Measuring the TinyFilter with a NanoVNA
I recently had the wonderful opportunity to work with Ara, N6ARA, on the TinyFilter - a 2m bandpass filter for hand-held VHF amateur radios. The filter is offered both as an assembled unit and as a kit. For both versions, many people have been wanting to measure the tuning and performance with a nanoVNA, and we've been getting a few questions on how to do this properly. 
For my testing, I set the start frequency to 100 MHz and the stop frequency to 200 MHz. You can use whatever you like, but it is important that the calibration steps that are described below are done for exactly the same range that you would like to measure the filter with. In the photo below, I'm entering first the start frequency, ...
One additional thing that is worth a quick mention: I have set up my display so that it is showing 10 dB per division on the y-axis. I find that is a nice amount for measuring the TinyFilter, but you can use whatever you want. You can change this by using the "SCALE" submenu under the "DISPLAY" menu. You can also change where 0dB is on the y-axis - that is in the "REF POS" submenu.
One area of difficulty that people have had is that measuring a filter is different than measuring the SWR of an antenna. I suspect many hams use their nanoVNA for antenna tuning, which requires a less extensive calibration procedure than measuring a filter. In this note, I'll walk through the steps of calibrating the nanoVNA and measuring a TinyFilter, with lots of pictures.
The equipment needed is shown in the photo below. There is the nanoVNA, of course (shown on the right). The picture shows the filters to be measured, both a BNC version with its orange label and an SMA version with a clear heat shrinking covering a green printed circuit board. Just to the left of the nanoVNA are two short cables with SMA connectors. I'm using two coax cables for this explanation. You can get away with one for the SMA version, and you can probably make it work for the BNC version as well, but it is starting to stress the cable, so I'll use two. Then there are the calibration standards on the far left: an open, a short, a load, and a through (this last is directly below the orange filter). These are all SMA pieces that came with the nanoVNA. Below left are two SMA-BNC adapters that are useful for measuring the BNC version of the TinyFilter.
The nanoVNA that I have is a relatively old and beaten up nanoVNA-H4, so that's what I'll show. Those with newer or nicer VNAs may have a slightly different appearance or menu items. Sorry for the smudgy screen! In retrospect, I should have cleaned it up before taking pictures.
Now that we have all the pieces, it's time to do the calibration. To begin with, I set the frequency. I use a range centered around the middle of the filter and encompassing areas where I want to see the performance. You can set the frequence in the menu under "stimulus".
... and then the stop frequency.
Next, I like to turn off all the traces except two. I find that if I have all 4 traces on at the same time, the display is a little too busy for me. To turn off and on various traces, go to the "DISPLAY" menu, and look under "TRACE".
Make sure that at least TRACE 1 is on. When the traces are on, their corresponding number will be highlighted.
I also like to set TRACE 0 to measure the reflection (S11) and TRACE 1 to measure the through signal (S21). S11 means the reflection coefficient - the nanoVNA sends out a signal from the port marked "CH0" and measures the amplitude and phase of the signal being reflected straight back at it. S21 means the transmission coefficient - the nanoVNA sends out the same signal from the port marked "CH0" and measures how much makes it through to the port marked "CH1". This is also called the "through" signal.
Double check that that traces are measuring what they are supposed to! In the picture immediately above, if you look closely, you can see that the blue trace (TRACE 1) is measuring CH0 (look at the text at the top of the display), which is the reflected signal - that is not what we want. We need to set it to measure the through signal, not the reflected signal. To do this, go to the CHANNEL menu,
... and select "CH1 THROUGH". If it is all good you should see the blue trace with CH1 in the text at the top of the display.
One thing to be aware of is that when you use the "CHANNEL" menu to alter the channel of a trace, it works on the current active trace. You can tell which is the current active one by the inverse-highlighted text at the top of the screen. It is generally whichever trace you have most recently used the "TRACE" menu on.
In the same way, you can set the yellow trace (TRACE 0) to "CH0 REFLECT". You may have to use the "TRACE" menu to select which trace is active before you try to change it. It may take a little back and forth in the menus.
For the formatting of the traces, I use "logmag" - that just means it is the logarithm of the magnitude of the signal - it is just the amplitude expressed in dB. To set the format for the traces, use the "Display" item on the menu. Click "FORMAT", ...
... then "LOGMAG". You can do this for both traces. You will have to alter the currently active trace to change the format for both of them.
Ok, well that was a lot of digging around in the menus. You may have already set up your nanoVNA to measure things that way, but if not, the steps above should be enough to guide you.
Now it is time to get to the heart of things, the calibration. For the first step, we are going to reset the device to clear any existing calibration. To do that, go into the "CAL" menu item:
... and now tap "RESET".
then, tap "CALIBRATE".
You should get a screen that looks something like that below:
Make sure all of the connectors are snug. No need to use a wrench, just make sure they are finger tight.
After waiting a second or two, the text on the "OPEN" menu will be inverse highlighted and the next item will be highlighted, telling us it is time to do the "SHORT" cal.
After connection and making sure it is snug, tap the "SHORT" menu item to let the nanoVNA to know that it needs to record this connection.
After you do that, the "SHORT" will be reverse highlighted, and it is time to move on to the "LOAD" connection.
Remove the short cal standard and connect the load standard. It goes on that same coax.
