Power Supply for DIY multi effect

[rockhorst]

I'm constructing a multi effect for a friend. I would contain:

Buffer > Flatliner Comp > Cardinal Tremolo > Phase 90 > Timmy > Plexi Drive > Arion SCH-1 Chorus > PT2399 Delay

For most of these I've designed a PCB (except for the trem). The guy would like to power it with a single power supply, like a 1SPOT. Therein lies a problem. Most of the effects play nice with each other on a daisy chain. The chorus, however, does not. Daisy chaining it with anything else introduces a pulsating whine into the sound. You can minimize it by dialing the tone knob on the Chorus all the way back, but that's hardly what you want to do.

So, I'm looking for suggestions. Should I build a custom power supply (a first for me) into the multi effect, including transformers to isolate the outputs? I saw a project that could do this on General Guitar Gadgets. An alternative I see to this is just house a commercial unit inside . Or would using some regulators on the separate chains be enough to keep the signal clean?

Also, would grounding all the effects to the same enclosure defeat the separation between them?

Any and all suggestions welcome! Thanks!

[GibsonGM]

Hi Rock,

Here's a LITTLE info, and I'm sure you're going to have more ppl chiming in with better for ya!

Have you tried to find out why the chorus causes the problems?  Sounds to me like there could be a decoupling issue there - and often, a simple R-C filter at the power input will fix that!    I try to keep things simple if I can, not have to re-invent the wheel for something like what you're doing...I would think you should be able to "plug and play" most any suitably rated power supply in there with no bad results.

Nothing wrong with DIY'ing the power supply, the GGG unit might be what you need - if that's what you want to do - it's up to you!   But again, why does the chorus not play nice?    If I were doing multi-effects, I WOULD consider having a built in supply, probably a wall wart >> regulator type deal, just like that. 

I often (constantly) have everything connected to the enclosure (using star grounding, of course), and have had no problems. In fact, for me, it makes more sense than trying to isolate everything, just my 2 cents...

[rockhorst]

For reference, here is the power filtering of the PCB, which already includes a LP filter. I copied it from the commercial unit, which has the same dislike of daisy chaining. It usually has to do with the clock the chorus uses.

[R.G.]

I run into this a lot in my day job. GibsonGM's got the right idea - find out the root cause of the issue, and fix that.

Some pertinent questions:
1. Does the chorus whine when powered by a 1Spot all by itself?
2. Is it possible to do a better job of grounding the pedals than by daisy chaining?

In many cases, the simplest (and cheapest!) solution is to use two 1Spots, one for the offenders that don't play well with others, and one for everybody else. Multi-output pedalboard power supplies can fix issues with isolation, but they're fairly expensive. And building your own many-output power supply is possible, but I took down the how-to info on the Spyder power supply at Geofex because I kept getting worrying questions from people who clearly didn't know how to wire AC power without endangering themselves. Music can be a life-changing cause, but in most cases musical power supplies are not worth dying for. 

[rockhorst]

1. Does the chorus whine when powered by a 1Spot all by itself?

No, only when connecting a second plug from the daisy chain to another pedal

2. Is it possible to do a better job of grounding the pedals than by daisy chaining?

Possibly...it might be enough to have all the effects grounded to the same enclosure (as in the multi effect). This should prevent a voltage drop between ground paths. But I don't want to wait until I've build up the whole multi to find out I'm not at the stage where all effects are in the same enclosure (close though), and I'd like to have some idea of how to prevent the whining befor I get there.

In many cases, the simplest (and cheapest!) solution is to use two 1Spots, one for the offenders that don't play well with others, and one for everybody else.

That was going to be my quick and dirty fix as well. I was also considering putting Jack Orman's power supply in the multi enclosure, the regulators might be enough.

