I have my eye on the TDA2822 dual audio amp. As with many popular audio amplifiers of recent decades it includes a Zobel network (Req and Ceq in the schematic above) to compensate for the reactance of the speaker.

My question is: In applications which do not include a reactive component like a speaker, using an IC like the TDA2822 (or for that matter the ever-popular LM386), such as when driving a diode bridge balanced modulator, be stable without the Zobel network?

I suspect some of us have been through this in more depth than I have, and I solicit your opinions as to whether or not such uses will be stable and relatively free from ringing.

Opinions, please.

 

Hello,

A zobel network is an impedance compensation network.
See this page for more info:
https://sound-au.com/articles/z-compensation.htm

Bertus

 

The voltage gain of the system is gm.Zload, so gain increases with frequency if the load is inductive. The Zobel network adds a parallel resistance at high frequencies (frequencies where the zobel capacitor is effectively a short circuit) and prevents the gain from increasing far enough to cause instability.
A special case is when R and L of the load are both known, and then the Zobel components can be calculated so that the load appears resistive at all frequencies.
This happens when Rzobel = Rload, and Czobel = Lload/(R^2)
There is a good explanation in Doug Self’s power amplifier book.
Old SGS-Thomson amplifier such as the TDA2822 IC are notoriously unstable and need all the help they can get not to be an oscillator.

 

The datasheet recommends the network so it's likely part of the device high-frequency stability compensation for most likely loads. The extra cost and space is trivial for the peace of mind IMO.

 

The RC network is required at the output of many audio chip amps for stability.
It started with the LM380, LM386 as their output stage has a tendency to oscillate at RF frequencies.
It is due to the lateral PNP output driver transistor as part of the Sziklai pair. Called "the bottom side fuzzies" when you look on a scope and see the bottom half of a sine-wave with HF oscillations. It's mentioned in the National Semiconductor Audio Handbook as I recall. It just sounds like distorted audio at higher volume.
So the RC network presents a load at HF on the output pin. The ST TDA2822 has a slightly different output stage but it's still required, TDA2003, TDA2030 etc. as well.

 

Thank you all for your comments.

It looks like the smart money is on “Just do as the data sheet suggests, this is no place to try and save a penny or two.”

 

For those old SGS-Thomson amplifiers, build it EXACTLY as shown on the datasheet, or it will oscillate. Use the same gain setting and power supply voltages.
If you look stare too hard at it, it will probably oscillate.