[G Georgopoulos]

A discussion on damping factor can also be found here:

 http://www.audiocircle.com/index.php?topic=117407.msg1231829#msg1231829

cheers

[art]

Pete,do you know of a formula for calculating the impedance of current sources and current mirrors...

Depends on current, and other factors. Walt Jung has a series of articles of CCS, on his site. No, I don't know the URL..............you will have to find that yourself.

They could range from 100k, up to over 1 G. The latter is for cascoded MOSFETs. If you are running tons of current, they won't be 100k. It all depends where you are using them, and what you want to do with them. If the Z is in the 1M region, a few pF of stray capacitance will eat into the frequency response.

[G Georgopoulos]

I found this link of a series of articles:

http://electronicdesign.com/author/walt-jung

havent yet found what I want,if anyone has better luck please post the link here

thanks art for contributing

in the meantime I will go through the articles of the above link

cheers...

[art]

No, look on his site. He did an article, I believe in 2 parts, that appeared in what used to be Audio Express. (I think..............senior citizen moment.)

Anyway, for a simple 1-transistor CCS, you need to know the "h parameters". Not sure anyone puts enough info in their datasheets, these days.

But..............raising current lowers Z. Adding emitter resistance raises it. Cascoding really raises it.

A long time ago.................

RCA had some app notes, that showed the I-V curves, for various 2- and 3-transistor current mirrors. More importantly than increased Z was how stiff the curve was. IOW, the current did not vary as voltage was raised or lowered. That is something you can do, with just a voltmeter. Measure the current, as voltage across a load resistor. Raise and lower the rail voltage, and see how stable it is. A good CCS will not vary, no matter what.

You can play around, and see if you like a 3-transistor current mirror, or a simple CCS, that is cascoded. Which only takes 2 transistors.

Enjoy!

[G Georgopoulos]

No, look on his site. He did an article, I believe in 2 parts, that appeared in what used to be Audio Express. (I think..............senior citizen moment.)

Anyway, for a simple 1-transistor CCS, you need to know the "h parameters". Not sure anyone puts enough info in their datasheets, these days.

But..............raising current lowers Z. Adding emitter resistance raises it. Cascoding really raises it.

A long time ago.................

RCA had some app notes, that showed the I-V curves, for various 2- and 3-transistor current mirrors. More importantly than increased Z was how stiff the curve was. IOW, the current did not vary as voltage was raised or lowered. That is something you can do, with just a voltmeter. Measure the current, as voltage across a load resistor. Raise and lower the rail voltage, and see how stable it is. A good CCS will not vary, no matter what.

You can play around, and see if you like a 3-transistor current mirror, or a simple CCS, that is cascoded. Which only takes 2 transistors.

Enjoy!

Hi Art, i have found the 2 part audio express articles,but they have nothing on
impedance of current sources/mirrors,your mention of cascoding current sources/mirrors
is something i have never done,in theory is a great concept and i thank you for telling me,it's
something i'll try and see what results that gives...

i want to thank you and Pete for telling me about these theoretical concepts,
please post again!!!...

cheers..

[art]

Yes, it does. Fig. 2 gives the impedance for several calibration values.

So, if you look at 3B, it shows that it is 45k.

6B shows how 6A is improved, at low frequencies, for a cascode.

OK............I'm going back to what I was doing.................enjoy!

[G Georgopoulos]

Here are the links of the articles art is referring to:

http://waltjung.org/PDFs/Sources_101_P1.pdf
http://waltjung.org/PDFs/Sources_101_P2.pdf

cheers...

[G Georgopoulos]

I have built current source/mirror amplifiers but when it
came to how they sound i didnt like it the sound, was like
open loop amplification,it was kind of too loose and extreme
whereas when i used resistor loads the sound was more
warm and natural.

This headphones amplifier uses resistor loads both in the
differential amplifier and the voltage gain amplifier,even
though it does not push in the extremes of nfb it has great
s/n ratio of >100dB,and with an open loop gain ~250 it
sounds good,i plan to use active devices and extremes of
nfb to see how they compare to amplifiers using resistor
loads...

cheers..