Can your arb generator do this?

[nfmax]

I had a play with my shiny new Agilent 33522B, and it is possible to do what you want, using a single channel (so a 33511B would also work).

Darn... that series of arbs looks better and better... maybe I'll keep an eye out for a used one...

Thanks for checking this.

I got mine used for 1600 USD (which seemed unusually cheap). It was in perfect condition, too! I hadn't been specifically looking to upgrade my ARB, but when an opportunity arises...

[Berni]

I have been using my Tektronix AWG2041 signal generator to generate sonar pulses.

It has what is essentially segmented memory where a set of waveforms is stitched together in a sequence. This lets me separate the active pulse area from the long silent area so that i don't need to fill the memory with silence (The silence is just a short segment of silence looped around 1000s of times). Requires no PC or software to set all of it up too.

Additionally it has two marker outputs. I was using one of them to generate a Enable signal (High during burst) for my power amplifier and to trigger my scope. The DAC sample rate can also be set up to 1GHz with 5 digits of resolution. But its only 8 bit so i guess it won't meet your requirements.

[nfmax]

The 33522B has segments as well, but you don't need to use them for the OP's application.

[precaud]

Correct, segments aren't needed in this app.

I got mine used for 1600 USD (which seemed unusually cheap). It was in perfect condition, too! I hadn't been specifically looking to upgrade my ARB, but when an opportunity arises...

Boy, you got a heckuva deal... used 11B's seem to go for that much. Is that an OLED display?

So... thus far, the 33500B series is the only capable replacement for the Pragmatic.
And the adaptive filtering puts them in a class by themself.
Despite all that, I really don't want to spend that much on an arb gen.
So unless/until another viable alternative arises, I'll keep using the 2414A.

[nfmax]

The display is described as "4.3" Color TFT WQVGA (480x272) with LED backlight" so not OLED.

The 16-bit DACs are nice, too. There are all sorts of other goodies, like variable bandwidth (pseudo-random) noise that can be used to modulate or add to an output. I've hardly started to explore it yet. This is an upgrade from an HP 33120A which was the first ever model of ARB I used, many years ago, to simulate a pulse-output transducer.

I now have a collection of 33522B, 33120A, 3312A, & 3311A generators...

[jpb]

Correct, segments aren't needed in this app.

I got mine used for 1600 USD (which seemed unusually cheap). It was in perfect condition, too! I hadn't been specifically looking to upgrade my ARB, but when an opportunity arises...

Boy, you got a heckuva deal... used 11B's seem to go for that much. Is that an OLED display?

So... thus far, the 33500B series is the only capable replacement for the Pragmatic.
And the adaptive filtering puts them in a class by themself.
Despite all that, I really don't want to spend that much on an arb gen.
So unless/until another viable alternative arises, I'll keep using the 2414A.

My 33522A cost around £1500 from Agilent/Keysight's ebay store but this was with 20% VAT so probably close to $1600 ex VAT. The Keysight ebay store do good deals from time to time and they accept offers but it all depends on what demo stock/trade-in stuff they have.

They have one for $1780 asking but would probably take closer to $1600:
https://www.ebay.com/itm/Keysight-Used-33522B-Waveform-Generator-30-MHz-2-Ch-w-Arb-Agilent-/172910474534?hash=item2842447526

What I'd really like would be the 33622 but that is even more pricey!
https://www.ebay.com/itm/Keysight-Used-33622A-33600A-Series-Waveform-Generator-120-MHz-2-Ch-Agilent-/173230335511?hash=item2855552617

[nfmax]

They have one for $1780 asking but would probably take closer to $1600:
https://www.ebay.com/itm/Keysight-Used-33522B-Waveform-Generator-30-MHz-2-Ch-w-Arb-Agilent-/172910474534?hash=item2842447526

That one is fully optioned up too, so definitely a good deal even at the asking price.

[precaud]

OK, the Instek MFG-2000 series appear to meet the criteria and are now back in play. Which is good because I like that they have isolated outputs, and they boast "true point-by-point arb" which I take to mean variable-clock, non-interpolated points in arb mode. It turns out that models with an "M" suffix (the Modulation option) have a parameter called "trigger delay" which sets the trigger frequency. So the one I'd be looking at is the MFG_2130M:
https://www.gwinstek.com/en-global/products/Signal_Sources/Arbitrary_Function_Generators/MFG-2000

One possible negative of this series is the lack of an output antialiasing filter. At least its not mentioned in their specs. Am asking their tech to confirm. Since I will not be operating at anywhere near the 200MHz clock freq, this may not be an issue. Your thoughts?

A rant: All of these "2nd-tier" (compared to the big boys) test equipment companies really need to learn how to write good user manuals. Half of the instrument's features are mentioned by name but not functionally explained. The Instek manual is better than many I've seen, but unless you read the Remote Programming section, you won't know what many named features actually do. Even their tech support didn't know this feature existed until I prodded them and told them where to look.

Interestingly, their tech told me that the MFG functionality "was benchmarked to the KS 33500 series".

[precaud]

The newer arbs have faster clocks and typically 14 bits. So I figgered, if I could add two bits to the DAC, double (or more) the clock rate, and double my waveform sample length, that would reduce the errors caused by the stairstepping by at least a factor of two, probably more. It would be a reasonable thing to try, as long as the cost was not excessive and other functionality was maintained.

Yes, I know... quoting yourself is dangerous, but I thought, before throwing money at this, I'd better test my figgering, and see if sample rate X2 or X4 actually does result in reduced errors caused by the stairstepping between points. The time domain waveform looks a little better on the screen. But FFT'ing each at the original sample rate and comparing them, the results are much worse. Putting the arb clock pulses in between every sampled data pair generates noise in the FFT. The attached graphs tell the story for X2 and X4. AS you can see, increasing the arb sample rate only distributes the noise more evenly over the spectrum.

So if you're FFT'ing the arb waveform (and that's about all I use an arb for), unless the waveform is filtered to remove the arb sample freq steps, "oversampling" the arb only makes things worse.

And the KS 33500B-series is the only arb mentioned thus far that does that.