[sdiy] Introduction and a couple of noob questions

Marc Nostromo [M-.-n] marc.nostromo at gmail.com
Thu Oct 4 14:48:12 CEST 2012

Thanks for the explanation and links Olivier.. I'll read that
carefully and try to digest it !


2012/10/4 Olivier Gillet <ol.gillet at gmail.com>:
> Hi Marc,
> Cool to see you here!
>> 1) Audio voltage: What is the range of an audio voltage and it's 'polarity' ?
> It really depends on the device... The oscillator of a modular
> synthesizer is going to be at a fairly "hot" level (say 10V peak to
> peak); while consumer audio equipment is more likely to run at less
> than a volt peak to peak. http://en.wikipedia.org/wiki/Line_level
> The "polarity" of the signal and whether it is referenced to ground or
> to a higher voltage (say 1.65V for a media player running at +3.3V) is
> irrelevant if you put AC-coupling capacitors at the input stage of
> your system.
>> I'm expecting it to be +/- a few volts around ground value. But what
>> happens when you use an op-amp where the V- is connected to the ground
>> (like in the single supply amplifier from forrest mims) ? Is it the
>> same ground reference as the audio input ? In that case, what happens
>> when the audio voltages is negative ? Can it be transformed by the
>> op-amp to a voltage lower than v- ? Or should the ground lead from the
>> audio be connected to a virtual ground point ? Somehow I feel I'm
>> missing a point there.
> In practice, it does not matter because the signals are AC-coupled.
> For example, if you want to design a pedal running on 9V with a 4.5V
> virtual ground, you will have an AC-coupling cap at your input stage
> (see for example the schematics on figure 3 of this:
> http://www.eng.yale.edu/ee-labs/morse/compo/sloa058.pdf - which I
> recommend reading), so it doesn't matter if the signal fed into your
> circuit is referenced at 0V or 4.5V or 1.65V...
>> A lot of circuit schematics I see expect a voltage reference of 15v.
>> But I also know some of them will work with 9v. How does one know how
>> low the voltage can be ?
> - If you see that one of the supply rails is used for anything else
> than powering op-amps (example: fed through a resistor to an op-amp
> summing node to offset a CV ; hooked to one of the pots...), you will
> have to recompute some of the resistor values.
> - The next thing to look for are headroom/signal levels. If you have a
> circuit powered by +/- 12V, it won't handle signals with an amplitude
> of more than 24V pp. Many op-amps start misbehaving when the signal is
> getting close to the rails, which reduces the maximum amplitude
> further. For example with a TL072 your headroom would be more
> something like +/- 10.5V.  So reducing the power supply, say from +/-
> 12V to +/- 8V is going to reduce the amplitude of the signals that can
> run through the circuit, and cause it to malfunction if it operates
> with large inputs.
> - Finally, check for fancy parts that might require high supply voltages!
>> +9/-9 as source and all I have is a 9v dc source, is there a way for
>> me to generate the +9/-9v from +9/0 ?
> Yes, there are DC-DC converter like the MAX7660/LT1044 or LT1054.
> Drawbacks of these parts: might introduce high frequency whine if not
> used correctly; expensive; somewhat fragile (can be killed by excess
> voltage); and limited current output (100mA for a LT1054 and at this
> current it delivers less than -Vin).
> The other option is to use a "single supply" design (V- to ground; V+
> to 9V) for your op-amp, and consider that all your signals are
> referenced to 4.5V.
> http://www.ieee.li/pdf/essay/single_supply_op_amp_design.pdf
> Risks: design is not as straightforward (you can't blindly use a bit
> of application schematic made for a split supply design); you will
> lose half your headroom; you will lose one important property of
> op-amps - power supply rejection (If you are not careful about what
> you do, any fluctuation of the supply voltage will affect your
> signals, resulting in a noisier output); you will lose the ability of
> using chips that requires higher operating voltages (eg: SSM2164
> requires at least +/- 4V); and finally you might have to make
> excessive use of AC-coupling between the different stages of your
> circuit, resulting in degraded performance in the lowest frequencies
> if you are not careful about what you do.
> Best,
> Olivier


More information about the Synth-diy mailing list