General Transistor Bias Question

[Phend]

Hello:
I built a Tube Screamer using the GGG schematic.
The circuit shown is using 2N4401 tranistors.
I am substituting them with 2N5088.

GGG suggests voltages as: (with 2N4401)
C 9.0 v
B 3.0 v
E 2.5 v
That is with R3 as 10k to gnd off the emitter.

With 2N5088 I get the following (R3 = 10k) (with 9.4v supply)
C 9.4 v
B 4.3 v
E 3.9 v

Changing R3 to 1k I obtain the following.
C 9.4 v
B 3.1 v
E 2.6 v

Changing R3 to 1k brings the voltages very close to the GGG suggestion.

Question: Have I done the "proper" bias ? (Both of the 2 transistors are the same)

Although the higher voltages sound darn good, making the R3 change to 1k seems better.



Thank you

[Dormammu]

In such schemes the biasing of transistors base is better to have on half of the voltage supply.  If you like widest dinamic range for signal.

[antonis]

the biasing of transistors base is better to have on half of the voltage supply.

Actually, Emitter should stand on half-supply voltage ..
(for almost infinite load..)

More presicely, Emitter(s) should stand at about 2/3 of supply voltage (for particular loading R5 & R13) for symmetrical +/- signal swing..
Suppose Q1/Q2 Emitters are biased at 2,5V (GGG configuration)..
For a load equal to their Emitter resistors value, they could go ali the way up to +9V but only down to 1.25V, due to 10k/10k voltage dividing effect..
Of course, for input signal amplitudes lower than 1.25V/A, there should be no problem..
(where A  = voltage gain - 0.99 here..)

[Clint Eastwood]

With the 1k resistors you have a lot more current running through the transistors: 2.6/1000 = 2.6 mA, times two. This may be a concern when using a battery power supply.

[Dormammu]

the biasing of transistors base is better to have on half of the voltage supply.

Actually, Emitter should stand on half-supply voltage ..
(for almost infinite load..)

More presicely, Emitter(s) should stand at about 2/3 of supply voltage (for particular loading R5 & R13) for symmetrical +/- signal swing..
Suppose Q1/Q2 Emitters are biased at 2,5V (GGG configuration)..
For a load equal to their Emitter resistors value, they could go ali the way up to +9V but only down to 1.25V, due to 10k/10k voltage dividing effect..
Of course, for input signal amplitudes lower than 1.25V/A, there should be no problem..
(where A  = voltage gain - 0.99 here..)

Actually, IC1B can give an amplification beyond supply voltage in an asymmetrical form.
And the repeater load is determined by Re multiplied by hfe.   IIRC   

[antonis]

Actually, IC1B can give an amplification beyond supply voltage in an asymmetrical form.
And the repeater load is determined by Re multiplied by hfe.   IIRC   

Unfortunately, I can't get what you're saying..

I was talking about Q1 & Q2 "correct" biasing which, IMHO, should be independent of any further intentional asymmetrical clipping..

[Dormammu]

Unfortunately, I can't get what you're saying..

For example, me too also do not see a reason for the existence of Q1 and Q2 in this scheme. But if someone put them there - let them be.   

[Phend]

So, I shall change R3 back to 10k.
Question,  what's going on with this circuit?
Shouldn't there be resistors going to the collectors for biasing ?
Like in the BMP, Foxx, etc .
Could I add them ?
I see in RG Keen's write up he says R2 510k is a biasing resistor, should that be tweaked ?
Or should I just not be concerned about the Q voltages ?
Thanks again.

[Dormammu]

So, I shall change R3 back to 10k.
Question,  what's going on with this circuit?

I see in RG Keen's write up he says R2 510k is a biasing resistor, should that be tweaked ?
Or should I just not be concerned about the Q voltages ?
Thanks again.

RG Keen - is right.
Tweaking with such circuit - almost useless. It`s just a buffer.

[FSFX]

Shouldn't there be resistors going to the collectors for biasing ?
Like in the BMP, Foxx, etc .
Could I add them ?

In the circuit you are referring to the transistors are just used as an emitter follower type of buffer and so all you need to do is bias the base to a voltage that will result in an emitter voltage of about half the supply voltage. Just be aware that low hFE transitors will require a higher base-emitter current and so that may load down any bias voltage derived through a high source resistance. 

