Resistor values for transistor biasing

[ReeceAblaze]

Hello everyone!

I was looking at how to bias a transistor online and it seems to always be calculated with two values to find the third. So if I have a fresh set of transistors and want to try to make something new from scratch, how do I find the resistors needed in each position. Which would need to be found first to find the others?

[GibsonGM]

Hi Reece, it's a good question, but a long one to answer!   And not easy to understand if you're new.   Paul R., PRR on here, wrote a lengthy and good, easy post about how to do exactly this, maybe you can find it!!  I can't find the link, but I did download the image to jog my memory when I need it.  When in doubt, COPY what others have already solved! 

Quickly:  you pick a current you want the transistor to operate at, could arbitrarily say 1mA if you like.  No need to get fancy.   You choose the collector resistor based on this by letting the collector ride at about half your supply voltage.    You give the base a 'stiff voltage', maybe 2V, and use that to reverse-calculate for your biasing resistors based on the voltage drop.  Note the 1V at the emitter, "close enough", and that lets you pick the E resistor using 1mA thru the device...

This pic was part of Paul's description; if you look at the collector calculation, then the base/divider calc's, you might see how this is done.  Base current is set at .01mA because typical BJT beta is about 100, and current in the divider is 10x that as a good rule...we're working from where we want to be backwards!   Does this help?  There are also UTube vids about this, but they can be more complex than we need.   If you keep it simple, this isn't hard to do at all, it can just be awkward to describe.

[amptramp]

I learned it the same way - a tenth of Vcc on the emitter, half of Vcc on the collector and ten times the base current in the base divider.  For maximum headroom, collector voltage should be equidistant from Vcc to the emitter voltage at minimum current, so a slightly lower collector resistor is often used.

[ReeceAblaze]

Hi Reece, it's a good question, but a long one to answer!   And not easy to understand if you're new.   Paul R., PRR on here, wrote a lengthy and good, easy post about how to do exactly this, maybe you can find it!!  I can't find the link, but I did download the image to jog my memory when I need it.  When in doubt, COPY what others have already solved! 

Quickly:  you pick a current you want the transistor to operate at, could arbitrarily say 1mA if you like.  No need to get fancy.   You choose the collector resistor based on this by letting the collector ride at about half your supply voltage.    You give the base a 'stiff voltage', maybe 2V, and use that to reverse-calculate for your biasing resistors based on the voltage drop.  Note the 1V at the emitter, "close enough", and that lets you pick the E resistor using 1mA thru the device...

This pic was part of Paul's description; if you look at the collector calculation, then the base/divider calc's, you might see how this is done.  Base current is set at .01mA because typical BJT beta is about 100, and current in the divider is 10x that as a good rule...we're working from where we want to be backwards!   Does this help?  There are also UTube vids about this, but they can be more complex than we need.   If you keep it simple, this isn't hard to do at all, it can just be awkward to describe.

This has been insanely helpful. I assume that to make a transistor work as it does in a fuzz these values need to be altered somewhat? I tihnk I remember reading somewhere that lowering the collector resistor can do this?

[R.G.]

This has been insanely helpful. I assume that to make a transistor work as it does in a fuzz these values need to be altered somewhat? I tihnk I remember reading somewhere that lowering the collector resistor can do this?

The water gets a little deeper in that direction. Making a transistor work as it does in a fuzz is not just one thing, it's a whole world of possibilities.
Fuzz, and distortion of all kinds, means that you're forcing the transistor to not amplify linearly for at least part of the signal it's amplifying. How much of the signal gets distorted, and exactly how it's non-linearly amplified is critical to what the distortion sounds like. Go have a read at Distortion 101 at geofex:
http://www.geofex.com/effxfaq/distn101.htm
for some conceptual pictures of what distortion does to a signal.
The fundamental of using a transistor, or any amplifier, to distort is to drive the device with enough signal so the signal forces the transistor to try to amplify outside the range that its biasing and power supply make possible. The transistor's inability to actually make an output signal that big or that shape is what makes fuzz.
In the simple four-resistor transistor circuit shown in the pictures, you can do that lots of ways. One way is to make either the collector or emitter resistors too large or small. This changes the no-signal operating point of the transistors towards either too high a voltage or too low a voltage to amplify properly, and it distorts.

You can do the same thing by modifying the base biasing resistors to shift the no-signal bias point towards the power supply voltage or towards ground. This lets smaller signals force the output signal to bump into a power supply limit earlier than it would otherwise, and you get distortion.

Yet another way is simply to make the gain of the transistor very large. For every transistor and power supply there is some size of incoming signal that is simply too big. In that case, the output signal the transistor tries to make is bigger than the power supply lets it make, so you get distortion.

