Can a AWG replace (or be used as) a RF Generator?

[Commander]

I recently purchased a used HP 8656B Signal Generator (RF). Its looks like a very expensive piece of equipment. But it weighs a ton and is too big for my bench.

The HP can generate a signal to 1Ghz, but I am only interested in SW and CB (<30Mhz) tuneups/repairs.

So, I was looking at these new Arbitrary Waveform Generators, particularly the Siglent SDG1050 (http://siglentamerica.com/pdxx.aspx?id=145&T=2&tid=16). It appears the AWG can easily generate a 28Mhz signal, and even AM modulate a tone with it.

Is there a real difference between the two generators if all I need is a signal up to 30Mhz? For example, can I do just as good/high-quality alignment/tuneup of a receiver with the AWG?

[pigrew]

The AWG definitely has advantages. It can go to higher power (about +24 dBm), lower frequency, generate more interesting modulations, and create signals other than a sine. It also probably has a more accurate output power, especially when performing modulation.

The two instruments have similarly specified harmonic distortion (-30 vs -32 dBc), but my feeling is that the 8656B would be slightly better. For small signals, the AWG probably has a lot of quantization-noise.

The frequency reference in the 8656B would be much better, within a few ppm per year drift. The SDG100 is +/-100 ppm per year. However, this could be easily rectified with an external reference. And, if you care about your frequency, you should be using a more accurate reference (like a GPS-DO, or OCXO).

Another big advantage of the siggen is its dynamic range. It can go down to -127 dBm. The AWG can only go down to -50 dBm.

I've never done receiver alignment, but I think that your major issue would be the frequency accuracy. However, for either instrument, you probably would need an external reference (unless you have the high-stability timebase option 001 for the 8656B).

That being said, I'm in the market for a siggen (I need up to 128 MHz), in case you're selling. [I'm in Indiana]

[amspire]

You will definitely want to get some BNC attenuators so you can get the levels down and this has the advantage of protecting the AWG.

There is a difference in the way the signal is generated. The RF gen probably has an oscillator actually running at the output frequency (or two oscillators that are mixed to produce the output frequency as the difference) so it is possible to have a very steady jitter free signal.

The AWG is generating the output waveform from a fixed clock running at a frequency that is nothing to do with the output frequency and at 30MHz, you are only talking about 2 1/5 A/D samples per cycle. The locations of the samples will be continuously changing from cycle to cycle, and it is totally depending on some kind of super filter to recreate a clean sine wave. I don't think the waveform will match the RF Gen for stability and purity.

If you get rid of the RF Gen, you will probably miss it.

[pigrew]

The AWG is generating the output waveform from a fixed clock running at a frequency that is nothing to do with the output frequency and at 30MHz, you are only talking about 2 1/5 A/D samples per cycle. The locations of the samples will be continuously changing from cycle to cycle, and it is totally depending on some kind of super filter to recreate a clean sine wave. I don't think the waveform will match the RF Gen for stability and purity.

Good point (although I think that the math is wrong?). The SDG1000 series outputs at 125 MSa/s. The nyquist frequency is half that, 62 MHz. You'll start to have images show up at 94 MHz. My guess is that it's output filter will be good enough to get rid of most of the 94 MHz signal. However, it'll stall be there. This is a "spurious signal". Unfortunately, Siglent only specifies the spurious signals when your carrier is less than 10 MHz, so we have no clue how bad it would be. Note that these are not harmonics, so they don't contribute to the harmonic distortion.

If you want a "good" AWG for this use, I'd suggest a sampling rate at least ten times your frequency of interest (so at least 300 MSa/s), if you want a fairly pure signal.

[amspire]

Yes - I did get the wrong sampling freq from somewhere, so my calculations are wrong. That would be 4 and a bit samples per cycle which is better, but not ideal.

[mmagin]

IMHO, any of the good (heavy) RF signal generators of the past 3-4 decades will certainly give you two things modestly-priced AWG/Function generators will not:

• The ability to attenuate the output to very low levels without leaking higher signal levels elsewhere (AC line conducted emissions, radiated emissions, etc.)
• Much better spectral purity

Now, there are a heck of a lot of radio repairs you could even accomplish with a cheap Heathkit/Eico two-tube RF generator from 60 years ago (people did!), and the modern AWG is at least a vast step up from that in frequency stability, among other things.  But if you do much RF stuff, you probably won't regret keeping the 8656B.

[SingedFingers]

Got a Marconi 2019A here.

The killer feature is the calibrated attenuator and display of the attenuator setting. -130dBm. SDG1050 quotes 2mV p-p into 50R minimum which is -50dBm, so you'd need a pile of attenuators to give low power outputs.

[DimitriP]

Can a AWG replace (or be used as) a RF Generator?

Time to remodel the bench

Here is an interesting thread on the subject https://www.eevblog.com/forum/testgear/rigol-dg4000-issue-with-output-level-of-modulated-signal/msg1097424/#msg1097424

[Commander]

Wow, that guy was not happy. I am surprised with Rigol's response, they could *easily* fix that problem with a software change. I am also a programmer and it would be kids play to simply add a lookup chart to the firmware that would auto-adjust the digital signal output as needed for each % of modulation so the carrier db level stays the same.

Thanks guys for all your replies. I have a lot to think about, but its looking like I should keep the HP.

[mmagin]

Wow, that guy was not happy. I am surprised with Rigol's response, they could *easily* fix that problem with a software change. I am also a programmer and it would be kids play to simply add a lookup chart to the firmware that would auto-adjust the digital signal output as needed for each % of modulation so the carrier db level stays the same.

I'm not sure.  Possibly the magnitude of the effect varies from unit to unit?
Anyway, that response from Rigol is complete garbage.  If they said that it maintains constant (RMS/average/peak, power/voltage/derp-a-aderp) when modulation level is varied or something as to the actual design goal, I'd say, "okay".  But this mumbling about time and frequency domain is pretty hilarious.

[eb4fbz]

AWG are usually the low frequency cousins of vector signal generators. Both could be used as RF (tone) generators with no problem in their frequency range.