Discrete Op Amp Design Help

thederke said:

what benefits would a PNP add?

Click to expand...

Aren't you making a discrete opamp? Have you looked at the schematics of many opamps on their datasheets to see all the PNP transistors and their current-source (high impedance) functions?

It's not an either/or thing, it's ability to have gain,
bias currents, buffer stages from either rail. For
example rather than all common-emitter NPN,
resistor loaded stages you could have a diff pair
of NPN and a PNP current-mirror load (or vice versa)
and realize much higher stage gains. You could do
folded cascode second stage and have all the gain
you're looking for, and only need to add a buffer
stage (for whatever output attributes).

Yes, waayyy back in the day people made analog
(and digital) parts with only one species of active
device. But there's nothing good about being so
retro.

Check out some old timey databooks for parts like
LM324 - simple as hell, and good enough for a lot of
things. See how they used PNPs.

The LM324 is low power (it has very low power supply current) so it has trouble with high level outputs above only 2kHz and its output does not have bias current so it produces crossover distortion.

Audioguru said:

The LM324 is low power (it has very low power supply current) so it has trouble with high level outputs above only 2kHz and its output does not have bias current so it produces crossover distortion.

Click to expand...

Yikes, yeah the low frequency would be a problem. I'm told by my peers that our professor really only cares about the gain, won't really be checking bandwidth etc., but Ill try to find some PNP schematics and look at how they're incorporated. Would like to do a good job on this.

Another thought: Would an NPN CE followed by a PNP CE have any advantages over two NPNs?

Complementary transistor combinations are needed to make effective DC coupled amplifiers (or in other words OP amps). AC amplifiers don't need it.

The extremely high impedance of a PNP current source allows it to be the collector load of an NPN common-emitter transistor for very high gain.
An NPN current-sink can be the collector load for a very high gain PNP common emitter-transistor.

Hi,

These are (very, very bad!) two-stage op amps, only good for voltage followers, plus their output impedance is very high so they can't carry a sensibly low value resistor load... )))) DO NOT USE THESE AT HOME, OR ANYWHERE ELSE ))))

Anyway, it's the schematic that appears in lots of op amp design notes, etc, and the point was (excusing my fetish for emitter resistors in the current sources), the resistors are replaced with current sources, which apparently boosts gain capability, and gives a feeble degree of PSRR. A full Wilson or a cascode version of that current source would be far better if you have the headroom.

I tried, amongst many other things, a three stage using a diff pair followed by another diff pair to boost gain, with a complementary pair as the output stage, but then the frequency response became dreadful over 2Hz, that'll be human failure/incompetence at placing additional Ccs.



Op amp books and so on do seem to explain current sources a lot as they are a big improvement on using resistors as current sources, unless your class is about designing the uA741, there's no reason to exclude PNP's or make archaic op amps, even if gain is the goal, things have improved since Jack Kilby's day in that respect.

"Current Sources and Voltage References" by Linden Harrison has a load of types you can assess as to their value in op amp design, but as I said, op amp stuff usually includes details about the current sources anyway.

Low voltage rail-to-rail op amps seem to move the learning curve from interesting hobby to self-inflicted punishment...

Hi, why don't you read the "Modern Techniques for the Design of BJT OP Amps by Sedra in his Microelectronic Circuits (chapter 13)" ?

Here's a neat history of op amps, and the first one shown
is NPN-only. Lots of schematics.

https://www.google.com/url?sa=t&rct...mp.pdf&usg=AFQjCNEu5k63m3WVSquWFVWTJlUM6z32Yw

I use this history each year to find out the age of the lousy old 741 opamp (today it is almost 49 years old).