Freedom 8100 here!

I prefer Porsche ;-)

@@w2aew yeah, Stuttgard over Maranello :)

You should see the ones used for 6m (50-54MHz) repeaters, the cans can be as tall as a small person.

@@w2aew thanks for the recommendation will check them out.

Some firmware versions for the H and H4 can do 401 points, but those don't have a low enough noise floor to see the notch filters.

I"m sure that DisLord's firmwre for the V2plus4 is very good, I will have to try it at some point...

To use a spectrum analyzer, you'd need a signal source (usually a tracking generator), and a directional coupler or return-loss bridge. With these, you can tune the passband. You won't need the coupler or bridge to tune the notches, but you will need to have sufficient measurement dynamic range to see the depth of the notch (90dB down or more).

You can set a narrow span, so the frequency resolution should be adequate for a 600kHz spacing. The dynamic range is another story - really depends on the generation of the NanoVNA - the -F and V2 units have better dynamic range than the original and -H -H4 units.

True, but the actual length differences are typically 1% or less given the typical splits for a given band, which is likely within the cut tolerance of the interconnect cables.

@@w2aew First - good presentation... the "Nano-VNA" products are a decent 'clue' but not sufficient for really accurate, in-depth work as-is-prudent in many cases. I just did a commercial 'bottle' retune with proper commercial equipment, then checked with my Nano-VNA... nice 'hint' but I wouldn't use a Nano-VNA as final proof for a repeater I cared about before taking the set of bottles to a commercial radio site. The '1%' works if the filter package is more or less within as-designed/as-built band-segment. 220 is 'easy.' I bring this up as hams are want and often acquire to re-purpose surplus commercial products that are in significantly different band-segments. Can you use a 450-460, 470-490+ set of 'bottles' at 440-450, or 420-430? 155-159 MHz bottles at 146? Yes... sort-of... but the results are not good/best. 10-20 dB of not-quite-best pass- and/or reject- performance difference is not insignificant. It is prudent/recommended to consider and re-configure coupling if you're taking a used product 10, 20, 40 MHz out of as-built band. Some filter manufacturer's filter-prep bench have an array of various length coupling lines to dial-in each product sent out. There can also be internal coupling loop differences which is also a factor. Two 'identical' products can leave the factory with different length coupling cables. This becomes more critical as hams are want to try to re-purpose less-stringent mobile products to repeater service. MORE critical as we have and continue to see some vendors re-packaging "off-shore" mobile products with 'generic' reject-only flat-packs - all of which wholly inadequate for commercial site installations. Many commercial and government sites do not allow... re-purposed mobiles as repeater or remote-control installations, reject-only filters, lack of isolator on the TX, 'ham' antennas, LMR-type feedlines, etc. Some of our site owners call at least one of us to pre-qualify potential new tenants, and we ask about the equipment, etc. Some vendor/brand names and products trigger an immediate NO! Even if what's stated is good enough to yield a Yes, post-install inspection and spectral analysis gets interesting. If your ham radio presence at a great site is 'essential', the equipment and engineering should be equal to the objective.

@@jimaspinwall7814 Jim - thank you for all of this great insight. You've clearly done this more than I have, and this real-world practical info is very valuable.

Yes, including the cables in the calibration is important when you want to remove the attenuation and phase (delay) of the cables. In our case where we are only looking at the relative depth of the return loss and insertion loss (no phase measurements, no absolute parameter measurements), you can get away without including the cables.

@@w2aew Thank you for the clarification. As I said I am no expert but always with a hunger to learn. 73's a d take care

It is a close assumption. At that frequency, and with this level of attenuation, it is possible that you could be getting leakage through the shield of the coax - which means it could actually be a little better than 90dB.

Yes, S21 is insertion loss.

Dynamic range is nearly the same, but it’s about 2-4x slower.

Well, if someone wanted to load me their 4401A VNA, I'd be happy to do a video with it!

If I needed to accurately know the phase at the filter ports, yes. In my case, the loss at these frequencies and the impedance rotation due to these cables, will have no effect on the filter tuning - so it wasn't necessary to calibrate at the end of the cables.

Point of detail: S11 is not Return Loss, it is the Reflection Coefficient. The difference between the two though, is just the sign.

@@w2aew Thanks Alan. Your videos inspire me to a lot of VNA uses beyond antenna analyzing. 73, F4HYZ

the nanovna can't do UOSM from what I understand. but if you're calibrating at the port instead of at the measurement plane, it all is rule of thumb anyways.

@@tommihommi1 Yep, but in the case of a true TOSM or UOSM at least one can de-embed the characteristics of the cables from the DUT if interested. 😀

@@stefanhummelink3662 Yes, but in this case of course, there wasn't a need to establish a measurement plane at the filter ports. I have another video that talks about the importance of VNA calibration and the establishment of a proper measurement plane (and another video on port extensions).

I think it might be generally more difficult to get a good, solid RF connection on the threaded rod - but I'm not the designer of these things so I can't say for sure.

