Matching the 2m Colinear for FM

Hi all,

I have described my efforts in building a colinear for 2m over several posts. The base unmatched colinear was sort-of ok at 146.5, but not very good at 144.1, soI set about performing a match for 144.1 using data obtained using BlueVNA and the mrs MiniVNAPRO vector network analyser. This has resulted in a fairly good match, with an SWR around 1.7 and the FT-817 showing only one SWR bar. I can live with that.

Considering a match for FM

Back at 146.5, the FT-817 shows 3 bars. This is not so good – its not a disaster match, but it is not too special either. My earlier analysis suggested that I could put a stub out a few cm further from the feedpoint, but this did not work. I think the reason was that the data I was getting from the MiniVNAPRO was not that good. Recently, I sought to rectify that, as discussed in this post, which the discussion relates not only to the extender but the base MiniVNAPRO as well.

With that in mind, I set up the 2m colinear and did some tests. Here is a screenshot of Zplots showing the data I obtained:

Zplots showing resistance and reactance of the 2m colinear at the BNC connector

Impedance at the BNC connector

Just for piece of mind, I ran the analysis with the extender, and this is what it shows:

Impedance at BNC connector obtained using extender, which shows very similar results

Impedance at BNC connector obtained using extender

The results are quite close, and any analysis done with one data set will pretty much be matched from the other.

The T piece for the 144.1 match

As well as the BNC connector for the radio (or a cable going there), I also have another convenient point to consider matching – the same place I have the stub for 144.1 I have put a BNC t piece here to connect the stub (and disconnect it when I don’t want it). I know I can’t use another stub for 146.5 right at this point on the transmission line, but I could use some surface mount components for a shunt, or add an extra piece of coax and put a shunt off that. We’ll first look at the data from this point:

Impedance at the T piece

Impedance at the T piece

The interesting thing about this is that this is an effective transmission line transformation. If I use zplots to consider the effect of 36cm of RG58 on the impedance, I get the very similar results, and there is about 34cm of RG58, plus the BNC connectors in the middle. The results are not exactly the same – the empirical results are a little flatter because of loss in the cable.

The impedance looks quite different, but if you consider the plot on the smith chart over on the right, it is simply rotated nearly halfway around the centre of the chart. This is the effect of a 50 ohm coax transmission line.

Now I can’t use a stub right here, but lets see what linsmith suggests I can do:

Match design in linsmitch showing the untransformed impedance then two steps of transformation

Match design in linsmitch

First up is to use 15cm of RG58 to get to a point where a series stub could match. The untransformed impedance is the pink dots – this is taken from the Zplots data. If you look carefully at the smith chart from “Impedance at the T piece”, you’ll notice a subset of that curve is shown here. The 15cm of coax transforms this to the green dots. Putting the series stub with a short 18cm away then transforms it to the red dots.

Series stubs are more difficult to create than parallel stubs. The inner conductor of the coax must be broken, but the outer conductor allowed to continue. The stub inner conductor must be connected (eg via solder) to one side of the line inner conductor and the shield to the other. This is logistically quite difficult.

I also provided about 5cm of extra coax with another BNC connector on the other end. This will allow me to easily put this segment into the colinear when I want to match to 146.5, or to take it out. When I am operating on 144.1, I do not want this segment with this stub.

Here is the data with the match in place:

2m Colinear matched in the FM band segment

2m Colinear matched in the FM band segment

I didn’t quite get the match quite right, the best match is around 147.3, however I can live with this. A FT-817 shows no SWR from 146.2 all through to 148.

Now why did it not quite match? I think it’s because I might have stubbed about 10mm too short. Such is life. Here’s a linsmith look at this assumption:

linsmith analysis of the match result

linsmith analysis of the match result

Note: I’ve added an extra bit of transmission line to account for the measuring point of the VNA – it’s not right at the stub point. This means the red dots are replaced with a second lot of green dots, with the new red dots after the extra coax transformation – note that these calculated red dots fairly match the empirical data shown in the zplots smitch chart in “2m Colinear matched in the FM band segment” above.

