I can add a few comments regarding the performance of this generators, some positives and some bugs (big ones for me
). My background for the last several years is as a power electronics engineer, first for a couple big companies, and now a small one for the last year or so. For the latter, I have been getting by with an old HP6012, but having pulse width change over time and from day to day has been a big problem, and sometimes I need signals for modulation, to test startup, transient behavior, etc., but a high-end ARB/pulse gen has been out of reach. The big problem with most of the budget ones has been pulse jitter that basic makes them unusable, along with limited duty cycle range. I'm happy to say that the SDG2042X meets these needs for me and I have been happy with that.
A very nice feature for me is the dual output, with the ability to delay or phase shift one pulse with respect to another. Many power converters have two switches that must be operated roughly out of phase with each other, but often with some small but important deadtime where both are off, or overlap time where both are on. It looks like the 2042X can do this, but...
There is are some show-stopper bugs for me in the synchronization of the two outputs during enabling of outputs, and during changing of pulse widths, and perhaps more. I did not dig deeper, since I get paid for other things, and I need to get paid.
Bug 1: With a pulse output on both channels, and channel 2 delayed with respect to channel 1, if you enable channel 1, then channel 2, the waveform initially looks good. However, a fast scope shows that when channel 2 is enabled, there is actually a small glitch where there is no delay. If these are controlling two switches in a half-bridge power converter, during the glitch you are shorting the power bus with your transistors. Not good. Hopefully this shows up in the first attachment, where I have the scope on infinite persistence. If you look closely, you can see the bright traces, and the faint trace. The bright trace shows a double trace in channel 2, and this is not jitter. This only happens as you enable channel 2. If you look more closely, you can see a faint green trace where channel 2 is not delayed. This only happens for a brief period after enabling channel 2.
Bug 2: If you adjust the delay setting on the fly, the same thing happens as with Bug 1. I spent some more time with triggering to try and capture this, in the second attachment. I start with channel 2 delayed from channel 1 by 10ns, and I have the scope set to trigger if the delay becomes longer than 15 ns. The top trace shows several ms of capture, and the bottom trace shows a zoomed in view prior to the trigger. When the delay is set to 10ns, the scope is not triggering, and then I change the delay to 20 ns, which causes the scope to trigger.
If you look at the zoomed-in waveform, before about -2.5454 ms, the delay is about 10 ns, as set. After this, the delay goes to zero, and stays that way for over 2 ms, with one dropout along the way. Finally, at t=0, the delay hits 20 ns as desired.
This basically means that the dual pulse output is of very limited use for me. 2 ms of the wrong setting will easily fry a couple of small power transistors (GaN power FETs in my case) when they are shorting a power bus. I really hope this can be fixed in a firmware update, and I've seen that Siglent hangs around here, so I hope they are looking.
I'm still keeping the generator. It meets or exceeds all my other needs. I just really wanted to be able to use the two channels for driving a power half-bridge without needing a bunch of other stuff to synchronize every thing. I will note that the Rigol 4162 that I got at my old job also could not do this properly, and it had so much jitter as to be much less useful.
John