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

D-STAR Modifications

Repeater Modifications

While the repeater modules look nice, there are always things that can be done to make them work better. :)

Of course, it should go without saying that if you undertake any of the modifications listed on this page, you do so at your own risk. Further, you will probably be voiding your ICOM warranty, so you're on your own.

While we will try to be as accurate as possible in the information we provide, there are no guarantees that it will work with your equipment. Something as simple as a manufacturing change may render this information inaccurate.

Under the Hood

There are some interesting things to notice when you peek under the hood of the repeater equipment.

This is a picture of the ID-RP2C (click for zoom):

Not really too much to see here. Below is the block diagram of how it works.

Interesting to note is that the ASSIST ports (ID-RP2L 10GHz link) connect via a 10MB/s ATM connection using LVDS (low voltage differential serial).

There is a 16Mbitx8 Boot ROM (flash) (IC12), 32Mbitx16 Flash ROM (IC11), and 4Mbitx32 of SRAM (IC2, 3, 6, 7) on board.

Also of note is that J6 is an RS-232 interface directly connected to the CPU (IC5). The pinout of this connector appears to be:

	Pin 1 - RTS
	Pin 2 - TX Data
	Pin 3 - RX Data
	Pin 4 - CTS
	Pin 5 - GND
	Pin 6 - GND
	

Now, the IC-RP2D. This appears to be nothing more than perhaps an ID-1 in a rackmount case, presumably with some slightly different firmware.

Some investigation will be required to see if the radio responds to other commands on the USB port (as defined in the ID-1 interface specification).

Since the RP2D can only operate half-duplex, it only needs an integrated transceiver.

On the other hand, the repeater modules that need to do duplex require seperate transmit and receive units.

This is what the IC-RP2V looks like inside:

RSSI

It seems that the pinout of the RJ45 on the rear of all the repeaters is the same. The pinout of this connector is:

	Pin 1 - TC (Transmit Clock)
	Pin 2 - TD (Transmit Data)
	Pin 3 - Ground
	Pin 4 - TE (Transmit Enable)
	Pin 5 - RE (Receive Enable)
	Pin 6 - RSSI (Receive Signal Strength Indicator)
	Pin 7 - RC (Receive Clock)
	Pin 8 - RD (Receive Data)
	

Now, something that is immediately interesting about this is that all of the repeater modules have an RSSI signal that they send to the controller.

This RSSI signal comes pretty much directly from the receiver IC (MC3356 in the RP2D, or TA31136 in the RP2V). It is NOT used by the RP2C.

This signal is an analog voltage, that appears to be in the range of 1-4.5VDC (depending on signal level).

An IC-RP4000V was tested by applying a CW signal to the antenna connector, and measuring the corresponding RSSI voltage as the level was changed.

With no input, the RSSI voltage was 1.078VDC. The signal was swept from -128dBm to -60dBm in 4dBm steps (source was an IFR 1200S). The resulting data was graphed to produce the following:

You will notice that the curve is basically linear right up to about -72dBm, at which point the detector circuit obviously saturates (4.5VDC).

The receiver was also swept with a CW at both +/-12.5kHz from the operating frequency. As you can see, it takes a stronger level from the adjacent channel interference to cause the RSSI to rise, but it still affects the RSSI.

The actual selectivity of the receiver was not tested at this time.

We performed a similar test with an IC-RP2000V (new out of the box 2010/01)

Again, the curve is similar. Measurements were made every 2dBm. We did not test with an off-channel CW.

The test point used for this graph on the RP2000V was not the RSSI pin on the rear connector, rather, it was an RSSI test jack that was installed in the chassis.

Using a piece of RG-174 and a couple of test probe jacks, we extended the RSSI signal from inside the RX unit out to the chassis. The only difference between the RSSI test point inside the RX unit and the one on the RJ45 on the chassis is that the one on the RJ45 goes through an additional buffer op-amp that has an output impedance of 10k ohms.

We already had the RX unit open to add the discriminator tap, so it was convenient to add the RSSI tap at the same time.

The advantage of having the RSSI available is that you can get an idea of how much site noise your receiver is dealing with. As any ham should already be familiar with when driving around with a mobile, there are times when you are in "intermod alley" and your signal strendth indicator is pegged, even though your're not receiving anything. This is an indication that your front end of your receiver is saturated with RF, often leading to all sorts of nasty things (desense, intermod, etc).

By watching the RSSI test point when you hook up the antenna system to the repeater, you get an indication of how much RF is hammering the front end of the repeater (these are just glorified mobiles after all).

Ideally, you'd want to see very little, if any, increase in RSSI voltage when you connect the antenna/duplexer to the radio.

Its also useful if you are checking for desense... using the manual key modification (below), you can key your repeater and see if the RSSI voltage increases. If it does, you may not have enough isolation from your transmitter in your duplexer.

