Homebrew Software to Control Multiple HPSDR Radios either Locally or from a Remote Location

by W3SZ 6/27/2015  most recently updated 3/24/2024

C# application running under Windows 10. It will control up to 8 HPSDR radios simultaneously. It is interfaced with N1MM Logger Plus.

My main interests in Amateur Radio are VHF/UHF/Microwave weak signal operations, and Software Defined Radios. This project involved setting up an SDR system so that the system appears to the operator as a single radio (well, two radios, actually: one Main Radio and one Liaison [Auxiliary] Radio, operating so one can both transmit and receive on two bands at once) covering all of the Amateur bands from 50 MHz through 24 GHz, inclusive.

The goal of the project is to be able to run, either from a single local computer or remotely via the ethernet, multiple software defined radios [SDRs]using the same screen/keyboard/mouse and also the same microphone/footswitch/speaker for all of them, without the need for any manual bandswitching. The goal is to have individual, always-on bandscopes that simultaneously view the 50, 144, 222, 432, 903, and 1296 MHz bands, as well as a selected microwave band on a seventh SDR that covers each of the bands 2.3 through 24 GHz inclusive, to have another bandscope/controller for the HF bands, and to have the "radio be the bandscope". This minimizes the operational complexity of the system, by allowing the operator to instantly switch the "active radio" frequency to be that frequency on which an interesting signal has been spotted, on any one of the eight "bands", by clicking on the appropriate radio/bandscope. Also, as noted above all eight SDRs are appear as two radios (SO2R-style, but with the capability of both transmitting and receiving on two bands simultaneously) to the N1MM+ logging program, and for N1MM+ to automatically select the correct SDR when the operator switches bands within the N1MM+ program. The screen grab above shows the system in operation. 144 MHz is selected as the main radio, and 1296 MHz is selected as the liaison (auxiliary) radio, and these bands are displayed as large, full-function windows. The N1MM+ entry window for each of these two radios is situated immediately to the right of the corresponding radio window. The other radio bands are not selected, and so they are being monitored in the smaller bandscopes visible on the screen above the two selected radios.

So the paradigm is to have 8 radio GUI-bandscopes, one for each "band" as defined above. When not selected by the operator, the individual bandscopes, as just noted, appear as smaller windows as you have seen in the image above. When selected as either [1] the main radio or [2] the aux (liaison) radio, the selected bandscope will enlarge and move to the preselected screen position for the main or aux radio, depending upon which spot was selected for that band, as seen above, and all of the necessary radio controls will appear for that radio on its enlarged GUI.

This software can be used either with the operator at the location of the radios or with the operator at a remote location connected to the radio site by an ethernet link, either wired or wireless. In either case, the operator interacts with the client software which interacts with the server software for each hardware radio.

The software I've put together to accomplish this is based on the fantastic KISS Konsole software by Phil Harman, VK6PH, which works with the excellent HPSDR hardware. Phil is one of the gurus and leaders of the HPSDR project. I wrote/modified/compiled the software using Visual Studio 2013 (and later, Visual Studio 2015) which is now a free product. The software was initially targeted at .NET Framework 4, and was begun on Windows XP. But because there were features that I wanted to use that required Framework 4.5, the project was moved to Windows 7 and then 8.1, then 10, and finally Windows 11. It will no longer run on Windows XP, but will run on Windows 7, 8, and 8.1, 10, and 11.

With the client software I wrote, there are 3 ways for the operator to select which band (radio) is active as either the main or the aux radio. The first way of selecting a band for operation is by mouse-clicking on the spectrum/waterfall of the small bandscope for a given band. Left-clicking will move the radio to the main radio position. Right-clicking will move the radio to the aux radio position.

The bands to be selected as main and aux radios can also be selected from the Main Controller Bar, which as you can see in the image above, is placed at the bottom of the screen. It is shown in more detail below.

Clicking a "Radiobutton" for a band on the row of the bar labeled "Main" will bring that radio to the Main Radio position on the screen, and the expanded radio window thus exposed gives access to all of the main radio controls. This action will also make this band the main radio in N1MM+. Similarly, clicking a "Radiobutton" for a band on the "Aux" row of the bar will bring that radio to the Aux Radio position on the screen, and the expanded radio window thus exposed gives access to all of the aux radio controls. This action will also make this band the aux radio in N1MM+.

The third way of moving a radio to the main or auxiliary radio position is through N1MM+. Typing a frequency into the main radio entry window of N1MM+ will select the appropriate radio and place it in the main radio position on the screen and put it on that frequency. Similarly, typing a frequency into the aux radio entry window of N1MM+ will select the appropriate radio and place it in the aux radio position on the screen and put it on that frequency.