Hm. The plunger and a threaded... gizmo... whatever it is... could be separate pieces. Even temporarily installed. I suppose if you did this sort of thing on a regular basis... either you develop ninja skills, or if you're incapable of such (like me), make more tools!

I use a small-ish wrench. To move the plunger down, i slightly tap the top of the rod with the wrench. To move the plunger up, I put the open ended side underneath the knob and tap upwards. Works like a charm every time.

Sure, if you have them, use them.

Yes, but there will be some interaction when they're connected together, so the notch will move.

Not really - the software wouldn't help with the measurement dynamic range.

As far as I know, the duplexer is not contained in the TKR-851 - it is external, so the tuning process would be similar to shwa is shown here.

Return Loss in dB should always be a positive number. It is simply the magnitude of the reflection coefficient, with the negative sign removed. It is common (but incorrect) to see the Return Loss expressed as a negative dB value. Return Loss is equal to -10*LOG(PR/PI). Since the reflected power PR is always less than the input power (PI), the result of that equation will always be a positive number. Larger values mean higher return loss - which is a GOOD thing - you want the reflected power to be very low - thus you want the return loss to be very high.

@@w2aew well I was thinking, the key would "switch" to a dummy load , so, back and forth, essentially

I don't have access to an SV4401A - so I can't perform a comparison.

@@alanwolke6253 Thank you.

Mainly by learning from more experienced mentors, reading application notes from manufacturers, we well as studying various source of information (websites, books, videos, datasheets, etc.).

Would you be willing to give me some assistance? My email address (scroll over email) is on my QRZ page.

I am trying to tune two Comprod duplexers, one UHF (66-40-46) and one VHF (66-13-46). I am trying to learn the process on a NanoVNA V2 and I will purchase a NanoVNA V2plus4 if I can figure it out but I am beating my head against the wall. My traces look very different than what you have shown in your video.

@@4150CHF I will reply to the email that you sent to me as well. But I also wanted to share this video from Sinclair (another duplexer manufacturer) that goes through the tuning process in detail for their type of duplexer. kzfaq.info/get/bejne/i52Bjd2rp5iwiok.html

This type of filter can't really be accurately conveyed in a schematic. There are very few traditional electronic parts - it is mainly built of mechanical structures to create resonant cavities, etc.

@@w2aew Thanks. Do you have simplified schematic? I want to see how the simultaneous transmit and receive work.

Duplexers, Diplexers can be modeled with lumped circuit elements but this would be very difficult. In the Feynmann lectures they show as the RF frequency increases a metallic can acts like an LC system. I can recall for example a coke can, becomes a resonant circuit at about 455Mhz. What you really need is a mechanical diagram not a electrical schematic. I have always thought of Duplexer design as a bit like black magic. very specialized area.

A diplexer is generally for separating frequency bands not as closely related as a duplexer, which needs fine notch filters to notch out very close in-band signals. Tuning method would be identical or similar, but how you actually do it would closely depend on the design of the unit itself.

@@ruhnet Diplexers/Duplexers are sometimes used from Receiving and Transmitting on the SAME frequency or Different frequencies using only ONE antenna and ONE receiver.

@@waynegram8907 That is correct (as are quite a few other uses of them in different forms).

@@ruhnet like what other different forms? The isolation from port#1 and port#2 has to be -80db to -100dB to prevent crosstalk of the transmitting and receiving

@@waynegram8907 One example: a simple diplexer in a receiver is used to separate low (audio) frequencies from RF in a mixer/detector output. This is one of the "different forms" I was referring to. A NanoVNA (the original design) can accurately tune just about anything as long as the required dynamic range is within 60dB or so. (But note that the dynamic range of the NanoVNA differs with the various frequency bands it covers.)

LOGMAG is just the display format - it is plotting the MAGNITUDE of the measured parameter in dB - that is what LOGMAG means. There are two traces that I configured - one of them is measuring S11 which is the reflection coefficient, the measure of how much of the injected signal gets reflected back off of the input, back towards the source (VNA CH0). The other is the S21, which is the transmission coefficient, the measure of how much of the injected signal makes it through the duplexer. Each of these two measurements are being displayed in the LOGMAG format.

@@w2aew Is S11 (CH0) connected to input of one side of duplexer (Lo/Hi) and S21 (CH1) connected to duplexer output (antenna port)?

@@4150CHF yes

@@w2aew What do I need to select in the menu to ensure that the second trace will be associated to S21 (CH1) and not a duplicate of S11 (CH0) or is it automatic ? Please forgive my ignorance.

@@4150CHF Under the Display menu, you'll see a Channel button that allows you to select which channel is being used for the selected Trace. Use the Display>Traces menu to first select the active/selected trace, then go back to Display>Channel to select the channel for that trace.

Most (but certainly not all) of the amateur radio repeaters in the US are coordinated with their output (transmit) frequencies between 1282-1288MHz and 1290-1294MHz, and the repeater inputs are between 1270-1276MHz.