If you do go for the SDG1025 keep in mind that there is horrible jitter on the square wave output.
Here some images about normal SDG1000 series generator. (in this example SDG1010)
Two images about Square wave. Old times there was high amount of glitches just before rising edge (ans also same before falling edge. (just there in corner) Glitches level was up to 50% of signal level and also theey start far before rising edge.
Corner zoomed 100x (note signal level)
Corner zoomed 1000x
Then about Square wave cycle time jitter.
It is well inside specifications (spec: typical 0.1% at 1kHz) Here max peak time error is around 0.05%. It need also note that many manufacturers do not tell anything about peak values. Mostly they tell RMS jitter. It also depends what is distribution of samples for calculation. I can not even imagine what is here true rms value of jitter. Horizontal scale is zoomed 1000X related to delay time (cycle time)
This is well inside specifications. I do not know what is horrible (what is definition for horrible?). Some times 1ns jitter is horrible and some times 1ms jitter is still ok. What other this price class function generator give better? Then, if use pulse mode for square wave (why not use) for example 1Hz frequency. Cycle time is 1s. Working principle jitter is 8ns and then some small jitter from reference clock etc. If we think it is total 10ns. 1us is 1ppm in this case. 10ns is 0.01ppm.
With Square function is is perhaps 0.1%. 1s cycle time it means 1ms (1000ppm)
1MHz cycle time is 1us. 0.1% is then 1ns! But with 1MHz Pulse mode still have 8ns jitter.
Here Pulse mode risetime. Well under 7ns. Also overshoot is well inside sppecifications.
Pulse mode do not have at all these rising start and falling start edge noise. Pulses are generated totally different working principle than Square. Only Square is derived from internal sinewave using classic comparator method what have some advantages but also disadvantages.
Using Pulse mode instead of Squere give less cycle to cycle time jitter with low frequencies. In some frequency point when going higher freq it turns opposite and Square give less time jitter. This is natural and normal when look circuits working principles.
Square wave is derived from internal sinewave using PECL comparators. Pulse is done different, directly using DAC's. For adjust frequency (inside some time interval) correct it need 8ns "hopping" in cycle to cycle and in pulse width.) Square wave derived from internal sinewave do not need this, but with low frequencies comparator method make jitter due to fact that internal sinewave is not ideal pure with zero noise and also comparator is not ideal component.
Here one example from very old test about pulse mode 8ns hopping.
Picture 11.
Pulse mode. Frequency 117.72Hz pulses are around over 8us wide.
Trigger position one pulse before centerline positioned pulse.
Center line zoomed (small picture inside) and zoomed more up to 2ns/div
Now here we can see that pulse mode jitter is bistable. (it do 8ns jumping forward and back for adjust freq. This is nessessary in this kind of "simple circuit working principle" and 125MSa/s)
Both channels jitter is 8ns (8ns jump in place or not jump. Both channels do this)