Friday, April 09, 2021

First Impressions of the Elad FDM-S3

I've had the FDM-S3 a couple of weeks, and physically it is a totally different beast than for instance the SDRPlay RSPdx or Airspy HF+. For better - and for worse. Here are my first impressions of the Italian flagship. 



Friday, April 02, 2021

The Bodnar Mini Precision GPS Reference Clock: A Perfect Tool for the Perseus SDR

Older SDRs aren't very stable. Well, compared to old boatanchors and even newer conventional receivers, +/- 1 ppm is excellent. But when keeping track of the offsets of stations is an important factor in MW DX-ing, we do not want the radios to drift. The Perseus SDR, though a magnificent radio in so many aspects, does drift. The stability vs. ambient temperature is +/- 1 ppm, while current SDRs do a lot better, 0.01 ppm or even 0.001 ppm with an OCXO and GPS reference.

At the KONG site, ambient temperatures routinely vary between 10 and 20 degrees C, and in extreme cases (also this winter), down to almost freezing. The Perseus drifts a lot during these temperature changes. We have seen 6-7 Hz drifting over a 10-15 degrees temperature change. Keeping track of offsets becomes challenging and manual calibration on known stable stations is often impractical.

The Bodnar Mini Precision GPS Reference Clock has gained some fame in the DX and radio amateur world for being a relatively low cost, high accuracy tool for reliable frequency readout. With it, you can insert a reference signal to your SDR.

There is no input for a reference signal in the Perseus SDR though. And since we only sample the MW band, we needed the reference signal to be present within the bandwidth we sample. The signal output from the Bodnar is +6 dBm even at its lowest setting, which would need some serious attenuation of the signal to avoid saturation. And with nine Perseus SDRs, did we have to buy nine devices?

No, we didn't. The nine SDRs are fed from four splitters, one for each beverage feedline. With the high signal level of the reference signal, it is in fact easier to wrap a wire from the signal output around the feedline. Induction rules, and the signal level is a lot lower. So, we still had to buy four devices?

No, we didn't! With such a strong reference signal, and the feedlines physically very close to eachother, we ended up with having to buy only one! A few turns of wire around each feedline provided us with a GPS reference on 1705 kHz for all our four beverages and nine Perseus.

It should be noted though, that our solution is based on the Jaguar software for Perseus - Jaguar enables us to manually or automatically calibrate readout.

The device needs a 5V power source, which I found on a spare USB port on one of my PCs. The connection also enables communication with the device itself. The only other requirement is of course access to open sky for the GPS antenna.







Saturday, March 27, 2021

Elad FDM-S3 - Very First Looks

 So I decided to buy one. At 780 EUR + Norwegian VAT I could easily buy three RSPdx for one S3. And probably fit six of not more of them inside the S3 cabinet. But it was intriguing to test a proper 16-bit ADC SDR with HF and FM coverage and 24 MHz sampling rate.

The money drain didn't stop there. To be able to sample 24 MHz, none of my PCs are even close in performance. So I bought a Lenovo ThinkCentre desktop which set me back as much as the S3. The all-cores passmark is 13000-ish so it's reasonably quick.

The S3 arrived two days ago, one day after the Lenovo. I set up the combo at my home here in Tana, where I have no antennas, for testing performance and compatibility. They seem to work very well together. I have been running the S3 with SDR Console at 24 MHz sampling for 30 hours now, with no issues at all. CPU usage is quite stable around 34 %. During IQ recording, the load increases to 40 %.

During the Easter holiday I expect to have lots of time to explore the SDR in depth, including doing the mandatory MW sensitivity measurements. Below some photos from unpacking and first run.

S3 in blue colour, "matching" USB 3 cable + Powerpole DC cable. No PSU supplied but three SMA to BNC adapters were most welcome.



Yikes! Boatanchor!



S3 running with a headless Lenovo ThinkCentre (behind), remote-controlled with the Dell laptop.




Tuesday, March 02, 2021

Winradio G31DDC - Now With SDR Console

July 2010: I received my first Winradio G31DDC from Waters & Stanton in the UK. I was looking for an SDR at least as good, or preferably better than the Perseus SDR, and this was it. Armed with a 16-bit ADC and slightly better sensitivity, it was Winradio's attempt to cater for the DX community rather than professional customers. 

I didn't fancy the graphical user interface though. And what's worse, for one used to making IQ recordings, both live and scheduled, its software failed miserably. First, the IQ files could not be date- and timestamped. Why on earth....? Second, the scheduling option was audio ONLY! Why, oh why.

Later on, the G33DDC appeared which solved these "issues", but at more than double the price. The GUI itself was still not what I wanted, and no third-party software was in the pipeline. Then came the Jaguar software for Perseus, a giant leap for DX-ers who want to schedule, record and analyze IQ recordings. So, both Winradios were sold.

Fast forward to 2021: SDR Console and Simon Brown decided to work on the G31DDC. By coincidence I had a G31 belonging to an estate, waiting to be sold, and decided to wait out Console support. I got a beta version in mid-February, and it worked well, although without an antenna. When I drove to the KONG HQ last weekend I brought the G31DDC with me, and got it playing with a beverage antenna. 

