Very simple and robust method for check and measure this level of oscilloscopes channels time difference.
For avoid all hassle and suspects use ONLY one signal. Take this same signal with Time matched cables to scope channels under test. You do also not need play with different frequencies etc or Phase angles. Just measure time, repeat, measure time.
Simple method is Y cable. Yes, just cables centers connected and shields connected. Do not think impedance matching here, this is simple split. Now it is also easy to fine adjust both output sides perfect time if want.
Third cable in this Y, input. This length do not matter, 1cm to 1 meter is just ok.
Connect it to gen out and then these Y cable time mached sides to scope channels under test.
Use what ever signal what give fast edge as possible. It can be pulse, square, triangle0% or even sine but twith sine you need high frequency. Using just fast edge waveforms you can use 1kHz or 1MHz, no matter... but keep this base freq so that period is least well over 2 times max time difference in channels under test... so even first time noob can not go to wrong.
Also I have seen many og SDS1000X-E scopes, 2 and 4 channels. I have never ever seen 10 or 30ns time skew between any channel. Also I ask how it can be FW problem if there 1000 scope without this problem and 1 scope with this problems. Just wake up and start thinking...
Also it is amazing what kind of hassle this looks when this is tested in this thread... how it can be this kind of total hassle.
Even if can not have/diy this kind of Y cable... I have told to OP how he can make very very easy and robust test just with all what he have... but no... just continuing this hassle like walking in circle where nothing is clear at once.
Here attached one image.
There is also simpliest possible DIY Y cable.
Done so that cable time skew is roughly 2.5ps so I just multiply it with 2 so that I can always trust it is <5ps when BNC are connected perfectly. I have hand adjusted lengths using scope when connect this Y center.
Note that RG316/U travel time 4.856ps/mm (ref. Belden) (yes I did many iterations for get as perfect match what I can easy detect with my tools. Note that 1cm difference mean still under 50ps error, so you can do this kind of cable just measuring length)
With this kind of simple split it is very important that cable outputs load impedance/reactance are as equal as possible. Still naturally it is not at all impedance matched signal transmission. But, this do not matter because only important is Time difference and just roughly.
No need hassle with 100MHz or 1MHz or 15MHz. Just what ever from 1s to 1us period edge is enough and measure channels time skev (including amount of Y cable time error). Do not play with probes or other things what can way or other suspect. Just reduce all unknowns to minimum.
If scope have roughly 29ns enormous time skew between channel pairs (different ADC), as previously told, no need anything but return it as DOA and get it changed or repaired. Have you contacted Siglent technical support.
This kind of thing need just one or two very simple and easy measure and thats all. Result is failed or ok. Until start playing "this and that".
-siglent-sds1000x-e-series-oscilloscopes/?action=dlattach;attach=1234186;image)