Multi-output pedalboard power supplies can fix issues with isolation, but they're fairly expensive. And building your own many-output power supply is possible, but I took down the how-to info on the Spyder power supply at Geofex because I kept getting worrying questions from people who clearly didn't know how to wire AC power without endangering themselves. Music can be a life-changing cause, but in most cases musical power supplies are not worth dying for. 

Agreed, I've seen some projects for something like that and think I could do the job properly (I've fixed some power supplies over the years and had a good look around) but I'd rather not.

[hymenoptera]

Not much to say without knowing the cause of the whine, but for starters, 100uF main filter cap might not quite be enough to run 6 or 8 pedals off of. That looks like the PSU for just one pedal. Also make sure the wall wart your using is over-rated for the job. 500mA or even 1A isn't out of the question.

Have you got a scope to help track down the source of the problem? Do you know if it's on the ground, or coming from the Vcc? Are all the pedals specced for 9vDC tip-negative supplies (I'm sure they are if it's working at all...)?

[rockhorst]

I could possibly get my hands on a scope, but not right now. I do know that this happens often with time based effects using the MN3207 - MN3102 combo. The power supply section I posted is indeed for one pedal. Each of the PCBs involved has it's own on board filtering and I might also add some filtering or regulators at the point where the power enters the enclosure.

[R.G.]

Possibly...it might be enough to have all the effects grounded to the same enclosure (as in the multi effect). This should prevent a voltage drop between ground paths. But I don't want to wait until I've build up the whole multi to find out I'm not at the stage where all effects are in the same enclosure (close though), and I'd like to have some idea of how to prevent the whining befor I get there.

Actually, that could also introduce problems. It would be far better to have the power grounds all come from one point (i.e. star grounding) that's NOT the enclosure to avoid having the ground-current generated voltages in the ground conductors have any chance to cause whining. Actually, it's better if the effect inputs and outputs are not connected to the enclosure either, but starred to the main reference point, and one-and-only-one conductor from the main star reference ground to the enclosure. This setup - tiring as it is to set up - forces no currents other than RF shield currents to flow on the enclosure, and no currents other than signal currents to flow in the signal grounding lines.

Good grounding practice forces you to think in terms of where the currents flow, not so much what the voltages are; this is not least because modern pedalboards have huge gains available and they can amplify up tiny noises to major irritations.

That was going to be my quick and dirty fix as well. I was also considering putting Jack Orman's power supply in the multi enclosure, the regulators might be enough.

Multiple regulators are a half-fix, although they do look fancy to the uninitiated. Notice that all of those regulators share a ground. I suspect that the majority of the good that many-regulator setups do is to force you to run a single power+ground to a pedal, and in a sneaky, underground way make you do star grounding. Regulating the non-ground side of a power supply does reduce noise carried ON THE POWER SUPPLY LINES but it can't do anything for you in terms of ground noise. Makes great advertising, though; and might even help in some limited situations, like in transistor-only circuit with no power supply rejection.

Agreed, I've seen some projects for something like that and think I could do the job properly (I've fixed some power supplies over the years and had a good look around) but I'd rather not.

Good. We don't need any of our forum participants to get a Darwin award.

[rockhorst]

So I should connect all grounds to a single bus line and on one side connect that to the enclosure? Other/additional suggestions?

[R.G.]

So I should connect all grounds to a single bus line and on one side connect that to the enclosure? Other/additional suggestions?

That might work, but that's yet a different scheme for grounding.

People used to think of "ground" as a metal rod driven into wet dirt - that is, electrically connected to the planet itself. That is still how the electrical utilities do it. But for electrical equipment that is lower power (and some other issues), we can make do with deciding that "ground" is anywhere we want to define as the ultimate source of 0.00000... volts. We may (and do!) define anywhere as "ground", and as long as we measure from that point consistently as a reference, any point will work. "But wait..." I hear you thinking, "You mean just anywhere can be ground?"

