AMZ Relay bypass popping

[gutsofgold]

I just built up a PT-80 delay from general guitar gadgets and used a relay bypass board from AMZ in it. I was getting a real loud pop with every engage and disengage so I switched my grounding up. I put the ground from the DC jack and the PT-80 circuit on the input jack gnd tab and then put the Relay pcb ground and the LED cathode on the output jack gnd tab. This quieted the pops down but they are still there every time I hit the footswitch. Is this the proper ground scheme for the AMZ relay board ?? Relay pcb and LED share one ground connected to chassis and everything else shares another ground connected to the chassis at a different location?

[amz-fx]

Does the PT-80 have pulldown resistors? If not, they may have to be added.

Edit: I just pulled up the schematic and it looks like there is not one on the input side. You should add 1M from the pcb input to ground.

regards, Jack

[gutsofgold]

That's surprising, I assumed all of the GGG circuits would have pulldown on in and out. I'll add that resistor to the input side and see where that gets me.

Thanks!

[R.G.]

It's worth noting that hard-metal "true bypass" can never be completely quiet, even if the input and output DC levels are the same and never change. Unless the metal contacts happen to open or contact at very nearly a signal zero crossing the signal itself is discontinuous and this can be heard as a click.

This was an issue in electronic organs, and they worked long and hard to do "keyer" circuits that would minimize this on their keyboards.

[Lurco]

Or they added the "key-click" feature! http://en.wikipedia.org/wiki/Hammond_organ "Click Okay"

[gutsofgold]

Interesting about Hammond.. I however can't pass this click off as a 'prominent attack' 

I added a 1M resistor to ground at the input of the circuit before the input cap, this did nothing to reduce the click. Other than perhaps lowering the 1M resistors at input and output to say 500k to see if that suppresses the click more, and changing up grounding, what else can I try? This pedal is running into an amp that is 3-prong chassis grounded and I am using a 1-spot power on it. And to clarify what I said before -

INPUT JACK GROUND TAB <--- delay circuit ground and power ground from DC jack
OUTPUT JACK GND TAB <--- AMZ relay circuit ground and LED cathode

Thanks for the help so far 

[R.G.]

Since this is a new build, have you checked for DC voltage on the input and output of the effect?

There are a number of defects in pedals that will let current leak to the output pad, and this can cause serious pops that pulldowns can't cope with. The first thing to look for is whether the input and output capacitors are electrolytic, and if so, are they inserted correctly? If they're not, they let too much current through for a pulldown to deal with well.

[gutsofgold]

http://generalguitargadgets.com/pdf/ggg_pt80_sc.pdf

The input of this circuit seems to be a non-inverting opamp with input pin at half supply or 6VDC. Now I have added a 1M resistor before the .047u input cap but could this perhaps not be enough to quiet the DC on other side of the cap? The output SHOULD be at zero potential but I will check that later on today when I get a break. Not sure where else to look, will also be probing away with a scope later today.

Input cap is .047u film and output is 1u film as well both non-polar.

[R.G.]

Films don't give the leakage that electros do, so this may not be the problem. Do the voltage check when you get a chance.

[gutsofgold]

There is negligible voltage at the output and input jacks (~5mV on input and 0v on output) with nothing plugged into the pedal and power applied. There is ~5.5VDC at pin 5 of the TL072 and ~6VDC at pin 1 of the TL072.

[R.G.]

There is negligible voltage at the output and input jacks (~5mV on input and 0v on output) with nothing plugged into the pedal and power applied. There is ~5.5VDC at pin 5 of the TL072 and ~6VDC at pin 1 of the TL072.

I'm guessing that you mean on the schemo of the pedal at http://generalguitargadgets.com/pdf/ggg_pt80_sc.pdf .

If that's so, then a 1M on the input would be major overkill on removing popping at the pedal input.

Here's a simple test: I'm not familiar with the relay bypass setup, although it can't be that complicated. But in the end, if it's working right, it approximates a mechanically actuated DPDT switch, I think. Is it possible for you to temporarily swap in a DPDT switch to make the same connections as the relay bypass circuit? If this pops, the problem is in the pedal or the switching setup (or in the nature of true bypass, which has its own problems). If this does not, it's either in the relay bypass setup or the combination of the relay bypass setup and the pedal.

Relays have their own sets of issues, which include coupling the coil voltage transient into the signal carrying metal parts inside the relay. This is trivial in 600 ohm phone lines in telecom practice, but can be crucial in 1M pedal practice.

[Mac Walker]

Relays have their own sets of issues, which include coupling the coil voltage transient into the signal carrying metal parts inside the relay.

A common practice in the industrial controls world is to place a reverse biased diode in parallel with the relay coil, when the coil de-energizes it generates an extremely high voltage for a short period of time, a result of the magnetic field of the coil breaking down.

Adding the diode "snubs" the induced breakdown voltage to ground, the term used is "snubber diode".

This high voltage can in some cases be enough to backfeed industrial computer output transistors (digital outputs), destroying them in the process....

Might be worth a shot, putting one in parallel with the coil, reverse biased.....

(Here's an example):

[R.G.]

A common practice in the industrial controls world is to place a reverse biased diode in parallel with the relay coil, when the coil de-energizes it generates an extremely high voltage for a short period of time, a result of the magnetic field of the coil breaking down...
Might be worth a shot, putting one in parallel with the coil, reverse biased.....

