Custom attenuator/level shift for unipolar analog synthesizers

 

Goal : Design and build a circuit capable of transmitting unipolar analog synth signal to audio devices such as effects or amplifiers

Specifications/Research :

    After finishing my first analog synth build and confirming its functionality, my first inclination was to send the signal to a small speaker. Upon doing this, the sound from the speaker certainly matched the synth's output frequency but was extremely quiet. What was not realized at the time was that the unipolar (0-10V) signal has a 5 VDC bias which likely destroyed the speaker. Once I finally realized this, I knew I needed to design a circuit which would level shift (making it a bipolar signal for audio circuits) and attenuate it to varying degrees (10V is way too high to input into any kind of audio filters or amplifiers I have access too). 

Progress log : 

• 4-24-23 : My first inclination was to just add a 5V DC shift at the output. The problem with this train of thought is the same one I had initially - the speaker/audio device ends up DC biased.
• 5-22-23 : Next, I considered just using capacitors to remove the DC shift. I eventually moved away from these designs because the spectrum of audible frequencies was simply too wide for the circuit to be optimized across them all. If I eliminated delay in the high frequencies, it would be adding it to the low ones. This is because the capacitors always take some time to charge, adding unwanted delay and momentary DC bias to my signal chain.
• 8-15-23 : Finally, I came across a design for a level-shifter that used a single op-amp and realized that it was by far the best solution I had found. Firstly, the design could be modified to be powered by the same +/-15 VDC that powers the synth. Secondly, the design had almost zero delay. Initial simulation with this comparator design was extremely promising.
• 9-20-23 : After breadboarding the simulated circuit, I noticed the outputs' peaks were somewhat flattened. I realized this was due to the capacitive nature of the output load (guitar amp input or guitar pedal input in my case); so I edited the output resistors to provide the same output voltage ratios with more output current headroom. Finally, I also edited the rest of the resistors to minimize the amount of standby current usage. 
• 9-23-23 : A desire to add volume control saw me edit the output resistors to instead branch into separate paths; each with individual 100K potentiometers to control the output levels. After testing this final design and getting the results I was looking for, I added the module to the synth's PCB. 
• 9-27-23: After encountering repeated issues with my LF411 chips, I decided to replace them with a TI TL071; whose low noise, high slew rate, and high bandwidth provided a good balance between economy and specification. It also is a drop-in replacement for the LF411 so no rewiring would be necessary. 
• 12-04-23: Having finally gotten a chance to test my module in the signal chain it was meant for, I realized an error in my design. The output impedance was far to high and therefore beginning to distort the waveform. While keeping the attenuation ratios similar, I reduced the output impedance to around 5k ohms. I also added a 10 uF coupling capacitor while I was at it to absolutely make sure the output has no DC shift. After testing and getting the results I wanted, I went ahead and modified my synth board with the changes. The output waveform now looked almost exactly like the input waveform, showing me the importance of maintaining a proper output impedance for a given task.

Conclusions:

    While a unity gain op-amp with coupling capacitors initially seemed like the ideal solution, I found that a simple level-shifting design with a single op-amp ( currently Tl071) and resistors eliminated both delay and potential sources of EMI.  Final schematic, spice sim, and PCB integration: