yes frank all inputs outputs are DC coupled so the offset does not occur there. That is where the misconception or lack of understanding starts. The offset happens at VG or VR or the voltage reference or virtual ground. Otherwise just a basic voltage divider to produce the reference voltage.
which can effect any non inverting op amp stages that do not use bypass capacitors in the gain setting voltage divider, or any inverting op amp stage which will always be connected directly to DC through the non inverting input tied directly to VR or VG
and yes oscillator frequencies in a charge pump is usually non issue. again the issue is at VR or VG. Also a problem with effects loops when the loop on the amp is from a higer voltage bipolar supply.
and then ties to a single ended effects chain and then returns to a bipolar effect return.
also with boutique/major manufactures the circuits are copied from bipolar designs, and then poorly converted to single ended designs. some connections that should be connected to VG are not and vice versa some connections to battery ground should not or need to be decoupled at certain points, or need additional pull down resistors to operate correctly or should be at VG not Battery Ground . likewise with the common LFO circuits that copied over and over, the original designs they copied are incorrect and include variable resistors to compensate for DC offset, otherwise a proper conversion from the bipolar designs they were taking from would have eliminated a majority of the problems. Likewise with almost every bucket brigade analog delay chip, and associated compander/expander circuits for feedback loops, the conversion from bipolar to single ended is incorrect. And the lower 5volt rail on a bucket brigade is highly affected by offset.
mainly causing distortion on the brigade and poor tracking for the compander.
I am really interested in examples of pedals that have the charge pump issues of noise/hum and create the DC VR VG offset problems that you have mentioned. I have been working with charge pump circuits for over 30 years and researching circuit noise issues for many years now, so your post really captured my interest. That is why I asked if you could please identify the pedals that have these issues, so I can examine them. Also, if you could tell me of some designs that have VG VR design problems and any of the other design mistakes you mentioned I would greatly appreciate that too.
I think I understand what you are referring to now with your DC offsets
ground loops. What threw me off was when you said down the line. I read that as everything after, but ground loops can be a problem for any device, anywhere on the shared power ground.
Some pedal manufacturers have quasi isolated the signal ground from the power ground with resistors, diodes or both that exacerbates ground loop problems. What usually happens with these pedals with resistors between the sig and pwr grounds, is that nearly all of the current that the pedal is using passes through the signal ground, through the all of the other pedals (that share the parallel power connection), back to the power supply. Since the power current will passed from that pedal to all of the other pedals, the current can be amplified as hiss, noise, oscillations, squeals, feedback
etc. These pedals are not improperly or poorly designed, because quasi ground isolation can substantially lower noise in some designs. I also dont think that the designer is necessarily lacking in single ended topology knowledge. It is a very cheap and effective way to reduce noise and it is especially useful for designs that have a clock for BBDs, LFOs etc. Pedals that have this quasi ground isolation should be identified (by the manufacturer) for use on isolated power only
sometimes they will operate with no issues at all on daisy chain power. But they are not improperly designed
they were just not designed to be used on daisy chain power connections. Connecting one of these pedals (using a daisy chain power connection) to another pedal that bonds the pwr and sig grounds effectively shorts the quasi isolation resistor and eliminates the noise reduction that the designer had originally implemented. A simple ohms measurement between the pwr and sig grounds will quickly identify these pedals.
The DC shifts that you have stated happens when there are unexpected ground loop currents and there is a resistance in the pedals ground. I have seen this several times. I am sure if you were to analyze the problem pedal with these voltage drifts, you will find substantial resistance in the grounds. Again, if there is a resistance in the grounds, it designed to be used with isolated power. If the pedal has its sig and pwr grounds bonded there is no chance that the micro ohms of resistance will develop enough of a DC voltage across it to drastically affect the circuit. The natural device tolerances and drift of offset bias currents and voltages within the circuit would be far greater. If the circuit is designed where it can not tolerate small drifts it will potentially be a problem regardless of the power source used. Especially since most analog pedals have no internal voltage regulation and will experience voltage drifts due to a battery or the variances between AC adapters.
If the power and signal grounds are bonded and you have issues with digital glitches the culprit will most likely be AC noise and/or spikes. AC noise propagates very efficiently thru ground loops and that is the biggest problem that we see (or hear) with audio ground loops. AC noise can also be transmitted and amplified across pedals even with isolated power, although this is much less of a problem.
The only way to minimize pedal ground loops is to use isolated power or break the signal ground with isolated signal coupling, like transformers. But, transformers can be great pickups for environmental noise, so they are rarely used in pedal designs.
Another source for noise with daisy chain power is that the power supply is not stiff enough for all of the pedals connected to it. What happens here is that the power supply can have ripple currents or audio signals reflected back onto the chain that can be amplified by some or all of the pedals connected to it.
-Frank
