Current Sources in audio amplifier

alvise

Member
2013-01-22 7:56 pm
Dear all,
I'm a newbie in the audio electronics, but I'm studying it.
I bought the Bob Cordell's excellent book (ed. 2011). With reference to figure 2.12 (pg. 42) I need to ask something to the experts. In particular I still have some mis-understanding about the current source used as active load of the CE amp (Q6/Q7, Q4).

I know that in order to bias a transistor, more options are available; one of those is a configuration with an emitter resistance (like that one in the figure, R6, 22Ohm) where we impose a voltage to the base and then across the emitter resistance. This will produce an emitter (and a collector) current decided by the imposed voltage (minus the Vbe drop) and the value of the emitter resistance. That said, in the figure, there's another actor imposing the bias: the current source Q6/Q7 that 'inject' 10mA in the Q4's collector.

Now the question: when designing an amplifier like that, with a common emitter and a current source, I presume that much care must be used when choosing components' values... in fact what happen if there's even a very small mismatch between the current imposed by the Q4's base voltage and emitter resistance and the current imposed by the current source Q6/Q7? Any resistance tolerance cannot guarantee any current match...

Could someone explain me some details about common emitter biased with current sources ?

thank you :)
 
Hi

Q6 & Q7 form what is known as a 2 transistor ring type constant current source. It uses one transistor (Q6) to form a feedback loop to the current source transistor (Q7). When R9 has 0.6V across it, Q6 turns on, keeping Q7 at the bias level so as 0.6V always remains across R9. Iout = 0.6V/VR9. R10 provides a conduction path for Q7 base current and Q6 collector current. It should be noted that it is possible for this feedback loop to oscillate. Placing a small 100pf to1nf cap across C-E of Q6 will dampen it

The constant current load allows for a constant Ic in the VAS transistor. This makes it more linear as it transistions across the load line in the graph of Ic vs Vce. It now only has to change Ic by what the driver transistors require in dynamic base current, and also the miller loop at high frequencies.

:2c:

I know that in order to bias a transistor, more options are available; one of those is a configuration with an emitter resistance (like that one in the figure, R6, 22Ohm) where we impose a voltage to the base and then across the emitter resistance. This will produce an emitter (and a collector) current decided by the imposed voltage (minus the Vbe drop) and the value of the emitter resistance. That said, in the figure, there's another actor imposing the bias: the current source Q6/Q7 that 'inject' 10mA in the Q4's collector.

Check out section 3.2 regarding transistor degeneration. This will explain R6.
 
Last edited:
Q4 is not biassed by the emitter resistor and the base voltage. There's 0.22V at the top of R6 (because of the 1mA coming down through the transistor). As long as the base is > 0.22+0.6 = 0.82V with reference to the negative rail the transistor is turned on and the current is controlled by the current source. There's 2mA coming down through Q1 & Q2. As long as half of it goes through Q1 (the balanced condition) then the top of R1 is at 1V re. the negative rail. Both Q1 & Q2 (PNP) are turned on because their bases are at ground and their emitters are at ~+0.6V. R2 ties the base of Q2 close to ground and the emitters are tied together and (ignoring R3) the current through R2 is only 1mA/hfe of Q2. The source has a comparatively low impedance to ground, so the current out of the (PNP) base of Q2 to balance the LTP flows with little hindrance.
 
Last edited:

kasey197

Member
2009-09-08 9:29 am
There is some very practical and useful info for solid-state CCS in the tube section for ex see this link. This thread: http://www.diyaudio.com/forums/tubes-valves/200264-ccs-ltp-please-help.html has some gems, and this is particularly useful: http://www.diyaudio.com/forums/atta...6-ccs-ltp-please-help-diyaudio-ccs-beta2_.zip

The Walt Jung articles are also a must read. I know there is also very cool stuff collected from a (i think) european magazine but I can't seem to find it.

Maybe our resident taxonomist and all-round archival genius tiefbassuebertr can post some of them....