Natural tone preamp

[Jeff Siddall]

This project is finally getting some more love, and one of the things I have been considering lately is op-amp selection. I was planning on using an OPA2145 (dual OPA145) which is pretty remarkable as a low noise, low power option. I even managed to snag a few before they went out of stock everywhere.

However, I just discovered today that it is remarkable in another way: it has an especially poor common-mode voltage range. It caught me off-guard because right in the title they call out "Rail-to-Rail Output" and on the front page of the spec sheet talk about how it is designed to run on +/- 2.25 V supply. But then they print this:

If you do happen to want to run it on a +/- 2.25 V supply then 0 volts at the inputs is well out of spec. In fact, the largest voltage you can apply is -1.25 V, and the largest AC signal you can apply is just 0.39 V RMS -- of course DC offset to -1.8 V.

Uhhh... really TI? You built a new op-amp that won't work on a ground-referenced AC signal over it's rated supply range? Sigh.

I might just stick with the OPA1692 since I have a few of those on hand, and they actually work decently at low voltages, though they are significantly more power hungry. There are also other options that are even lower power, and work really well at low voltages (ex: the OPA2314 at just 150 uA per amp that can go rail-to-rail on both inputs AND outputs), but definitely noisier. I could still try to use the 2145, but will need a ground reference of about -0.75 V. Even then, there may be pickups that put out more than a 0.4 V RMS signal, and since the first stage of the preamp is a non-inverting buffer, that could overdrive it.

Hmmm...

 

[bowthing]

I ended up dropping the bias voltage to about 35% of supply. The 145 still works if you don't, but harmonic distortion starts going up. It will handle most pickups down to 7V, but I recommend much more for hot ones. The 1692 is great but it has a large input offset current, so make sure your design handles the resulting voltage differences. JFET vs bipolar requires different designs.

 

[bowthing]

35% for single supply, that is!

 

[Jeff Siddall]

@SubSpec mentioned DeepPCB | Pure AI-Powered, Cloud-Native PCB Routing in another thread so I figured I would give it a go at this board.

Unfortunately it didn't end as well as I had hoped:

Time Spent:
1days 15min
Status:
Done
Connected / Total
212/216
Number of Vias Used
104
Total Circuit Length
24133785

After more than a day, and 410 "solutions" it hadn't actually routed everything and used about 50% more vias than FreeRouting to do it.

Oh well...

 

[John Burgess]

Dave Jones of the EEVBlog did a video trying out DeepPCB and was likewise totally unimpressed. I don't think any of the "AI" or even professional autorouters can match the logical flow of a person - yet. They don't split the design into tight functional blocks, or have an understanding of which nodes should be kept small, or separated, for example feedback nodes, or separation between fast switching digital traces and your high impedance analog signals.

I spent lots of time shuffling parts around before laying any traces.

Given the size of your design, I wouldn't shy away from trying a 4 layer board. Yeah you can get it onto a 2 layer, but I'm curious about the noise reduction and shielding that can be achieved with ground and power planes. 4 layer can be had quite economically now.

 

[Jeff Siddall]

Dave Jones of the EEVBlog did a video trying out DeepPCB and was likewise totally unimpressed. I don't think any of the "AI" or even professional autorouters can match the logical flow of a person - yet. They don't split the design into tight functional blocks, or have an understanding of which nodes should be kept small, or separated, for example feedback nodes, or separation between fast switching digital traces and your high impedance analog signals.

I spent lots of time shuffling parts around before laying any traces.

Given the size of your design, I wouldn't shy away from trying a 4 layer board. Yeah you can get it onto a 2 layer, but I'm curious about the noise reduction and shielding that can be achieved with ground and power planes. 4 layer can be had quite economically now.

Yes, I agree with all your points about autorouters, there is relatively little to configure in terms of telling the autorouter about "special routing" so no way for it to even know about how to handle certain traces. For now I have only hand-routed the decoupling capacitors. After testing I may do more.

Speaking of testing, I finished populating the prototype board and repaired it (unfortunately noticed a couple of issues with the schematic after receiving the prototype boards) so it is ready for testing. However, the test mule bass I need to stick the preamp into isn't done yet. To do this summer hopefully.

