Voltage supply rating on datasheets

[Koreth]

I've been looking at datasheets for things like op amps and chip amps the past few days, and some of the specs in them confuse me a bit. Some say something like 18V, others say +-18V. Are these meaning the same thing or something different. I assume that 18V means it can have a total of 18V across it's power supply terminals before taking damage, and that +-18V would mean that the V+ terminal can be +18V, and the V- terminal can -18V, for  total of 36V across its power supply terminals, but we all know what happens when you assume. Am I on the right track here or am I off in la la land?

[phatt]

Hi Koreth,
YES! correct, +- means *postive and negitive* DCVolts
(with respect to Zero volts)
Phil.

[Koreth]

So then an opamp, output chip or whatever else with a supply voltage rating of 18V used as a voltage amplifier could only swing a max of 18Vpp assuming it received an 18V supply, but one with a +-18V supply rating could swing 36Vpp, again assuming it received its maximum rated supply voltage.

In the various schematics I've looked at, I don't know that I've ever seen any op amp or chip amp run at its maximum rated voltage supply. This makes sense to me as running things at their max all the time leaves no room for error. If ever there was a momentary surge from the power supply for whatever reason, you'd risk blowing the device. I'd like to think that the manufacturer's ratings are somewhat conservative, but better safe than sorry. So, that said, is there a general rule that's considered good design or best practices when deciding what voltages to run devices at, (e.g. so many volts under the max, a fixed percentage of the max, etc.)?

[J M Fahey]

Yes, using not more than 80% of max. ratings is good engineering.
We are talking voltage and current.
Speaking of power transistors, we often can't use even 25% of its power dissipation, because, opposite to voltage and current, dissipation is specified for a ridiculous and practically impossible case: the transistor *perfectly* mounted, no insulation and with the best grease in the universe, to an infinite block (or at least a 2.5 TON one, no kidding) aluminum cube, kept at an eternal 25ºC.
Have you *ever* seen such a heat sink?
Yes, I thought so, me neither.
Op Amps usually support +/- 18V, regularly are used with +/-15V.
Some used in Fender power amps support +/- 22V and "float" to stand being used with +/-40 to 45V, that old Fender trick of using special components to make servicing difficult and cloning impossible. Oh well !! :grr

[Koreth]

I know it's not a voltage question, but it is a datsheet question and I figured better here than to start a new thread.

What about the closed loop gain rating seen on the datasheets for various chip amps? For the TDA20xx series, it's specified with a min, typical and max in dB. Is the minimum gain specified on the sheet mean that's as low as you can set the closed loop gain and any attempt to set it lower will result in some internal configuration in the chip taking over and keeping it at the max closed loop gain? Or is it more like the max voltage ratings, IOW, go too low and you risk damaging the chip, having stability problems, or other badness? Or is this a thing which can vary from manufacturer/chip family/lunar cycle to the next?

[J M Fahey]

Thanks God, no Lunar cycles here.
We already have enough trouble at home because of them.
We must be aware that these are (Power)Op Amps.
Usually three gain levels are specified:
*Open loop* gain is very high, but unusable for us; besides OpAmps are compensated, meaning there is a treble killing capacitor stuck in.
*Typical* gain is a recommended one, guaranteed not to cause problems.
"Preamp" OAs such as TL072, are compensated for unity gain, but some specifically audio ones and most chipamps are not: they have "too small" compensation caps (or the treble response sucks) so:
* Minimum* gain is the value below which OAs are *guaranteed* to oscillate and make our life miserable,