"Discriminator Tap" (Audio Output)

While there isn't currently any test equipment on the market that can encode and decode D-STAR, there are some work arounds.

Right up to the modem chip, the RX radio is still an FM receiver. So, all the standard tests can be done to check the RF performance of the radio by utilizing the receiver audio prior to the modem chip.

In the RP4000V, this is relatively painless. If you remove the top cover of the RX radio, you will see the RX interface board. Near the middle of that board is the CMX589A modem chip. Right near that chip is a test pad labelled "DET". Yep, you guessed it, that's good old FM Detector audio.

If you look on the main board, right near the front, you'll see some more test pads, "MIN" which appears to mean "Modem IN", and "RSSI" which is the receive signal strength indicator pin.

The RP2000V is the same.

So, after getting our hands on a service manual and having a look... The MIN signal is the AF output of the receiver IC, amplified, and heading to the input of the Modem IC. This test point is missing off page 9-3 of the RP2000V service manual (but it is there in the RP4000V manual). It connects to the junction of R9 and C12 (except on the parts layout on the pcb diagram, C12 does not appear populated).

R12 is a 100 ohm resistor that appears to be a protection resistor to protect the output of the op-amp. Likewise, there is another 100 ohm resistor, R15, that protects the input of the Modem IC. Therefore, you should be safely be able to use the MIN test point directly without any further protection.

A note about the op-amp itself. It is actually a low pass filter, with a cut-off frequency of 2258Hz and a de-emphasis time constant of 70.5uS (pretty close to the standard of 75uS). The gain of the amplifier is set at 2.7. This means that the MIN test point is flat (de-emphasized) audio.

We modified our repeater, and this test point is wired out with a piece of RG-174 to a chassis mount BNC connector on the rear of the repeater.

With an on-channel test signal modulated with 1kHz at 1.5kHz deviation, the output is about 700mVpp. This level is actually sufficient to make SINAD measurements directly.

Measured sensitivity (bench) of S/N 0601328 is -120dBm for 12dB SINAD (50Hz HPF/3kHz LPF), or -117dBm for 12dB SINAD (50Hz HPF/15kHz LPF) using an HP8921A.

With the receiver audio output, you can do all the standard FM repeater tests, such as checking for desense.

While we had the RP2000V on the bench, we decided to sweep the audio response of the receiver.

An on-channel carrier at -80dBm was modulated at 1.5kHz deviation with an audio tone that was swept from 10Hz up to 4500Hz. The resulting output at MIN was measured on an oscilloscope and graphed:

Transmit Enable/Isolate/Manual Key Switch

There are no provisions in the repeater for manually isolating the transmitter (other than turning it off), or manually keying it for making measurements such as spectral purity, power output, desense, etc.

This modification adds a SPDT, center off, toggle switch to the rear of the chassis to perform these functions.

Unfortunately, ICOM chose to use positive keying for the TX EN control, so a source of 3.3VDC is also required to make this work.

The TX EN control comes in from the RJ45 connector on the rear of the repeater, through a 100 ohm protection resistor, and heads to the TX radio module. In there, there is some additional protection (3.3k ohm resistor to ground) before the signal is presented to a TC74AC541 buffer. From the buffer it heads to the microcontroller and controls the radio.

The easiest place to get at the TX EN line, and find a source of power, is on the interface board in the rear corner of the chassis that also has all the power connections on it.

The TX EN line is the purple wire on J1. It goes through a via in the board and comes back up through another via near the RJ45 connector.

The traces are very tiny, but with a sharp knife you can cut that trace to open it up. I'd rather do that then cut apart a harness... its easier to repair the pcb trace.

Then, using a piece of 3 conductor stranded ribbon cable, you can solder one wire to the via on each end of the trace. If you use small enough wire, you can even insert it through the via and solder on the top side (one to the connector pin and the other to the SMD resistor).

That takes care of isloating the TX EN line.

Now for power, we need 3.3V, or a logic HIGH to be able to key the transmitter manually. This is easily accomplished with a 1k ohm resistor and a 3.3V zener diode.

Following the input power cable into the main connector, you will see two other red wires come off that connector near the other end, those are switched B+.

Solder one end of the 1k resistor to switched B+, the other end to the cathode of the zener, and the anode of the zener to ground. This makes the junction of the resistor and the cathode of the zener your 3.3V reference (or there about).

Run the 3.3V source to one side of your switch.

Run the wire you soldered to the end of the TX EN line attached to the TX module to the middle of the switch.

Run the last wire that connects to the resistor by the RJ45 connector to the other side of the switch.

Heatshrink it all up, and you're done.

The 3.3V source is inherently current limiting because of the 1k resistor. And the most you should be able to blow up in the TX module if things go wrong is the buffer chip (that's its job, to protect the micro). So this modification should be pretty safe.

So, after all the above mods.. we end up with something that looks like:

RP4000V Internal Jumpers

It has been reported by others, and confirmed also on our unit, ICOM in their infinite wisdom indeed used RG-58A/U for the jumper cables between the RX and TX units to the back of the chassis.

