voltage clamping with a battery as power supply.

[Eddododo]

I am working on an onboard preamp- it will be powered nominally at 18v by two 9v, but in reality the op-amps themselves are 16v max. (I'm using these Op amps (TLC2264) largely to help battery life in the first place). I'm trying to figure out a good way to choke down the voltage, with battery life in mind.

Because of the battery, I'm starting with the assumption that an active regulator is off the table..
Similarly, I thought of something like a reverse zener to ground to choke it like the Dr Quack bias trick, but this is again bleeding precious current

The obvious choices are a) series diodes, or b) a series resistor to form a voltage divider at the 'top' of the Vbias voltage divider, to lower the 'source' voltage.

Is one of these better than the other? Is there a better way?

Edit: I should add that I'll be using the buffered bias 'trick,' IE feeding a voltage divider into an op amp  buffer to create the Vbias.. not sure if that changes anything

[PRR]

Instead of doubling the voltage, why not run half the gain? I'm sure 3Vrms is ample to overload your main amplifier.

[Eddododo]

To elaborate/clarify a little bit while I give people time to respond-

The left side of the schematic is a battery indicator- if the zener voltage is exceeded by the supply voltage, it will briefly flash the LED.

It's an onboard preamp/eq for a bass I'm building for someone.
I'm obviously using two 9v for the added headroom; I don't feel like I need 18v or anything, I just never feel like 9v is enough to comfortably work with for a versatile EQ..

As far as the op-amp, I wouldn't say I'm married to that one, it was just a good candidate for battery life, and had good characteristics otherwise.. I'm just likely to get lost in the weeds if I try to start over choosing one



For over voltage protection, I was also thinking of a 16v zener to shunt any spikes from static or perhaps an especially excitable battery

[Eddododo]

Edit: something went haywire and I multiposted..

[Eddododo]

Instead of doubling the voltage, why not run half the gain? I'm sure 3Vrms is ample to overload your main amplifier.

Hmm, it's certainly worth consideration, but a pretty big design goal is that the preamp would be dead unity when flat, such that there's no volume change if the preamp is disengaged.

It's for a bass, and one that will be run direct regularly- on the one hand, while we don't need to worry about nicely driving an amp in that way, it certainly would also help to not overdrive a DI or whatever else.. but the flip side of that is that we're dipping closer to the noise floor on something that will be used completely clean, and in very sound-exposed environments (he plays a lot of theatre gigs, for example)

[Eddododo]

I guess I should really put it this way- I'm not as clear on managing/budgeting current as I'd like to be.. I'm good to go on the algebra of ohms law and power, but a little murky on the behavior of the components

-for series diode drops to lower the available V+, am I somehow 'wasting' any current draw? Or do they just behave as a short (minus the tiny inherent resistance of all components), just with a voltage drop? The power dissipated is the forward voltage times the current THROUGH it, as drawn by whatever is next in series?

-for the voltage divider approach, is there a downside? Seems that I'm drawing LESS current, with a teenytiny bit of added noise.. is there some kind of stability trade off or something else that's flying over my head? The only reference I ever see is people AVOIDING voltage drop from excess resistance in power filtering etc, but it's the one thing I want!


Because even when the batteries are at EOL, minus another ~3v from the choke... I'm still at 12v, which beats the hell out of 9.. I'm not looking for huge gain outright, I just want clean headroom to boost frequencies and attenuate to perceived unity.. it would also allow me to either frontload the gain adjustment stage, or at least lead with a gain bump for noise without clipping everything after, even if I end up with makeup gain/atten at the tail end.

Edit: crap, this is probably important to add, I'll add to the top post, but I'll be implementing buffered bias 'trick' (voltage divider vref into an op buffer).

[Rob Strand]

The simplest way of dropping *small amounts* of voltage is a silicon diode.   For 2V use three diodes in series.   For low currents you can even use LEDs (watch out for start-up surges through the supply bypass caps).  The advantage of diodes have over a dropping resistor is the voltage drop is substantially constant with current.  For a change of 10 times the current each diode will rise less than 100mV in voltage.  Diodes can handle quite a bit of current.  A resistor voltage dropper is all over the place and won't work with some circuits.   There's also the option of a VBE multiplier but you need to put some care in the choice of base resistors.
https://en.wikipedia.org/wiki/Rubber_diode

[antonis]

I should add that I'll be using the buffered bias 'trick,' IE feeding a voltage divider into an op amp  buffer to create the Vbias.. not sure if that changes anything

It depents both on divider resistors values (for entened battery life) and current draw..
For currents larger than 5mA (or so) the use of a series pass transistor should be suggested..

