Searching the "best" CCS

[Terry Demol]

Originally posted by jcx


This is easily done with a R divder with the same ratio as the feedback network, the bootstrap (small signal) voltage can match the diff emitter V to better than 99%, giving ~ 100x multiplcation of the ccs effective dynamic Z

JCX, has any commercial amp used this arrangement?

It's definately a smart idea and worth a try.

Cheers

Terry

[jcx]

Originally posted by Terry Demol


Not so JCX.

I know and have -used-, (not simmed) both. Some bootstrapped
circuits get their origins from Hawksford.

Terry

How about changing "you're confused"

to

"you're confusing terminology that others in this thread have already agreed on"


when talking about ccs circuits I believe it is useful to make the distinction between the unambiguous term cascode to describe 2 transistor composites, and bootstrapping, which should be readily differentiated from the active cascode as a "passive" circuit technique that relies on a low impedance circuit node that has nearly unity gain with respect to your ccs load point

"C bootstrapped" resistor VAS ccs is the most commonly seen example, dates back to vacuum tube era, same with active cascode

it is fine to consider the underlying commonalty of the 2 approaches too but confusing to have to parse which circuit example you are referring to if you the term” bootstrap” for both


my literary friends solve crossword puzzles as entertainment, I sim discrete transistor audio circuits


my day job is designing precision instrumentation and usually the shortest distance between a transducer and a ADC is through one or more op amps

knowing how much time and effort I spend on debugging, testing, and verifying objective performance measurements in my professional designs I can only be envious of those who can just listen to their circuits and tell that they've improved it by tweaking a single subcircuit

[vax9000]

I read through this thread and I got confused. So what is the best CCS, as a conclusion?

[mikeks]

Originally posted by vax9000
I read through this thread and I got confused. So what is the best CCS, as a conclusion?

Amplified negative feedback (ANF) current source (or sink).

Its output impedance is virtually identical to that of a cascode source, but uses only two transistors and two resistors.

[darkfenriz]

How about these two?

[Wavebourn]

Originally posted by tvi
I posted this sometime ago, you might find it of interest
Constant Current Source article (Old)

Regards
James

James, I used such sources to load vacuum triodes, istead of resistor between emitters I have sometimes to use a bleeding resistor ftom base of upper transistor to the ground, because when transistors are too good compared to Zeners the source won't start.

[MikeB]

Originally posted by vax9000
I read through this thread and I got confused. So what is the best CCS, as a conclusion?

Refering to post #1, following are my favourites:

1, ANF, simple and works good, all parts easily already lying around, can be improved with zener or bootstrap.
2, very good, but needs selected jfet and has low max voltage
3, simple and very good, should also work with leds
7, seems to be best, winner in all disciplines, stability/speed, but needs selected jfet.

The 2 diode version (No4) is quite "useless", all others discarded because of speed/stability issues.

Mike

[mikeks]

Originally posted by darkfenriz
How about these two?

Both are ANF current sources with improved PSRR due to zener regulation, with the first possessing enhanced loop transmission and thus greater output impedance.

This of course is just as dependent on frequency as the forward path gain of the op amp used.

[forr]

To my experience, incestuous CCS's like MikeB's #5 and #10 are very prone to oscillations and disappointing.

The circuit at the right of Darkfenriz's picture is one of the most dependable I know. It has a large bandwidth PSRR. For a temperature elevation of 50°C, the current will decrease about 20%. Using an LED instead of the current sensing transistor gives a better thermal compensation. Supplying the zener current with an other CCS is a quite unnecessary complication.

PSSR is not all. Rejection of the voltage variations at the CCS output may matter. A common base bipolar in a cascode configuration can be used to enhance it. However, once again, too much sophistication may give oscillations. My prefered scheme is to include a 1 Kohm resistor in series above the reference voltage and to connect the base of the common base ouput transistor to its other side. The 1 Kohm resistor has the double benefit of providing a sufficient voltage above the voltage reference and acting somewhat as a base stopper. Sometimes after I concluded that this simple scheme was really good, I found that Morgan Jones uses it in its valve amplifers.

More sophisticated circuits may give marginally better results but are less reliable.

[mikeks]

Originally posted by forr
The circuit at the right of Darkfenriz's picture is one of the most dependable I know. It has a large bandwidth PSRR.

Did you know that its output impedance is just as high as that of a cascode current source?

Originally posted by forr
For a temperature elevation of 50°C, the current will decrease about 20%.

This is a ''good thing'', for it enhances reliability.