Now tap the "LOAD" menu item.
Now it is time to do the "ISOLN" and "THRU". If you have only done antenna reflection measurements before, this may be new to you. My nanoVNA is old, so I need to do the "ISOLN" cal manually. Some newer and most fancier VNAs do this automatically, but mine does not. "ISOLN" means isolation, and it tells the nanoVNA there should be no signal going from CH0 to CH1.
To do this part of the cal, remove the load standard from the coax attached to CH0 and attach it to the other coax on CH1. This should prevent any signal from making it into the CH1 port on the nanoVNA.
Tap the "ISOLN" menu item,
and it should accept that and move on to the next menu item. The final item is "THRU", and it needs to have that barrel connection to connect one coax to the other. This gives the nanoVNA a reference to how strong a signal will be if there is neither attenuation nor amplification when a signal passes from one port to the other. Remove the load from the coax, and connect the two coaxes together with the barrel connector.
After that, the "THRU" should turn to reverse highlight. It is now time to hit "DONE".
As I said earlier, it is possible to do all this cal and measurement with only one coax, but it tends to stress the coax with a very sharp bend. I avoided that by using two cables.
And this filter shows about a 2.75 dB insertion loss at 146 MHz.
Technically, this is the loss of not only the filter, but also the SMA-BNC adapters as well. Now that you have the system calibrated, and are measuring the filter, you can check other things like how much loss there is in the FM radio broadcast band down at around 100 MHz, or other areas of concern at higher frequency.
That will clear the existing calibration. Now, we need to apply a new calibration. Start by tapping the "CAL" menu,
If you have previously been focused on antenna reflection measurements, you may have only used the top three items on this screen, the "OPEN", "SHORT", and "LOAD" procedures. The other two are optional for reflection measurements, but they are very much required for the through measurements we'll be doing on a filter. My nanoVNA came with calibration standards for open, short, and load. They look nearly identical, so I have used a Sharpie marker to mark the tops of them.
It is important to keep track of which is which. If you get confused about which is which, you can examine the underside. The open standard has no pin in it (it is open!). The short standard has a pin and that pin is connected directly to the shell (it is shorted!). If you measure with an ohm meter, you should get very close to 0 ohms between the center pin and the outer shell. The load standard also has a pin, but it is not directly connected to the shell. Instead, it is connected via a 50 ohm resistor.
If you have trouble telling the difference between them, you can double check with an ohm meter. We will also use two coax cables, as I mentioned above, as well as a "through" connector, which is simply a straight through connector adapter.
When using two coax cables like this, we will measure the filter between them at something we call the calibration plane. I won't go into detail on that, but if you are interested in some of the technical details, I will give a link to Alan W2AEW's video on this at the bottom of this blog post, where he describes some of this.
For the first step of the calibration, we are going to use the open cal standard and apply it to the end of one of the coax cables. It needs to go on the cable that is attached to CH0, as shown in the picture below.
After that, click the "OPEN" menu item.
Remove the open cal standard from the end of the coax, and connect the short cal standard to that same coax:
Now tap the "LOAD" menu item.
After you hit the "DONE" menu item, you should see a screen that looks something like the picture below. There are a number of memories where you can save your recently completed calibration.
I'll tap "SAVE 0" to save the calibration parameters in memory 0.
After this, you have completed the calibration! Take a look at the blue and yellow traces in the picture below. The blue trace, which represents the "through" amplitude, listed as "CH1" is hovering very near 0 dB. That means there is essentially no amplification or attenuation of the signal going from one port to the other. The yellow line for me is hovering around -30 dB. This is the level of reflected signal the nanoVNA sees at CH0. Since we have just completed the calibration, this number would ideally be a lot lower. The fact that it isn't represents the limitations of the calibration accuracy. My calibration connectors are old and dirty; maybe you will do better than I'm showing here.
You can also notice some text along the left side in the picture below. The "C0" in that text at the left edge of the screen indicates that the nanoVNA is using the calibration parameters that we just stored in memory 0.
Here, in the picture below, is a closer view of the text at the top right. It is showing 0 dB at a frequency of 146 MHz, representing neither gain nor loss. If you are not getting something very close to 0, there was an error somewhere in the procedure.
If you are getting a flat blue line close to 0dB, it is now time to remove the through cal standard and replace it with the TinyFilter.
I had to use a connector adapter on one side of the TinyFilter as shown in the picture below. I just used the through connector.
You can now take a look at the display. Marker 1 is at 146.000 MHz and shows a loss of 2.70 dB. (The logmag of S21 is -2.70 dB). A number between -2.7 and maybe a tad over -3.0 is typical. Sorry the screen is so dirty in my picture!
Time now to measure the BNC version of the TinyFilter.
Well, the coax has SMA connectors, so we'll need to use some BNC to SMA adapters to make this measurement. They are shown attached here. I still needed that SMA F-F adapter to make it all work.
The picture below shows the setup for measuring the BNC version of the TinyFilter.
If you want to go into some further technical detail about the nanoVNA, I highly recommend checking out the in-depth video series from Alan, W2AEW. You can find it at this link.
I hope this note was useful. Let me know if you have any comments!
73,
Mike K6STR
Great post, thanks!!
ReplyDeleteGood Stuff. Helpful beyond the TinyFilter checks. W0KU
ReplyDelete