As you can see above, on the right side of the Main Controller Bar are Radiobuttons in both the Main Radio and the Aux Radio rows for "L Foot" (the footswitch), the Microphone, and the (CW) Key.  If the "L Foot" button on the Main Radio row is selected, then the footswitch will be assigned to whichever radio is assigned as the main radio. Similarly for the "Mic" and "Key". Of course, if the "L Foot" Radiobutton on the Aux Radio row is selected, then the footswitch will be assigned to whichever radio is assigned as Aux Radio. For the operator using N1MM+, the experience is that of having a single radio with always-on bandscopes for all of the bands described above that covers every band from 50 MHz through 24 GHz, inclusive, with automatic bandswitching not only of antennas, but also of footswitch, microphone, key, and receive audio. To the right of the "Key" Radiobuttons are additional controls. The leftmost trackbar sets the audio level for the transmit audio monitor. To the right of this trackbar is a GroupBox labeled "N1MM+". Checking either the Main or the Aux "Tx" box will sent N1MM-generated audio to the respective radio. Checking either the Main or the Aux (or both) Rx box will send the receive audio from the respective radio[s] to the N1MM+ add-on QSOrder. In this manner all contest audio or specific QSO audio or both can be saved. To the right of the N1MM GroupBox is the Digital Groupbox. The checkboxes serve similar purposes to those described for N1MM, but in this case they send received audio from the selected radio to a digital program such as WSJT-X, WSJT, etc. and they receive transmit audio from WSJT-X or WSJT, etc. and send it to the appropriate radio.  Each radio can be connected to a separate instance of WSJT-X, and this software can transmit simultaneously on the Main Radio using one instance of WSJT-X, and on the Aux Radio using another instance of WSJT-X.  The illustration at the top of this page shows the software simultaneously calling CQ using MSK144 on both 50 MHz and 144 MHz. To the right of the Digital GroupBox is a GroupBox labeled "Mute". Clicking the appropriate checkbox in this GroupBox will mute the speaker/headphone received audio for the selected radio[s]. It will have no effect on audio being sent to N1MM+ or to WSJT-X, WSJT, etc. To the right of the Mute GroupBox is a trackbar for adjusting the audio level of the Main and Aux radios. Additional volume controls are provided on each radio, but this provides a convenient place to adjust the global received audio level.  To the right of this trackbar are checkboxes to activate CAT control for the Main and Aux Radios, and to the right of those controls are buttons to reset the Mic and Receive audio steams.  The Setup button is above one of these buttons, and brings up the setup form that will be described below.  In this setup form the bands to be displayed can be selected and many other parameters can be selected or changed. If a band has not been selected for display, then it will not appear on the control bar, and its radio GUI will not be displayed.

I placed a video on YouTube that shows me running the client software through its paces. The software has been updated and expanded since this video was made, so this video does not show all of the features currently available with this software.

The video is here.

Basically, this software makes the following changes/additions to the original KISS Konsole software:

1. Splits KISS Konsole (written by Phil Harman VK6PH with important contributions by others) into a server and a client, designed for multiband VHF and microwave use.
2. The client controls multiple [up to 8 as configured] HPSDR Radios/servers:
50 MHz
144 MHz
222 MHz
432 MHz
903 MHz
1296 MHz
2-3-5-10-24 GHz
HF.
3. The server sends Spectrum/Waterfall/receive audio data to client.
4. The client controls the server, to provide remote control [via ethernet] of all important radio functions.
5. The client interfaces with N1MM+, appearing to N1MM+ as one multiband radio covering HF through 24 GHz.
6. The client uses either a wired or wireless ethernet connection to send CW key input, microphone input, footswitch input from the computer at operator's site running the client software to the radio servers at the remote site. Bandwidth required for both microphone audio and receive audio is substantially reduced through use of the Opus codec.
7. The 2-3-5-10-24 GHz radio server switches transverters-antenna combinations among radios for these bands based on band data sent to this server by the client.
8. Added Adjustable FFT sizes for the graphic displays from 4096 to 524288 to server.
9. Increased receiver audio FFT size from 2048 to 32768 to provide for steeper filter skirts.
10. Added "Wisdom" optimization of FFT calculations, which is performed the first time program is run.
11. Added multiple waterfall palette options to client.
12. Added computer receive audio so that audio can be obtained without connection to radio headphone jack, via Windows sound (available at either the server computer if the server is used in stand-alone fashion as it might be for testing/single-band operation, or at the remote computer if, as usual, the client software is used).
13. Added zoom of spectrum/waterfall to both server and client.
14. Added CW sidetone to to both client and server. CW input can be either from a straight key or WinKeyer output, or with typed CW from N1MM+ (which usesWinKeyer).
15.Up/Down arrows will move frequency up/down by "Step Size". This allows the ShuttlePRO v2 to act as a "knob" for frequency control.
16. Step Size can also be set from the ShuttlePRO. The software is set up with the following key assignments: a=1 Hz; b=10 Hz; c=100 Hz; d=1 kHz.
17. Mode can be set from the ShuttlePRO. The software is set up with the following key assignments: e=FM; f=USB; g=LSB; h=CWU; i=CWL.
18. The client will "kill" transmit automatically if , while transmitting, either the Main or Aux radio is deslected from its Main or Aux position. This feature is included to make the system more "fool resistant", this necessity becoming apparent as a result of extensive contest use, sometimes after little or no sleep for 24 hours or more.
19.Direct digital audio input from N1MM+ or other contest logging programs is provided, as is digital PTT. Ability to send digital audio from the HPSDR radios directly to N1MM+/QSOrder is provided.
20. Direct digital audio capabilities to and from digital programs such as WSJT-X, WSJT, etc. and digital PTT from these programs is provided.  The Main and Aux Radios can each run a separate instance of WSJT-X, and both can transmit simultaneously if desired.
21. Monitoring of transmit audio from the microphone, N1MM+, WSJT-X, etc. is provided. Also provided is monitoring, at the client site, of the output of the transmitter at the remote site via the receiver portion of the server program, whose audio is sent to the client program via the ethernet as noted above (full duplex mode).
22. Addition of frequency-domain frequency-specific AGC, because the usual time-domain AGC is poorly suited to use with digital signals, and because some AGC is unfortunately necessary with the nowadays usual combination of extremely strong local signals and very weak distant signals in the same audio passband.

