开云体育

I recently tried to measure the loss in my cables for HF ham radio. I did the calibration steps, but when I do the THRU calibration, after calibrating the CH1 LOGMAG line isn't flat at 0dB. It looks like this:
What could be causing this?
Thanks.

Check your cables with an ohm meter (DMM). Your photo is fuzzy but it looks like you have a short....

Roger

I had tried with another set of cables with the same result. I measured the cables not connected to anything and on DC they measure fine. Connected to the VNA both read 50 ohms (DC)

Are you doing a RESET before Calibration. I find sometimes I have to touch RESET a few times to perform the operation.

Roger

贬别谤苍á苍,

Ensure that you do a "CAL->RESET" before you do the OPEN, SHORT, LOAD, ISOLATION, THRU calibration. Skipping that step can result in a bad calibration state even though all screen indications look normal.

- Herb

Yes, I was doing that too. I also did as Roger suggested and hit RESET several times with no luck.

One more thing to try. You can do a thru calibration without doing an OSL calibration if all you are interested in is the S21 loss/gain. I often do that if I am checking an attentuator or filter over a different range than I calibrated. Try doing a "THRU" and "DONE" and see if you get a zero reference across the screen.

-Herb

Hmm that seemed to work! I wonder why?

On Mon, Mar 16, 2020 at 06:37 PM, Hernán Freschi wrote:

Hmm that seemed to work! I wonder why?
=============================================
贬别谤苍á苍,
That shows that your cable is o.k., CH0 is putting out a signal, and CH1 is receiving it. Either something is amiss with your calibration procedure or you need to perform a "clearconfig 1234" from a terminal program to get rid of a corrupt configuration.

- Herb

clearconfig 1234 did not seem to help

To calibrate I:
Connect open to CH0, OPEN
Connect short to CH0, SHORT
Connect 50 ohms load to CH0, LOAD
Remove 50 ohms load from CH0 and connect to CH1, ISOLN
Connect CH0 to CH1, THRU

I tried both connecting the loads directly to the VNA, and through male-male cables with female-female adaptors in the ends (then remove one of the adaptors to perform THRU).

It is my understanding that after you do a Clearconfig 1234 you must turn off the power and after turning the power on do a touch calibration and then a calibration with a save to Channel 0.

Maybe Heb can confirm this sequence...

Roger

On Wed, Mar 18, 2020 at 10:32 AM, Roger Need wrote:

It is my understanding that after you do a Clearconfig 1234 you must turn off the power and after turning the power on do a touch calibration and then a calibration with a save to Channel 0.
=============================================================
If "clearconfig" didn't fix your situation then you might try to confirm the calibration is not going wrong when you use the on-screen display. Try performing a calibration using a software package such as Rune's NanoVNA-saver. A cautionary note is that NanoVNA-saver may not be able to properly interface with your unit if it is not properly calibrated.

Other questions you might consider are:

1. Did the device perform properly when you first received it?
2. If so, what firmware revision are you using and have you upgraded since then?
3. If you've upgraded, have you tried a more recent firmware from either edy555 or hugen?

- Herb

Nanovna-saver seems to show a flat line.

1. I'm not sure now. I never noticed this before, because I always used this at over 100MHz. The problem appears when the sweep starts at low frequencies.
2. I upgraded, but I'm not sure what the original firmware was. It was the one that came with the device. The PCB say NANOVNA-H 3.2 but the firmware was NANOVNA. It didn't have a D2 diode. I tried with the latest available versions of NanoVNA and NanoVNA-H. Both give the same problem
3. I tried the latest. Should I downgrade?

On Wed, Mar 18, 2020 at 08:08 PM, Hernán Freschi wrote:

Nanovna-saver seems to show a flat line.
==========================================================
If NanoVNA-saver is giving you a correct calibration then your NanoVNA is operating correctly. The difference between its guided calibration and your manual calibration is you are using the touch screen to input each standard.

1. If you still have the protective film on your NanoVNA, remove it as it may interfere with the touchscreen input.
2. Re-check and save your touchscreen calibration, then repeat your OSLT calibration (If the touchscreen cal is off your inputs may not register correctly).
3. When you perform your OSLT calibration ensure you wait a couple of sweeps for the trace to settle between selecting each standard.

If you calibrate the unit above 100 MHz and try to use it a low frequencies the results will be indeterminate because there are no calibration points in that range. If you intend to use the NanoVNA at low frequencies perform a calibration including those frequencies for best performance.

Good luck.

- Herb

I think you should not remove the 50 ohms load to CH0 when doing ISOLN.

VR2XHQ

I found the problem. My 50 ohm load was OPEN. I don't know if it can be opened to solder something inside. But tapping it on the table made it work again.

So weird.

Many 50 Ohm loads are made with two 100 Ohm smd resistors in parallel. I've seen many loads fail after intensive use, all the sudden they are 100 Ohm when connected. Unconnected 50 Ohm. We solved it by using male/female adapters, so the loads will never be disconnected. By the way, there are adapter without internal thread (male side), you just push them over a female connector. We use the loads in the final product test of RFID readers.
But for calibration of a VNA, I'd use a good quality load without an adapter. Currently I use the cal kit supplied by sdr-kits, based on Rosenberger connectors. They also sell kits based on Amphenol connectors, these loads are junk and will become 100R soon. Never had a problem with the Rosenberger load. Also the load from minicircuits is excellent (ANNEQ-50X+)


Op 21-3-2020 om 05:46 schreef Hernán Freschi: