On a recent activation on Mt Torbreck as part of JMFD, I attempted to create a stub to tune a 70cm colinear where the antenna had been built physically too short. Tests at home showed that the feedpoint impedance was on a point on a Smith Chart where a shunt could bring it to around 50 ohms resistance.
It proved not to be. I’ve done a number of matches before, and all the theory in the world can’t be wrong. What could be wrong is the data obtained.
Overlap between the base MiniVNAPRO and Extender
The Mini radio solutions MiniVNAPRO is a vector network analyser device that operates from 100kHz up to 200MHz. It is a two port device that works in both transmission and reflection mode. Measuring antennas only makes use of the reflection mode. Transmission mode is useful for baluns and anything that does any kind of transformation.
The Extender is a device that like a transverter for the MiniVNAPRO. It effectively extends the operating frequency up to 1.5GHz. It can operate both in reflection and transmission mode, so like the base MiniVNAPRO, it is possible to both analyse antennas in reflection mode (s11 data) and transfer devices to obtain the full s11, s21, s12 and s22 data sets.
The Extender can actually report from about 50MHz up to 1.5GHz, so there is an overlap where both the base device, plus the extender can both report from 50MHz to 200MHz. This is quite useful because it can be used to check the performance of the extender. The data that is obtained should be similar to what is obtained from the base MiniVNAPRO. If the data obtained in this 50MHz to 200MHz range agrees between both devices, then data from the extender at higher ranges has at least more of a chance of being close to the mark.
Checking results on the 2m colinear
I have discussed the construction of this device before. I used the MiniVNAPRO and obtained s11 data which was then used with a smith chart tool called Linsmith to calcuate a required stub to match the antenna at 144.1 MHz. This was deployed my recent field day trip and worked well. This tells me that the data I am obtaining from the base MiniVNAPRO is good and useful for designing matching solutions.
While up at Mt Torbreck, I ran tests on the colinear which showed quite a substantial difference between the results that the base unit and the extender obtained. I did a quick test with a 50ohm standard and both the base unit plus the extender both agreed that there was a 50 ohm load. We’ll that’s fine but I need to have a good idea of what an unmatched impedance is so I can do something about it. I’m not going to need to do something that already is 50 ohms!
It would need to wait until I was back home to look at this issue further.
Environmental sensitivities in calibration
After sleeping on it for a day, I thought this problem could arise for several reasons:
1) Could be a fault with the Extender itself. Some on the net have a poor opinion about this device, but we have to rule out everything else before going here.
2) A software problem with BlueVNA. If this was true, then I could use VNA/j on the laptop and this would give substantially different results if this were true.
3) A problem with calibration in some way. It is interesting that the 50 ohm test worked ok, but non matched impedances showed substantial variances – this made me think that perhaps if something in proximity to the device when doing a calibration could impact on the open or short circuit results that could have impact practical results.
I did some tests using VNA/j and that showed some significant differences on a 2m antenna between the base MiniVNAPRO and the extender. My old calibrations were done with the unit lying on the table, but when I use it in the field, it is hanging off some antenna in free space. Perhaps I should try to do calibrations again holding the unit up in the air.
After doing this, I noticed that there was a shift in the results obtained. Calibrating the MiniVNAPRO with extender lying on the table does not seem a good idea. My table at home is wood, but has metal bracing underneath the surface. Still after holding the unit in the air did not bring agreed results, but one set was within 10%. Good, but could be better.
In the end, I grabbed on of those in car phone holders that have a suction cup to go on a car windscreen. I put one of these on my desk and used it to hold the device during calibrations. I also found that there is benefit doing an average calibation – 3 to 5 passes. Most of the benefit is obtained with 3. There is little point to a greater number of passes such as 10.
2m Jpole results using VNA/j with USB interface
Here is a smith chart plot of results obtained on a 2m jpole that is a little long – both in the results obtained here, and in real life.
We can see that the results are quite close. They aren’t exactly the same but they are much closer than the results I was previously obtaining. One thing I could observe is that the better dynamic range of the base unit does show through. Also, I used 20,000 calibration steps for both calibrations. On the extender, the steps are nearly 8 times further apart than for the base unit.
Still these results are close enough that these extender results are useful.
2m Jpole results using BlueVNA with Bluetooth interface
With the VNA/j results giving me greater confidence, I then moved over to look at BlueVNA. BlueVNA does not allow averaging of calibration data. I still tried doing the calibration in the phone holder. Here is what the Jpole looked like in BlueVNA.
Firstly, the results broadly match, both with the base unit and with the extender. The basic line on all four charts is the same. One thing that I have noticed is that the “noise” in the data is higher with these results. In terms of quality I would rank them from best to worse as follows:
1) MiniVNAPRO with USB
2) MiniVNAPRO with Bluetooth
3) MINIVNAPRO Extender with USB
4) MINIVNAPRO Extender with Bluetooth.
Having said that, all four results are quite usable, so I think I have got on top of this problem for now. Obviously I cannot compare the base unit at 70cm as only the Extender works there, but perhaps I could compare USB mode against Bluetooth. We’ll see how that goes with the 70cm colinear that I’ll still see if I can match it at 439 MHz.
73 and regards, Wayne VK3WAM