SDR Equipment

 I decided to pull together here my experiences with this new branch of radio development (currently dotted around my website) to help spread the word about this completely different branch of radio and show you various examples of equipment.

You get what you pay for in this branch of radio and of course you'll need some sort of computer and suitable software to make it work, although some relatively expensive offerings include a built-in computer with firmware.

First the UK-developed SDR Play

click the box below to read more

 

 Next the Lime SDR

 The Lime SDR is a bit different to the SDR Play because not only is it a receiver but it can also transmit. It can run from a standard PC using appropriate software over a publicised frequency range of 100KHz to 3.8GHz. You'll notice from the pictures below it has three Receive and two Transmit input/output options as well as two separate receive/transmit connectors. It also carries a special USB connector and a socket for using an external power supply. The USB connector is the type used on external USB 2.5 inch hard drives and DVDs which carry dual USB plugs for power+data and power. Ideally external power should be used in transmit. Click either picture to read more.

 

 

 

Now, my favourite SDR the Andrus Mk1.5

This, now very scarce SDR, is described below, complete with my attempt at repairing it.

 

 This rather fine looking "Software Defined Radio" was made in Estonia and was designed for reception of radio frequencies from 5KHz to about 32MHz. It's on its way here but I bought it untested at about 20% of its usual price so whether it will work or not I don't know... It was described as "house clearance" and "can't test because I don't have a power supply". A rather risky purchase. Sure enough.. I plugged it into my computer and there wasn't a request for a driver. After a few minutes the case began to get warm so I unplugged it. The next day I removed the circuit board from its case and found that it has three LEDs inside. One lit green and so did the second. Pressing the two buttons "Reset" and "HWBE" caused the second LED to flash three bursts of five which looks to me like an error message? I measured the internal power supply (there are several) One was about 3.3 volts and another 1.8 volts but the other was trying very hard and failing to reach 3.3 volts. I checked and found the circuit drawing too much current and causing the regulator to run exceedingly hot was labelled VCC3V3ETH/AVCC3V3ETH and fed almost exclusively the ethernet controller, a chip labelled KSZ8851SNL. This fed an ethernet socket incorporating two LEDs. I removed this and also a tiny choke feeding the two ETH supplies. Connecting a power supply to the ETH circuit proved it to be consuming 1.5amps at 1.5 volts with the chip running too hot to touch. Maybe someone connected a cable carrying DC (power over network cable) and blew up the chip? The chip is dreadfully difficult to remove and just as tricky to fit a new one. See the picture below..

Without the ETH supplies connected, I plugged in the USB cable and after pressing Reset and HWBE the third LED lit and the computer requested a driver...

click the picture to see its schematic

 

 Although I've swapped lots of surface mounted chips this one will be the trickiest so I've ordered a new rework station. Up to now I've struggled with a normal soldering iron but the so-called QFN package with 32 pins is not going to be easy. It's size is 5mm square x 0.85mm high and the 32 pins are on the underside of the chip. Some QFN chips do not have side access to their pins but this one may because it's defined as a Micrel MLF package. Above, you can see the code number in full (often this isn't the case and you're left puzzling over exactly what it is from the sparse information printed on it). In fact the KSZ8851SNL is no longer made because the KSZ8851SNLI (the version with a wider temperature spec of -40/+85C) has displaced the 0/70C version.

Why did the original fail? As it connects to an ethernet network it might have suffered from a high voltage from lightning or even an experimenter playing around with Power-over-Ethernet. I removed the ethernet socket in case this was responsible for the problem. This is coded TE-6-6605851-1 (click to see the spec) and, because it includes some parts I'll need to check it out for damage before refitting it.

 

 
 The rework station arrived and I detached the faulty chip without any trouble. To see if the chip had been preventing the board from booting up I plugged the board into a USB port. Two green LEDs came on but it was only when I pressed the Reset and HWBE that one of the green LEDs went out and a red one come on that it was detected by the computer (exactly as before). My guess is the ethernet chip is checked by the main processor and failure to respond will prevent completion of bootup.

 

 

 Before fitting the new chip.

As there was a decent quantity of solder present on the pads I didn't add any fresh solder but a good smear of flux was added before applying hot air at around 380C to the board to preheat it especially the solder on the centre pad which I checked was fluid before accurately placing the new chip above its pads.