Yes. As long as all measurements AND SIGNALS are considered to be referred to that point. If you decide ground is on one of your pedals, then the only way you can carry that "ground" voltage to one of the other pedals is if truly, no-fooling zero current flows through it, because all wires are really resistors, and any current flowing through a wire causes a voltage drop across it, however small the voltage. There are two ways to get a low voltage drop transferred to some (electronically) remote location - lower the resistance between them, or lower the current going through that resistance. Well, OK, three ways: lower both resistance and current.

A buss wire scheme tries to lower the resistance. You hook up all the "ground" points per circuit to that buss wire, and count on the buss wire to have "very small" resistance and hence voltage drop. Sometimes, especially in small, low gain, low current setups it works fine.

What I had in mind was actually a one-point ground. You decide that one point, usually somewhere on the power supply output, is the One True Ground. All other "grounds" are not really 0.000000... volts, but something different, offset by the current flowing in the wires to those points times the resistance of the wires. If you take one "ground" wire from the One True Ground to the pedal circuit and it has to carry both signal ground and power supply return current, then the signal ground out at the pedal is offset from the One True Ground by the pedal's power current times the wire resistance, and you now have a replica of the power supply current being supplied to the input as a signal.  If you take *two* wires to the pedal circuit, one carrying signal ground and precious little current, it arrives at the pedal NOT offset from the One True Ground voltage very much; the second wire carries used-up electricity back to the power supply, and has a voltage drop from that, but since it's in a different wire, it cannot pollute the signal reference voltage.

I had in mind a star ground, with a central One True Ground point and independent ground wires going from that to each effect. That is not perfect, but it at least keeps the pedals from cross-contaminating their signal grounds with their power supply return current.

[rockhorst]

What I get from RG's post is that it would be wise to separate signal ground from RF ground. I should use a star ground for that purpose. I'm not entirely sure what would be the best way to implement that and what the best way would be to route all thee power lines to the effects. I usually twist the red and black wires together in a single pedal, but that may not be a good idea here, if I want to do the star ground thing. Should I separate power to status LEDs from the signals also? A good time for a little more info on the design: it would roughly look like this



The enclosure is a 50 cm wide Hammond 1456WL1BKBU.

Some more guidance?

[antonis]

For reference, here is the power filtering of the PCB, which already includes a LP filter.

It is allways a good idea to place a 100nF ceramic decoupling cap (disk) as close as posible to IC's Vcc..
(and Vee in case of bipolar supply..)

[GibsonGM]

For reference, here is the power filtering of the PCB, which already includes a LP filter.

It is allways a good idea to place a 100nF ceramic decoupling cap (disk) as close as posible to IC's Vcc..
(and Vee in case of bipolar supply..)

Often, literally across their power pins...

[Kipper4]

I've seen Dave use those little rice grain caps under the ic

[rockhorst]

For reference, here is the power filtering of the PCB, which already includes a LP filter.

It is allways a good idea to place a 100nF ceramic decoupling cap (disk) as close as posible to IC's Vcc..
(and Vee in case of bipolar supply..)

Thanks for that. Just as a double check: you mean 100 nano, not pico, right? I could easily solder those underneath the boards, from Vcc to ground to see if that helps.

[hymenoptera]

I've seen Dave use those little rice grain caps under the ic

+1.

Probably the only place I ever use MLCC caps. And these are perfect for that! I get mine from an ebay vendor in Virginia called huskerfan. He's got a pretty big selection to pick from. 100 packs, on tape. It ends up being slightly cheaper than Mouser. 

[antonis]

For reference, here is the power filtering of the PCB, which already includes a LP filter.

It is allways a good idea to place a 100nF ceramic decoupling cap (disk) as close as posible to IC's Vcc..
(and Vee in case of bipolar supply..)

Thanks for that. Just as a double check: you mean 100 nano, not pico, right? I could easily solder those underneath the boards, from Vcc to ground to see if that helps.

Doublecheck correct...

(In Greece we call them "lentils" and we also put them in soup..)