[ElectricDruid]

As far as the op-amp, I wouldn't say I'm married to that one, it was just a good candidate for battery life, and had good characteristics otherwise.. I'm just likely to get lost in the weeds if I try to start over choosing one.

Given that it's the op-amp choice that's giving you this problem, having another look at that seems like a wise way to deal with it. That op-amp is intended for +5V or +/-5V operation. +/-8V is the "absolute maximum" supply, and it wouldn't be a great idea to run it permanently at that level. If you ran it in it's "comfort zone" of +/-5V, you'd hardly gain any headroom over a single 9V supply, so it doesn't seem like a good choice.

I'd look for low current audio op-amps. They'll all work comfortably at +/-9V and then you won't have a problem. Back in the day (1980's) I used to use The TL06x series and LF444 for battery-powered pedals because they didn't use much power, but these days there are probably better options.

I agree that *finding* those better options is pain in the neck. TI have a great parametric tool for finding op-amps. Here's the page for audio op-amps:

https://www.ti.com/amplifier-circuit/op-amps/audio/products.html

The only problem is...it doesn't seem to include supply current as a parameter! Argh!

[Christoper]

I also agree that running it at 9v might be the play here. Double the batteries for added headroom that you likely won't even use isn't the biggest appeal to me personally. Why won't your circuit run at unity when flat? That's usually a tunable parameter that's independent of supply voltage, provided there is enough gain to achieve unity in the first place.

[amptramp]

There are other ways to do this.

If you don't want the limitations of battery operation, you can have the cable go to a box external to the guitar and the circuitry of a 4 to 20 mA process link outside of the guitar.  If you are using a pedalboard, this is just another box to go on it but it would not need a stompswitch since it would always be on.  Since you are not providing power on board the guitar, you can use ordinary power supplies (although I would select them for low noise or add a low-noise regulator).  This eliminates having to carve up a guitar to add a battery holder.

You could also use an RF link from the guitar to the box via a carrier current transmitter and receiver.  This uses a transmitter in the guitar and a receiver in the box to modulate the power supply so the data can be recovered.  Carrier current is the modulation of a power line with RF to enable signal recovery in the receiver at the box end.  Once again, no need for a battery in the guitar but the box would need to be there.

[ElectricDruid]

What level is the output from the pickups?

I'm feeling like we need some facts here. Whether 9V is enough headroom depends on (a) what output the pickups produce, and (b) how much gain and/or boost the EQ provides.

My experience is that bass pickups produce a much higher level than guitar pickups when you start getting into some serious slapping and such like and that you need a lot more headroom for bass as a result. But that's in general, and we're talking about a specific instrument that this EQ is going to be fitted to.
 

[Eddododo]

There are other ways to do this.

If you don't want the limitations of battery operation, you can have the cable go to a box external to the guitar and the circuitry of a 4 to 20 mA process link outside of the guitar.  If you are using a pedalboard, this is just another box to go on it but it would not need a stompswitch since it would always be on.  Since you are not providing power on board the guitar, you can use ordinary power supplies (although I would select them for low noise or add a low-noise regulator).  This eliminates having to carve up a guitar to add a battery holder.

You could also use an RF link from the guitar to the box via a carrier current transmitter and receiver.  This uses a transmitter in the guitar and a receiver in the box to modulate the power supply so the data can be recovered.  Carrier current is the modulation of a power line with RF to enable signal recovery in the receiver at the box end.  Once again, no need for a battery in the guitar but the box would need to be there.

A bit too exotic and limiting for the end user.

What level is the output from the pickups?

I'm feeling like we need some facts here. Whether 9V is enough headroom depends on (a) what output the pickups produce, and (b) how much gain and/or boost the EQ provides.

My experience is that bass pickups produce a much higher level than guitar pickups when you start getting into some serious slapping and such like and that you need a lot more headroom for bass as a result. But that's in general, and we're talking about a specific instrument that this EQ is going to be fitted to.
 

Working currently to build up the blocks together and test for clipping when boosting bands. I've found that simply big fat low notes are way more of a headroom issue than slapping (if you don't play like a high schooler imitating flea).. On the bass side, it's pretty idiomatic to boost the lows and/or mids pretty dramatically, and lower the output to compensate; in that context, you can pretty easily clip at those bands and downhill stages, whether or not the resultant output sounds quieter than a 'dry' signal. There are considerations like placement of the gain control, but let's not get sidetracked

There's also the added advantage of a much, much longer usable battery life.