Yeah, I just kind of assumed that was on the relay board. It's kind of Relays 101, and I'm reasonably sure it's on the PCB, although I have not looked.

Relay coils are inductors and inductors embody V = L di/dt. If you're trying to reduce the current, di is negative and then the voltage inverts and goes to whatever is needed to make di/dt come true.

Snubber diodes are also called "catch diodes" or "flyback diodes", in that they prevent the ugly parts of the flyback pulses. What's not normally appreciated is that the basic application of such diodes actually slow down the decrease of current in a relay by clamping the relay voltage to one diode drop in the raw form. V = L di/dt works both ways. If the voltage is clamped to 0.7V, then the time taken to reduce the current to zero is dt = L*di/V = L*di/0.7V.  Advanced applications of catch diodes will use a zener or some other mechanism to let the flyback voltage increase and turn the inductor current off faster.

... and that makes it worse for audio.    When a relay makes, the coil voltage is switched semi-instantly to the power supply voltage, so this voltage transient appears across the relay pins - and is coupled into the signal carrying metal parts by the stray capacitance. That's on the make. On the release, turning off the voltage to the coil lets the coil voltage fly to the power supply plus a diode drop if there's a flyback diode, or the power supply plus the extra recovery voltage if there's fast turnoff stuff there. That increases the voltage transient over the power supply voltage, and couples more into the signal parts.

Anyway, I bet there's already one there.

[gutsofgold]

No such diode on the AMZ pcb. I understand the need for the flyback diode but I ignorantly assumed that it wasn't needed in this application. Perhaps I will email Jack tonight to see what he says about it, I will also be swapping this setup out for a regular mechanical DPDT tonight to see if it is any more quiet.

[duck_arse]

the circuit board in question?



I didn't see a part number when I looked, but I woulda thort the little "+" shown on one of the relay contacts would indicate it had the flyback diode fitted internal. is this why some relays have polarised coils?

[edit: ] first search return was the obsoleted board. image corrected.

[gutsofgold]

Im using the obsoleted board, not sure why this new board was made but it looks like the same circuit as the old board. I'd post the schematic as simple as it is but its not something Jack publishes publicly. I am using a Pana TQ2-L-5V relay which does not seem to have any internal flyback protection.

[R.G.]

Some relays do have internal catch diodes. The circuit would then be highly specific to that relay part number.

Is that a simple pull-and-hold relay, not a latching one? One of the beauties of using a uC is that you can use latching relays and they do all that messy latching on and off stuff for you, so you can save the current used by leaving the relay on all the time.

... of course, modern use of pedals is nearly all on a power adapter, not on batteries, so this may not be critical. Is the schemo for this on line somewhere? I see a G6H relay, which could be one of several voltages, but it's probably either 9V or 5V; a three terminal regulator (IC2) to provide power to the uC, the uC itself, and some filtering components.  The G6H from Omron comes in both latching and non-latching versions, so it's hard to tell which version it is, although the "+" symbol does semi-indicate an internal diode. The G6H is obsolete, I find. Probably replaced by something in the G6A line.

A catch diode does slow down release time, so some have it and some don't.

Im using the obsoleted board, not sure why this new board was made but it looks like the same circuit as the old board. I'd post the schematic as simple as it is but its not something Jack publishes publicly. I am using a Pana TQ2-L-5V relay which does not seem to have any internal flyback protection.

The insides of the uC do have clamp diodes to the + supply internally, and that might work OK. I always worried about relying on the chip's internal protection and tended to protect the protection.

I did a similar thing a ways back, posted some of the info here. I'll go look at my version and see what I did.

[ubersam]

It uses a single coil latching relay. I'm using NEC EA2-5SNJ relays in mine and so far no pops. I wondered about the lack of catch diodes too but from what I understand, the coils are only energized for a few micro seconds. Not long enough to produce a significant flyback current. My understanding could be wrong, though.

[Mac Walker]

How close is the relay coil itself to the audio path?  Are the boards physically stacked on one another?  Is there room to separate them to some degree?


There is a Craig Anderton trick of wrapping a timing LFO electrolytic cap in wire, and landing the other end to ground.  To protect LFO pulses from getting into the audio portion of the circuit.....  Separation and or/shielding in this case might improve the condition.

[R.G.]

It uses a single coil latching relay. I'm using NEC EA2-5SNJ relays in mine and so far no pops. I wondered about the lack of catch diodes too but from what I understand, the coils are only energized for a few micro seconds. Not long enough to produce a significant flyback current. My understanding could be wrong, though.

OK, so it's using two uC pins, one on one side of the coil, one on the other. The way single coil latching works, you pull one side up and the other down to set the relay, and then reverse the polarity to reset the relay. The coils have to be energized for some milliseconds, not microseconds. The EA2 datasheet says 2mS, which is pretty fast. But not microseconds. Well, OK, 2000 microseconds.

The relay coil must be energized enough to make it flip state, and that's enough to make as much flyback as it would if it were left powered for a day.

How close is the relay coil itself to the audio path?  Are the boards physically stacked on one another?  Is there room to separate them to some degree?

The audio path goes *through* the relay body itself. I switches the audio.

There is a Craig Anderton trick of wrapping a timing LFO electrolytic cap in wire, and landing the other end to ground.  To protect LFO pulses from getting into the audio portion of the circuit.....  Separation and or/shielding in this case might improve the condition.

Some relays have internal shields between the coils and the switching contacts for exactly this. But it has to be inside the relay to be any good to you.