It's funny - one part of my EE college courses was biasing transistors to make them NOT distort. It's a highly specific set of conditions to make a transistor biased so it can handle the biggest possible signal without distortion. Anything else distorts.

[GibsonGM]

Yet another way is simply to make the gain of the transistor very large. For every transistor and power supply there is some size of incoming signal that is simply too big. In that case, the output signal the transistor tries to make is bigger than the power supply lets it make, so you get distortion.

It's funny - one part of my EE college courses was biasing transistors to make them NOT distort. It's a highly specific set of conditions to make a transistor biased so it can handle the biggest possible signal without distortion. Anything else distorts.

I can't stress enough that you learn about the above, Reece...learn to read load lines!  Look it up on UTube.   You'll see how 'fuzz' or distortion can be created by pushing an input signal into a device that can't amplify it above (or below) its power supply...running out of headroom.     For me, learning load lines on Merlin B's site (same concept, but for tubes) made it somehow much easier to understand...

You might have a look, it could help. Transistors ARE a little different, but the CONCEPT is quite similar.   You can't use the tube principles to bias a transistor, but you CAN see how any active device will distort if you try to make it amplify more than what its power supply can deliver.   http://www.valvewizard.co.uk/Common_Gain_Stage.pdf

Shooting for a bias point around 1/2 the supply is how one gets the most headroom (R.G.'s biasing NOT to distort), and is what is shown in the graphic above.  If you cascade stages, you ARE going to start distorting anyway, because you'll violate the 'rule' and ask a stage to output more than the power supply...they avoid that by reducing the signal between stages, actually!  ("Gain" control...)    It's all really interesting stuff.

[ReeceAblaze]

Yet another way is simply to make the gain of the transistor very large. For every transistor and power supply there is some size of incoming signal that is simply too big. In that case, the output signal the transistor tries to make is bigger than the power supply lets it make, so you get distortion.

It's funny - one part of my EE college courses was biasing transistors to make them NOT distort. It's a highly specific set of conditions to make a transistor biased so it can handle the biggest possible signal without distortion. Anything else distorts.

I can't stress enough that you learn about the above, Reece...learn to read load lines!  Look it up on UTube.   You'll see how 'fuzz' or distortion can be created by pushing an input signal into a device that can't amplify it above (or below) its power supply...running out of headroom.     For me, learning load lines on Merlin B's site (same concept, but for tubes) made it somehow much easier to understand...

You might have a look, it could help. Transistors ARE a little different, but the CONCEPT is quite similar.   You can't use the tube principles to bias a transistor, but you CAN see how any active device will distort if you try to make it amplify more than what its power supply can deliver.   http://www.valvewizard.co.uk/Common_Gain_Stage.pdf

Shooting for a bias point around 1/2 the supply is how one gets the most headroom (R.G.'s biasing NOT to distort), and is what is shown in the graphic above.  If you cascade stages, you ARE going to start distorting anyway, because you'll violate the 'rule' and ask a stage to output more than the power supply...they avoid that by reducing the signal between stages, actually!  ("Gain" control...)    It's all really interesting stuff.

The idea of the transistor not being able to handle the single is really interesting. It makes the sound make a lot of sense. Over the next few days I am going to setup a none distorting transistor amplifier and then change some values up and down until it either distorts or there is no signal (I am betting this will be the case most of the time) and hopefully doing this will help me understand it a lot better!

[GibsonGM]

Here, Reece - sounds like this guy already did it!  https://www.youtube.com/watch?v=c6cmkm3UPUI

You can see in this vid what 'clipping' is...that's what we call 'distortion'.  Top (or bottom, or both) of the signal gets cut off.  That happens either because you biased the transistor 'too close' to the upper or lower limit of the power supply, or the signal itself is too large, the BJT can't amplify it above its power supply, and the portion(s) that exceed it are clipped off the output.   Bzzzz.....fuzz.    How hard you make this happen = how raw the distortion is. 

The collector resistor is just setting the quiescent current thru the device...the biasing has much more to do with where on the curve you end up.

[ReeceAblaze]

Here, Reece - sounds like this guy already did it!  https://www.youtube.com/watch?v=c6cmkm3UPUI

You can see in this vid what 'clipping' is...that's what we call 'distortion'.  Top (or bottom, or both) of the signal gets cut off.  That happens either because you biased the transistor 'too close' to the upper or lower limit of the power supply, or the signal itself is too large, the BJT can't amplify it above its power supply, and the portion(s) that exceed it are clipped off the output.   Bzzzz.....fuzz.    How hard you make this happen = how raw the distortion is. 

