Noiseless Biasing, 470k or 4.7M series resistor?

[R.G.]

I couldn't get it working at all until I added C6 from R2 to ground as you suggested. I had thought this was an optional part.. Thought I'd seen schems without it.
Maybe they had bi-polar supplies instead of biasing? Do you need this cap just for the circuit to function?

That cap makes the gain go down to 1 at DC. Some bipolar supplied opamps can do without it, but it makes the difference between the circuit amplifying DC or not.

Do I need to lower the value of the resistors in the voltage divider? (R6,R7) - If I use op-amps with bi-polar inputs later in the circuit, would I need more bias current than these can provide?
(Let's say 4 op-amp sections, from 2 dual devices, a TL072, & an LM833.)
I could make R6 & R7 100k each? (Or should I make a reference voltage using an op-amp??)

There is a set of reasoning for the values, but while you're learning, just make them 10K each. Get the rest of the circuit doing what you want, then worry about changing them.

I don't expect current to flow in D1,D2 under normal conditions, it's just a protection clamp. I just guessed at a value for R9 though.
I was thinking: if someone plugs the output of another pedal into this input, rather than just a guitar, there could be a few volts of signal. - I don't want these diodes to clip/clamp audibly with such a signal, nor the op-amp, so I added R9 to raise the threshold, & soften any clipping. What value should I make R9?

Make R9 be zero ohms.

The diodes will then be between the two inputs. The inputs will have to be more than one diode drop (500mV - 700mV) apart before the diodes can possibly conduct. The opamp will maintain the inputs within 10-20mV at all times when it's working normally. If they get more than the few millivolts apart, you're no longer worried about softening clipping, you're trying to prevent damage to the inputs. You don't need R9 to soften anything. The opamp will follow even volts of input without letting the inputs get more than tens of millivolts apart.

[PRR]

> D1,D2 ... a protection clamp. I just guessed at a value for R9 though.

R9 makes this a soft-clamp. It works against R3 and R2/R1, and a large fraction of overvoltage at the jack is imposed on the opamp input pins. It is nearly-no protection.

R9 could be zero.

But there's a better way. Nearly all common chips can take 10mA fault current. Guitar inputs can have 10K series resistance (your R3) without degrading the signal. With 10K at R3 it takes 10mA*10K= 100 volts to blow the chip. This is generally ample. And 2 fewer parts.

> if someone plugs the output of another pedal into this input..., there could be a few volts of signal. - I don't want these diodes to clip/clamp ..., nor the op-amp...

Think. Put 2V signal in. R1 R2 program a gain of 6. The output is 12V. However the power is 9V. 2V input WILL clip the opamp. Clamping the inputs does not change that.

If you do expect 2V inputs, put a switch to break the R2 leg. Now it works at unity gain. 2V in is 2V out. 3V in will probably clip slightly. More than 3V in is unlikely in a 9V-power world.


> If I use op-amps with bi-polar inputs later in the circuit, would I need more bias current than these can provide?

As drawn now, the bias-divider ONLY feeds opamp bias current. (Some other plans connect many other things to Vref).

The Input Bias Current is on the opamp datasheet.
http://www.national.com/ds/LM/LM741.pdf
Page 2: 80nA typ 200nA max at room temperature, 0.8uA at extreme temperature.

> Let's say ...an LM833

http://www.national.com/ds/LM/LM833.pdf page 3.... Hmmmmm. LM833 is a real pig for bias current, and not real suited to inputs needing "high" bias resistors.

For tutorial, let's stick with '741 or better, 80nA to 0.8uA Input Bias Current.

This works against the bias network. This is 1Meg plus 250K||250K or 1.125Meg.

Too many zeros!

I know that 1uA in 1Meg is 1V.

Ha, 0.8uA in 1.125Meg is not quite 1V.

"Typical" is 10 times less, so 0.1V error.

The base idea is to hold the input "Half Way" between the +9V and ground. Same as you put your car in the garage halfway from each wall, for maximum room both sides.

But a 6 foot car in a 10 foot garage, you can be a foot off "halfway" and still be fine. Likewise the "half of 9V" or 4.5V can probably be 3.5V or 5.5V and the circuit will work.

With a "typical" chip in decent temperatures, the error is more like 1/10th of a volt. Un-important.

The LM833 with anything like a Meg of bias resistor will hang a volt away from Vref. That's pretty sloppy. Go ahead and stick one in. Report back with the difference between Vref and the opamp's output DC voltage.

[matt239]

Thank you guys so much!
I'm getting closer. I read stuff, work on it, you guys explain it, then I read & work more, & it starts to make sense.
I'm still working on basics obviously, but jumping in & building stuff helps me absorb it.

OK, so the diodes won't conduct except in real FAULT conditions, not with just some volts of signal. R9 is superfluous.. (actually, counter-productive..)
Is it good practice to have the diodes? When do you need them vs. not?

10k series resistor won't affect signal/noise too much?
Hmm.. I should try to calculate that myself...
______________________________________________

I'm planing to build a device with 1/2 a TL072 as input, LM833 as subsequent stages; bias resistor for LM833 can be much smaller than 1M..

