Question about differntial transistor amplifers

[wanna_eat_1kg_of_elecons]

Hello,

I hope you can help me out. I made this circuit with a pair of BC337 transistors:

**broken link removed**

I am getting strange results so first i will describe what i did and then i'll ask my questions

I made a current source and measured the current @ 2mA that was what i calculated so that part went fine.
Then i calculated the bias point. Vcc is 6.2V and there is 2 mA for both RC resistors so one resistor gets 1mA. The resistor i choose was 2k7 so the bias point is Vcc - (2k7 * 1mA) = 3V5


Then i calculated the gain. As G_diff = Rc/2*(re + RE) --> 2700/2*(25 + 100) = 10.8

(Question 1)
So i expect that if one base is 100mV higher than the other base the output will be Gain * 100mV + bias_voltage. So if i make one base 100 mV higher than the other one i suspect 3.5V + 1.08V = 4.58V
Is this assumption correct?

so i did put one base at 3.5v and added Delta V to the other base with a PSU the values i measured are:

Delta (mV) ------ Output
0 ------- 3.29v
100 ------- 4.1v
200 ------ 4.77v
300 ------ 5.54v


if i calculate it like this for delta = 300mV:

5.54 - V_bias = 5.54 -3.5 = 2.04
Gain is Output/ input ---> 2.04/300mV = 6.8

(Question 2)
Did i calculate the gain wrong? I Calculated 10.8 and measured a gain of 6.8


I want to use this amplifier to multiply a small delta voltage. But the common between 4 and 2.5 volt so. If i put 2.5v on one base and 2.6v on the other. Or 4v on one base and 4.1 volt on the other. I want an output of bias + gain * delta_v = 3.5 + 1.08 = 4.58

Question 3

Can i do it with this circuit?


Question 4

When i measure the other Collector it is always about 3V except for the case when delta = 0 then it is 3.6V. It looks like it is clipping what am i doing wrong i think i am missing the point here and i am running in circles.

Can someone please help me out?

I know op amps are easier but i want to learn transistors so please no opamp remarks.. (maybe i'll try them in a few wile but for now it is transistors)

i am tired now so i hope my text makes sense thanks for the effort !!

 

[MrAl]

Hi,

If no one else gets to help you out i'll try to get back here tomorrow to see if i can help. Too tired tonight to do any more analysis
Tomorrow is another day.

 

[wanna_eat_1kg_of_elecons]

That would be really great MrAl, I have been working for hours on it and it still goes wrong

It doesnt have to be extremely accurate when i can predict it is 5% it is fine .

eventually i want to use this circuit to measure a shunt and to see if the current gets above a current X.

As the shut is high end (about 12V) and the rest of the circuit is low end (Vcc = 6v ) i want to use two simple resistor voltage dividers to get the voltage of the shunt down well within the 6v range.

When the voltage dividers have divided the signal by 3 i want to use this amplifier to multiply it by 30 or so to and i hope to lose the large common mode signal and be able to use the differential signal.

I would be really happy if you helped me out i dont understand what is going wrong here

thanks

 

[MrAl]

Hello again,

First of all, 1kg of electrons? That's a lot of electrons. Maybe consider some protons in with your diet next time

Ok, so did you build the circuit that they showed on that page? It looks like they used 2N2222 transistors so that's what i used too. I used all the resistors they showed, except i changed the lower resistor (RTAIL) to 7.4k instead of 7.5k. That's a change that made the amp more linear for low level inputs.
The results i got were promising even without any other modifications. The amp seems linear to some degree and the accuracy with different inputs seems not too bad. Take a look at the numbers yourself in the following table and see what you think.
Note however that some inputs will not be valid because they saturate the system, in particular 0.100v input with a 4v common mode input, but it is quite usable up to 3v common mode input and 0.1v differential input the way it is i think.

Take a look at these numbers. The common mode input voltage level is given, and the table shows values for the input differential voltage and the differential output voltage. It's not super good, but not super bad either.

IN     DiffOut(minus)

Common mode voltage=0.000v
0.000  0.0000
0.001  0.0558
0.002  0.112
0.005  0.279
0.010  0.558
0.020  1.12
0.050  2.78
0.100  5.52

Common mode voltage=1.000v
0.000  0.0000
0.001  0.0566
0.002  0.113
0.005  0.283
0.010  0.566
0.020  1.13
0.050  2.82
0.100  5.61

Common mode voltage=2.000v
0.000  0.0000
0.001  0.0574
0.002  0.115
0.005  0.287
0.010  0.574
0.020  1.15
0.050  2.86
0.100  5.69

Common mode voltage=3.000v
0.000  0.0000
0.001  0.0581
0.002  0.116
0.005  0.290
0.010  0.581
0.020  1.16
0.050  2.90
0.100  5.76

Common mode voltage=4.000v
0.000  0.0000
0.001  0.0587
0.002  0.117
0.005  0.293
0.010  0.587
0.020  1.17
0.050  2.93
0.100  4.97 [extreme inaccuracy only with this combination]

 

[wanna_eat_1kg_of_elecons]

Thank you for the effort! It looks nice indeed! Did you simulate this or did you use real transistors? When i have got some time i'll try again

 

[wanna_eat_1kg_of_elecons]

I think the problem i had was that the transistors although both bc337's differ a bit so one leg amplifies different from the other leg.

 

[MrAl]

Hello again,


I simulated it so the two transistors were a perfect match. For critical applications you would normally used a matched pair in a single package.