 

 
 Once the chip was placed in position with its marker dot correctly orientated I continued to apply hot air. At no stage can you see whether it has been soldered correctly as the pads are mostly underneath the chip. After around 10-15 seconds I removed the hot air gun, waited for the board to cool then refitted the ethernet socket which I'd removed for access then refitted the inductor removed to break the short-circuit on the power rail. Just a few spots of flux to remove and the radio looks as good as new...

 Now came the proof of whether (a) the chip has been soldered properly and (b) whether this was the only fault...

The board is powered via a mini-USB lead fitted with two USB plugs for USB3 sockets. Because the SDR program deals only with the Andrus via its ethernet adaptor I also plugged in a patch lead. Plugging in the mini-USB lead made the computer respond which it had failed to do previously. With the ethernet chip removed I had to press the reset button to enable the board from being detected and at this point it couldn't find the right driver. The aerial plugs into the inner BNC socket and then I opened SDR Console and searched for and selected Andrus... whereupon it was detected correctly, and starting the receiver proved all was fine. Radio 4 came in loud and clear with 40m noisy but full of stations.

 

 Here's a couple of screenshots, the second an 80m QSO between G3OQD and G4HJW..

 

 

  The Adrus SDR is best operated using its RJ45 interface because this enables it to be remotely located and minimise local interference. There are two means of supplying power.. first via its mini USB connector and secondly via its circular coaxial power socket. Ideally the former should use a USB lead having only power connections. Commonly these cables are supplied with external DVD or hard drives where a Y-cable splits their required power between two USB plugs, of which only one carries signals. What's the power requirement of the Andrus SDR? I measured this using the power plug and found that about 350mA was drawn at 5 volts for a few seconds, then once the RJ45 LEDs came on this increased to about 550mA at 5 volts.

Add-on Hardware 

 Now some ideas about practical aspects. Anyone trying out an SDR will be rapidly become aware of its drawbacks as well as its fascinating features. Most radios are designed to be tuned precisely to a broadcast of interest and simulateously "detuned" from broadcasts that are of no interest. Most SDRs are wide open with no tuning other than their local oscillator (except perhaps some sort of passband filter). Ideally then, to make the most of an SDR you should use an appropriate filter in the aerial lead. For example, if you're listening to amateur radio broadcasts you do not want interference or degraded reception due to strong local Band II FM transmissions or a strong local medium-wave transmission.

Another add-on, which mainly concerns the Lime SDR, is the use of a converter. The Lime SDR is optimised at frequencies in the VHF and higher bands and is pretty poor at MF and HF. To receive frequencies below say 30MHz an up-converter is recommended and if you'd like to transmit below 30MHz a downconverter is desirable.

 VLF Reception

 

 Frequencies below 100KHz are quite interesting although nowadays these can be plagued by interference. Many transmissions are associated with the Military and are relatively narrow band. These can be identified by their musicality, if that's a sensible term, rather than the random hash normally associated with chopper power supplies.

click the picture to read more

If you don't have an SDR you can use your PC sound card instead!
 

 To get the best reception at VLF you need a decent aerial and the easiest to make will be a long "random" wire. Unfortunately this will pick up any strong local signals outside the band of interest so you can make a low pass filter to improve matters. Without a filter you may get unwanted AGC action which reduces signal strength or cross modulation resulting in jittery reception or even speech appearing on say a teletype broadcast.

click the picture to read more

 

A simple Upconverter for 0 to 30MHz 

 

 

 Here's a converter that transposes zero to 30MHz to the range 50-80MHz. I chose that VHF range because it's relatively free from strong broadcasts.

click the picture to read about it

 Transmit with a Lime SDR

 

 Read about transmitting experiments with a Lime SDR

 

 Reject strong FM broadcasts

 

 If you're close to a Band II transmitter you can easily build this filter

 

 Build a Loop Aerial

 

 

 To minimise interference and to get really excellent reception you can build a loop aerial. This can be located at a point furthest away from local interference and, because it's directional, it can be rotated to get the best results. I carried out a few experiments and found medium-wave varicap diodes are ideal for the job and still available. Essentially one of these diodes can take the place of a 500pF tuning capacitor and can be adjusted remotely allowing you to tune a station from in front of your PC.

My example has three large coils selectable via a couple of miniature relays so I can select and tune Long Waves, Medium Waves and the 80m band. Connection between the operating position and the aerial is via multi-way CAT5 cable which gives me enogh conductors for coil selection, varicap tuning, power and RF output.

Another solution which I haven't tried is to use a ferrite rod, again tuned by a varicap diode. This would be a lot smaller than the loop and could even be mounted high up on a cheap rotator.

click the circuit to read more
 

 Return to Reception