[antonis]

You could add Collector resistor but it shouldn't result into anything usefull..
(on the contrary, it should lower availiable headroom)

Q1 & Q2 are configured as IN/OUT buffers (search for CC amp or voltage follower) and serve for impedance matching..
(high input impedance - low output impedance)
https://www.quora.com/Can-you-explain-biasing-in-a-common-collector-transistor

510k resistor indeed serves for bias purposes, in a way where its value could be ingnored for Q1/Q2 high DC gain but taken into account for Q1/Q2 input impedances..
On upper leg of R2 the voltage level is half the supply voltage where on lower leg (at Q1 Base) the voltage level is half the supply voltage minus 510k times Base current..
(which Base current equals Collector current divided with Q1 DC gain)
That's the reason for getting higher Emitter voltage for 2N5088 compared to 2N4401 - the higher the gain the lower the Base current (for a given Collector current) hence the lower the voltage drop across R2, resulting into higher Emitter voltage..

[Phend]

Student driver here, will keep my eyes on the road and hands upon the wheel.
Shall leave it be, it works great thru my tube amp.
Thanks all

[PRR]

You hang a swing from a tree. It sits too high off the ground. Do you--
1) Lengthen the ropes? (Larger Rb)
2) Add load to stretch the ropes? (Smaller Re)
Either answer can work.

[Phend]

You hang a swing from a tree. It sits too high off the ground. Do you--
1) Lengthen the ropes? (Larger Rb)
2) Add load to stretch the ropes? (Smaller Re)
Either answer can work.

Great analogy PRR.
So keep 10k at E as Clint says to save battery.
Increase 510k at B to approach what the voltage should be.
Since these Q's are buffers, what should the V's b ?

[Dmel]

I have noticed your R21 POT right-hand side at the output stage.   It should connect to the ground, NOT to the bias.  The schematic is incorrect and cannot work correctly this way.  These 2n4401 are buffer NPN transistors or "emitter followers"..... as such they need to be biased just as the schematic shows with the two resistors and capacitor developing 4.5 v which is 1/2 the 9v input.   Nothing else is needed.    When the voltage supply input drops to say 8.8v  the bias is still optimized at 1/2 of 8.8 IE 4.4v.   

[Dormammu]

I have noticed your R21 POT right-hand side at the output stage.   It should connect to the ground, NOT to the bias.  The schematic is incorrect and cannot work correctly this way. 

On the one hand — you're right, if you want to build the engineeri-ish right amplifier.
But, since it's a distortion, maybe it's some kind of solution that the schematic designer found. It will work because according to the AC component is closed to the ground through C7.
Perhaps this solution adds some more distortion through feedback loops.

[antonis]

I have noticed your R21 POT right-hand side at the output stage.   It should connect to the ground, NOT to the bias.  The schematic is incorrect and cannot work correctly this way. 

I can assure you that it can work perfectly this way..
(Bias point is considered AC ground due to C7, as Dormammu already pointed out, so from signal point of view there should be no problem..)

I'm on your side only from  garbage-free bias point viewpoint..

P.S. Welcome..

[antonis]

So keep 10k at E as Clint says to save battery.
Increase 510k at B to approach what the voltage should be.
Since these Q's are buffers, what should the V's b ?

IMHO, answer shouldn't be so easy..

10k at Emitter is good for low current consumption and high input impedance (dominated by 510k resistor) but bad for load driving (R5)..
510k inrease should only result into Emitter voltage decrease, which voltage already is low..
510 decrease should result into Emitter voltage increase but also into input impedance decrease..

For particular bias configuration, the only "improvement" could be an even higher beta (DC gain) BJT in conjunction with a lower Emitter resistor value..
(the former for less voltage drop across 510k resistor, the later for better load driving capability and both for maintaining high input impedance - at least, the part of impedance they contribute to..

edit: If you want to tweak it a bit, add a capacitor and alter R2 & R5 values..

[Dormammu]

but bad for load driving (R5)..
[/quote]
What do you mean ?

[antonis]

but bad for load driving (R5)..

What do you mean ?[/quote]

There are two cases for Emitter follower load driving..

1. Emitter resistor itself  is the load..
(DC and AC load lines coincide..)

2. Emitter reistor is effectivelly set in parallel with external load..
(different DC and AC load lines)

We here deal with the 2^nd case..

The lower the Emitter resistor value, in relation with load (R5 here), the better the signal negative waveform undistorted replication..

P.S.
Simple Emitter followers are inherently non-linear configurations due to various factors..
One of those is the "asymmetrical" current sourcing & sinking capability..
While Q1 can, whithing limits, source as much current as required into the load, its sinking ability is limited by the Emitter resistor which forms a voltage divider with the load resistance..
For present case, negative clipping should occur at about half the Emitter biased voltage, due to Emitter resistor and load (R5) equal values..