Onboard bass preamp / convoluted effort

[Eddododo]

I've been working on this onboard preamp for a bass build for awhile now, and as I'm tying up the loose ends I'd love to get some eyes to see if I'm missing anything (and also y'know, post about it).

A big area of concern is managing the DC conditions; I think I have taken the needed care for the current / battery management, DC offset / decoupling decisions, etc.. but I'm operating on a foundation of a kludged-together knowledge stemming from stompboxes, and we often don't have to do much past the bare minimum with regards to DC..

Also watching for things that may be not showing up in LTspice.. I'm using universal op amp models for most of my sims, since I don't have a working model.. So I've been squinting at datasheets and trying to apply the math and specs, I'm just not always sure I'm applying it correctly.

I'm using OPA2990 for the opamps; super low current draw in particular. It's being run off at 18v. Link to component page

The circuit has Gain Bass Mid Treble, as well as 4 'contour' options;
-The bass and treble are peaks when boosted, shelves when cut, tuned to frequencies I felt emphasized the instrument well and in a natural way.. no snake oil!
-The mid control is rather broad and general, sort of just managing the meatyness of the signal.. 
-contour 'a' is a 3-way with two different scoop/bump contours derived from the 'pulltec trick' bass circuit
-contour 'b' switches between a 2k boost 'presence,' and a low mid-boosted signal with a HUGE Rolloff of the highs

The switches make it look kind of messy.. I'll try to follow up with a simpler version.
The switches are 3pole 3-position slider switches.. the 3 positions connect 1-2, 2-3, and 3-4

A single pole of the switch would be


So here's the circuit




The battery indicator is neat.. when the circuit is powered, and it is at least 17.5v, two LEDs (red and yellow) will flash briefly, letting you know that it has a healthy full pair of batteries. When it dips below 17.5, but above 16v, only the yellow LED flashes, letting you know it's time to make sure you have a battery and/or change it. With no flash, you're playing 'how long does my car go on E.'

[antonis]

A big area of concern is managing the DC conditions;

Indeed..

e.g. U2 & U3 outputs have DC gain of X3.2, resulting into about 15mV offset (10pA bias current X 470k bias resistor X DC gain)..
This might seem of no significant importance but, IMHO, a couple of blocking caps in series with R26 & R29 should give you peace of mind..
(the above offset should be equally settled with caps in series with IC2 & IC3 respective outputs but it should call for IC1 extra bias resistor and signal loss..)

edit: Forget all about it..
(wrong decimal points, wrong voltage difference - more coffee, plz..)

[fryingpan]

Why R13 = 300 ohms? It could/should be much lower.

[antonis]

Why R13 = 300 ohms?

As far as affordable voltage drop across it is OK, let it be..

It could/should be much lower.

On the contrary, C5 should be 10 times bigger..!!
(despite opamp's high PSRR)

[fryingpan]

Why R13 = 300 ohms?

As far as affordable voltage drop across it is OK, let it be..

It could/should be much lower.

On the contrary, C5 should be 10 times bigger..!!
(despite opamp's high PSRR)

I presume that, being the circuit battery powered, ripple is low/non-existent.

[antonis]

I presume that, being the circuit battery powered, ripple is low/non-existent.

In case of battery (2X9V) powered preamp, many items, like D5, R13, C5, R14, R15, C6 & U5, are redundant..

[sergiomr706]

-The bass and treble are peaks when boosted, shelves when cut, tuned to frequencies I felt emphasized the instrument well and in a natural way.. no snake oil!
Sounds interesting, which frequencies did you choose? Or could you please provide formula to calculate this tonestack? Thank you

[JTEX]

At first glance:

C9 & C10 are way too large. 100nf would be plenty enough to pass the full audio band. With 10uf, the circuit will take a long time (some seconds) before reaching a stable DC operating point on powerup. Calculate the RC time constant between C10&R24 to see. In general, always a good idea to calculate RC time constat of caps that charge through resistors on powerup. Or you might have to wait for a while before you get any sound...

Also, I'd really consider OPA2196 instead of OPA2990. Half the noise, similar power. Unless cost is a significant factor.

Couldn't the bass and treble be done with a single op amp? In general, it seems to me there are more opamps being used than would be strictly necessary to achieve the circuit's functionality. But that's just a quick look.

R13/C5 is redundant for battery power. There's no ripple to worry abount.

[duck_arse]

is R13 and D5 part of your scheme to reduce the increased supply voltage to stay within IC specs?

[Eddododo]

I presume that, being the circuit battery powered, ripple is low/non-existent.

In case of battery (2X9V) powered preamp, many items, like D5, R13, C5, R14, R15, C6 & U5, are redundant..

I'm curious why the Vref voltage divider resistors would be redundant? I chose to use a buffer for vref largely because I had an extra op amp, but I also typically use one just because it's so easy, lol..

I was using C5 and C6 because I'm stuffing SO many wires into tight routes that it seemed like a decent idea to help with crosstalk noise, but I'm not certain if it's actually effective

The reverse protection felt smart because it's easy to apply reverse voltage while feeling out a battery snap in the dark etc..though you'd have to do it with BOTH batteries, since they're in series, lol

I know R13 is HUGE, and I don't have a great reason for that; honestly, I was just nervous that I misunderstood how the IC specs apply, Or that I was otherwise missing or miscalculating the current draw..  does seem unnecessary with a battery though

[Eddododo]

At first glance:

C9 & C10 are way too large. 100nf would be plenty enough to pass the full audio band. With 10uf, the circuit will take a long time (some seconds) before reaching a stable DC operating point on powerup. Calculate the RC time constant between C10&R24 to see. In general, always a good idea to calculate RC time constat of caps that charge through resistors on powerup. Or you might have to wait for a while before you get any sound...

Also, I'd really consider OPA2196 instead of OPA2990. Half the noise, similar power. Unless cost is a significant factor.

Couldn't the bass and treble be done with a single op amp? In general, it seems to me there are more opamps being used than would be strictly necessary to achieve the circuit's functionality. But that's just a quick look.

R13/C5 is redundant for battery power. There's no ripple to worry abount.

Thanks for the tip on the op amp! It was a big task for me to find one, I've never had to source one that wasn't just a common stompbox chip or stuff like Burr Brown etc.. Ill be ordering some shortly

Good catch on C9 C10- I have them as I think 220n on the board now, for exactly that reason. I didn't do the math, but is was several seconds for the circuit to 'power up' with the 10u

For the bass and treble- even on a standard baxandall I just didn't like the interactivity. Granted, its so ubiquitous that its self-evident that it's a non-issue in practice, but when A/B testing, I preferred them with independent stages..

BUT because of how the Bass stage is modified from bax to integrate the peak boost/ shelf cut response, it doesn't play nice with the treble stage on the same stage anymore, and I never got a version worked out that played nice.

[Eddododo]

is R13 and D5 part of your scheme to reduce the increased supply voltage to stay within IC specs?

Why R13 = 300 ohms? It could/should be much lower.

I was nervous about miscalculation of current drawn or of the Op amp specs, so 300 was kind of a goofy gut-based middle ground, almost a superstitious choice lol! I chose it, then just checked if it got warm at all when powered... One of the things I was hoping to clarify from this thread, for sure


I also realize I misspoke in the original post; The treble does not have an asymmetrical response. It initially did, and sounded nice on my own bass, But once the actual bass build was far enough to be tested with the circuit, it just sounded better with a shelf response in the highs

[duck_arse]

"it got warm" is somewhat less precise than using the meter. measure the voltage from end to end of that resistor, work out the current into the rest of the circuit. also figure the power dissipated and see if you should expect any [detectable] heat.