I had been looking around for some time if someone had already managed to get I/Q data from Icom’s IC-7610 HF transceiver without going through the binary-only hdsdr driver provided by Icom, but couldn’t find anything.

First attempts at doing that on Debian Linux using libftdi1 didn’t work, so I resorted to using the (again binary-only) libftd3xx driver from FT, and succeeded after some tinkering around.

The program writes raw I/Q data to a file in int16-int16 format.

• ic7610ftdi: https://github.com/df7cb/ic7610ftdi
• Get libftd3xx from https://ftdichip.com/drivers/d3xx-drivers/
• IC-7610 I/Q port reference: https://www.icomjapan.com/support/manual/1792/
• The IC-7610 needs to be connected using a decent USB 3 cable, preferably without any hub in-between
• If the SuperSpeed-FIFO Bridge disappears after some time, re-plug the cable or power-cycle the transceiver

$ lsusb | grep IC
Bus 002 Device 024: ID 0c26:0029 Prolific Technology Inc. IC-7610 SuperSpeed-FIFO Bridge

$ make
cc -Wall -g   -c -o ic7610ftdi.o ic7610ftdi.c
cc   ic7610ftdi.o  -lftd3xx -o ic7610ftdi

$ ./ic7610ftdi iq.cs16
Device[0]
	Flags: 0x4 [USB 3] | Type: 600 | ID: 0x0C260029
	SerialNumber=23001123
	Description=IC-7610 SuperSpeed-FIFO Bridge
fe fe 98 e0 1a 0b fd ff
fe fe e0 98 1a 0b 00 fd IQ data output: 0
fe fe 98 e0 1a 0b 01 fd
fe fe e0 98 fb fd ff ff OK
RX 42 MiB ^C
fe fe 98 e0 1a 0b 00 fd
fe fe e0 98 fb fd ff ff OK

$ ls -l iq.cs16
-rw-rw-r-- 1 myon myon 44040192 26. Aug 22:37 iq.cs16

$ inspectrum -r 1920000 iq.cs16 &

To talk to my IC-7610 via wireguard, I set up UDP port forwarding for ports 50001 50002 50003 on a raspi:

domain (ip) {
  table filter {
    chain FORWARD {
      policy ACCEPT;
    }
  }

  table nat {
    chain PREROUTING {
      interface (wg0) {
        proto udp dport 50001 DNAT to 192.168.0.6;
        proto udp dport 50002 DNAT to 192.168.0.6;
        proto udp dport 50003 DNAT to 192.168.0.6;
      }
    }

    chain POSTROUTING {
      outerface (eth0) {
        proto udp dport 50001 SNAT to-source 192.168.0.3;
        proto udp dport 50002 SNAT to-source 192.168.0.3;
        proto udp dport 50003 SNAT to-source 192.168.0.3;
      }
    }
  }
}

LoRa APRS iGate

The official documentation of https://github.com/lora-aprs/LoRa_APRS_iGate uses the PlatformIO plugin of MS Visual Studio Code. Here are the commands to get it running without the GUI:

git clone https://github.com/lora-aprs/LoRa_APRS_iGate.git
cd LoRa_APRS_iGate

• Edit data/is-cfg.json with your station info
• Edit platformio.ini: board = ttgo-lora32-v21

pip3 install platformio

pio run
...
Building .pio/build/lora_board/firmware.bin
...

pio run --target upload
...
Uploading .pio/build/lora_board/firmware.bin
...

pio run --target uploadfs
...
Building SPIFFS image from 'data' directory to .pio/build/lora_board/spiffs.bin
/is-cfg.json
...
Uploading .pio/build/lora_board/spiffs.bin
...

LoRa APRS Tracker

The procedure for the tracker is the same, but the GPS module might need a reset first:

git clone https://github.com/lora-aprs/TTGO-T-Beam_GPS-reset.git
cd TTGO-T-Beam_GPS-reset
pio run -e ttgo-t-beam-v1
pio run --target upload -e ttgo-t-beam-v1
# screen /dev/ttyACM0 115200

… and then upload https://github.com/lora-aprs/LoRa_APRS_Tracker.git

Classic ham radio transceivers have physical connectors for morse keys and microphones. When the transceiver is a software defined radio (SDR) device, voice operation is easy by attaching a headset, but solutions to connect a morse key, be it a straight key or paddles, to a modern PC are rare. In the old times, machines had serial ports with RTS/DTR lines, but these do not exist anymore, so a new interface is needed.

I am using a LimeSDR as ground station for the QO-100 satellite, and naturally also wanted to do CW operation there. I started with SDRangel which has a built-in morse generator, but naturally wanted to connect a CW key. At first sight, all the bits are there, there’s a tune button that could be used as a straight key, as well as keyboard bindings for dots and dashes. But the delay key->local audio is almost a full second, so that’s a no-go. I then went to hack my K3NG keyer to output ^ (high) _ (low) signals on the USB interface, and have a smallish Python program read that and send SDRangel REST API requests. Works, but that solution always felt “too big” to me, plus the sidetone from the buzzer inside the Arduino case could be heard in the whole house. And the total TX-RX delay was well over a second.

Next I tried building some GNU Radio flowcharts to solve the same problem but which all had the same trouble that the buffers grew way too big to allow the sidetone to be used for keying. At the same time, I switched the transceiver from SDRangel to another GR flowchart which reduced the overall TX-RX delay to something much shorter, but the local audio delay was still too slow for CW.

So after some back and forth, I came up with this solution: the external interface from the CW paddles to the PC is a small DigiSpark board programmed to output MIDI signals, and on the (Linux) PC side, there is a Python program listening for MIDI and acting as a iambic CW keyer. The morse dots and dashes are uploaded as “samples” to PulseAudio, where they are played both on the local sidetone channel (usually headphones) and on the audio channel driving the SDR transceiver. There is no delay. :)

DigiSpark hardware

The DigiSpark is a very small embedded computer that can be programmed using the Arduino toolchain.

Of the 6 IO pins, two are used for the USB bus, two connect the dit and dah lines of the CW paddle, one connects to a potentiometer for adjusting the keying speed, and the last one is unconnected in this design, but could be used for keying a physical transceiver. (The onboard LED uses the this pin.)

            +---------------+
            |            P5 o  -- 10k potentiometer middle pin
        =====  Attiny85  P4 o  -- USB (internal)
   USB  -----            P3 o  -- USB (internal)
        -----            P2 o  -- dah paddle
        =====   78M05    P1 o  -- (LED/TRX)
            |            P0 o  -- dit paddle
            +---o-o-o-------+

There is an extra 27 kΩ resistor in the ground connection of the potentiometer to keep the P5 voltage > 2.5 V, or else the DigiSpark resets. (This could be changed by blowing some fuses, but is not necessary.)

For years, the Xmas contest has virtually been my only ham radio activity. This year, after I had to take down the antenna in my parents’ yard, I put up the 40m part of a 80/40m dipole in the attic of my flat in Saarbrücken. The first tests were satisfying so I drove back to Saarbrücken for December 26th.

The contest started nice with a S9+50 signal from DR5S - of course DOK Q10 meant it was just very close. Unfortunately, things turned out much worse. The antenna wasn't as matched as I had hoped and I had to put my small MFJ tuner in. But, whatever I would tune to, the 1:1 would only withstand some second of transmission and I had to retune. The 1:1 there would pop back up, and I had to retune back to the original settings. GOTO 10.

After a bit more than one hour I gave up. The final log is here:

I don't know where the fault was, in the TRX, the tuner, or the antenna. I hope it is not the TRX.