This analysis shows that I might actually be able to take off about 5mm from the stub and I’ll further improve the match. I’ll not do that straight way – I’ve got a big trip to the Australian Capital Territory coming up and the match is already fairly good – good enough that further improvement might have only marginal effect.

Anyway it’s pleasing to have a good match now at 146.5, and I’ll be using this a lot in the next week.

73, best regards, Wayne VK3WAM

MiniVNAPro Extender with Bluetooth

Hi all,

After my earlier post which considered use of the MiniVNAPro Extender using a USB connection, I wanted to give the Extender a go with the BlueVNA program.

Using BlueVNA with the extender

When running BlueVNA, a picture of the device will show. This could be a a MiniVNA, a MiniVNAPro or a MiniVNAPro with the extender. I swiped left, which went through these three devices. As the pictures suggest, use of the extender requires the MiniVNAPro with extender to be shown on the screen.

As I discussed in my earlier post, I have found that I need to use a specific startup sequence to get useful data from the extender. When using USB, the USB cable needs to be connected before connecting the CAT5 cable between the MiniVNAPro and the extender. In the case of bluetooth, the CAT5 cable should not be connected when turning the MiniVNAPro on. I wait for the Bluetooth blue light to come on on the MiniVNAPro, and then connect the CAT5 cable. The extender’s sole orange light then comes on.

After that, the calibration sequence is as per usual. After calibrating with a open, short and 50ohm standard, I ran a test on a Diamond RH771 2/70 dual band whip antenna. Side note: This thing is not crash hot on 2m, it is better on 70cm.

Here’s a pic of me holding the MiniVNAPro connected to the extender with the Diamond on the RX port:

Diamond RH771 on the RX of an extender connected to a MiniVNAPro

Diamond RH771 on a MiniVNAPro with Extender

Here’s a look at a screenshot:

Diamond RH771 data at 70cm using BlueVNA

Diamond RH771 data at 70cm

With a return loss of over 16dB at 439MHz, this antenna will work well at the 70cm FM calling frequency. It is still respectable at 432MHz.

An interesting thing about this data is the presence of a few artifacts in the data. They are the “square wave” type formations that can be seen. These arise from the calibration process. BlueVNA does not allow the number of steps to be set during calibration, unlike VNA/J. With VNA/J, I set the calibration steps to the max value, which is 25000, rather than the default 2000 steps. It makes the calibration process take longer, but then actual use of the device is not any slower. What it does do is reduce the size of any “square wave” artifacts.

Suspect Bluetooth interference

If I calibrate in VNA/J using a lower, say 500, calibration steps, using USB, I also notice the presence of artefacts, but they are much less than what we see here. We are seeing the Return Loss being moved up to a full 1dB in these results shown above. This then flows through to everything else, SWR, resistance, reactance, because all of these things are calculated by the software using the return loss and phase (I have the phase switched off on the screen shot).

When using the MiniVNAPro with the extender in USB mode, things are getting a little “hairy” around 1.3GHz and above, I don’t think the dynamic range of 50dB is true at 1.5GHz. It’s still not bad, and certainly still useful in doing antenna analysis for 23cm.

I’m not sure I could say that about Bluetooth mode. I’ll do some 23cm analysis in the future, but here at 70cm, we are seeing these 1dB artefacts that are affecting the results. We can still look through them, and we can still draw conclusions about the antenna.

Perhaps there may be some value in BlueVNA allowing several calibrations and averaging the results. Same goes for the scans. Doing an average scan would slow things down, but it may help to mitigate some of these effects. Perhaps there also might be ways to better shield the Extender from Bluetooth activity.

Even with all of this, it is still better to have the flexibility of the bluetooth mode. Also it looks like the artefacts have a predictable behaviour (either adding or subtracting from the return loss an amount up to 0.5dB). It might be possible to correct for this in software.

Regards, Wayne VK3WAM

ADDENDUM: I used the extender to do some testing on a 70cm colinear I have been building. I performed a calibration outside, and the results I obtained did not show the effects that are seen above. So, there is another noise source on my desk at home that affects the extender/MiniVNAPro combination in bluetooth mode, rather than it being bluetooth itself.

