What's the function of the "BIAS-R"?

[the_floyd]

Specifically, the one labelled R1 here:



What purpose does it serve? I understand that it "biases the op-amp" - could someone please clarify what that actually involves? Is the voltage being dropped across the resistor? The current?

I'm trying to make the jump from simply switching values based on what I read online, to learning what parts of the circuit actually do. As such, help would be much appreciated! 

[twabelljr]

The two 100k resistors in series to ground with the 1uf cap tied between them and grounded is the part of the circuit that actually creates the bias voltage for the opamp. Since they are equal value and are connected in series from the 9v source and ground they each drop 4.5v leaving the bias of 4.5v for the opamp where the two 100k's connect to each other. As for R1, it connects from the input to the bias and I believe sets the input impedance of the circuit. Hopefully somebody can explain R1 better than me. there are some VERY intelligent people here. You came to the right place. Welcome!

[the_floyd]

Hmm... so it sets the input impedance, then? Why wouldn't it just go to ground, as I've seen in most designs? I'm still curious as to how it would affect the voltage / current going to the opamp... Thanks for your help, though!

[ashcat_lt]

It would limit the bias current available at the input of the opamp.  Luckily it don't take much. 

Mostly what it does is moves the audio reference voltage (the voltage at the opamp input when the source is silent) to a point halfway between the power supply rails.  Without it, the signal swings around 0V (the bottom of the battery).  The opamp can't provide an output less than 0V, so without this bias voltage the "bottom" of the signal swing would be clipped off in half wave rectifier fashion.  Might sound cool, but usually not what we want.

This resistor is parallel to the opamp input, and since the opamp's impedance is extremely high, the bias resistor is pretty darn close to the impedance your guitar sees.

This circuit only has one active stage, but in places where there are more which need biased, the bias resistors help to isolate the input of one stage from another.

[Gurner]

Hmm... so it sets the input impedance, then? Why wouldn't it just go to ground, as I've seen in most designs? I'm still curious as to how it would affect the voltage / current going to the opamp... Thanks for your help, though!

Becuase the circuit has a 9V supply, you want the opamp to operate in the middle of this (normally)....setting such a mid point (4.5V) is called biasing. (edit: sorry just re-read your post & you know what  that is!)

Think of ground & the 9V supply as two extremeties your signal *cannot* go past ...so you want your signal to be centred/based around the mid point. If a drunk guy walks down an alleyway, it's best that he tries to stay away from crashing into  the walls - so best to have him start off walking down the middle - the R1 in the schematic biases the opamp at mid point...therefore the signal has about 4.5V leeway (headroom) before it crashes into a wall.

If you connected R1 to ground (to set the impedance), the bias would be 0v (give or take) ....so, your signal is crashing into one of the walls as soon as it comes into the circuit - so in that particular circuit that's a no-no. The input impedance at audio frequencies (assuming the opamp has a very high input impedance), is R1 and one of those 100k resistors combined in parallel (about 90k)

There will be some small amount of unavoidable current flowing through R1 (which is a big value, so even small amount of current will amount to a voltage drop) so in actual fact you'll get slightly more than 4.5V as your bias.

[ashcat_lt]

The bias resistor s in series to ground with one of the 100k's and the other has nothing to do with the input circuit.  In-Z is 1.1M parallel to the opamp.  90k would be completely unacceptable for passive pickups.

[Gurner]

The bias resistor s in series to ground with one of the 100k's and the other has nothing to do with the input circuit.  In-Z is 1.1M parallel to the opamp.  90k would be completely unacceptable for passive pickups.

Quite correct - it's in series, it's late & I've had one tinny too many - goodnighsh! (off to walk down an alley)

[the_floyd]

Thanks for all the explanations - they've actually been quite helpful.

I now understand the purpose of biasing, but there's one thing I still don't quite get - is it R1, or the voltage divider network between R1 and the power supply, that sets the bias? Or the combined effect of the two? My understanding of it is that the resistor connected to ground in the voltage divider network 'dumps' the voltage difference between Vin and Vout to ground - is this the case? I'd think there's more to it, if R1 plays a role in setting the bias voltage...

Or is current part of the biasing process, as well? Because R1 limits current, I do know that.

And while I'm at it (sorry for all the questions), I'd like to get to the practical application for my question. I'm working on a Rat clone, which I've recently started running at 18 volts. Nothing exploded or smelt bad (all the caps are 25V rated, at least), and there was a noticeable loss of low frequency 'sag' at higher gain settings. So what led me into this whole train of thought was wondering if I could achieve the same effect by changing the bias resistor to the op-amp - is this so? It is my understanding that sag is due to insufficient current available to amplify the more current-intensive lower frequencies - but raising the voltage solved the sag issue, not raising the current... Or could that be due to the increased voltage preventing the op-amp from distorting and inducing sag? The more I know, the less I know, it seems...

[Hides-His-Eyes]

The dividers set the bias, and the big resistor delivers it to the signal whilst limiting the current that can be drawn from Vref.

