Several basic power supply and opamp reference voltage questions

[Marmoset123]

Hi,

I am trying to combine multiple pedal circuits in one box (to fulfil all my musical needs) and have a few questions regarding the power supply. I've done a lot of reading and I'm starting to get a bit more of an understanding of these things but I still have a load of questions (some relating to my practical needs at the moment, and some just for my understanding).

I understand the concept of a voltage divider to achieve a reference voltage for an opamp, but If for example I wanted to use multiple opamps in one circuit wouldn't I be reducing the overall resistance to ground for each voltage divider I add?

Is the correct course of action to:

Adjust the values to compensate?
Not worry about it and just add them as shown?
Share the reference voltage between multiple opamps?


From looking at other circuits I can see that c is probably the answer, but I have a few questions about it's implementation..

Say for example I already have a power supply that's putting out a 4.5v reference voltage, and I wanted to use it to also power this booster circuit (see pic). Presumably I could just get rid of the 2m2 resistors and attach the 4.5v to the point where the 2m2 resistors met, but looking at various designs, I see that the reference voltage is always going through a resistor.

What is the purpose of this resistor? Is it to stop the audio signal from travelling up the 4.5v line? Is the value important and how do I decide? I have seen 100k and 1m resistors in this position. Should the value be different if I am using a buffered 4.5v supply? Is there anything else I need to add to make this (or any other noninverting tl072) work from a separate 4.5v supply? Can the resistor be omitted in an inverting opamp setup?

Sorry about all the questions.. I'm not finished yet!


Now my other questions are with regards to power supply smoothing/filtering capacitors:

How are the values determined (or in a more practical sense, are there standard values that people use)?

Looking at a couple of schematics I see a 100uf capacitor between 9v and gnd, a 47uf capacitor between the 4.5v reference and ground, and either a 47uf or 100uf between the output voltage of a regulator and ground.

If I want to combine 2 pedals into one circuit, I'm guessing I only need 1 x 100uf capacitor between 9v and ground etc? Or would would I gain (or lose) anything by keeping both?

Hopefully someone can help explain or point me in the direction of some reading that will help.

Thanks for your time.

Jamie

[antonis]

<What is the purpose of this resistor? Is it to stop the audio signal from travelling up the 4.5v line? Is the value important and how do I decide?>
Yes..!! 4.5V bias volatge coming directly from a voltage source is "seen" by signal as a short..
(DC voltage source exibits an almost zero impedance for AC - not to refer on, always there, big filter capacitor.. )

Bias resistor value is important mainly for HPF, created by Input cap and bias resistor, cut-off frequency..
(secondary for Input off-set current, but let it be for the moment..)
In case of input signal shunting (non-inverting input) see P.S.1 below..

<Can the resistor be omitted in an inverting opamp setup?>
According to the above said, Yes..!!

<Now my other questions are with regards to power supply smoothing/filtering capacitors:
How are the values determined (or in a more practical sense, are there standard values that people use)?>
Roughly speaking, only by desirable ripple voltage..
You should search about power supply filtering  to find a formula for capacitance calculation, more convenient than ΔI = C x ΔV/Δt ..
(I could write down a couple of them to choose which fits to your need but I strongly recommend you to understand how they occur..)

<Looking at a couple of schematics I see a 100uf capacitor between 9v and gnd, a 47uf capacitor between the 4.5v reference and ground, and either a 47uf or 100uf between the output voltage of a regulator and ground. >
Quite right due to 100μF cap has to filter 9V (of some specific current) and 47μF has to filter 4.5V (half the supply voltage) of the same current..

<If I want to combine 2 pedals into one circuit, I'm guessing I only need 1 x 100uf capacitor between 9v and ground etc?>
Only if 100μF cap is capable to filter double current consumption..
(e.g. IF 100μF are marginally OK for 1 pedal specific voltage regulation, doubling pedal will result in halving voltage regulation due to doubling current quantity..)

