Discussions/Suggestions on some issues for KC901V Network Analyzer

[Mechatrommer]

i just got this (Deepace/Kechuang/Measall KC901V Network Analyzer) recently and still learning. its good item i think mainly for its advertised purpose, and shariar's review has persuaded me to get this unit i cant get out the temptation out of my head for months ... i've been eyeing 2nd hand VNA in ebay for a lot longer. for the same price, the used working unit mostly only goes 3GHz, calibration status is uncertain. if it goes beyond 3GHz at the price, then it must be either non working sold as is, or we are going to need additional part/unit to make it work as a whole. so i think the KC901V must be it, my benchmark is no lesser than 5.8GHz because i have 5.8GHz antenna and wireless video system to tune and optimize for... and i had experience of mysterious 2.4GHz LNA and transceiver damage on my rc remote controller, so without this kind of thing, i'm just as good as a blind man in the dark... well enough story...

but i have some issues (mainly negative) i want to highlight/discuss/suggest. whichever i want to suggest and hope for firmware improvement, i will try to get it through their bbs forum. but first i will post it here for the purpose of learning and discussion in case i did something wrong...

first issue: impedance not showing the right thing at very low frequency (DC - 100KHz or 500KHz), my setup is as in the first picture.
2nd picture for open load, it should show Inf ohm from the beginning.
3rd picture for short load, it should show 0 ohm from the beginning.
4th picture for 50 ohm load, ok this one they got it right  .
5th picture for 100 ohm load., it should show 100 ohm from the beginning.

so you see, everything will start from 50 ohm reading no matter what resistance i put at termination, the only right thing they do it if we put 50 ohm terminating resistor, even that it shows a mysterious bump at the beginning. am i missing something? is there any good reason for this? for practical purpose it can be confusing, why cant they do this right? if they cant do this right, my suggestion is that either in firmware:

1) cheat in firmware to show flat horizontal line on the screen from say 100KHz or 500KHz (where the line showing false impedance) down to DC.
2) or better yet, just dont show (blank line) in the region where they cant do it right.

if this is beyond firmware, ie hardware limitation, is it too difficult to do impedance test at DC (or near DC)? in this $2K device? it should be simple isnt it? just a resistive divider and passing Vdc at the test port, and then sense the divided voltage, no? any better suggestion/insight?

[hendorog]

Perhaps the internal bridge has poor directivity at the extreme end of the frequency range?

I see the manual recommends using an external bridge for better accuracy.