Constructing a 2m colinear Part 2

Hi all,

This is a continuation of Constructing a 2m colinear Part 1.

Repairing the Colinear

On a recent SOTA activation trip, I tested the colinear in the field. As discussed in that post, the lower SWR point seemed a little high. At the second and third activations on that trip, I was unable to use the colinear due to a persistent short. This short turned out to be in the BNC connector.

After coming back from the trip, I set about repairing the antenna and upgrading the connections to make them more robust. This was done using some spare foam dialectic from some RG6 cable I had lying around. I had previously used the braid from this for another project. This dialectic could now make itself useful as a physical shunt to give the joins of coax some structural integrity. I simply inserted the dialectic into the join area and secured it with tape. Seems to work well, and the colinear does not feel anywhere near as fragile.

Analysing the Colinear

Once getting the colinear fixed, I put it up at home in the backyard to have a look at its frequency characteristics. I used a MiniVNA Pro with the BlueVNA app running on a Samsung Galaxy S2, using bluetooth connectivity. Here is a screenshot:

Colinear screenshot from BlueVNA showing SWR, impedance, resistance and reactance

Colinear screenshot from BlueVNA

Several things to note from this picture is the SWR is unacceptably high. The FT-817 reports a lower SWR as some is being burnt up by warming the coax. Also, the resonant point is at 147MHz, when I was expecting 145. Before we get too excited, one thing to remember is this is a picture of what is at the BNC connector, not what is actually at the “feedpoint” which is through the common mode choke, up to the first coax connection, about 910mm of coax away.

Here is the same data using Zplots, which is an excel spreadsheet. The data from BlueVNA was saved to a s parameter file, which can be loaded into Zplots.

A Zplots render of the Colinear at the BNC connector, showing a graph and a smith chart.

Colinear at the BNC connector in Zplots

This zplots chart shows much the same information as the screenshot from BlueVNA, which is no surprise given it is based on the same data.

Zplots has the capability of compensating for the effects of transmission lines. I could then algorithmically have a look at what is going on at the actual feedpoint, which would be very difficult to physically access. Here is a screenshot from Zplots at the “feedpoint”.

A Zplot, consisting of a graph and Smith chart with compensation for the transmission line.

Colinear characteristics at feedpoint

This graph shows the antenna is resonant at 144.7MHz, which is not far from what the antenna was designed for: 145MHz. The gotcha is that the resistance at the point of resonance is about 14 ohms. No wonder there is bad SWR on this thing. This is a prime example of how a resonant antenna does not necessarily have low SWR.


What to do about about this antenna that is resonant where planned, but effectively gives me an SWR of 3 to 4 over the entire 2m band? I am going to have to do some impedance transformation to match to this antenna. What to do?

There is another Smith chart shown below. The purple dots represent the impedance at various points from about 144MHz up to about 146.5MHz. These points are effectively the same as the “feedpoint” data calculated by Zplots, with 910mm of 52ohm coax (RG58/U) being added to the data obtained at the BNC connector. Along the coax from the connector to the feedpoint, they essentially rotate on the Smith chart around the centre point. The centre point is, by the way, a 50ohm 1:1 SWR point. This is where we want to be, or at least within the red circle on the Smith chart.

I have more choices than right at the BNC connector or at the “feedpoint”. The first 500mm from the feepoint is a bit hard, as it is inside the aluminium sleeve of the colinear. This bit of coax is inaccessible. About a 290mm segment is also not accessible because it is in the common mode choke. There is a section about 350mm (360mm from where MiniVNA would have measured from) from the BNC connector that looks nice. This is 545mm from the “feedpoint”. This is represented by the smaller green dots. From here, I can use a parallel shunt bit of coax. If this bit of shorted shunt coax is 160mm long, then the impedance is transformed into the “matched” zone.

Impedance matching the colinear antenna using a Smith chart tool. The Smith chart shows a transmission line and then a shorted coax parallel shunt

Transformation of feedpoint impedance

One draw back is the cable will have a DC short because of the parallel shorted coax shunt, but this is a “nice” short.

73 de Wayne VK3WAM