HOME / KC901 serial VNA

Temperature drift in KC901V

By Andy •  2016-11-01 00:31:17 •  1748 clicks
I've been testing the new KC901V for a couple of weeks now, and find that in most respects this model is a significant upgrade to the KC901S, which I also have. Kudos to the engineers who designed it.

One problem that I noticed though is that the unit needs 10-15 minutes of warm-up time to generate accurate measurements. Otherwise S11 values for a calibrated open or short can be off by 0.5 - 1.0 dB across the entire frequency range. This was not the case with KC901S.

Please see the image below. The unit was SOL calibrated and shut down for more than an hour. On power up, I attach the same short that I used for calibration, expecting an S11 reading of 0 dB. Instead I get the curve shown in blue (by pressing the hold button). 10-15 minutes later the curve shifts to the yellow line, as expected.

I performed the same test with 901S and only see a 0.05 - 0.1 dB shift there. The large drift can of course be a problem if the unit is designed to be operated in the field, where waiting for 10 minutes for the unit to warm up may be impractical.

Is this a known issue for the beta version of 901V? Is there a possibility of correcting this issue with a firmware update?

Thank you.

frequency  designed  generate  testing  minutes 

10 Reply | until2018-10-21 07:50Add a reply
    EA1DDO  2017-01-08#1 -Reply

    Hi gents,
    Here it the result of my test.

    Regards, Maximo

    VK3NH  2017-01-06#2 -Reply


    As promised here are my results for S11 with the cal short fitted. Thanks for doing the S21 test comparison for me. It is quite a hot day today around 35 degrees which is hotter than average outside when I started my first test. I have previously performed SOL at a lower temperature about 3 weeks ago but I did not change anything.

    I followed your instructions except I used my camera to take screen pictures as its easier to mail and upload. I did save the displays as well so I can send them.

    The KC901V on power up said 31.5 deg C. The first trace was after turn on and temp check and the second 15min later.

    This picture shows the KC901 settings and temperature at the end of the test.

    I then turned the KC901 off and put it in my freezer to bring the temperature down to where you tested. When I took it out it started at 0.5 deg C but after 2 minutes it was up to 4 deg C as shown when i did my first sweep. I expect the amplitude swings are because my original SOL cal was done at an ambient temp around 23 deg C.

    So there you have it. My unit does not drift in the manner yours does but temperature still affects the smoothness of the cal sweep.

    I have included a new item on the spectrum analyser 30kHz RBW symetry problem.

    regards Noel

    BG8NPK  2017-01-06#3 -Reply

    We will test the unit before send back to you to make sure it will work well. If your instrument issue caused by the filter we will fix it and do a full factory calibration then send back to you,if it has other problems we will send a new one to return.

    Best regards,

    Andy  2017-01-06#4 -Reply

    Thank you for your replies, VK3NH and BG8NPK. I will email Anna later this week to arrange for a replacement.

    I powered on the device this morning at 5C, and saw the error double to 1.5 dB. It appears that there is nearly 1 dB of error for every 10C of temperature change.

    If there is temperature calibration performed at the factory, could it be that it is missing or corrupt on my unit? Could the faulty pre-amplifier filter really account for 1 dB drift over 10C?

    BG8NPK  2017-01-05#5 -Reply


    My colleague have contacted you many weeks ago to send our address information that you can return the KC901V, but it seems not arrive your mail.

    Our team trying to repeat this drift to know how this happen,until yesterday our engineer find a filter before a rf amplifier may get a drift when KC901V warm up. it cause few KC901V will drift the measurement result.

    VK3NH  2017-01-05#6 -Reply

    I am away from home today but I can do that check for you tomorrow night.

    Like you I am very impressed with the KC901Vs capabilities and I love what it does well but I would like to improve what it does not do so well.

    can you possibly check what I have asked Maximo to check in my message to him under I can't log in under VK3NH regarding is there a step in the S21 cal and is the 30kHZ bandwidth centred and symetrical
    Regards Noel

    Andy  2017-01-05#7 -Reply

    Deepace kindly agreed to replace / repair my unit, but before I spend $70 to ship it back and risk losing the device in transit, could someone please try the experiment below and report the results? I am curious to see how other 901Vs perform in this situation.

    1) Power off 901V for several hours (e.g. overnight). Leave it unplugged from the charger, as charging generates heat.
    2) Turn on the unit, go to S11 mode, choose ReturnLoss format, and select 1db/div scale.
    3) Set center frequency to 3500 MHz and span to 7000 (to see the entire range)
    3) Press Run and then HOLD to capture the signal
    4) Wait 15 minutes, then save the screen image to a flash card (Shift - 1).

    In my case, the yellow line (signal after 15 minutes) is significantly different from the original signal (blue line). The difference can be 0.75 dB or more. The temperature change from startup to 15 minutes later is only about 10-15 C (as reported by FUNC -> About). As 0.75 dB is around 20%, I am surprised that I see so much variation.

    It would be very helpful to see how much signal shift there is in other 901V units. Thanks!

    Andy  2016-11-01#8 -Reply

    Serial is 2016322154. Software version 1.0.0. Hardware version MB-V1.2 RB-V1.04.

    I did reset system calibration (via Func Fn 7). Perhaps that invalidates temperature calibration somehow?

    BG8NPK  2016-11-01#9 -Reply

    The temperature will efficient the measurement result, but we do many calibration to fix this,it should not be like this 1db error.

    please give us your KC901V No. and the software version.

    EA1DDO  2016-11-01#10 -Reply

    Good one !!
    Have you got any information about those comparatives KC901S/KC901V ?


LoginPublish or Still no account?Register


Theme List

KC901 serial VNA

Relevant topic