Some documents linked from this page may require Adobe Acrobat PDF reader.

D-STAR RF Hardware

1.2GHz Duplexers

We are using all the same duplexers (Sinclair) as we did with the analog repeaters. They perform well, and have good isolation. There is the issue of using the 1.2GHz DD and DV repeaters with the same antenna, a solution is outlined below.

Background

The IC-RP2V 1.2GHz Voice Repeater is like most others, it has seperate TX and RX antenna ports. The IC-RP2D Data Repeater, on the other hand, uses the same antenna port for TX and RX (since it operates half duplex).

This presents a challenge when trying to use both repeaters with a single antenna.

Normally, you would be able to use a TX combiner and RX multicoupler system, albeit designed for 1.2GHz. With the DD repeater transmitting and receiving on the same frequency, this would wipe out (and probably blow up) the RX multicoupler.

There are a couple of ways to solve this problem. The IC-RP2D can be set to use an "offset" frequency to split the TX and RX frequencies. But you still have a single antenna port to deal with. That could be rectified by attaching a circulator to the output of the repeater, which would seperate the TX and RX paths.

Another option (the one we worked on) involves seperating the DD's operating frequency far enough from the DV's TX and RX frequencies to allow more conventional filtering techniques to be used.

Using this option, the existing duplexer can be used for the DV machine, and the DD gets coupled onto the transmission line using techniques developed by Sinclair for their C-Series Combiners.

C-Series Multicouplers

C-Series Multicouplers utilize a series of bandpass filters on the equipment to be coupled, and a notch filter, tuned to the same frequency as the pass filter, to isolate the rest of the equipment on the transmission line.

See the Sinclair document for further information.

Do a patent search for US Patent 3,124,768 if you want more theory.

Parts and Pieces

It just so happens that a 420MHz duplexer actually performs not too bad at 1260MHz. This is because it is an odd multiple wavelength, so if its designed for 1/4 wavelength at 420MHz, its working at 3/4 wavelength at 1260MHz.

It also just so happens that we had a Sinclair Q3220E duplexer kicking around, as well as a bunch of other compatible parts and pieces from bandpass filters.

There is a Sinclair Res-Lok design for the C-Series Multicouplers, so we figured we'd try and make something similar work.

We ended up taking the Q3220E duplexer, cutting in holes for three more coupling loops (to make bandpass cavities), modifying all the coupling loops for higher frequency, and building new harnesses.

Note, you will not be using the coupling loops with the capacitors, so you actually need 7 plain coupling loop assemblies in the end.

This is the UHF duplexer, before we started in on it:

Disassembly

First, put some reference marks on the top lid of the duplexer, and the body, so that it all gets oriented correctly later on.

Next, you will need to take out the 60 or so screws in the top lid that holds it on to the body.

Be careful in handling the plunger assemblies, you don't want to get finger grease all over them and cause the silver plating to corrode. You also don't want to scratch, bang, or marr them.

Remove the existing coupling loops and harness assemblies.

Here is the top assembly of the UHF duplexer:

Bandpass Modifications

In order to create bandpass cavities, you will need two ports on each cavity... right now there is only one.

We chose to use three of the cavities for bandpass, and the fourth as a reject (notch) filter. That meant we would need to cut three additonal holes for three more coupling loops to be installed.

Layout is pretty critical. You will want to cut in the new holes diagonally opposite from the existing ones. Other considerations to make are that you will also need to drill retainer holes for the retainer screws to hold the coupling loops in place. Try and make everything symmetrical for the most efficient operation. Also, consider where the casing screws are going to be when you are planing where to cut.

It would also be prudent to check that you have the proper coupling loop assemblies that fit your duplexer, before you cut more holes. Depending on where you salvage parts from, they may be a different size (we picked ones from our collection that were the same diameter as the existing ones).

Here is the top plate, after cutting the new holes with a hole saw:

Note that the three cavities on the left will now be the bandpass ones.

Coupling Loops

The shape and size of the coupling loops affect the performance of the filter (its Q).

Too big, and the filter response won't be sharp enough, but you'll have low insertion loss; too small, and you'll have sharp response, but high insertion loss.

There surely is a calculation to figure out the right balance, we cheated and peeked inside our existing Sinclair 1.2GHz Res-Lock, and use the same size loops (25mm).

The "short" sides on the loops are about 5mm, with the remainder the "long" side. Make them so that they form nicely between the existing terminals.

This is what the completed loops look like:

Remember, you need 7. We used the same copper stock that was on the salvaged ones we had (they were from an assortment of VHF and UHF filters.

Reassembly

Its pretty straight forward to put it all back together. Install the 60 or so screws (don't forget the orientation of your reference marks).

Install all your coupling loops and their retainer screws, and you get something like this:

Harnesses

In the normal C-Series Multicouplers, inter-cavity coupling is done inside the cavities. That was too much like work, which is why we opted to cut in the extra coupling loops.

In order to couple the cavities together, you will want to create a harness that is an odd multiple of the electrical 1/4 wavelength at your operating frequency, including the length of the coupling loops (25mm each).