The radio and computer hardware used with the current system is similar but not identical to that used with the older system that this system replaced and which is no longer in use:

The client software runs on a 4.2 GHz I7-5770K when I am operating from the remote location where the radios are. At home, my main machine is an I7-4790K running at 4.0 GHz, and so I use that to run the client software as a matter of convenience when operating from home, via a dedicated wireless link.

50 MHz SDR: HPSDR Atlas/Ozy/Penelope/Mercury with the server running on a 3.2 GHz I5-4460
144 MHz SDR:HPSDR Atlas/Ozy/Penelope/Mercury stack running on a 3.1 GHz Core 2 Duo
222 MHz SDR: HPSDR Atlas/Ozy/Penelope/Mercury stack running on a 3.1 GHz Core 2 Duo
432 MHz SDR: HPSDR HPSDR Atlas/Ozy/Penelope/Mercury running on a 3.1 GHz Core 2 Duo
903 MHz SDR: HPSDR Atlas/Ozy/Penelope/Mercury stack running on a 3.6 GHz Intel I3
1296 MHz SDR: HPSDR Atlas/Ozy/Penelope/Mercury stack running on a 3.1 GHz Core 2 Duo
2.3-24 GHz SDR: HPSDR Atlas/Ozy/Penelope/Mercury stack running on a 3.1 GHz Core 2 Duo

I used to run HPSDR Hermes units on 50 and 432 MHz, but the Hermes was extremely glitch-prone, whereas the stacks that I am currently running are free from this problem.

The individual HPSDR computers running the server portion of the software are linked to the client/controller portion of the software using a Gigabit Ethernet switch, and all of the servers are running Gigabit links with jumbo frames capability. This bandwidth capacity was necessary with the old OSX system, but is far greater than is needed with the new system. With the new system I only need to run a 100 Mbps NIC on the client computer, and only 0.5% of the 100 Mbps bandwidth is used by the system, as you can see in the image on the left. This low bandwidth was achieved by using the Opus codec to substantially reduce the streaming bandwidth. With two-radio receive audio streaming without the use of Opus, the bandwidth used is 2.2-2.5% (2.2-2.5 Mbps). With the addition of the Opus codec, each audio channel takes less than 30 kbps, and the total receive audio bandwidth for two radios is only 50 kbps, on the order of 2% of the bandwidth without the use of the codec. In fact, when the Opus codec is used, the total ethernet bandwidth for the entire program with its control signals, 7 bandscopes, and simultaneous two radio receive audio streaming is only 0.5 Mpbs, as noted above and as shown in the illustration above.  The Opus codec also provides Forward Error Correction (FEC) of the audio streams.

The network connections between the client and server are set up using setup menus contained in the client and server software. The setup menu shown here and labeled "Server IP" is one of six menu pages available from the Main Controller Bar of the client software. With this page the user [1] sets the IP address for the HPSDR-server for each band (radio), [2] selects the radios that will be displayed, and [3] assigns the COM ports that are used for various functions. Assignment of the IP addresses for each radio server is self-explanatory. The virtual COM ports provide for WinKeyer Control, CW Key/Keyer and Manual PTT input, N1MM+ CAT control for the main and auxiliary radios, WSJT-X CAT control for the main and auxiliary radios, and PTT control from N1MM+ as well as the WSJT-X instances associated with the main and auxiliary radios.

The second setup screen available from the Main Controller Bar of the client software is labeled "Transverters", as shown here. It allows the user to set the LO frequency and offset for each transverter. Because my transverters are GPS-locked on all bands, the LO offset is consistently zero. The third setup screen, not shown here, allows the user to set the transmitter Band Gain for each HF, VHF, UHF, and microwave band. I have a separate, software controlled HP attenuator that is used to set drive levels and receive attenuation levels automatically for each band, to provide optimal transmit drive for each band and a consistent noise floor on receive for each band, and so the third setup screen is not needed in my current setup. It is however maintained in case of future need.