P.S.1
2M2 bias resistors are considered of relative high value, both from "noise" and input leakage current point of view, together with input cap HPF (1.5Hz -3db cut-off frequency)..
You may safely reduce their values about 10 times having an iput HPF of about 15Hz cut-off frequncy..

Their parallel combination equivalent resistance sets pin 3 input impedance so, if you worry about lowering it 10 times you may alter signal input to inverting and use a single resistor of 510k say, for non-inverting input bias..

P.S.2
Less than 7db shouldn't be considered as a boster, should it..?? 
You might have to replace R2 with a 500k pot, wired as variable resistor, to be able to set gain from unity (0 db) to 6 (15.5 db)..
(and name that pot "Boost" - instead of 10k pot which could be renamed something like "Volume"..)

P.S.3
Of course, a single answer post can't make clear all of your queries so don't hesitate to further ask whatever you need to undestand in more details..
(like the reason for creating a high pass filter of subsonic cut-off frequency (160mHz  ) together with the absence of a R2 shunting cap for low-pass filter in negative feedbcak loop..) 

[Marmoset123]

Thanks for getting back to me. That helps a lot with my understanding. I will try do do some more reading about power supply filtering.

I had already modified the resistor values in the 'boost' circuit I showed from the last post and it serves my purpose as a cut/boost, I was just wondering if it was better to power it from the buffered 4.5v supply that I already had elsewhere in the pedal.

I guess my remaining question is would this work (see pic) ?? 

(green bits show what is currently there, red bits show how I'm thinking it would attach to 4.5v supply with the green bits removed).

If so is it the 100n capacitor at the input and this resistor that form the HPF? So I could use 100k resistor for approx 15hz hpf? (according to a RC filter calculator I just found..)

Thanks

[antonis]

Of course it will work..!!
(but let's face the "quality" of work..)

Bias resistor first: a value of 470k - 510k should be fine, fed from a voltage divider of 10k - 47k pair of resistors..
(you may use even lower values if you don't mind about current consumption..)

That bias resistor value should create HPF with input cap..
(more correctly, that resistor in series with parallel combination of voltage divider resistance, in case of 4.5V filter cap absence..)

Feedback/gain resistors: 220k/33k form a gain of 7.66 (1 + 220k/33k due to non-inerting stage) or about 17.5 db..
IMHO, this is marginally acceptable for a Booster, always depended on max incomming signal voltage..
(i.e. for a maximum undistorted swing of 4.5 volts with a gain of 7.6 your input signal voltage should be marginally lower than 600mV..)

Now about that 10μF oversized cap: You should set its value in conjunction with 33k for a HPF cut-off frequency derived from
C = 1/6.28*33000*f, where f is your personal taste cut-off frequency..

<green bits show what is currently there)
If that stands, you've biased non- inverting input at 9V..!! 
(lower 2M2 resistor doesn't find any DC path  - I let you find the reason for that..)

<is it the 100n capacitor at the input and this resistor that form the HPF? So I could use 100k resistor for approx 15hz hpf?>
Sure if you don't mind about 100k input impedance..
If you REALLY mind, you can make that resistor 5 times bigger and make that cap 5 times smaller..

10μF output cap should be consider as a "safe" value for a next effect input impedance such as low as 1k..

About Feedback/Gain resistor values: Feedback resistor value (220k) should be a compromise between maximum op-amp total output current and noise impementation due to its resistance value..
Gain resistor value (33k) is then set in accordance to Feedback resistor value and desirable Gain..

It should be a good idea to also add a feedback resistor shunting capacitor (parallel to that resistor) to form a LPF at a cut-off frequency of your taste..
(forming a band-pass booster..)

[ElectricDruid]

Not to disagree with anything that Antonis said (all good advice from what I read) but rather to offer a different view, or another way to think about it.

Say for example I already have a power supply that's putting out a 4.5v reference voltage, and I wanted to use it to also power this booster circuit (see pic). Presumably I could just get rid of the 2m2 resistors and attach the 4.5v to the point where the 2m2 resistors met, but looking at various designs, I see that the reference voltage is always going through a resistor.

