10 common best practices and mods we should use in every circuit?

[EBK]

I'd use the OUTPUT jack for any stereo-jack switching, as putting it on the input jack runs all the device's current through the input ground, and inserts another place to introduce ground noise.

Along these lines (maybe), I use fiber shoulder washers to isolate my input jacks from the enclosure, and I star ground everything to the DC power jack (at least everything that I consider a separate noise source gets its own ground wire).  Maybe it improves things, and maybe it doesn't (it's possible that I'm just crazy).

[tenser75]

I'd use the OUTPUT jack for any stereo-jack switching, as putting it on the input jack runs all the device's current through the input ground, and inserts another place to introduce ground noise.

Along these lines (maybe), I use fiber shoulder washers to isolate my input jacks from the enclosure, and I star ground everything to the DC power jack (at least everything that I consider a separate noise source gets its own ground wire).  Maybe it improves things, and maybe it doesn't (it's possible that I'm just crazy).

why do you want to isolate the jacks? Shouldn't be the opposite?

I feel the enclosure should become a interference field, but maybe i'm wrong...

[EBK]

I'd use the OUTPUT jack for any stereo-jack switching, as putting it on the input jack runs all the device's current through the input ground, and inserts another place to introduce ground noise.

Along these lines (maybe), I use fiber shoulder washers to isolate my input jacks from the enclosure, and I star ground everything to the DC power jack (at least everything that I consider a separate noise source gets its own ground wire).  Maybe it improves things, and maybe it doesn't (it's possible that I'm just crazy).

why do you want to isolate the jacks? Shouldn't be the opposite?

I feel the enclosure should become a interference field, but maybe i'm wrong...

I only isolate the input jack from the enclosure, not the output jack.  Enclosure is still grounded via the output jack.

[tenser75]

I'd use the OUTPUT jack for any stereo-jack switching, as putting it on the input jack runs all the device's current through the input ground, and inserts another place to introduce ground noise.

Along these lines (maybe), I use fiber shoulder washers to isolate my input jacks from the enclosure, and I star ground everything to the DC power jack (at least everything that I consider a separate noise source gets its own ground wire).  Maybe it improves things, and maybe it doesn't (it's possible that I'm just crazy).

why do you want to isolate the jacks? Shouldn't be the opposite?

I feel the enclosure should become a interference field, but maybe i'm wrong...

I only isolate the input jack from the enclosure, not the output jack.  Enclosure is still grounded via the output jack.

ohh ok! and why is that? what is supposed to do/prevent?

[EBK]

ohh ok! and why is that? what is supposed to do/prevent?

I'm trying to be careful and not claim that does anything. 
In theory, it prevents the enclosure from becoming a ground current conductor, which could create ground loops, and it prevents extra noise from being dumped into your circuit's input (and possibly being amplified by your circuit).  In practice, it probably does little more than make me feel better.   

[samhay]

100R resistor in series with the input wire to the PCB, and a 47pf ceramic cap from input to signal ground after the 100R on the PCB....

Good advice when all circuits used a few transistors at most.
But, if you do this with a circuit using a couple of NE5532s or a PT2399, or lots of LEDs, or etc, then your circuit is likely to be sucking 10's mA and dropping 1V+ across the resistor. At this point, I would argue that it is no longer 'best practice'.

[blackieNYC]

Are we confusing the signal input and the DC input above?

[electrosonic]

We are talking about too different RC (lowpass) filters with very different frequencies.

The 100 ohm / 100uF combination in the power supply is has a -3db point of about 16Hz - reducing power supply ripple.
Every 10 mA of current draw will drop 1 volt of power supply so the value of R might need to be lowered to accommodate high current draw circuits.


The 100 ohm / 47pF combination in the audio input has a -3db point of  about 33kHz - shunting RF to ground to keep radio station noise out of the audio (this is ignoring the output impedance of the guitar or pedal which is in front of the filter)

Andrew.

[R.G.]

As noted - this is on the input signal line, not on the power supply line.

All circuits need good power decoupling.

[GGBB]

The 100 ohm / 47pF combination in the audio input has a -3db point of  about 33kHz - shunting RF to ground to keep radio station noise out of the audio (this is ignoring the output impedance of the guitar or pedal which is in front of the filter)

100R & 47pF = 33Mhz actually. There's a great deal of RF below that. I don't see the value in this, unless there's something going on with the output impedance of the prior pedal or guitar. Unless that was supposed to be 47nF (=33kHz)?

