It's not necessarily cost prohibitive either, I just don't think anybody has taken the time to properly design something of this nature. What would be really cool is if somebody designed an SMPS that could generate all of the rails without such a bulky transformer. If designed properly, it would be darn near as quiet as any of the linear designs I've seen.