What is the purpose of this resistor?

The 4.5V is a "virtual ground". You don't want your input signal shorted to ground, but you *do* want the input to sit at that ground level when there's no signal. So you put a resistor there. It'll bias the input to ground when there's no signal, but when there is, the input won't be shorted to ground and the input voltage will be developed across that resistor.

Is the value important and how do I decide? I have seen 100k and 1m resistors in this position.

This resistor is the biggest determining factor in the input impedance of your device. It's in parallel with the input impedance of the op-amp input, but that's usually (or at least "often") very large and can be ignored. 1M is good value for input impedance, although you'll see far less, particularly in old pedals from the 70s. Once you've got that, you need to realise that it forms a highpass filter in conjunction with the input capacitor. The roll-off point for that RC filter is 1 / (2*Pi*R*C). You want to set that for your lowest frequency of interest, maybe 100Hz if you want to avoid a bit of mains hum, or 20Hz if you really want a full audio spectrum or you're playing bass. I use this tool to avoid doing the sums:

https://electricdruid.net/rc-filter-calc/

Should the value be different if I am using a buffered 4.5v supply?

No. The same rules apply.

Is there anything else I need to add to make this (or any other noninverting tl072) work from a separate 4.5v supply?

No, it should be fine.

Can the resistor be omitted in an inverting opamp setup?

Yes. The inverting op-amp set-up is biased by the positive input of the op-amp, which should be tied to 4.5V, and the negative input is biased by the feedback from the output.

Now my other questions are with regards to power supply smoothing/filtering capacitors:

How are the values determined (or in a more practical sense, are there standard values that people use)?

Again, it's a 1/(2*Pi*R*C) calculation. You've got a small resistor and a large cap, and you're making a lowpass filter this time. You want it to remove as much audible noise as you possibly can and just leave DC, so the frequency should be as low as possible and definitely below mains hum at 50/60Hz. In practice, you also want to avoid large and expensive caps in a little pedal, so 47uf or 100uF is a good compromise. Run that value and (say) 47R through the filter calculator and see how low the filter is.

If I want to combine 2 pedals into one circuit, I'm guessing I only need 1 x 100uf capacitor between 9v and ground etc? Or would would I gain (or lose) anything by keeping both?

Capacitors in parallel add their values, so 2 x 100uF would be like 200uF.

HTH,
Tom

[PRR]

Go to basics. A car starter motor is a 100 Amp load, needs a big battery. A fuzzbox is a 10mA (10/1000A) load, can use a small battery. An op-amp input pin is <1uA (1/1000000A) so can use a very low-power supply. We commonly take it from a voltage divider across the main supply.

You "can" use two 2Meg resistors, supply to pin, pin to ground. But power supplies are often crappy. This wants to inject half that crap right AT the input pin.

Also if you have multiple op-amp pins to bias-up it can get to be a lot of resistors.

So a popular plan is a couple 10K to 100K resistors as divider. With a several-uFd cap to ground to muffle supply crap. In most cases this can supply any reasonable number of op-amp input pins.

You obviously can not connect your precious input signal and your solid bias together; the bias would be a near-short on the signal. There are several ways to mix a DC voltage with an AC/Audio voltage. The elegant way is a coil and a capacitor. However Audio-frequency coils are BIG and costly. And our audio amplifying devices can usually tolerate significant DC resistance in their DC bias feed. If you use FEt-input TL072 the DC feed resistor can be 1Meg or more. This will not load guitar. It also allows a low-price 0.1uFd cap for 16Hz bottom.

When you get to '5532 and other BJT input chips, and many BJT buffers, 1Meg may be too much drop, and you see 470K, 100K, sometimes lower.

[antonis]

If you use FEt-input TL072 the DC feed resistor can be 1Meg or more. It also allows a low-price 0.1uFd cap for 16Hz bottom.

Much closer to 1.6Hz, I would say...