[R.G.]

As a practical matter, the 100R/47pF combination works very well. The values are based on field experience, not calculation.

I suspect, but cannot verify, that the incoming wires become less and less effective at being antennas as the RF gets lower in frequency - they're just too short. 33MHz (if that's the rolloff, I haven't calculated it) is a third of the FM band, so antennas for that range would be three times as long.

And the technique does need a series element. A 100R resistor works well, a ferrite bead does too. But simply putting a small cap to ground at the input can actually make RF worse if you happen to get an unlucky combination that tunes in the station more readily. I've seen it happen. The series impedance forces attenuation of the RF, not just tuning it in mysterious ways.

But if you happen to live under an antenna farm, it's possible that nothing will really help. Too much RF power flying around.

[electrosonic]

100R & 47pF = 33Mhz actually.

Right.

33 MHz not 33kHz, my mistake

Andrew.

[bluebunny]

a ferrite bead does too

This appears to be a very popular technique in the Synth DIY world.  You don't see it around here so much.

[Phoenix]

The 100 ohm / 47pF combination in the audio input has a -3db point of  about 33kHz - shunting RF to ground to keep radio station noise out of the audio (this is ignoring the output impedance of the guitar or pedal which is in front of the filter)

100R & 47pF = 33Mhz actually. There's a great deal of RF below that. I don't see the value in this, unless there's something going on with the output impedance of the prior pedal or guitar. Unless that was supposed to be 47nF (=33kHz)?

You've got to take into account the source impedance of whatever you're driving the input with too. In the guitar world this could be anything up to 125k with a 500k volume control turned down to half resistance. Because of the massive variability in source impedance, you can't rely on a predictable low pass corner frequency right at the input. The values of 100 ohm and 47pF work pretty well in this regard. If the source impedance is 125k, +100 ohm, the corner frequency is right around 27kHz. You could probably bump up the value of the cap a bit if you want, especially seeing as we're rarely interested in much above 7kHz for guitar.

Anyway, the moral of the story is that if you want predictable results, you have to perform your low-pass AFTER the input stage, where you have a known source impedance and so can get predictable results.

[tenser75]

100R resistor in series with the input wire to the PCB, and a 47pf ceramic cap from input to signal ground after the 100R on the PCB.

Add a 0.1uF ceramic cap from +V to ground where the power comes onto the PCB.

With the circuit laying open on the bench, this may not be enough, depending on how far you are from the antenna farm.

any chance to have a quick sketch of this 100R/47pf circuit, just so I don't make mistakes?

[GGBB]

          100R
In -> ---/\/\/\--- -> rest of circuit
                  |
                  |
                ----- 47p
                -----
                  |
                  |
                 Gnd

[tenser75]

thank GGBB

im sending a PCB to print tonight and I'll try this solution! thanks

[tenser75]

          100R
In -> ---/\/\/\--- -> rest of circuit
                  |
                  |
                ----- 47p
                -----
                  |
                  |
                 Gnd

question: somewhere I found this solution mentioned to be used after the input cap... not sure if I understand correctly, but
say I have a fuzz with a 100nF input cap... your 100R/47pf where does it go? before or after?

thanks!

[GGBB]

question: somewhere I found this solution mentioned to be used after the input cap... not sure if I understand correctly, but
say I have a fuzz with a 100nF input cap... your 100R/47pf where does it go? before or after?

thanks!

I'm not sure it will actually matter. They might sound a wee bit different but I doubt it. Some circuits may have other things going on - e.g. an input resistor - that could have an affect, so experimentation is worthwhile. I typically see it done before the input cap, and that's where I'd do it in a typical fuzz.

[tenser75]

question: somewhere I found this solution mentioned to be used after the input cap... not sure if I understand correctly, but
say I have a fuzz with a 100nF input cap... your 100R/47pf where does it go? before or after?

thanks!

I'm not sure it will actually matter. They might sound a wee bit different but I doubt it. Some circuits may have other things going on - e.g. an input resistor - that could have an affect, so experimentation is worthwhile. I typically see it done before the input cap, and that's where I'd do it in a typical fuzz.

THANKS!!