They probably could not have picked a worse cable for this application. It isn't the loss that is a problem (about 0.2dB), its the lack of shielding.

Put a 25W UHF radio right beside a sensitive receiver, and you're asking for desense when you use cable with only 95% braid (if you're lucky).

These are the factory installed jumpers:

In order to eliminate the RG-58, and improve the shielding, we replaced the factory jumpers with some semi-rigid coax jumpers instead. These jumpers happened to be surplus out of another application, and fit the bill. They are N(f) bulkhead on one end, and SMA(m) on the other. A high quality SMA(f)-N(m) adapter was used to attach to the radio.

The result:

On a side note, these are obviously nothing more than modified mobiles that make up the repeater... if you've used a ham mobile in a high RF environment, you know how the front end is wide open like a barn door and will get hammered by strong adjacent channel interference.

Since there is so much room inside the chassis, you would be well advised to add a tuned RF preselector in front of the receiver to try and keep some of the "crap" out.

A surplus out of a Daniel's Electronics UHF receiver would work perfectly inside the chassis. We use an external Sinclair Laboratories receiver multi-coupler ahead of ours that contains a bandpass filter and pre-amp. It makes a BIG improvement.

Output Power

The RP4000V was checked for its output power by applying 3.3V to the TE line on the RJ45 connector (you can find 3.3V on one of the accessory connectors inside the transmit radio).

On low power, the radio puts out 2.2W (seems low, should be 5W?).

On high power, it puts out 24.7W (on spec).

We did the same check with our RP2000V.

On low power, the radio puts out 2.6W.

On high power, it puts out 24.8W (on spec).

Input Current

The RP-4000V draws about 400mA idle (RX). During TX (low), it uses about 2A, and on TX (high), about 5A.

TX/RX LED

One of the biggest things that the repeater modules lack is an indication of whether the repeater is transmitting or receiving.

Instead, all you have on the front panel is a nice, green LED to tell you that the power is on.

However, all is not lost, it seems that ICOM did include all of the circuitry on board to give you a TX and RX LED, its just not brought out to the outside world.

This modification was first tested on the IC-RP2D 1.2GHz Digital Data repeater. It is confirmed to work with this unit. It has also been proven to work on the IC-RP2V 1.2GHz Digital Voice repeater. The difference on the IC-RP2V is you will need to pull the RX from one radio unit and TX from the other.

Open up the repeater to expose the radio unit. You will also need to open the top cover of the radio unit to expose the logic board.

On the logic board, near the front of the repeater, you should see three series of three through-hole plated vias.

The set of holes on the left are connected to the bias resistors of the TX/RX LED circuit.

Note that the picture of the board layout is actually of the under side of the board, when you are looking down from the top, R114 (TX) is actually the LEFT hole, and R113 (RX) is the RIGHT hole.

Since the bias resistors and drivers are already installed, all you have to do is install some LED's!

Our prototype used a bi-color, common cathode LED. We connected the green anode to the RX line, and the red anode to the TX line.

This is a picture of our prototype in "proof of concept":

Once we proved that it actually worked, it was time for a more permanent solution.

The factory power LED is pretty much an idiot light. As long as the fan is running on the back, there is power. We re-located the power LED from the top of its PC board to the bottom (you'll need to rotate it 180 degrees so the pinout is correct when you do this).

This freed up room behind the little frosted hole in the chassis to install our TX/RX LED without having to drill holes in the case.

Extend the LED with a nice piece of shielded wire, hot glue it in place, and you get this:

On the IC-RP2D it flickers nicely red and green as data is being moved. On the voice repeaters, you will usually find the bi-color LED will be amber/orange since both TX and RX will be active at the same time.

On the RP2000V and RP4000V, ICOM did something different. Unfortunately, the above modifications won't work.

However, after reviewing the service manual, something interesting was noted. On the "REAR" interface board, there is an RE signal. When traced back, this actually comes from a pin on the microcontroller called "RX LED"... go figure.

It would appear that RE only goes active when a valid D-STAR signal is present, since it was not active when an on-channel CW signal was applied to the repeater.

This is a 3.3V active high signal that heads to the controller. You should be able to buffer this signal and drive an LED of your liking. Just tap the RE signal off the REAR interface board and go for it.

Likewise, the TE is the transmit enable (PTT) signal. It too is 3.3V active high. You could buffer that signal as well and drive an LED to indicate TX.

Note that the RE signal is buffered through a gate in a 74AC541. It could probably support driving a RX LED directly, if desired (remember that its going to be 3.3V, so figure your current limiting resistor accordingly).

If you don't want to go through all that trouble, N5EBW has made a drop-in replacement REAR board that brings out the LED signals for easier installation. You can find his information here. Unfortunately, it does not appear that he is actively producing them at this time (the comments on the page are pretty stale as of 2010/01).