1. Home
  2. Electronics Design
  3. Op-amp behaviour with a failing battery

Op-amp behaviour with a failing battery

I have recently had a very embarrassing episode where one of my prototypes failed when it was being used at a public event.

It appears, on post-mortem, that the battery voltage had fallen to well below the normal limits and the op-amp output became inverted. This turned the negative feedback loop into positive feedback; the resulting screamy-farty noise through the P.A. system nearly deafened 100 people.

I have heard short bursts of similar noises on mains-powered kit when the supply rails fall after the power has been switched off, but the noise was sustained on this occasion due to the more-gradually falling battery voltage. In all these cases the NE5532 was the op-amp involved.

As the equipment is designed to run at microphone level, it has to use low noise op-amps. The obvious solution is to find an alternative low noise op-amp which doesn't misbehave when the supply voltage falls below the manufacturer's usual limits. I have looked at many data sheets but they don't say whether or not this inversion behaviour is likely to occur.

Does anyone have any practical experience of a low-noise replacement for the NE5532 whose performance degrades gracefully as the suppy voltage falls towards zero? The high output current of the NE5532 is not needed in this particular case.

--
~ Adrian Tuddenham ~
(Remove the ".invalid"s and add ".co.uk" to reply)
www.poppyrecords.co.uk
Reply to
Adrian Tuddenham
Loading thread data ...

"inverted"? Please explain. Perhaps a schematic?

I'd suspect "motor-boating" due to the high impedance of the failing battery.

...Jim Thompson

--
| James E.Thompson, CTO                            |    mens     |
| Analog Innovations, Inc.                         |     et      |
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    |
| Phoenix, Arizona  85048    Skype: Contacts Only  |             |
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  |
| E-mail Icon at http://www.analog-innovations.com |    1962     |
             
I love to cook with wine.     Sometimes I even put it in the food.
Reply to
Jim Thompson

.
d

There probably isn't one. The RC4580 is a dual that's spec'd for

+/-2V, but there's no telling how such severely discharged batteries will interact with /any/ circuit. There are lots of examples of using things like LDRs to fade circuit gains, you might consider working something like that into the equation so that the amp is completely dead by the time you reach 0.8V/cell on the batteries.
Reply to
Fred Bloggs

Would the addition of a few bypass caps across the rails near the op-amp be helpful here?

--Winston

Reply to
Winston

Perhaps. Sometimes all it does is lower the motor-boating frequency ;-)

We need to see Adrian's configuration. Otherwise we're stuck with making suppositions. ...Jim Thompson

--
| James E.Thompson, CTO                            |    mens     |
| Analog Innovations, Inc.                         |     et      |
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    |
| Phoenix, Arizona  85048    Skype: Contacts Only  |             |
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  |
| E-mail Icon at http://www.analog-innovations.com |    1962     |
             
I love to cook with wine.     Sometimes I even put it in the food.
Reply to
Jim Thompson

formatting link

It is not unknown for signal inversion to occur in 'overloaded' amplifier stages (this was one of the problems identified with some configurations of valve phase splitter). Usually this results from very high input signals, but it could occur equally well in a stage which is handling normal signal levels but is starved of current.

It also happens with a low-source-impedance bench power supply if the supply voltage is reduced to about 3v. Surprisingly it does not happen if the voltage is reduced by interposing a resistor in series with a 9v supply; instead the system just goes 'plonk' and shuts down.

I was recording the event, which fortuitously has given me a recording of the moment it happened:

formatting link

The recording mic was about 3ft away from one of the P.A. lousdpeakers. The faulty mic pre-amp was plugged into one of the channels of the band's amplifier. As you can hear from the recording, they cut the power to the band's amplifierto kill the disturbance. It was eventually traced to my mic pre-amp, but at the end of the evening we re-connected it and found that it worked perfectly.

It was only today, when I started bench testing it, that I discovered the battery was nearly dead; that gave me a clue where to start looking for the fault.

