I've been trying to figure out how to prevent a non-inverting op amp amplifier from clipping to its rails so as to avoid signal inversion or latching if it's gain is set too high for it's input signal.
Ok.
Well, "duh" solutions first:
1. Use an amp that's fine with this. Even TLV2371 will clip without unexpected behavior (phase reversal, slow recovery). (It's probably too slow, noisy and low voltage to be directly applicable, but there are plenty of similar parts out there.)
You don't need to use ancient POS amps like LM741 anymore! (Not that you are, but old and funky still matters.)
TL071 is still rather popular, but there are better options without costing much more. Shop around!
2. Clamp the input instead. Back-to-back diodes? Low level op-amp clipper stage? (Maybe with a fixed gain preamp in front, to keep noise down.)
The clamping might have to be adjustable, or set to a worst-case range, to make this work for any output gain.
A little background, I'm hoping to use this as an input for passive electro-magnetic guitar pickups or other musical instrument signals, so the input impedance must be high (the traditional value is >1MEG). The input signal may be expected to be anywhere between about 100mV peak to up to 4V peak depending on the output amplitude of the passive pickups themselves, or any preceding effects/pre-amplification.
The voltage rails will be +/-9V, I plan to use a TL07x, and the load impedance on the op amp will be 10k, so maximum swing that the TL07x should be able to handle is around 7V peak. I was hoping to limit the maximum output swing of the stage to somewhere below this using diode clipping in the feedback network to prevent the op amp from ever approaching it's rails.
As shown is fine, but you should use an inverting configuration instead, so the diodes clamp against virtual ground. As shown, output is input plus gain, so you'll still saturate the amp on very large inputs.
My problem is, as the gain of the stage is increased with the potentiometer, the capacitance of the zener diodes becomes a problem, and high frequency content is lost. I could use a string of small signal diodes such as 1N4148, but with their low forward voltage, it would take a long string to achieve an appropriate limit. Another issue that I have run into is that I have not been able to find any zener datasheets that specify their junction capacitance, which strikes me as a little odd, but means I am not able to find zeners which may have an acceptably low junction capacitance.
1.
http://www.mouser.com/ds/2/427/1n5221-240386.pdf Fig. 6
Note it's at Vr=2V and the varied parameter is the Vz rating. Actual C for an individual part will follow a C vs. DC bias curve typical for diodes, having quite high capacitance at low voltages, somewhat less by 2V (how much depends on the voltage rating), and a minimum value at just below Vz.
2. Use a smaller feedback resistor. Which means you need an even smaller resistor against the divider, but that's okay. (Not as okay for the inverting configuration, but you can use an input buffer or gain stage to solve that.)
3. Use smaller diodes. Although I don't know that you can really find much under 100pF, anyway.
4. Put one zener inside a FWB of regular (1N4148 ish) diodes. Zener goes cathode to anode across the DC+ to DC- terminals, and the AC terminals go across the op-amp.
This is frequently used when low average capacitance is needed; high speed data line ESD chips do this, for instance.
Tim
Home
Help
Search
Login
Register
