DIY digital synth unhappy to share 9v supply with other stompboxes

Started by beammy, May 27, 2015, 08:08:22 AM

Previous topic - Next topic

beammy

Hi!

I've nearly completed a digital synthesizer project. It sounds great plugged straight into an amp.

It runs off a 9v power supply, same as most stompboxes. However, one thing I'd like to be able to do is plug its signal into other stompboxes, and have it share a 9v power supply with these other stompboxes via a daisychain. At the moment, if I try this, there's a lot of hum.

It seems that this issue is related to grounding. So here's how the grounding in my device looks: Ground (including ground of the output signal) is tied to the negative power rail - i.e. it's at 0v rather than 9v or anything else.

A tiny bit more info, although I don't think it's relevant - the final audio output stage is through an AD8616 opamp which is run from an internally generated +2.5v and -2.5v. The -2.5v is generated using a charge pump. Prior to that opamp, a DAC's signal which initally would have had something like 0.5v DC offset is run through coupling capacitors so that it's anchored around 0v, and the opamp's amplification leaves it anchored to this 0v ground when it is outputted. All of this performs very well when not sharing the 9v power supply with any other equipment.

My first guess as to the cause of this problem would have been that the pedals I'm trying to use it with have their ground sitting somewhere other than 0v, and that this is causing current to want to flow down the ground wires in the audio cable I'm using to connect my device to whatever pedal. I've tried this with a Boss DS-1 and a Boss CE-2. Sure enough, by measuring, the ground in these pedals sits at around +2v.

So I guess my question is, clearly other manufacturers know how to make pedals which can share a 9v power supply with the majority of other pedals and not have hum introduced if these various pedals' ground voltage is slightly different - so what are they doing differently that I need to do to my device in order to make it behave properly too?

Many thanks :)

slacker

When you say the Boss pedal's ground is at 2 Volts what is this relative to? The grounds of everything in the chain should be connected together through the shields of the audio cables so they are all effectively the same point, so must be the same voltage. My guess whould be either that the sleeve connection on the audio jack on your synth isn't connected to ground but to some other rail or the point you're measuring from isn't ground.

beammy

Yes when I say the Boss pedals' ground is at +2v, I mean that the "sleeve" of the input jack on either of the Boss pedals sits at around +2v relative to the negative power input (the centre of the 9v power jack). I measured this with a guitar plugged into the pedals' input jack, and also with just a cable plugged in with nothing on the other end.

On my synth, the ground (i.e. the sleeve of the audio out jack) is tied to the negative power input (centre of 9v jack). Should I have designed this differently so that my synth is more happy for its ground to sort of "float around" in some way? My brief bit of googling on stompbox grounding suggested that it's normal for stompboxes to have their ground tied to the negative power rail, so I wonder what's wrong with my design.

bool

Boss pedals have a ground-loop breaker designed in. The solution is to group your Boss pedals on one PSU and everything else (or what you have) on another separated PSU.

slacker

Oh yeah I'd forgotten that, old Boss pedals that were designed for the ACA power supply, have a resistor and diode between the negative DC and ground and will do exactly what Beammy measured, sorry for assuming you'd done something wrong. Newer Boss pedals don't do this, negative DC jack is wired to signal ground.

According to this page though http://www.bossarea.com/other/aca.asp plugging one of those pedals into another one where DC negative is ground should short out the resistor and diode, grounding the DC negative. That makes sense if you look at this schematic http://www.hobby-hour.com/electronics/s/boss-ce2-chorus-schematic.php if the negative DC jack and input sleeve are both grounded then D5 and R53 are shorted, so there shouldn't be a problem sharing a power supply.
It could be noise from the charge pump, have you tried powering the Boss pedals from a separate supply to see if that makes any difference?

beammy

Thanks for the info! I measured and yes, with my synth plugged into a DS-1 and sharing the power supply, the DS-1's ground goes to 0v.

I'm now thinking that the hum / noise I'm experiencing is more due to my digital synth polluting the power supply and causing the other pedals to introduce noise. Basically all digital units do this, from my experience. The problem is especially (probably unusably) bad with my synth, but then it does have a lot of bright LEDs which are pulsed on and off very rapidly, so the amount of current being drawn from the power supply will be fluctuating a lot, all the time.

Curiously (and this is what caused me to think there was something else going wrong), even if I use a double-adapter to take just the signal from my synth to an amp so I'm hearing just it, then the mere presence of another pedal (e.g. DS-1) additionally (via the double adapter) connected to my synth's output and sharing the power supply, is enough to cause this same noise *on my synth's output*. This seems odd. However, I tried this same configuration but with my synth replaced with another digital stompbox (EH Stereo Memory Man), and yes - although it doesn't introduce nearly as much noise, the same problem existed.

So at the moment it looks like the only thing wrong/different with my design is that it draws an unusually high amount of current in a noisy fashion.

Transmogrifox

If you have space for more PS filtering in your digital synth then add a bank of capacitors.  You'll need a bunch of bulk capacitance at the front end (like several hundred uF to a couple mF), and then you want your bank to contain several 0.1uF caps to eat the high frequency, sharp clock edge currents.  The more tightly that can be coupled to your chips the better.

The charge pump seems a likely culprit for eating up the PS. These take sharp pulses of current, and often around 10 kHz typical switching frequency -- perfect zone to introduce all kinds of noise cross-products. If you can get some capacitance tightly coupled around the charge pump power inputs that might do the trick.
trans·mog·ri·fy
tr.v. trans·mog·ri·fied, trans·mog·ri·fy·ing, trans·mog·ri·fies To change into a different shape or form, especially one that is fantastic or bizarre.

bool

You weren't paying attention.

The mechanism for *?##* injection is that when Boss pedal standard way of power supply delivery is broken, the power is delivered to the pedal via a "jack cable" shield.

IOW, colloquially, you are performing "dirty ground signal insertion" essentially.

Just do what I said to do before. There is no reason to reinvent the wheel, that's just how these things work. Ie., the same as 30 years ago.