And what should we call this ?

[tautech]

Recently a member was looking for advice on which of 2 sig gen instruments he should select however one being capable of just over 1 GHz and the other 2.4 GHz I suggested they should be more accurately called RF gens.

To my amazement many don't seem to think so and IMO recklessly refer to Function, Signal, Arbitrary Waveform and RF gens all as Signal generators.   
Really great way to introduce more confusion into this industry where IMO each is a quite differently capable instrument operating in greatly different frequency bands.

IME with instruments I've owned and some I still do a 2 MHz Phillips PM 5131, 30 MHz Siglent SDG1032X, 520 MHz HP8654B or a 2.1 GHz Siglent SSG3021X are each so different from one another they cannot all be referred to as just a Sig gen. 

Customer contacts tautech for a sig gen  what exactly do you need sir ?
A sig gen, you know, something that generates signals. 
tautech begins producing his own smoke signals ....from ears.
Sir, what sort or signals and in what frequency range and amplitude requirement ?
Oh, something like what an arbitrary waveform generator can do.
Deep breath, budget sir ?

Please discuss.

[xrunner]

Endless debates coming ... 

RF generator - used to test receivers or intermediate stages of such. May output from 100 kHz to GHz range. Ideally can be modulated with a baseband signal ... internally or from a function generator.

Function generator - used to generate useful baseband stimulating signals such as square, sine, triangle, step for testing circuits and components. Can be used to modulate RF generator (see above).

Arbitrary function generator - used to create signals not available from the standard signals on the function generator.

[pqass]

Although there are different names for more or less the same thing,
I've always understood it goes something like this...

If you can choose the general shape of the wave (square, triangle, sine, ramp, sawtooth), then it's a function generator.
If you can design your own wave shape, then it's an arbitrary waveform generator.
The above can both still have sweep or modulation (AM, FM); no name change needed.
But if the device only generates sine waves and has modulation (incl. exotic stuff) then it's a signal generator.
The above assumes that it can only produce a small* signal at its output.   (* <10Vp-p into 50ohm; milliWatts)
Otherwise, it's a transmitter. :-)
RF or microwave only figure into it as the starting frequency, say 500KHz or 1GHz, respectively.

[tautech]

Thanks those brave enough to answer thus far. 

However we should also add another in my mind determining factor, the signal output connector type as if further indicates specific use types, Banana, BNC, N type or some HF x mm type, 2.92mm for example.

[vk6zgo]

Historically, most "Signal Generators" were either Audio Generators or RF ones.
As such things as Function Generators, Analog video waveform Generators, & Sweep generators became mainstream, they were normally referred to by their proper names.

The term "Signal Generator" normally referred to devices, of varying degrees of accuracy, which produced sinewaves, at selectable frequencies & amplitudes.

"Audio Generators" were commonly designed to possess as low a level of distortion & noise as possible, especially if they were used to supply the input signal to a DUT when testing for noise & distortion.

"RF Signal Generators" normally were designed to offer a high degree of frequency & output level accuracy, & in many cases provided
modulation facilities, such as AM or FM.

In places which used both types, it was usually obvious by the context which device was required, so "Signal Generator" was sufficient.

"RF Generator" in a Radio transmission site could mean the basic RF exciter for the transmitter.
Some companies called them "exciters" & others "RF Generators".

I would suggest the least ambiguous term for the test instrument is "RF Signal Generator".

[bdunham7]

To my amazement many don't seem to think so and IMO recklessly refer to Function, Signal, Arbitrary Waveform and RF gens all as Signal generators.

I don't see the problem, they all generate signals of one sort or another.  "Signal generator" is just a generic term like "vehicle".  If you want to be specific, be specific.  And do remember that the 8662A that started this issue produces 'signals' down to 10kHz and actually says "Signal Generator" right on it's face...

[Bud]

"RF Generator" does not make sense. Got to be "RF Signal Generator if you are picky. Otherwise just a "Signal Generator" followed by the design frequency range.

[ebastler]

I dunno; seems like you are trying to create some artificial nomenclature there. Why would a 2 MHz or 40 MHz generator not qualify as an "RF generator"? That's RF in my book. And pretty much all modern signal generators can modulate their output, from external and/or internal sources, so that does not provide a distinction for "RF generators" either.

Just call them all "signal generators" and mention the key specs you are after -- max frequency, stability, signal/noise, modulations capabilities, arbitrary waveforms or whatever.

[tautech]

Just call them all "signal generators" and mention the key specs you are after -- max frequency, stability, signal/noise, modulations capabilities, arbitrary waveforms or whatever.

Lab senior says to junior, please go to the stores and get me a sig gen. 

[ebastler]

Just call them all "signal generators" and mention the key specs you are after -- max frequency, stability, signal/noise, modulations capabilities, arbitrary waveforms or whatever.

Lab senior says to junior, please go to the stores and get me a sig gen. 

Fixed the emphasis in the quote for you. Why would you willfully misinterpret my quote and stop reading after six words? For a rather lame joke?

[The Soulman]

Just call them all "signal generators" and mention the key specs you are after -- max frequency, stability, signal/noise, modulations capabilities, arbitrary waveforms or whatever.

Lab senior says to junior, please go to the stores and get me a sig gen. 

Terrible lab senior.

[DimitriP]

https://www.keysight.com/us/en/assets/9018-05715/user-manuals/9018-05715.pdf

some are even high performance signal generators
"Spectral purity is the key contribution of the Keysight 8662A signal generator, making it ideal for many radar, satellite communication, and phase noise measurement applications. Typical absolute phase noise performance of this generator at a 1 kHz offset is as low as -135 dBc/Hz, depending on the band of operation. The frequency range of the Keysight 8662A is 10 kHz to 1280 MHz. It offers versatile AM/FM, using either internal 400 Hz and 1 kHz rates or externally applied modulating signals."
https://www.keysight.com/us/en/product/8662A/highperformance-signal-generator-12-ghz.html

Done.
You are welcome.
The End!

[switchabl]

Traditionally, the term "signal generator" has mostly referred to what you seem to call an "RF generator". The earliest example I could find is the General Radio Type 403 "Standard-Signal Generator" from 1928 ¹.

However, there has always been some ambiguity. It has been mentioned that the term has also been applied to audio generators as well. This HP presentation from the 70s maintains that, among other things, "[...] signal generators must include some provision for calibrated modulation" and complains that "other sources which don't meet this definition are often mistakenly called signal generators". ²

These days, it seems, it is also used (maybe somewhat confusingly) as an umbrella term: https://en.wikipedia.org/wiki/Signal_generator

I generally consider it important to use precise language where it matters. In this case, I cannot really bring myself to care much. The meaning is usually clear from the context and where it is not, qualifications like "RF signal generator", "vector signal generator", "audio signal generator" or "video signal generator" are available.

Consider this analogy: if your were to enter (i) a power electronics lab, (ii) a microwave lab and (iii) an optics lab and ask to see a "power meter" you would in all likelihood be shown three very different types of equipment. And yet, this does not seem to cause any real problems in practice.

¹ https://worldradiohistory.com/hd2/IDX-CATALOGS/IDX/General-Radio/General-Radio-Catalog-E1-1928.o-OCR-Page-0093.pdf

² I suspect that this is not just due sloppyness but reflects the fact that people can't quite agree on what a "signal" is either. Depending on who you ask, almost any changing quantity might qualify, while others will insist that a signal carry "information".