I have consistently, for about 20 years, preferred 18v operation for preamps in my basses (and have also consistently preferred either 18v operation for EQ pedals, or else required EQ that has both input and makeup gain/level controls)


But really, really really, I'm mostly just trying to understand if there is a significant drawback to diode-dropping the supply voltage dramatically.  Does it eat current in some way I didn't understand? Is the heat/dissipation more of an issue than id guess? Radical noise issue?

 I do get that the voltage drop will change a bit in response to changes current, but is that actually something that will be happening in a meaningful way? (I guess I don't understand whether amps draw more current when there is a signal being amplified... my gut tells me yes)

[PRR]

whether amps draw more current when there is a signal being amplified... my gut tells me yes

Depends on the _load_. Some amps, un-loaded, do not increase supply current. If it does, it is probably a "class A or A/B" output stage and a fairly low load. Loudspeaker amps nearly all suck more current when beating a speaker loud. Dinky amps you'd put in an instrument: it depends, and the designer/maker may not want to come right out and say.

But "stage loads" are nearly all 50Kohms or more. A 5V peak in 50k is 0.1mA. The average of a 0.1mA peak sine is 0.03mA. The idle current of most listenable opamps is more than that, so you will hardly notice.

Still in all, with your concerns, you should probably be prototyping with a mA-meter in line. You may find that a 200% test tone does raise the supply current, but real runs at real levels don't raise it high enough long enough to matter.

Meanwhile, just diode-drop. It is simple and has only the obvious inefficiency (2V of your 18V is lost).

[Eddododo]

SO.. im hesitant to post this while I'm still figuring out if i did it right, but perhaps someone is bored enough to check my work..

I'm using REW (room eq wizard) which has a scope function and a tone generator. I calibrated my soundcard, and then set up a tone into the scope. I generated a 1v peak sine- a .707V RMS signal- into the scope, and scaled the scope to a .707V RMS FS sine... after having done this wrong a couple times, I have what should be a 1v signal peak showing up as a 1v signal peak in the scope.

From there, using the same calibrated input, I plugged in my bass and hit things hard.. different ranges, techniques, stroke direction. For the most part, the biggest peaks were between about .8-1.3 V, and that's only the initial transient.. after that, the rest of the signal was .5-1V peaks at most, most of the reasonable playing was more like .5 peak...

I also recorded the sine output into the looper, played it back to check that it matched the generated sine ( it did) and went back and forth between the looped sine and the live bass, just to keep the wheels on the track.

I ALSO took the multimeter and read the voltage at 60hz from the output of the generator/interface, which showed ~2V (again a 1v peak), and measured the same from the looper, showing the same

So.. I do feel somewhat confident about the measurements, leading me to assume a max PtP of about 2.4V, a practical max of more like 2, and a real working signal level of more like 1.2 V PtP..
I don't think I need the ABSOLUTE max in the plan, as it look a lot of effort to even get there, but I think the 2V line is a pretty good rule of thumb for a variety of techniques etc.

Back to the lab to figure out some numbers from the circuit.. I guess I could really post the circuit at this point.. I was initially trying to keep it pretty general so I could use the information devoid of context, but it starts to feel silly being overly hypothetical


I have started to pick out some op-amps that seem good.. my main issue is that basically nothing I've found seems to come in a through-hole package.. If things look bleak headroom-wise, I will probably use some converter boards for the SOIC chips, but I have a feeling that I'll be more than fine with a ~12v operation of this chip, if not 9.

[ElectricDruid]

Ok, so if we're looking for a 2Vpp signal to go through this circuit with no distortion, you can only boost it about x4 on a 9V supply. Even that might be a bit much depending on the op-amp choice. Some go closer to the rails than others, but most won't produce a 9Vpp output. TL072 will only give you about 7Vpp on a 9V supply, for example.

x4 is +12dB, which sounds reasonable for an EQ, so if your EQ bands don't give a lot more boost than that, I'd say that you should be pretty much ok. You might get a bit of "bite" on particularly loud transients, but that'd be a *feature* not a fault!
On 12V, you'd get a little bit more breathing space and *might* be able to keep a really clean sound even for more dynamic signals. Which is why people go to 18V - to *ensure* that's the case. Depends what you're after.

HTH

[Eddododo]

Appreciate the help across the board!