Changes in tone at higher voltage settings are probably partly due to increased headroom (RAT circuit includes distortion from the op-amp trying to amplify beyond the supply voltage). No amount of biasing can change that. Look into charge pump solutions to get 18v from a 9v supply.

[ashcat_lt]

+1  I think you're hearing less opamp clipping.  Only way to change that is to either reduce the rail voltage or increase the amount of gain. 

More resistance in series with the 100k gain pot might help.  Of course then you'll be hitting the diodes at the end harder...

The input to the opamp is extremely high impedance.  It wants almost no current no matter what the source is because I = V/R and as R approaches infinity I approaches 0.

Changing the bias resistor will change the in-Z, which will change the treble response, especially when connected directly to passive pickups.

[the_floyd]

Okay, this has cleared up a lot for me. I'll look into a charge pump circuit - I could probably wire this: http://www.muzique.com/schem/vdub.htm in between the power jack and the PCB, and mount it on the side of the enclosure; shouldn't be too big a deal. Thanks for all the help!

[stringsthings]

Specifically, the one labelled R1 here:



What purpose does it serve? I understand that it "biases the op-amp" - could someone please clarify what that actually involves? Is the voltage being dropped across the resistor? The current?

I'm trying to make the jump from simply switching values based on what I read online, to learning what parts of the circuit actually do. As such, help would be much appreciated! 

here's a link (to the link) to the datasheet of the IC ... http://www.datasheetcatalog.com/datasheets_pdf/L/M/3/0/LM308.shtml

if you take a look at page 2, over in the right-hand column are the numbers for the LM308 ( the IC that our rat likes ) ... we see that our chip can run on +_ 18 volts .... and Proco decided to run it on +9 Volts ... a common practice ...

if you skip to the 4th line from the bottom of page 2, under Electrical Characteristics, you'll see the numbers for "Input Bias Current" ... R1 limits the amount of bias current going into the IC ... for the LM308, the maximum amount of bias current is 10 nA ... if we replaced R1 with a wire jumper, we'd probably exceed this amount .... and the IC might exhibit "unusual" performance 

[PRR]

We are NOT "limiting bias current".

When the opamp is happy, it pulls whatever bias current it was designed to pull. (If it's unhappy, we don't care.)

> numbers for "Input Bias Current" ... for the LM308, the maximum amount of bias current is 10 nA ... if we replaced R1 with a wire jumper....

You are halfway there.

* If you replace R1 with a jumper, the opamp will be happy, but what happens to the SIGNAL coming in? It is dead-shorted to Vref's large capacitor!! "NO" signal will be present at the opamp!! (In this case, Vref's cap is not huge, you will get some bass sneaking in; but not what we want.)

We need enough resistance so the signal is not shorted-out, not even significantly loaded.

Guitar, to pass full treble, needs a load well over 100K, and preferably over 470K, often 1Meg.

At the other extreme: too large a resistor makes trouble because bias-current times R1 makes a voltage drop, and the input is no longer at or near Vref.

Let us pick a VERY large resistor: 1,000 Megs. 1,000 Megs times 10nA is 10V drop. If one end of R1 is at +4.5V, the other end is at negative 5.5V. Actually it will fall to about +1V and then stop working. No harm done, but not working happy.

In this case we must be somewhere between the zero V and 9V supply rails, 4.5V is the obvious point, but 3.5V would work just about as good. We can accept up-to 1V of drop in R1. So the maximum value of R1 is 1V/10nA= 100Megs.

In other cases we may need to hold the opamp output much closer to Vref than 1V error. But we rarely need to be closer than 0.010V (10mV). Now the maximum value is 10mV/10nA= 1Meg.

So we need much more than 100K, much less than 100Meg, preferably over 470K, and maybe not over 1Meg. And 1Meg resistors are common items, while 5Meg and 10Meg are traditionally rare.

So the first-dart value is 1Meg.

If you change chip, you should re-compute. 741/4558 pull much more bias current, will usually want something smaller than 1Meg. TL072 pulls much less bias current (at stage temperatures), can use a larger bias resistor, but there's no compelling reason to go over 1Meg or maybe 5Meg, it won't sound better.

I see that Rat-mutants try 220K here. This loads guitar treble slightly. That's not uncommon for fuzz circuits. Excess input treble makes tizzy fuzz. You could design a low-pass filter. However if coming from a naked pickup, a heavy load cuts treble, and is cheap/easy. (However it has no treb-cut effect when fed from buffer or other active box.)

[stringsthings]

We are NOT "limiting bias current".

When the opamp is happy, it pulls whatever bias current it was designed to pull. (If it's unhappy, we don't care.)

> numbers for "Input Bias Current" ... for the LM308, the maximum amount of bias current is 10 nA ... if we replaced R1 with a wire jumper....

You are halfway there.

* If you replace R1 with a jumper, the opamp will be happy, but what happens to the SIGNAL coming in? It is dead-shorted to Vref's large capacitor!! "NO" signal will be present at the opamp!! (In this case, Vref's cap is not huge, you will get some bass sneaking in; but not what we want.)

very cool ....  (so the circuit would exhibit the undesired response .... )