Resolving Softrock Ensemble RXTX transmission problems #3

After the past attempts to get the system working successfully there’s been a bit of a lull in activity due to Christmas and the pressures of modern day life!
By way of a recap, the theory of the fault finding and proving of this system has involved breaking down the radio to its component modules on the system diagram, checking the functionality in various transmission modes, then adding modules back in and repeating until hopefully a functional system is achieved.

Radio setup variations
So far the attempts to date have involved trying to get the SoftRock and HDSDR to transmit a clean CW signal. The test rig receiving this signal has been a Realistic DX-440 linked to a SoundBlaster Live 24 USB soundcard feeding into Fldigi running on my Dell Netbook.

A lot of time has been spent trying to fathom out why the received signal on the waterfall within Fldigi is about as straight as a banana and it was at this point I’d just about given up.
This time the test instrumentation and encyclopedic knowledge base came courtesy of Andy G7UHN, who gave up an afternoon of his time, for which I will always be in his debt.
Using Andy’s Elecraft K2 hooked to a Tigertronics Signalink feeding into Fldigi running on a netbook, the CW signal produced by HDSDR is clean and straight. The DX-440, as good as it was in principle is, is causing the bendy signal and all of the confusion!
Anyway, in a Top Gear style, moving on, the next step was to test a voice transmission from the SoftRock/HDSDR rig.

By adding a cheap microphone to the onboard Realtek soundcard microphone input an SSB audio test call was made and monitored on the K2 using a set of cans. After a little tweaking of levels, voice was received load and clear! Huge steps forward in all of 20 minutes!

Now the next step was to calibrate the RX and TX channels within HDSDR. Bearing in mind you get absolutely no help in HDSDR as to what this actually entails, other than a very bland dialogue box buried away in a sub menu, it transpired not to be as onerous as I’d thought.

Screenshot 2015-02-22 14.42.38

The K2 generates a very nice tune signal which was used as a reference. Checking the waterfall clearly showed a spurious signal at 10 kHz down frequency. By adjusting the amplitude and phase raw and fine sliders this was reduced to zero.
One down, one to go! The next head scratcher was what to test the SoftRock’s TX against. The advice from Guru Andy was to test against another SDR to give a comparable bandwidth of radio spectrum. But I don’t have a second SDR? Oh yes you do! My first foray into the world of radio was playing with RTL-SDR dongles and using one as an ADSB receiver to pick up the transponder broadcasts from passing aircraft as virtual radar.

Take said dongle, plug it into an HF upconverter courtesy of Nooelec and Robert’s your fathers’ brother!

Thankfully Andy had one in the boot of his car where it had been lurking since he’d been using it at work so there was a strong possibility this problem child may get sorted in one sitting!
With a laptop running SDR# and a jury rigged RTL-SDR / upconverter / every conceivable gender changer the pair of us possessed / dummy load strewn across the bench we were good to go.

Corrections used for the upconverter and dongle within SDR# (upconverter shift = -124,997,870, R820T = 136)

The answer to the next question wasn’t so obvious “How do you key the TX button in HDSDR with the calibration dialogue and slider adjustment screen open?” As I’ve said before, HDSDR really isn’t my favourite piece of software because it’s written possibly from the perspective of a software engineer rather than a radio operator and these weird little nuances can be a pain. Anyway, believe it or not, you can control the TX functionality of HDSDR with other menu windows open, which flies against every other Windows program I’ve ever used!

Again, gentle tweaking of the amplitude and phase raw and fine sliders within HDSDR resulted in a nice clean signal and waterfall on the SDR# test rig.

Remember, this isn’t about being anal when it comes to dotting the I’s and crossing the T’s. Your licence stipulates that “the licensee shall ensure that the emitted frequency of the apparatus comprised in the Radio Equipment is as stable and as free from Unwanted Emissions as the state of technical development for amateur radio reasonably permits.”

Apart from hacking off other amateurs with your poor quality splattery signal, having Ofcom banging on your door and inviting themselves in for tea and biscuits is never good form!

At this point, everything is working! It’d be very easy to start reaching for the Cubans and San Miguel at this point as victory is close and we have a fully functioning and proved SDR radio, in effect the left hand side of the system diagram.

Radio setup variations
The next step is to get Fldigi working with the radio so digimode transmissions can be sent and decoded.
Thankfully, no one had reached for the beer yet as the minute Fldigi was run, both the TX and RX signals reverted to a complete pile of mush on the waterfall in both Fldigi and HDSR. What the hell was going on?

Fldigi uses Virtual Audio Cable to pipe its feed to and from the radio. RX input is via VAC1 and TX output to HDSDR via VAC2. The settings on both were revisited over and over in my initial fault finding months ago with no effect and we spent a good 20 minutes of educated guessing as to what was going on before we gave up. VAC is a proven product used the world over by radio enthusiasts who have documented its use. Admittedly a lot of enthusiasts are still using Windows XP and a lot of the software in use was designed on that platform. Despite Microsoft’s cross platform compatibility modes and so forth, nothing is going to be 100% comparable on a brand new high end machine running Windows 7 x64 and a Windows XP machine from several years ago. Just consider the driver differences for starters.

With that in mind we switched the TX and RX piping to VB Audio Cable which markets itself as a fully functioning virtual audio cable solution. With two cables set up, one for input and one for output things were good to go and a test call made. Guess what? It only bloody well works!

Now at this stage you’ve got two grown men giggling like school girls and wondering if what’s just happened was a fluke of technology.
Test again, it works. Test again; it works, again, again, again. It works! Crystal clear CQ calls originating from Fldigi in PSK31 and being picked up on the K2 rig sat next to it on the bench.
This thing needs to be on a real antenna and talking to people!

With the ATU plugged in and a live antenna on the back of it, all that was needed was minor tweaking of the microphone gain and Windows audio levels to give a good 1W TX and 1.1 reflected power on the ATU’s power meter.

Time to do this for real! Thankfully at quarter to five in the afternoon there was a healthy number of amateurs chatting away on PSK31 and with a little bit of patience, someone eventually calls CQ and we’re in!

First contact on my SoftRock SDR rig was with Giovanni (IN3GNV) in the Dolomite region of Italy at 1711 hrs on Monday 16th February 2015, reporting a 599 signal.

What can I say?

Has it been a pain? – Yes it has.
Was I ready to launch the whole thing out the window? – Yes I was.
How much have I learnt? – More than I could ever imagine!

This has been one of the biggest and steepest learning curves of my life, which if it had worked first time when I took it out of the box, would never have been attempted let alone conquered.

Admittedly I could never have done it alone and my heartfelt thanks have got to go to Andy for his patience, insight, advice and support. Also, things wouldn’t have been possible without my sponsor (you know who you are!) There have also been a load of people on various forums who have offered advice in one form of another, but if nothing else the encouragement and the offer that “when you get there and are on air it’ll have been worth it.”

They’re not wrong!