OPA1652: Question on the rail performance

Hi Craig, 

Cool project! You are considering some important tradeoffs on the devices. Unfortunately I can't seem to find a plot showing the Vos vs Vcm data for the OPA1652 but I have experience with the device. The OPA1688 does have the Vos vs Vcm curve (see below) and the OPA1652 has a similar non-linear increase in Vos when going beyond the Vsupply-2V point. 

Due to the non-linear increase in Vos you can see where the THD would go up when entering this region. Only a listening test will tell if this really matters however. 

When I wrote my application note (link below) on "How to Measure Total Harmonic Distortion of an Op-Amp and THD + N Fundamentals" I would visually watch the FFT change as I pushed the devices into these non-linear regions while playing music and listening on headphones. I really had to push the signal into harsh clipping for it to really stand out. This can be taken with a grain of salt however since I wasn't in a perfect listening environment etc. Nonetheless as someone who mixes my own music in a calibrated home studio I was pleased with the device performance. When listening to something very clean I could slightly hear the distortion and it was subtle. When listening to guitar parts it wasn't obvious until the signal was harshly clipped. For my application note I used the OPA1656. You may also consider this device. It does have the same type of common mode limitation however. 

How to Measure Total Harmonic Distortion of an Op-Amp and THD + N Fundamentals:

https://www.ti.com/lit/an/sboa580/sboa580.pdf?ts=1725379948785&ref_url=https%253A%252F%252Fwww.google.com%252F

OPA1656 Vos vs Vcm:

I recently did an application note that does apply here even though mine is differential inputs. A different design could be done with single in and single out with an op amp instead of an INA. The application note is titled "Ultra-Low-Noise JFET Preamplifier Design for High Impedance Sensors." Your thinking is correct with the Humbucker pickup. In addition with the case of a humbucker pickup you can see below that the source impedance increases greatly over frequency (see below). For this reason noise should also be considered.

https://www.ti.com/lit/ab/sboa589/sboa589.pdf?ts=1725381426521&ref_url=https%253A%252F%252Fwww.ti.com%252Fproduct%252FINA849

Humbucker Impedance vs Frequency:

From my app note I compare sensor source impedance with the bipolar INA849 alone vs replacing the front end of the INA849 with our discrete JFET JFE2140. The same type of analysis can be performed in Tina Ti simulation as well for other configurations. 

We also have a single discrete JFET called the JFE150. I wrote an app note and created an EVM preamplifier for it as well that is available online. 

JFE150 Ultra Low Noise Preamplifier

JFE150EVM

All in all there is no perfect solution however I think with some listening tests you will be surprised on the performance of the devices. The noise may end being higher on the list for consideration over THD. 

I hope this information helps. Let us know if you have any further questions. We have a few of us here that are hobbyist musicians and enjoy these projects. 

Best Regards, 

Chris Featherstone

Hi Craig, 

I should also acknowledge it is easier said than done to set up a bunch of boards for listening tests or other op amp tests. For this reason we recently released an op amp evaluation platform that is solderless. This can be found below. I only mention this incase you find it useful for your evaluation. 

https://www.ti.com/tool/AMP-PDK-EVM

Best Regards, 

Chris Featherstone

Thanks so much, Chris, this was the sort of info I was hoping for. It's always good to study the available literature and the efforts of others before I head to the bench, but you're right, it's going to come down to listening tests, and I have a few parts to try a few approaches. The PSRR you get for free using some of these audio TI op-amps is so handy, and not knowing what power supply people might use 3 years from the time they buy the product, that has to be a good thing. But everything is a trade-off, as you say. I typically like having discrete front ends with variable gain to handle a range of inputs. I had seen your JFE150 preamplifier design, and you get great performance from that. I guess I thought that would best apply for microphones or very low-level signals, but it could be configured for less gain and tested out with a guitar pickup. Interesting that TI are still flying the JFET flag when so many models have been discontinued. I'm not aware of any low-level MOSFETs that TI produce targeted at audio, but correct me if I'm wrong, because a simple MOSFET follower stage could work too, with maybe a smaller voltage drop than a JFET.

Hi Craig, 

No problem! Thanks for checking out my work. Yeah the JFE150 circuit is closed loop so you can get predictable gain and set the gain as desired. I did a gain of 1000 to show it off. One caveat to that circuit is that you need clean power supplies because as you stated you get free PSRR with the op amps as opposed to the discrete solution. We recently started doing the discrete JFETs as we found a need for them in the market. I am not too familiar with our MOSFET line as they are handled by a different team. We are also doing discrete matched resistors now such as the RES11 if you find a need for them in a discrete solution. 

https://www.ti.com/product/RES11A

Best Regards, 
Chris Featherstone