[PRR]

> is it good practice to have the diodes?

Opinions vary. In _this_ case, _I_ say they are a total waste of 20 cents and 2 minutes time. You can put some resistance between jack and chip, a reasonable resistance will protect against insane external voltages.

There are similar cases where you _may_ want a diode-pair across the inputs. If a BJT pair breaks-down even "non-destructive", its DC offset changes and noise may rise a bit. But that matters mostly for much more precise measurements than guitar amplification.

> 10k series resistor won't affect signal/noise too much?

Millions of gitar amps have 34K in series.

Can you calculate hiss? (It is a rare skill.)

"Too much" is a key point. The naked guitar pickup is 5K resistance. Another 10K of dead resistance thermal hiss would be audible. However the guitar volume pot tends to pad this up to 10K-40K (and nobody objects to vol pots on the guitar). As an extreme the pickup and cable resonate to over 100K around 3KHz, and hiss at 3KHz is pretty audible. Another 10K in series with 100+K is teeny.

And 10K adds 1.4uV across the audio band. That TL072 alone is 2uV audio hiss. Less than 2db added hiss, TL072 alone versus TL072+10K.

OTOH some series resistance buys a lot of protection against Bad Things. Speaker voltage (oops). Static zap. Nearby AM and CB radios.

[PRR]

Here is an extract from a datasheet showing pretty complete protection:



In this chip, all six diodes are built-in. (Some are inherent in usual chip fabrication.) We do know that '741 and '072 typically do not have the cross-base diodes (because if we push them we can force them 7V apart). These chips are not known to die easy.

As the text says, series input resistors are a key part of protection from external abuse.

[matt239]

Would it be worth using a more expensive hi-spec chip? OPxxx? One with a high input impedance AND lower voltage noise spec?
Or have we reached the point of diminishing returns?
I mean a well designed circuit with the TL072 & LM833 should be able to be vanishingly quiet right?

[matt239]

Can you calculate hiss? (It is a rare skill.)

Not yet.. I'm working on it. I've been reading a lot, but I get lost in some of it.

Let me ask:

- What is a good final noise figure??
I've been using some sims, & calculators, but I don't really know what to make of the results; how many dB is a micro-Volt of noise?
What is a good SNR?

- Is there any reason not to use even smaller value resistors in the feedback loop?
Say, 1k for "R1," 250R for "R2?"
Similar gain, much less resistor noise. Is there a downside?

- Would it be worth using a more expensive hi-spec chip? OPxxx? One with a high input impedance, (& low current noise..) AND lower voltage noise spec?
Or have we reached the point of diminishing returns?
I mean a well designed circuit with the TL072 & LM833 should be able to be vanishingly quiet right? - I need to learn how to figure this out myself.. 

[matt239]

P.S.
Is there any noise advantage to using a unity-gain buffer on the input?

I've been operating on the theory that once you've designed a buffer with the desired input impedance, may as well have it contribute some gain, & this will result in less overall noise. (assuming your circuit will have gain..)
Is this right? wrong?

- Can a buffer achieve a high input Z w/o using high value input/bias resistor(s)??

[sault]

- Can a buffer achieve a high input Z w/o using high value input/bias resistor(s)??

In between stages you don't need the pull-down resistors, but on the input you do... it has something to do with keeping the pedal from making a 'pop' noise when you do something like turn a pedal on/off.

Its probably worth noting that a buffer does not contribute gain... by definition it can't. If you're going to want gain, might as well just use an inverting amplifier circuit to do it - both Jfets and opamps will still have very high input impedances, so no worries there.

EDIT: Okay, I think I misunderstood what you meant by the feedback resistor remark... ie, you're talking about an inverting amplifier opamp, with gain determined by -R2/R1, and you're asking why we couldn't lower these resistors, right?

At the very least, the input impedance would be determined by Rin, or R1, or whatever the first resistor is. A lower input impedance means you could experience voltage loss between stages... and that would be more significant than adding some resistor noise!

[matt239]

Right, a buffer by definition has no gain. What I meant when I said "once you've designed a buffer with appropriate input Z, may as well give it gain.." was to make it an amplifier, no longer a buffer...
(& assuming the larger circuit will have an overall gain.)

-I understand the function of buffers in some circuits, but is it really an advantage on a guitar input? (& specifically does it help with noise? (I would guess NO, but I may be wrong?))
(I am also considering this case by itself, not getting into pros/cons of always-on buffers, & multi-pedal effects chains..)

- I've been talking about a non-inverting configuration the whole time.
I see now, the limiting factor on how low the feedback resistors can be, is that they will load the output of the op-amp.

I would probably want to keep the total load on the output of a TL072, above 2k ?
& the feedback resistor string is in parallel to ground with the input resistance of the next stage..

[R.G.]

Would it be worth using a more expensive hi-spec chip? OPxxx? One with a high input impedance AND lower voltage noise spec?
Or have we reached the point of diminishing returns?
I mean a well designed circuit with the TL072 & LM833 should be able to be vanishingly quiet right?