Let there be no doubt: This is a first class SDR. Console is a very good overall SDR software, and the combo works extremely well. Since I DX the MW band almost exclusively, the Perseus/Jaguar combo will remain my favourite setup. But for most DX-ers, the G31DDC + Console combo is hard to beat. Not least because it's an active product from an active manufacturer who is in the radio game for the long run. 

I just had to buy the G31DDC.

And support for the G33DDC is nigh!



Friday, January 08, 2021

6 Watts, 1485 kHz, 2540 km

There's been a few DX tests this season, and rather by coincidence I stumbled upon another one the other day. The IRCA and MW Circle reflectors reported that Mittelwellen-Sender Joe, transmitting from Erlangen in Germany on 1485 kHz, was sending periodic 1000-Hz tones - one second on, one second off.

Now, that sounded like a nice challenge. The first evening I tested I heard the transmission right away at the KONG HQ, and a very good signal at Smøla. DX-ers in Scoland and Ireland also reported the signal. The initial enthusiasm faded a bit when I later learned that the effect was in fact 100 watts, not 6. 

What I hadn't noticed from the reflector posts was that from 17:00 to 05:00 UTC the power is 100 watts, the rest of the day 6 watts. I had been listening in the evening.

So I decided to try again, and focus on the period before Germany's sunrise. I monitored the signal from 04:58 (100 watts), and then at 05:01:35 the relatively strong signal disappeared...almost! A very faint signal was still there, and I got my 6 watts, 2540 km reception! The image below shows the weaker tones above the red line as the signal dropped.

Perseus/Jaguar recording, SDR Console playback

I chose the 540-metre Asia/Pacific beverage antenna for the reception, because its back lobe crosses western Europe. Interference mostly from SER Radio Santander.

More informasjon about Joe can be found here.

Ordinary programming may resume soon, so I don't know for how much longer it's possible to hear the tones 24h. The tone will be broadcast from Midnight to 05:00 (presumably UTC), but at what power I don't know.

I really think this 1000-Hz test should be able to travel quite far at 100 watts. The eastern shores of North America should be possible. At least worth a try.


Saturday, January 02, 2021

DX Test From WCPH Etowah TN 1220 - Heard!

WCPH is a 1-kW day, 109-watt night oldies station in the south-east corner of Tennessee, not far from Chattanooga. The station owner George Hudson agreed with the NRC/IRCA Courtesy Program Committee to set up two two-hour DX tests with their day facility. The first on December 26, the other on January 2, both tests running from 07:00 to 09:00 UTC.

1220 is certainly not the easiest frequency with powerful stations like WHKW Cleveland, CJRB Boissevain and often XEB Cd. Mexico with potentially strong signals. 1000 watts isn't much to compete with. But we know that morse code and sweep tones are extremely efficient methods if you want to reach far.

Reception of the December 26 test was a failure. Geomagnetic conditions were very poor. If I remember correctly the only signal we heard was a very faint CJRB.

Propagation forecasts for January 2 were far more optimistic though. And true enough, at 07:00 we did hear the sweep tones on our Smøla island rigs. Later on, sweep tones and morse IDs were heard at the KONG site, and with quite a bit better signal than Smøla. At KONG, there was little interference except CFAJ fading up a few times, and CJRB/XEB lurking in the shadows.

The distance to Smøla is 6660 km, to KONG 7128 km. WCPH was also heard in Lemmenjoki, Finland, and Scotland. And maybe elsewhere in Europe for all I know.

Many thanks to the CPC and George Hudson for making this opportunity to catch a new station! I hope George is satisfied with how his station got out. A good start for 2021, DX-wise!



Thursday, December 31, 2020

Zen And The Art of Antenna Maintenance

It's a while since I read the original book with a slightly different title (1978 perhaps?), but the title struck me as relevant when I was heading out in the nautical twilight to inspect some of our beverages. Very little wind, -4 Celsius and almost no snow - evidence of this extremely mild last quarter of the year. Not often does the weather permit working with bare hands.

Anyway! A few photos from my inspection follow below. A Happy New Year to everyone.

My house, the KONG HQ, in the foreground.

Turning the camera a bit to the NNW where the 310 beverage ends.

310 beverage starting point.

Inside the box, antenna, ground, transformer and feedline all ok.

310 beverage end point.

50 beverage starting point (just 15 meters from the 310 end)

Community and my travel powered by nearby renewable energy


Tuesday, December 29, 2020

QSL: KABN Kenai AK 960

I was alerted by OJ Sagdahl earlier in December about a "Moody Radio" on 960, apparently being KABN Kenai, as suggested by Paul W. Walker. The signal was heard in a typical "Alaskan part of the day" i.e. after our noon, and a "Cabin Radio" ID indeed confirmed that a new Alaska station was heard. A very friendly email response confirmed my 40th Alaskan station. Its license was granted on December 3. It's running 1 kW ND and is not particularly easy to hear due to the massive signal from CFAC.