We designed for use at 1250MHz, which is 240mm in wavelength. 1/4 of this is 60mm. We used RG-400, which has a VF of about 69.5%. So a 1/4 in RG-400 is 60mm x .695, or 41.7mm.

Odd multiples of this would be 125.1mm, 208.5mm, 291.9mm. We chose to go with 208.5mm, which was a reasonable length, which made our inter-cavity harnesses 208.5mm - 25mm - 25mm = 158.5mm from connector face to connector face. Build two.

Now, you also need a 1/4 wavelength harness from the output of the bandpass section to the "antenna" port, and between the "antenna" port and the reject (notch) filter (pass through port).

We used "T" connectors for the "antenna" port connecton, and at the "pass through" port.

To better understand how this is configured, this is what the finished product looks like:

After some initial testing, we found that the harnesses between the bandpass section, the antenna port, and the notch filter worked well when they were 180mm from connector face to connector face.

Coupling loop length and the adapters all play in this.

Your IC-RP2D repeater will connect on the left side at the input of the bandpass section. Your antenna connects at the "T" that is "floating" in free space. The antenna port of your existing duplexer from your IC-RP2V will connect to the other "T" connector that is on the notch filter section.

Tune-up

We designed this system so that our existing DV repeater could operate at 1291MHz -20MHz, and the DD machine would operate at 1251MHz. That's a nice 20MHz spread between each transmitter and the DV machine's receiver.

Fire up your tracking generator and spectrum analyzer and connect to the first bandpass cavity (sweep through it).

Hopefully you put the coupling loops in the same orientation, and made reference marks on them and the top of the combiner before you start tweaking.

Adjust the plunger to get your cavity on your chosen frequency. Then, adjust the coupling loops, each by the same amount, for minimum insertion loss. We were able to get less than 1dB IL out of each cavity.

Put in the harnesses between the bandpass sections and sweep all the bandpass section to check that the response is even and that the insertion loss is the sum of the individual cavities.

If you're good, you can move on to tuning the notch section.

Set up your equipment to sweep through the "T" connector by itself (the one you're going to connect to the cavity in the final configuration). Connect it to the notch filter and you should be able to adjust the plunger to your same pass frequency. This time, adjust the coupling loop for the deepest null.

Now you can add in the rest of the harnesses, and start sweeping the entire unit.

This is a sweep of ours with a detector on the Equipment Port, the source on the Antenna Port, and the Pass Through Port terminated:

As you can see, we've got about 2.5dB of insertion loss between the Equipment Port and the Antenna Port. Remember, all coupler systems have loss, its the price you pay for not having to put up another antenna.

Here is a sweep with the a detector on the Equipment Port, the Antenna Port terminated, and the source on the Pass Through Port:

You will notice that at our operating frequency, we have a pretty decent 21dB notch to isolate equipment further down the line.

Here is the Equipment Port terminated, with the detector on the Antenna Port, and the source on the Pass Through Port:

You will see that at the combiner's operating frequency, there is 19dB of isolation, and at the operating frequencies (1271 and 1291MHz) of the DV machine, there is 0.4 and 1.2dB of insertion loss, respectively. Not bad.

Just for completeness, here is a sweep with the Equipment Port terminated, the source on the Antenna Port, and the detector on the Pass Through Port:

Again, pretty much the same results as the other sweep.

Putting It All Together

So the combiner seems to be working, and doing what we want. It lets us couple the half duplex 1251MHz signal on to the same transmission line as the 1271MHz and 1291MHz signals.

Since we had the duplexer from the DV machine around, we put everything together and swept it.

Here is a sweep of the complete system with the source on the Lowpass Port of the duplexer, the detector on the Antenna Port, and the Highpass and Equipment Ports terminated:

You will see that we've got better than 33dB of isolation on the DD port, 1dB of insertion loss on our frequency of interest, and better than 55dB isolation on the other DV repeater port. Note that the 55dB is actually the limit of the detector head being used, the actual isloation is MUCH better.

This is a sweep of the complete system with the source on the Highpass Port, with the Lowpass Port and the Equipment Port terminated, and the detector on the Antenna Port:

Again, good isolation on the terminated ports, and just under 2dB of insertion loss on this port.

Finally, here is the sweep through the combiner's Equipment Port, with the Highpass and Lowpass ports of the duplexer terminated, and the detector on the Antenna Port:

So, we've got about 2.3dB of insertion loss on this port, and isolation that is beyond the limit of our detector on the other ports.

Conclusions

Could we get better performance out of this system, probably. However, one of the goals was to prove the concept works, use materials on hand, and get the repeaters on the air!

As you can see, it indeed does perform pretty respectfully. It also lets us use our GP-21 antenna that is fed by 120 feet of LDF-7 for both repeaters.

So far, on air tests seem positive, it appears to work quite well. Of course you're taking a 2.5dB hit on the DD machine, versus having it connected directly to the transmission line, but that's what compromises are about.

So, if you've got lots of parts laying around, the time, and the equipment, give it a go!

We've also heard recently that TX/RX Systems is now offering a solution to do just this. But, you can bet you're going to "pay to play".