Hey all, this is my first post at diyAudio, so apologies if this is the wrong forum for this but it seemed most appropriate given what I'm trying to do.
I'm working on the second revision of a DSP-based instrument (guitar/bass) effects board, and while I understand the digital side of things quite well, I'm in a bit over my head as far as the analog parts go. I'm intending on using an ADAU1361 as the codec to get the audio data in and out of the DSP (an ADAU1452), and thus need to be able to properly interface the raw input signal from the instrument, get it into and out of the codec, and then be able to drive whatever comes after in the signal chain - an amp, another effect, etc. The previous revision of this board used a pretty simple opamp-based input and output, but I'm not sure how well it actually performed, and this is the reason for my post.
At the moment, the input side has the following goals:
The output side I'm not quite as confident about, but is basically the mirror of the input:
The ADAU1361 codec supports both single-ended and (psuedo-)differential inputs and outputs, and I plan on having the input driven pseudo-differentially (to have it go through the PGA), and the output driven single-endedly.
Now, I can build up these circuits - opamp based, sallen-key topology for the filters, etc. - but I'm not really sure how to properly analyze them to determine ahead of time how well they're going to perform. I can build up the circuits in SPICE, and do some analysis, but I'm not really sure how to determine what the analysis is telling me or whether or not the changes I make are going to correspond to a better final result.
For the first iteration of this board, I used an input buffer/filter that I found from some bass guitar effect and the output buffer/filter as specified in the datasheet for the DSP (ADAU1701) I was using. As mentioned before it DID work, but sounded very tinny and hollow. Ideally, I want it to be as transparent as possible; there should be very little difference between having this in the signal chain (with the DSP just doing a pass-through of the input data to the output) versus not having it in the chain at all.
I'd prefer not to mess things up and have to respin the board, so any advice you all can give me on how to do this right the first (second?) time would be greatly appreciated. What properties should I look for in the opamps I'm using? What analysis should I do in the SPICE sim? How can I make sure that the design is going to behave correctly before I commit the design to PCB? I could potentially breadboard the input and output filters, but I'm not really sure how to properly measure that they're doing what I want them to do. I do have an oscilloscope and a signal generator, but I don't know what to measure beyond checking that the signal goes in and comes out biased correctly.
This schematic is what was used for the first revision.
Any suggestions or tips in getting this design up and running are greatly appreciated!
I'm working on the second revision of a DSP-based instrument (guitar/bass) effects board, and while I understand the digital side of things quite well, I'm in a bit over my head as far as the analog parts go. I'm intending on using an ADAU1361 as the codec to get the audio data in and out of the DSP (an ADAU1452), and thus need to be able to properly interface the raw input signal from the instrument, get it into and out of the codec, and then be able to drive whatever comes after in the signal chain - an amp, another effect, etc. The previous revision of this board used a pretty simple opamp-based input and output, but I'm not sure how well it actually performed, and this is the reason for my post.
At the moment, the input side has the following goals:
- High input impedance, to avoid loading down the pickups of the instrument
- Unity gain; the codec contains a PGA with between -12dB and 35.25dB of gain so I can amplify the signal after the fact
- Single supply, as the entire board is 3.3V
- 20-20kHz low-pass filter, to keep things inside the audio range
- 1Vrms full-range input to the ADC
- Single-ended, going pseudo-differentially into the ADC
- Low noise; the source signal could be very weak and may need to be amplified significantly by the PGA
The output side I'm not quite as confident about, but is basically the mirror of the input:
- 1Vrms DAC full-range output
- Low output impedance, to drive whatever comes next in the signal chain
- Filtering? Not sure what (if anything) would be needed here that the DSP couldn't do itself
- Single-ended
The ADAU1361 codec supports both single-ended and (psuedo-)differential inputs and outputs, and I plan on having the input driven pseudo-differentially (to have it go through the PGA), and the output driven single-endedly.
Now, I can build up these circuits - opamp based, sallen-key topology for the filters, etc. - but I'm not really sure how to properly analyze them to determine ahead of time how well they're going to perform. I can build up the circuits in SPICE, and do some analysis, but I'm not really sure how to determine what the analysis is telling me or whether or not the changes I make are going to correspond to a better final result.
For the first iteration of this board, I used an input buffer/filter that I found from some bass guitar effect and the output buffer/filter as specified in the datasheet for the DSP (ADAU1701) I was using. As mentioned before it DID work, but sounded very tinny and hollow. Ideally, I want it to be as transparent as possible; there should be very little difference between having this in the signal chain (with the DSP just doing a pass-through of the input data to the output) versus not having it in the chain at all.
I'd prefer not to mess things up and have to respin the board, so any advice you all can give me on how to do this right the first (second?) time would be greatly appreciated. What properties should I look for in the opamps I'm using? What analysis should I do in the SPICE sim? How can I make sure that the design is going to behave correctly before I commit the design to PCB? I could potentially breadboard the input and output filters, but I'm not really sure how to properly measure that they're doing what I want them to do. I do have an oscilloscope and a signal generator, but I don't know what to measure beyond checking that the signal goes in and comes out biased correctly.
This schematic is what was used for the first revision.
Any suggestions or tips in getting this design up and running are greatly appreciated!