The collector resistor is just setting the quiescent current thru the device...the biasing has much more to do with where on the curve you end up.

How is a signal made to be too large in the first place? Also raw distortion sounds good. I am after some pretty nasty fuzz noises, I am a big fan of Death by audio and their signal destroying boxes that they make so well!

[GibsonGM]

You make the signal too large by 'cascading' gain stages.  So, you might have a 1V P-P guitar signal at your input.  You run it thru one transistor stage (like an LPB-1 or something...) and it's cranked up enough to just start to clip a bit, so it sounds overdrive-like if you feed that to a clean amp.

Then you feed THAT signal into ANOTHER gain stage (maybe ANOTHER LPB-1 stage)...now you're throwing something like 7V input into a stage that will try to multiply it by a few 10s of times.   It can't output 50 or 75 volts....it can only output close to the supply, 9V.    So the signal begins to get cut off at top and bottom.    This is clipping.  In this case, it WOULD be pretty raw.    The more 'cut off' the tops are, the more square-wave the distortion is shaped, which "sounds" buzzy/fuzz.     Too much clipping in 1 stage can sound like @ss, though...so you see several gain stages in amps and things....to do it more gradually.  This encourages the distortion to be more touch-sensitive and have more 'character'.  It can still be metal-hard, just not buzzing and farting every time you touch the strings.    This is what most ppl are playing with - that balance between drive and touch-sensitivity.

Check out the Electra distortion...it uses diodes after a transistor gain stage to 'artificially' reproduce that limitation which causes clipping (for lack of a better way to describe diode clipping, ha ha).      That design may be useful to you...do a gain stage, into an electra, some Germanium diodes, and another gain stage to bring the volume back up, something like that!!    Don't forget - you have to do some tone shaping too, in order to make the fuzz/dist. that you generate sound good!   

[PRR]

> learn to read load lines!

Bah. Don't need lines.

If you put a 2-inch or 10-inch box in a 12-inch shelf, it fits without crushing, damage.

If you put that super-size 24-inch box of SugarPuffs in the 12-inch shelf, it has to crush.

So the key trick is to collect enough Gain so your box is bigger than the shelf (signal made bigger than supply).

Some side-tricks. Most HIGH-gain transistor amp stages have LOW input impedance. This loads-down the guitar pickup, more on highs than lows. While several classic pedals do that, it is part of the effect and an artistic choice.

When you put chords through distortion there is nasty intermodulation. Many distorters cut bass so mostly the highest note of the chord distorts alone.

You can distort both top and bottom of a wave; ultimately this always happens. However you can have an intermediate signal level which is only dented on the top (or bottom), and that is a different sound.

#1 rule: Plagiarize plagiarize plagiarize !!!! 99.9% of useful distortion mechanisms were known in the first days of vacuum tubes over 100 years ago. 60 years of "fuzz" has sorted-out the main tools for fuzz. There is no need to go back to the stone hammer and try square and oval wheels.

[antonis]

How is a signal made to be too large in the first place? Also raw distortion sounds good.

You can't get signal larger than 20 times power supply value from just a single stage BJT amp..!!! 
(which value is ideal, negleting next stage collector loading & various losses..)

[GibsonGM]

> learn to read load lines!

Bah. Don't need lines.

No, not to work with them (or the other active devices).  But maybe just to understand what's going on      I thought Merlin had a pretty fast & simple description, might help him 'get it'.     Maybe faster than I did!  ha ha


Absolutely no need to re-invent the wheel, Reece! Paul is right - there are 100s of BJT circuits you can "steal"...and they all were already copied from others!

Build a couple, mess with them and see how they work, then you can move on to understanding 'impedance' and 'loading' and 'asymmetrical clipping'...and what you do to the tone after (or before) you clip the signal has a huge effect on the sound (tone controls, small filter 'pieces', and impedance matching or mis-matching (loading)). 

[R.G.]

#1 rule: Plagiarize plagiarize plagiarize !!!!

Are you a Tom Lehrer listener too? From "Lobachevski", in a Russian accent:
Plagiarize, plagiarize, plagiarize
Let no one else's work evade your eyes
And plagiarize, plagiarize, plagiarize!
Only please always to call it research.

[PRR]

> Are you a Tom Lehrer listener too?

I have his first, 10-inch, disk. (Probably from not the initial run of 400, but certainly not the 1960 Revisited.)

[tubegeek]

Paul and R.G., "You should have a statue in bronze."

[GibsonGM]

Paul and R.G., "You should have a statue in bronze."