--
~ Adrian Tuddenham ~
(Remove the ".invalid"s and add ".co.uk" to reply)
www.poppyrecords.co.uk
Reply to
Adrian Tuddenham

Is there a DC path thru the microphones? Otherwise you've got the inputs right up against rail, which, if I read the NE5532 data sheet correctly, is a no-no. ...Jim Thompson

--
| James E.Thompson, CTO                            |    mens     |
| Analog Innovations, Inc.                         |     et      |
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    |
| Phoenix, Arizona  85048    Skype: Contacts Only  |             |
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  |
| E-mail Icon at http://www.analog-innovations.com |    1962     |
             
I love to cook with wine.     Sometimes I even put it in the food.
Reply to
Jim Thompson

[...]

The mics are electret capsules, so they develop about 3v D.C. and act as the 'half-voltage' supply for all the op-amps. The signal level is very low (around -40dBu) so it doesn't matter if the potential output voltage swing is assymetrical because only a tiny proportion of it will normally be used.

--
~ Adrian Tuddenham ~
(Remove the ".invalid"s and add ".co.uk" to reply)
www.poppyrecords.co.uk
Reply to
Adrian Tuddenham

OK. How "dead" was the battery? Potential? Resistance if you can measure? ...Jim Thompson

--
| James E.Thompson, CTO                            |    mens     |
| Analog Innovations, Inc.                         |     et      |
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    |
| Phoenix, Arizona  85048    Skype: Contacts Only  |             |
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  |
| E-mail Icon at http://www.analog-innovations.com |    1962     |
             
I love to cook with wine.     Sometimes I even put it in the food.
Reply to
Jim Thompson

It is a Panasonic PP3 zinc/carbon battery, the pre-amp draws about 9mA and it had been on load for about 90 minutes when the oscillation occurred. I can't say what the battery voltage was at the time of the oscillation, which happened three days ago, but I have measured it today as follows:

O/C 7.820v

On a 1k load:

1 min 6.662v 2 min 6.523v 3 min 6.421v 4 min 6.343v 5 min 6.281v 6 min 6.229v

At the end of the load test, I measured the internal impedance by means of a 1mA current at 1 Kc/s; it is approximately 13.5 ohms.

--
~ Adrian Tuddenham ~
(Remove the ".invalid"s and add ".co.uk" to reply)
www.poppyrecords.co.uk
Reply to
Adrian Tuddenham

How about the voltage at the microphone terminals? ...Jim Thompson

--
| James E.Thompson, CTO                            |    mens     |
| Analog Innovations, Inc.                         |     et      |
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    |
| Phoenix, Arizona  85048    Skype: Contacts Only  |             |
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  |
| E-mail Icon at http://www.analog-innovations.com |    1962     |
             
I love to cook with wine.     Sometimes I even put it in the food.
Reply to
Jim Thompson

Oooops! Datasheet says supplies are +5V and -5V MINIMUM!

I'm not quite sure why... maybe that FET bias start-up?

Take data on Ivcc versus VCC... it may go up abruptly at some low voltage. ...Jim Thompson

--
| James E.Thompson, CTO                            |    mens     |
| Analog Innovations, Inc.                         |     et      |
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    |
| Phoenix, Arizona  85048    Skype: Contacts Only  |             |
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  |
| E-mail Icon at http://www.analog-innovations.com |    1962     |
             
I love to cook with wine.     Sometimes I even put it in the food.
Reply to
Jim Thompson

If it's a 9V battery those are often not very good quality and can go hi-Z on you. Also, AFAIK only the NE5532 is endorsed to be operated down to 6V, while the others aren't.

An alternative might be the LM833 but only the one from ST that goes to

+/-2.5V or 5V single supply. Mind the common mode range, if you exceed that you can get all sorts of weird behavior.

