Fwd: WSU Repeater Isolation Analysis Help Needed

[John Metcalf]

---------- Forwarded message ----------
From: John Metcalf <johncm...@gmail.com>
Date: Sat, Sep 13, 2014 at 3:45 PM
Subject: Fwd: WSU Repeater Isolation Analysis Help Needed
To: kz...@arrl.net

Mike -

Luke Jenkins, John Karras, and I were talking with you this morning about the WSU repeater and our bounty of duplexers. I have an question that might be amateurish to ask but I thought you might be the guy to help.

You recommended a configuration of 4 notch, 4 bandpass on each RX and TX radios for our frequency pair (if I recall correctly) on our repeater. This got us thinking about the cavities we have at our disposal and what the optimal configuration might be. We were tempted to follow your advice and tune up all of our cavities and put the system into that configuration but I said that I wanted to do the analysis to see what we really neededWe may not need 8 cavities, while that would provide lots of isolation, it also adds insertion loss so less power from the user transmitting station gets in, and less power from the repeater transmitter gets out.

Currently our system is operating at up to 50W TX on 145.250 (Yaesu FT-1900R) with three notch cavities @ -80 dB on the RX frequency 144.650 MHz, and the RX radio (FT-1900R) has 3 notch cavities @ -80 dB on the TX frequency 145.250 MHz. Each cavity provides about 23 dB of isolation at the tuned frequency. I'm a bit rusty on the use of the terminology used to describe the typical waveform shape, but from the graph, when the amplitude (log scale) is halved there is a frequency spread of 300 kHz (3 dB point or FWHM?). You can see this in the following Excel graph:

Combined_Cavity_2010329-BigGraph.xls - Tab: TXRX_Graph: Shows the total isolation on the TX and RX frequencies

The transmitting power curves (I may be wrong here). The 1900R spec sheet (see link above) says +/- 2.5 kHz (Narrow FM). What I am imagining the general shape of the transmitter frequency spectrum will be a Gaussian shape with peak at 50 dBm (50W) and 145.250 MHz, with a Full Width at Half Max or -3dB down point there will be a 5kHz spread (that should fully describe the Gaussian shape, although I realize it is not Gaussian but at least this gives us something to visualize). I don't have any graphs of what this waveform looks like, but I did get a capture of a VX-8R spectrum the following Excel graph:

2014-913_Comm_Link-Budget.xls - Tab: VX-8R_TX_Profile

We could fit a Gaussian curve to this and get an approximation, it doesn't really matter the point is that the TX freq spectrum looks something like a Gaussian and that to some degree is good enough for this discussion.

This is the part I am having a hard time imagining. At one point (four years ago) I decided to interpret the sensitivity of 0.2 mV (-107.7 dBm) and selectivity of -60 dB / 28 kHz as another Gaussian-like shape that had a peak at 0 dB with width of 28 kHz at -60 dB. You can look at these shapes on the Excel graph (although now I'm thinking these should be upside down):

2014-913_Comm_Link-Budget.xls - Tab: All_Gaussian

Today I was describing the shape to Luke but I started to doubt myself. Now I am imagining it to be upside down from what is in that graph, instead it would be a notch shape response curve with the lowest point being -107 dBm and at the -60dBm points it would be the +/- 28kHz (with no upper limits).

Basically what I have described above is:

The RX response curve is centered at 144.650 MHz with low point at the spec sheet sensitivity (in dBm) that has a frequency spread of +/- 28 kHz at -60 dBm. (this is where the guassian shape comes in handy to visualize).

The TX transmit curve is centered at 145.250 MHz with a high point at the spec sheet max transmit power (in dBm) that has a frequency spread of +/- 2.5 kHz at what I'm assuming is the -3dB down point since the spec sheet did not state where that was.

Overlaying those two curves I can determine what frequencies there are overlaps of positive transmitter power into the receiver resonance 'bath-tub'. Where there is overlap, that is where I need to place duplexer cavities. I have several to choose from: the notches or bandpasses. I can use the data sheets or actual measurements from the cavities using a spectrum analyzer to determine their filtering capabilities, and add in the minimum number of cavities (since each has a non-zero insertion loss) required to keep the transmitter power from spilling over into the receiver resonance frequencies thereby overloading the receiver (desensing), and then using best judgement on the user community and our expected output power from the antenna - what configuration of filters would be best. More filters on the RX side will make it harder for 5W handhelds to get into the system but make more power available to pass to the antenna. Alternativly, we could put more filters in front of the transmitter so the RX radio would be the most sensitive to incoming RF but the TX radio would have to deal with more insertion loss due.

That is the path that I am going down for the isolation requirements analysis. Before I got any further, I wanted to ask if that is what you would do. For my isolation link analysis these are my questions.

What does the frequency response curve for a receiver radio look like and how to I interpret the specification sheets (sensitivity and selectivity parameters?) to get that curve?

What does the frequency transmit curve for a transmitter radio look like and how do I interpret the specification sheets (transmit power and maximum deviation?) to get that curve?

From those two curves, I should be able to overlay them which will allow me to determine how much and what type of filtering needs to be added?

That was a lot, and I don't expect an answer but I would really appreciate some guidance on this.

Thanks,

John Metcalf - ke7vvt

[John Metcalf]

Sent from my iPhone

Begin forwarded message:

From: Mike Fullmer <kz7o...@gmail.com>
Date: September 17, 2014 at 6:59:09 MDT
To: John Metcalf <johncm...@gmail.com>
Subject: Re: WSU Repeater Isolation Analysis Help Needed

Little Mt uses a common antenna for transmit and receive. The power is about 15 watts if I remember correctly, coming out of the transmitter.

Mt Ogden uses a separate receive antenna about 40 ft above the transmit antenna and still uses the duplexers also. It used to have a common antenna but something changed on the mountain a year ago and we started getting interference signals that we could not find. We think it is coming from another service there. The only solution that cured it was to install a separate receive antenna. The transmit power is about 20 watts.

Mike

On Tuesday, September 16, 2014, John Metcalf <johncm...@gmail.com> wrote:

Sorry... I got excited!

Shared rx and tx antenna on both mt ogden and Ltl mtn? What are the transmit powers set at? 

Thanks Mike!

Sent from my iPhone

On Sep 16, 2014, at 21:24, Mike Fullmer <kz7o...@gmail.com> wrote:

Wow, that is a technical writeup!

I think we misunderstood each other. There are 4 cans total on our repeaters. 2 in the receive line, 2 in the transmit line. They are set up with one band pass and one notch on transmit side and one band pass and one notch on the receive side. 

yes, there is considerable loss, especially on the band pass ones. We figure about 3 db total on both the receive and the transmitter. 

Mike Fullmer