bjtdiff-amp
I am working on the second stage of an operational amplifier which involve a BJT differential amplifier. I am trying to increase it's bandwidth (parts of the requirements) and was told by a TA that a simple cascode should increase the BW. I am running simulations and It seems to me me that the cascoded version actually decreases the BW. I have tried to change the biasing voltage with no success and cannot find any source about cascoding increasing the bandwidth.
You can try buffering the output of the filter with a pair of darlington emitter-followers, which will replace the four diodes with four transistors.
simulate this circuit – Schematic created using CircuitLab
You can try buffering the output of the filter with a pair of darlington emitter-followers, which will replace the four diodes with four transistors.
Check that the no-load bias current in the output transistors is not too high.
It might be easier to use an op-amp though, if you can get some that are suitable.
Your first two transistors look incorrectly biassed so that Ve is about (0.5*Vcc - 0.7V) in Q0, Q1 maybe slightly better.
This means (thanks to symmetry between Re and Rc (e.g.R0 and R3) that Vc is about Vcc/2+0.7V, in other words Vce=1.4V approx, and the transistor is close to saturation.
First step : run a DC analysis to confirm this for both common-emitter stages (the emitter follower is fine).
To maximise the linear dynamic range you want to maximise the possible collector swing, so as a first pass, re-bias for Vc=9V, Ve=3V, i.e. Vb = 3.7V, calculate the base resistors for that - and see if it's improved.
This gives you up to 6V p-p swing at the collectors. It's not the best you can do as you have less AC swing at the emitter therefore you can increase headroom further, but I'll stop here.
Best Answer
Examine this circuit; each of the diffpairs has gain of 8X at low frequencies; but their F3dB (45 degree point) are not consistent; why?
And Q7/8 do not have gain of 8X. Why?
Also notice the leftmost diffpair has no source resistor, thus CMiller can be infinite but the SPICE source handles the input current (Ibase, Cob) just fine; on the other hand, the 3 right circuits include Rin, and that R along with Cmiller impacts the circuit bandwidth.