Why not use a voltage supervisor chip and turns the audio path off when the battery sags below a certain limit? You might want to use two where the first one comes on a bit earlier to turn on a blinky-blink LED so the audio guys see some adcance warning that the amp is about to call it quits. Or have it cause a mild fart :-)

--
Regards, Joerg

http://www.analogconsultants.com/
Reply to
Joerg

The NE5532 data sheet says ±5V MINIMUM :-( ...Jim Thompson

--
| James E.Thompson, CTO                            |    mens     |
| Analog Innovations, Inc.                         |     et      |
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    |
| Phoenix, Arizona  85048    Skype: Contacts Only  |             |
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  |
| E-mail Icon at http://www.analog-innovations.com |    1962     |
             
I love to cook with wine.     Sometimes I even put it in the food.
Reply to
Jim Thompson

Not this one:

formatting link

But at +/-3V it's probably like molasses.

--
Regards, Joerg

http://www.analogconsultants.com/
Reply to
Joerg

l

That differential thing in there has positive feedback unless your schematic is in error. It may be a conditional stability issue that usually appears at lower open loop gains/ lower supply voltage.

Reply to
Fred Bloggs

Aha! That rang some bells. These NE5532 are low noise spins on the (various) 741 architectures. I say "various" because there are at least 3 different schematics calling themselves 741... and they have different common-modes ranges near the positive rail. I got burned in the early '70's when I was buying 741 die... Fairchild was the only version which would go to positive rail. ...Jim Thompson

--
| James E.Thompson, CTO                            |    mens     |
| Analog Innovations, Inc.                         |     et      |
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    |
| Phoenix, Arizona  85048    Skype: Contacts Only  |             |
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  |
| E-mail Icon at http://www.analog-innovations.com |    1962     |
             
I love to cook with wine.     Sometimes I even put it in the food.
Reply to
Jim Thompson

Sorry, that was a drawing error - now corrected.

--
~ Adrian Tuddenham ~
(Remove the ".invalid"s and add ".co.uk" to reply)
www.poppyrecords.co.uk
Reply to
Adrian Tuddenham

In practice it is quite satisfactory down to 6v total (+-3v), which corresponds to 1v per cell of the battery. The trouble only appears at around 3v total, by which time the battery should have been replaced. Under the fraught conditions which prevail during a live performance, it is only too easy for the musicians to ignore the falling gain or even a blinking pilot light which warns of a failing battery - until it runs right into the ground.

That might be one to follow-up, but there is a very nasty hiccup in the supply current around 2.5v total which might indicate a similar problem to the one I have already experienced. I did wonder whether

*increasing* the supply impedance in a controlled way might make the chip shut down gracefully at low supply voltage.

That is exactly the problem, the fart did occur after the other warning signs had been ignored - but it wasn't mild, it was almost rail-to-rail on an amplifier input which was expecting -40 dBu.

formatting link

This disagreement between data sheets seems to be an indicator of where future problems might arise. I really need a chip where the low voltage performance is part of the overall specification for the type and is followed by every manufacturer.

--
~ Adrian Tuddenham ~
(Remove the ".invalid"s and add ".co.uk" to reply)
www.poppyrecords.co.uk
Reply to
Adrian Tuddenham

well

The microphone supply point (top of the 100uf capacitor) tracks the supply voltage as follows:

V supply V mic

9.00 4.44 8.00 4.34 7.00 3.53 6.00 2.83 5.00 2.08 4.00 1.44 3.00 1.08 2.00 0.66

The mic capsules die completely at 1.45v on the capacitor, so there is actually a warning that the battery has gone dead some time before the op-amps reach 'farting' point.

I have listened through the recording of the event again and can hear that the hammered dulcimer (which was using this mic) had gone quiet about an hour before the farting event and hadn't been audible at all for at least the last 20 minutes. I was among the dancers and had noticed that the dulcimer was low in the mix, but for diplomatic reasons I didn't want to interfere.

--
~ Adrian Tuddenham ~
(Remove the ".invalid"s and add ".co.uk" to reply)
www.poppyrecords.co.uk
Reply to
Adrian Tuddenham

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.