It all depends. On stage? No need. In a very quiet studio, with very good equipment and "studio quality" effects if any, maybe.

It is possible, but difficult to get better than the TL07x series and similar workhorse chips. That's why they're the workhorses.

Worth it? It all depends on what your definition of "worth it" is. The quiet per dollar is definitely going down.

Is there any noise advantage to using a unity-gain buffer on the input?

I've been operating on the theory that once you've designed a buffer with the desired input impedance, may as well have it contribute some gain, & this will result in less overall noise. (assuming your circuit will have gain..)
Is this right? wrong?

It's complicated. Best noise practice is always to use a low resistance source, get your gain as early as possible with a low noise input stage. Even this is complicated. There are some devices better for some source impedances. However, an electric guitar is far from a well behaved source, so the normal considerations of low noise amplification get bent. The guitar's big issue is that its pickups are inductive, and hence much higher impedance at treble than bass, which forces all the loading to be high impedance, including the controls on the guitar, which add their thermal noise to any the pickup might have. This boxes you into using high input impedances to keep from losing treble, and that boxes you into the external noise pickup issues of high impedance cables.

The best situation for a guitar would be to have a high impedance buffer with some gain right at the pickup, before the guitar controls. That ain't gonna happen. So we are stuck with compromises.

- Can a buffer achieve a high input Z w/o using high value input/bias resistor(s)??

Yes. The classic way is with bootstrapping. This is a controlled application of positive feedback and is, well, delicate.

Right, a buffer by definition has no gain. What I meant when I said "once you've designed a buffer with appropriate input Z, may as well give it gain.." was to make it an amplifier, no longer a buffer...
(& assuming the larger circuit will have an overall gain.)

Buffers with gain are OK.

-I understand the function of buffers in some circuits, but is it really an advantage on a guitar input? (& specifically does it help with noise? (I would guess NO, but I may be wrong?)) (I am also considering this case by itself, not getting into pros/cons of always-on buffers, & multi-pedal effects chains..)

The screaming advantage of buffers for guitar are to keep from the selective loss of treble from loading of the highly inductive pickup (2-4 henries is common) by either resistive loads or the cable capacitance. Pickups really need to be buffered right at the coil to keep treble loss down. It's not so much a hiss issue as treble loss and pickup of external noise.

You're making noises like you may have or be getting some technical training. If this topic interests you the way it seems to, you need to find and read a copy of Henry Ott's "Noise Reduction Techniques for Electronic Systems".

As a practical matter, there is not a whole lot to be gained in terms of lower hiss from diverging from a somewhat carelessly designed TL07x input buffer. There is a lot to be gained in not losing treble by using some kind of buffer early in the chain, the closer to the pickup coil the better.

[sault]

There is a lot to be gained in not losing treble by using some kind of buffer early in the chain, the closer to the pickup coil the better.

Onboard buffer preamp +1 !

[matt239]

Buffers with gain are OK.

The screaming advantage of buffers for guitar are to keep from the selective loss of treble from loading of the highly inductive pickup (2-4 henries is common) by either resistive loads or the cable capacitance. Pickups really need to be buffered right at the coil to keep treble loss down. It's not so much a hiss issue as treble loss and pickup of external noise.

So, you're saying it's ok to make the input stage amplify, & there's no added advantage to a unity gain buffer? ..right?

...may have or be getting some technical training

It's been 20 years since i took 2 basic electronics classes, it wasn't my major, I just took them for fun. It's strictly a hobby @ this point. I've been reading everything in the online D.I.Y. community, & reading data sheets, app. notes & other docs, from some manufacturers. I'll look into that book.

I do feel the need to try to understand some of the theory, rather than just copy circuits, or build by trial & error. I don't have a natural affinity for maths, but I'm willing to work through some to try to get a grasp  on the situation.

I'm not on a holy quest for ultra-low noise, (& I'm not planning on buying any $10k speakers, & oxygen-free 8ga. speaker wires..)
I'd just like to build stuff that's basically quiet. Some musical instrument gear has noise that is pretty obnoxious.
- I must admit though, it would be pretty cool if the only way to know your pedal is on was by the L.E.D. !

[PRR]

> no added advantage to a unity gain buffer?

We often equate "buffer" with "unity(-voltage)-gain", because there are some super-simple good-performing unity-gain ways to buffer.

But with modern chips, if you really need to tamper with a signal, you may as well ponder if added gain might be nice; the cost is trvial.

I'm convinced that hiss in guitar-chains is NOT so very important. If it was, we'd _never_ have a volume pot between a weak source and the first amplifier..... yet ALL guitars have them. Contrast with microphones, where all condensers and now some ribbons have a head-amp IN the microphone. And the gain control is well inside the amplifier, after the first stage.

> basically quiet

Keep the gitar pot FULL up. Don't use gear which requires dropping level into the hiss-swamp. Don't use gear which amplifies low hiss to high level (this is any distorter or compressor with dramatic action).