I did refactor the board layout several times to do the basic "best practices": physically separating the analog from digital. To that end I put the microcontroller and voltage regulators on one end of the board, mixed signal (digipots, analog switches) in the middle, and pure analog (opamps) on the opposite end.

And yes, 4 layer was always in my back pocket. It isn't inexpensive (doubles the price) to go to 4 layer so I wanted to do the prototype board as a 2 layer to see if it was possible to get all the traces in a fairly small footprint and then test how well it worked. If there were noise issues the next plan was to go to 4 layers.

If I eventually need to do noise reduction I expect it will be by adding 2 ground planes. Given the very low current requirements of this board I think a dedicated power plane would be less useful than a nearby ground (layers 1-2 and 3-4 are typically about 8 mil apart vs. about 40 mil between 2-3). I am not sure if I will move the grounds to the outside on layers 1-4 and put the signal/power as layers 2-3, or the reverse, put the ground planes as 2-3.

 

[BQFS]

Hi Jeff -

I'm one of the people who followed Charlie's OS preamp thread from almost the beginning. I fab onboard preamps based on the original Stingray ones. I was reviewing the OSP thread and found your reference to this thread. How did things turn out?

As a retired electronics tech, I wouldn't mind trying to put something like what you are proposing in one of my own basses. Did you complete the project? Inquiring minds need to know....

Bud

 

[Jeff Siddall]

Hi Jeff -

I'm one of the people who followed Charlie's OS preamp thread from almost the beginning. I fab onboard preamps based on the original Stingray ones. I was reviewing the OSP thread and found your reference to this thread. How did things turn out?

As a retired electronics tech, I wouldn't mind trying to put something like what you are proposing in one of my own basses. Did you complete the project? Inquiring minds need to know....

Bud

Sorry for the late reply Bud. For some reason I didn't get an email about your post. Need to check my settings.

Sadly things have not progressed at all. Other life (mostly cars, house) have gotten in the way of all bass related progress. But I do hope to get back to this project this winter. I will definitely post updates once there are some.

 

[BQFS]

Sorry for the late reply Bud. For some reason I didn't get an email about your post. Need to check my settings.

Sadly things have not progressed at all. Other life (mostly cars, house) have gotten in the way of all bass related progress. But I do hope to get back to this project this winter. I will definitely post updates once there are some.

Fair enough. I've had enough disasters in the past month myself. Hip replacement surgery, my wife passing away in early September, and catching Covid shortly after that. Keep in touch....

 

[Jeff Siddall]

Fair enough. I've had enough disasters in the past month myself. Hip replacement surgery, my wife passing away in early September, and catching Covid shortly after that. Keep in touch....

That is terrible news Bud. I am sorry for your loss.

EDIT: I liked your post in sympathy

 

[Jeff Siddall]

I will add a bit of a teaser: I have actually made some small progress on this project. Initially I powered it from a bench power supply on the XLR instead of phantom but quickly discovered that caused a voltage spike on the 5 V rail due to protection diodes on the op amp outputs. That fried the voltage regulators so I needed to order some more of those which delayed things a bit.

While waiting for parts I also started thinking about another preamp design, still computer controlled, but using only digipots with no switches at all. This allows for the obvious benefit of being able to blend pickups. I have always been interested in Alembics so decided to also implement a tunable resonant filter, somewhat like Alembics use (and also the @Passinwind preamp for that matter). I dubbed this the "Unnatural Tone" preamp and now have some boards fabricated and waiting for population and testing. I will probably make a separate thread for that project eventually but for now here is a photo of the new "unnatural" board on the right, along with an expansion board that can be soldered onto a header on the base board to allow for another 4 pickup inputs (9 coils total) for particularly crazy setups:

For comparison the original natural tone board is on the left, and is almost 19 mm (3/4") longer.

Also, I got some tiny boards done to allow mounting small 2020 (2 mm or just over 1/16" square) addressable LEDs that would be suitable for horizontal mounting as side dot markers:

That's actually 2 boards stacked (still needs to be cut in half) with a 0.5 mm mechanical pencil for scale.

Someday!

 

[BQFS]

Nice work! Please stay in touch. If you need a beta (or even an alpha) tester, keep me in mind, as I am a retired electronics tech with over 50 years working on this stuff.