A capacitor is inert when its voltage is constant. Any attempt by a CCR to retain current as a constant, by dynamically varying its voltage drop, then has a capacitor in parallel that draws current every time voltage is dynamically varied. The net effect for the common cathode node is a non-constant current to ground.
RIGHT you are, and I'm tickled pink to hear someone besides myself cite this. And just as right, you could also cite that technically, constant current sources are not desired on what is essentially a current amplification output stage.
To whit: for the most part, all stages before the final(s) are voltage amplifier stages. Some of course are current-amplification stages, especially those of the long-tailed-pair LTP topology. They count on a CCS to deliver a finite and essentially invariant current, shared between a pair of tubes (or FETs or transistors), where signal comes into one (or differential signal into both), and the (1 - x) : (x) relationship very strictly defines the symmetry of the pair of plate/drain/collector resistances that end up turning the current back into amplified voltage. Gain.
Most of us know this.
But the whole point of the OUTPUT or final stage is not-so-much to amplify voltage (and for “sand” amplifiers, not at all to amplify voltage), but to dramatically lower the output impedance, to pass a large, changing CURRENT as well as the attendant voltage swinging. This is 'cuz the output either directly (sand-amps of both MOSFET/FET and transistor designs) drive speakers, or an impedance-matching transformer of substantial power handling. Again, to drive both current and voltage to the cones.
Therefore, if you view a "CCS" as the ultimate self-adjusting resistor, capable of finely delivering what ever the design-point current into its load, then bypassing that current regulator in turn dramatically lowers its variable impedance, making it a constant voltage device. Where the quiescent voltage biases the tube or MOSFET to exactly that bias point required by the design regardless of tube aging and manufacturing variations, and the bypass capacitor takes care of "bypassing" all audio-frequency modulations tracking the speakers and driving signals.
Of course I could be missing something. But a fairly long lifetime of experience reassures me that the above is a good way to think about it.
Just saying, or braying,
GoatGuy
Hi Andy,
not a schematic at hand. Do you have some bipolar transistor on hand we can work with?
Have a look at this document showing how to set up a NPN CCS, specially pages 3 and 4 talking about the LV NPN CCS and HV NPN CCS. The document is very well written, so I won't write much in addition
http://www.pastisch.se/faktiskt/diyAudio-CCS-beta3.pdf
Have to admit that I am also not very familiar with SS, but these CCS are really no secret. The above document talks about lower currents, so the Bipolars mentioned there won't work on an output stage. I would therefore look to use a high voltage MJE340 as the upper transistor (Q2) and a BD135 for the lower device.
Any further question, please post!
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.