Some chips already have the double diodes. The resistor? Opamps are designed to be multi-purpose. Many purposes (including audio) do not require that resistors. Some may be okay with 1k but some other (e.g. wide bandwidth ones) not.
In my experience, if you religiously follow the datasheet limits for a device, then I have never found an IC to fail, especially now in this age of advanced IC processing.
Before you blame the ICs, you have to make certain that you have not exceeded the voltage ratings for each pin of the IC in your circuit. A certain danger is when you have external circuitry that uses excessively high voltages, beyond the ICs supply rails, but your IC circuit does not have any clamping or limiting circuitry to make certain that these high external voltages will not be imposed upon the IC.
As for your assertion that ICs "decay", there are a number of really old recording consoles (and other devices) that use op amp ICs from what I call the "dark ages" of silicon processing, and they're all running happily, despite the sloppy problems that were common to ICs of the late 70s. ICs from the mid 80s onward are generally well built and should last for 100 years or more, unless their application circuit exposes them to failure modes like ESD discharge or terminal voltages beyond the supply rails.
Before you blame the ICs, you have to make certain that you have not exceeded the voltage ratings for each pin of the IC in your circuit. A certain danger is when you have external circuitry that uses excessively high voltages, beyond the ICs supply rails, but your IC circuit does not have any clamping or limiting circuitry to make certain that these high external voltages will not be imposed upon the IC.
As for your assertion that ICs "decay", there are a number of really old recording consoles (and other devices) that use op amp ICs from what I call the "dark ages" of silicon processing, and they're all running happily, despite the sloppy problems that were common to ICs of the late 70s. ICs from the mid 80s onward are generally well built and should last for 100 years or more, unless their application circuit exposes them to failure modes like ESD discharge or terminal voltages beyond the supply rails.
Reverse biasing of junctions is one reason I never recommend these popular component testers that claim to identify a device no matter which way around you connect it.
Its also a matter of degrees, you are never going to impair a 2N3055 by reverse biasing it with a DVM test, but a device designed for ultra low noise, perhaps having an Ic rating of just a few milliamps and its a different story.
Its also a matter of degrees, you are never going to impair a 2N3055 by reverse biasing it with a DVM test, but a device designed for ultra low noise, perhaps having an Ic rating of just a few milliamps and its a different story.
In pre IC days,when Dinosaurs roamed the earth,a popular gain block consisting of a DC coupled pair ,sometimes followed by an Emitter Follower,was used in audio designs.I have a circuit for a 1970s Goodman's Hi-Fi . Mullard's "Transistor Audio and Radio Circuits" ( mine dates back to 1969) and the "G.E. Transistor Manual" make for interesting reading for those who prefer not to have ICs in their designs.
I think Monte is right,failure might be due to ESD.
I think Monte is right,failure might be due to ESD.
Recently RS and others supply most parts in antistatic bags. One theory is it is easier than teaching staff which to protect. A friend says not at all. He thinks all devices are senstive to damage. Julian Vereker said the FT of output transistors degrades if precautions are not taken even at usual current limits. Alas he didn't take the arguement to the hows and the whys of it. All the same it sounds about right.
Possibly you are correct but the IC would have to be removed and tested. Like the geofex link stated its a degradation not a fail. Ic mfgrs put a lot of pressure on production to continuously reduce die size.
Those old parts did not suffer same.
My devices failed when HP generator was left attached outputing a 10vrms square wave and dut was powered down.
10ma was then exceeded opening the input.
So I have three choices, add current limiting on inputs, make the parts easy to replace or go discrete.
Discrete can suffer same fail but I have more control on the part.
Currently the IC cannot be replaced due to packaging.
The entire assy is thrown out.
A technical hypothesis without any proposed underlying mechanism, and without any supporting measurements either, is utterly worthless. It's the equivalent of saying "My car's engine makes more horsepower when the moon is full", or "Mpingo disks make my Hi-Fi sound better" ( The Magic of Mpingo ).
Taking a worthless hypothesis like that seriously is the first step towards superstition, which is a very poor substitute for engineering and science.
As far as I know, Ft of a device depends on internal capacitances (fixed at manufacture), and on the doping density of the silicon (also fixed at manufacture). Electrode voltages affect it as well, since they change the thickness of the depletion regions that act as the dielectric in the internal capacitances. There is nothing mysterious about any of this, as it all fits perfectly with what we know about capacitors and semiconductors.
Incidentally, today's semiconductors are so fast that even if Ft really did mysteriously degrade slightly every full moon, there would be no impact within the audio bandwidth!
-Gnobuddy
There is none that I have ever seen short of destruction. The generation of hot carriers by forcing Vebo and degrading beta is well known and this does not require much current (back to back diodes are standard).
Over the years I had to replace broken Op-amps in several analogue mixing consoles. I noticed that each broken op-amp had the non-inverting input tied to system ground directly. Without any scientific explanation it is my belief for decades that series resistors of several kOhm protect these.
btw. the last client had a bunch of broken NE5532 inside - which should be protected by their internal clamp diodes.
btw. the last client had a bunch of broken NE5532 inside - which should be protected by their internal clamp diodes.
Interestingly, a (protective) 2k resistor at room temperature, i.e. 20 degrees C, generates a noise voltage of about 6 nV/root Hertz.
This is as big as the equivalent input noise voltage of an NE5532 itself, at nominally 5 nV/root Hertz!
If you have seen many opamps with input directly connected to ground fail, then there is at least some experimental data to support the hypothesis. And it is relatively easy to think of possible causes of the failure, such as brief drastic fluctuations in the nominally 0-volt ground line, caused by electrostatic discharge when someone walks across a carpet and then touches in input jack, or plugs in a 1/4" cable. Or even RF noise picked up from a powerful radio transmitter in the vicinity.
So we don't have to invoke any mysterious unknown mechanisms to account for such failures. We already know that if you pulse too many joules into the base of a BJT, it will fail!
I have a Fender Super Champ XD, a hybrid guitar amplifier that includes BJT op amps, CMOS DSP chips, and a pretty traditional "classic" vacuum-tube power amplifier section. Mine is still alive eight years after I bought it, but online user reports suggest that a number of these amps have failed suddenly and for no apparent reason. The failure symptoms almost always point to immediate and complete failure of the CMOS DSP/digital section.
I think it takes good engineering and design (including good PCB layout) to get good reliability out of a PCB that lets 400 volts co-exist on the same board as CMOS devices that fail at anything above 5 volts.
-Gnobuddy
Yes I have tube on board as well. But mine is for hifi not instrument.
The fender has a 10k protect on non inverting.
So still I don't see a solution.
Attachments
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.