But is it new? A long time ago, KZXX was broadcasting from Kenai on 980. It went silent, and the last time it was heard at our KONG HQ was in 2001, apparently active only for a short while to keep the license. The logging was met with some skepticism from other DX-ers.

Anyway, further information from KABN confirms that 960 is now using the 980 facilities. So, for those who have a previously confirmed reception of KZXX 980, KABN may not count as a "new" station. But what does it matter. A new DX-able station on the MW band is not a bad thing.

The KABN call was chosen because it had an Alaskan "touch", and indeed there was another KABN, on 830 from Big Lake north of Anchorage in the 1980's, and it too used the "Cabin Radio" moniker. My hand-written QSL from that station was from 1987.


And before you know it, maybe another Alaskan MW station will hit the airwaves...

Thursday, December 17, 2020

The KSEN Shelby Montana 1150 DX Test

Note to self: Maybe I should post more often...

As many of you know, KSEN ran a two-night DX test on November 7 and 8. I have heard (and QSL-ed) KSEN on several occasions before, but I try to never miss a DX test.

November 7: Despite reasonably good conditions, in the beginning I heard nothing at all. Several other stations came and went on 1150, CKOC, WHBY, KKNW and CJSL were noted. Several North American DX-ers reported nothing heard, so I wondered if it was on at all. Then, at 07:35 I could hear occasional sweep tones, and finally just before 08:00 a morse code ID. No voice audio was detected. Why so weak? Well, it turned out that they had technical issues resulting in an output of only around 800 watts! So, they didn't reach out very well at all that day.

The reception was made with a Perseus SDR connected to a 340-metre dual (staggered) beverage antenna. The distance was 6402 km, not bad for 800 watts.

On November 8 everything was working fine, and reception was good at my spot. Today I received the DX test QSL. Thanks to Paul Walker, Les Rayburn and Joe Miller for their time and trouble.



Saturday, November 14, 2020

The RX888 SDR

(Updated DEC-5 re. loss of connection after PC reboot)

Late summer I became aware of the RX666, a Chinese build of Oscar Steila’s BBRF103 design. Basically, a design with a 16-bit ADC capable of sampling 32 MHz, and with a R820T2 chip “thrown in” to allow reception up to 1800 MHz. I did a write-up which you can read here. The antenna connectors were not up to standard, so after a while I returned it for a refund.

The RX888 is very much the same design, but with a smaller PCB and apparently better build quality. Unlike the RX666, the RX888 has an LNA built-in for better sensitivity.

As a MW DX-er I’m mostly interested in how it performs on MW. And for that, three things matter more: Sensitivity, large dynamic range, and a minimum of 1200 kHz alias-free sampling. And of course, software to match.

The RX888 footprint is quite a bit smaller than the RX666. It measures 72 x 83 x 43 mm (excluding antenna connectors) and weighs in at 160 grams – less than half that of the RX666. Heat dissipation panels make up three sides and of course adds to the physical size.


A 16-bit ADC (Analog Devices LTC2208 in this case) is always better than 14-bit or 12-bit ADCs, especially when negotiating the huge signal levels on the MW band. In that respect, the RX888 performs very well indeed, even if the only filter is a 64 MHz low pass filter. The downside, as with the RX666 is sensitivity. Although the RX888 has an LNA built-in, sensitivity figures are still in the lower part of the list, 3-4 dB better than the RX666. I measured -95, -98 and -99 dBm on 500, 1000 and 1500 kHz (AM mode, 6 kHz bandwidth, a 400 Hz signal with 30 % modulation). This is 5-6 dB worse than the Perseus, and 10-12 dB worse than the Airspy HF+. The good part though, is that the ADC is quite forgiving if you decide to use external amplification. A 16-dB preamp from DX-Engineering increased sensitivity by 13-14 dB, which is a massive improvement. I did test this setup on a very high signal level MW band, fed from a beverage antenna, and noted no overloading. If you do, you can always reduce RF gain by 10 or 20 dB.



Available sample rates when using the dlls developed for HDSDR are 2, 4, 8, 16 and 32 MHz. In addition, there is support for 64 MHz sampling rate. SDR Console also supports the RX666 and RX888 and offers a 1-MHz sampling rate in addition.

It does require a USB 3.0 port on the PC, and of course the wider sampling rates, the more it demands from your PC. Since the Cypress driver was already installed for the RX666, the RX888 connected right away.

Alas, to my disappointment both the RX666 and RX888 suffer from a serious flaw: After a PC reboot, the devices will not restart. You need to disconnect and reconnect the USB cable for the PC to recognize the devices. This may not matter too much if you always sit in front of the PC. If you want to control the PC via remote control, my only advice is: Don't restart the PC.

So, while the Perseus + Jaguar combo is still my choice for MW DX-ing, the RX888 appears to perform very well if some external gain is added.

Ebay prices are in the USD 230-250 range, not much more than the SDRPlay RSPdx but a lot more resilient to large signal levels and only slightly less sensitive. And not very much more than the Airspy HF+/Discovery, but able to sample the entire MW band.