Daisy-chaining two charge-pumped pedals?

[lars-musik]

Hello friends,

some time ago I built a KLON Clone for a friend and he's got trouble with a high pitched squeal on his board. As the noise wasn't there before the KLON was, the culprit was to be my pedal. So I changed the ICL7660S (yes, the "S" version and pin 1 and 8 connected for frequency boost) for a LT1054, hoping that was helping somehow. It didn't. Still squealing just more constantly as before. "Something like a kettle shortly before boiling" he describes it.

So I asked him to pull all the other pedals one at a time and see: When he disconnected his "Weehbo Morbid Drive" the sound went away. A quick investigation brought up a manual for the pedal and reveals that it also sports a charge pump (not saying which and  I cannot find a schematic either). But assuming it is a 7660S or a LT1054 as it does not whine when playing independently, here comes my question finally:

Is it a problem to have to carge-pumped pedals in a daisy chain. If yes, why and what to do (any filtering possible)?

Thanks for your help - again

Lars



EDIT: My friend opened the pedal and analysed the SMD chips with a magnifying glass. Result: One TC1044S(EAO), four JRC4580 and one OPA2604. I *think* the TC1044S handles 20 mA in boosted mode. If I add the current consumption of the opamps I get a typical consumption of 34,5 mA and a max consumption of 48 mA. Even if some opamp parts are unused, that doesn't look good to me.

Advice? Going into a 265 EUR boutique pedal with a soldering iron and change the charge pump for a LT1054?  Send the pedal in?

[anotherjim]

I think it could happen even if both pumps are supersonic. It's heterodyne noise. What you hear is the difference between 2 frequencies that are inaudible themselves.

Have you tried independent power supplies? Because I see 2 ways the pump frequencies could get mixed - audio path or common supply - or both! Supply is highly likely, since the frequencies should be beyond the pass band of average op-amps. But then people go and use fancy amps that can talk to bats in a 5Khz guitar rig .

This is why I stick 100nF ceramic disc caps on power rails for HF filtering. You don't know you need them most of the time, so an awful lot of designs omit them. This could be a time that you do need them.
The Klon schemes I've seen have no hf power decoupling, not even close to the charge pump chip. Tantalum caps are no substitute for adding 100nF ceramics.
You can add an RF filter to the output op-amp (there already is one on the preceding stage). I see 100k feedback resistor on the final amp. Try adding a 100pF or 220pF capacitor across it.

[R.G.]

AJ is right. The problem here is that the two charge pumps place ultrasonic noise on the shared power supplies, and the difference frequency is picked up and amplified to be the whine you're hearing.

It's most likely ground shift noise that you're hearing. Charge pumps by their nature pull big pulses of current from the power supply AND ground at the pump frequency. That causes the "ground" conductor to have a voltage shift because the conductor is really a low-value resistor.

Sadly, it is VERY difficult to filter ground, unlike a power supply line. AJ is also correct in that the best after-the-fact solution is to add good RF filtering very close to the noise source. This really doesn't so much filter the power supply as it does provide a very close-in "bucket" of charge that can absorb the pulses locally, not demanding pulses through power and ground.

At a more philosophical level, this entire issue, like so many others, is caused by the fact that pedalboards have grown by accretion, like a coral reef. There is no system planning in pedalboards at all. In this case, the problem presenting itself is that power ground is the same as signal ground, so noise gets mixed up with the signal ground. But it is what it is.

A good CERAMIC 0.1uF, possibly paralleled with a ceramic 0.01uF will probably help.  Using a different wall wart for one of the pedals to avoid the shared power ground will probably help, and may be the most cost-effective and expedient solution. It's also possible that one other pedal is what is most sensitive to the noise. Something like a Fuzz Face or other non-opamp, discrete distortion pedal will be very sensitive to ground noise, so putting it on a separate power supply or filtering its power, or both, may also help.

[Mark Hammer]

So here is the logical question that needs to be asked (even moreso considering the current prevalence of charge-pump powering and digital pedals):

WHAT NEEDS TO BE INSERTED BETWEEN "LINKS" IN A DAISY CHAIN TO PREVENT PEDALS FROM BEING "INFECTED" BY LINE NOISE ISSUES STEMMING FROM AN ADJACENT POWER-SHARING PEDAL?

[R.G.]

As always, good question Mark. Sadly, I don't know, and I've tried. It would be great if there was a little circuit fragment that could be potted up into a lump that had a male 2.1x5.5 connector on one side and a female on the other. If I had a good one of those, at least one company would be selling them today. That doesn't say they are impossible, just that I can't think of one yet.

Inserting any impedance into a ground line is a bad place to start. I've tried common mode inductors, separate inductors, resistors, and so on. If there is a big enough impedance to stem pulse noise, they lead to DC offsets.

Using a local charge bucket - power-to-ground caps, and good high frequency response caps - helps. It's not perfect, but may get you buy.

One radical choice is star power wiring. If shared ground impedance is a big part of the problem (and it is) then you can semi-fix that by star-wiring the power net, not series-chaining it. This is inconvenient on a big pedalboard, but would cut a lot of the cross-feed noise, as now the grounding impedances help attenuate noise, not spread it.

Another radical choice that seems to be winning is isolated power supplies. The Pedal Power (and Spyder) and other similar setups isolate power ground, and that helps a lot because polluted power ground doesn't spread noise. That's no help if the pedal pollutes signal ground inside the pedal and that's passed on to other pedals, but there's no help for that. I'm not a fan of many-output isolated power setups, even though my employer makes and sells them, largely because they are expensive unless you have a BIG power problem. Very often, a second wall wart, if it's a quiet one, can be attached to a problem pedal, or even two of them, at a much lower cost. But there are mixtures of high power digital pedals that can force you to buy so many separate power supplies that a multi-output power brick with lots of current capability may be a better solution.

So - sorry. I can't think of one yet.

[anotherjim]

I wonder how effective this would be...

I know this is commonly done to deal with RF, but I wonder if they would be efficient at ultrasonic?

[lars-musik]

Thanks for the ideas.

Some question: 1. What makes a ceramic 100nF a good ceramic capacitor in terms of HF filtering?

2. Where would I place that contraption if I was to try the effects?

.

[R.G.]

I wonder how effective this would be...
I know this is commonly done to deal with RF, but I wonder if they would be efficient at ultrasonic?

That is a common-mode choke. It is effective when the same current flows on both wires, as it does when RF currents use both power and ground, with the currents going in the same directions. The differential currents created in the circuit are unaffected, so this does not cut down on ground-born noise like a charge pump or digital pedal generates. I tried them.

[anotherjim]

Ceramic has better performance at high frequencies. Electro can be worth 1/10th of their value over 1Mhz. 100nF happens to be a big enough & cheap enough Goldilocks choice.

Thankfully, RG has offered real-world experience on the ferrite idea - which is don't bother!

[lars-musik]

I just went for a visit to my friend's and packed my soldering iron along with some 104, 103 and 101 ceramics.

As always, you've been spot on!

I put a 100 nF (actually measured 84 nF) with a 10 nF in parallel on the power rails along with the 100 pF on the output op amp's 100K.

Problem solved, noise gone, schematic updated!

Thanks again! This forum and especially you guys are fantastic!