Not correct. With sufficient emitter degeneration, the current ratio is almost only determined by the resistors. Better than 5% is easily done, and this is already more than good enough in most cases.
Paying attention to power supply rejection, and in some cases output Z and noise, is of much greater concern IMHO.
Samuel
Lousy bandwidth, you'd do much better with a single transistor circuit. A two transistor cascode will be even a better CCS, but the extra performance of an improved CCS will do next to nothing for the overall performance of the circuit.
I put the wrong question. I meant the behaviour at higher frequencies.What would one respond to this question Franzm? I don't know of any thing that has a static behavior.
This type of current source can be made with a very high impedance, but in an amplifier it must also be able to follow the voltage variations of the signal very precisely, and these in general can't, certainly not when made with the types of op amps you find in application notes.
In practice the single and double transistor current source are all that's needed, and in some high frequency large swing application a cascode current source is very useful.
rcw
In practice the single and double transistor current source are all that's needed, and in some high frequency large swing application a cascode current source is very useful.
rcw
WJ On current sources from his AX series:
http://waltjung.org/PDFs/Sources_101_P1.pdf
http://waltjung.org/PDFs/Sources_101_P2.pdf
-- note that you can modulate a depletion MOSFET
http://waltjung.org/PDFs/Sources_101_P1.pdf
http://waltjung.org/PDFs/Sources_101_P2.pdf
-- note that you can modulate a depletion MOSFET
Op amp CCS are good when it works, but probably not so good when it doesn't or when the rest of the amplifier is not in a normal operation condition as the CCS is intended to work with. CCS constructed with transistor can have very predictable behavior during transitional moments such as powering up/down, or in the circumstances the circuits they are serving go into non-linear status such as clipping, saturation, or when the protection mechanism kicks in. One probably could hardly say the same to an Op amp based CCS.
I proposed an Op amp based biasing control circuit for a diamond buffer triple output stage in the thread of "Bob Cordell's Power Amp Book", Bob did indicated his preference towards transistor based "passive" bias control over an Op-amp based design although he encouraged me to further work on my idea. I understand the motto of better dealing with a known devil than with an unknown one. The likelihood to go wrong with a "passive" circuit does seem lower than an "active" one.
I proposed an Op amp based biasing control circuit for a diamond buffer triple output stage in the thread of "Bob Cordell's Power Amp Book", Bob did indicated his preference towards transistor based "passive" bias control over an Op-amp based design although he encouraged me to further work on my idea. I understand the motto of better dealing with a known devil than with an unknown one. The likelihood to go wrong with a "passive" circuit does seem lower than an "active" one.
A cascaded FET is so many times better than other issues in a power amp, it is just no the place to worry about things. There are also quite a few folks that think a resistor sounds better. Probably because it behaves the same at all frequencies, something active current sources do not do. I would be far more concerned with the outputs.
Hi,
In #6 You suggested that the OPamp CCS would be very accurate and Sam gave in #21 a first reply to that. i'd say that the CCS presented in #1 is not acurate at all. The accuracy of that CCS depends mainly on the reference voltage at the OP's input pin, here generated by a simple voltage divider from Vcc. Any noise, ripple, temperature tracking or other disturbance directly affects the output current. If noise is a major factor a simple JFET/Source-R CCS may be best, even if it's current stability and output impedance are just mediocre and temperature dependant.
Also the cost of a OPamp based CCS is more than just the pennies for an OP, but also the cost for more resistors, capacitors, a high precision VCS and layout board space and design time
jauu
Calvin
In #6 You suggested that the OPamp CCS would be very accurate and Sam gave in #21 a first reply to that. i'd say that the CCS presented in #1 is not acurate at all. The accuracy of that CCS depends mainly on the reference voltage at the OP's input pin, here generated by a simple voltage divider from Vcc. Any noise, ripple, temperature tracking or other disturbance directly affects the output current. If noise is a major factor a simple JFET/Source-R CCS may be best, even if it's current stability and output impedance are just mediocre and temperature dependant.
Also the cost of a OPamp based CCS is more than just the pennies for an OP, but also the cost for more resistors, capacitors, a high precision VCS and layout board space and design time
jauu
Calvin
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.