Tuesday, 14 January 2025

Low cost SDR update (4) One is not enough!

 Having been impressed by the receiver I could not resist building another one, especially at the price! My supplier of the A/D board had a spare board and as G4BVY also wanted to build one a receiver for skimming purposes so we combined orders

I also discovered GI4DOH had designed (after many iterations) a 3D printed box that was a custom fit to the board stack so I managed to get the .stl file from him and I printed one (took 4 hours (box and lid) on my 3D printer

The printed box was much smaller than my diecast box and was unscreened. I also could not see a way of getting a switcher PSU  in the box and heatsinking it so i continued with the external 5V PSU supplied with the FPGA board



The box also has provision for a fan. I fitted one but have not needed it yet

Friday, 3 January 2025

Low Cost SDR update (3): Multimode skimming

Having got all this data from the Receiver it obviously needed skimming. My current desktop, a 5th generation I5 would not have the power so I sarted to look around. I consulted GI4DOH what he used but G4BVY discovered an Intel I7 Gen 11 NUC at a good price so we ordered one each

When mine arrived I was stunned at how small it was! Its features were 
Intel NUC 11 NUC11PAHi7 32GB DDR4 RAM,1TGB SSD,Win 11 Pro Mini PC,Corei7-1165G7 Processor

  
Initially I decided to set it up for 8 bands of cw skimming and 8 bands of RTTY skimming. The instructions were at https://pavel-demin.github.io/qmtech-xc7z020-notes/sdr-receiver-hpsdr-77-76/  and were easy to follow. It worked fine but not much RTTY action

Skimmer server CW (8 bands) + RTTY skimmer (5 bands) +Aggregator




To stress the computer a little mor, for the ARRL RTTY contest i set it up the same way (8 x CW 8 x RTTY) and decoded a lot of RTTY signals but CPU useage never got above 35%

Friday, 27 December 2024

Low cost SDR update (2): Encapsulation

The next task was to put the receiver in a box. As I wanted to make the receiver powered by 12V I needed room for a DC to DC converter and possibly some external filtering (to get rid of 198kHz droitwich for instance). For the same reason I wanted a screened metal box

Searching the mouser diecast box lists I could not find on  just big enough but i did find one that would do, so i ordered it. I did have to 3D print some 3mm threaded spacers of appropriate lengths as the 2 boards were at different heights above the bottom of the box.

Initially I used an LT1084 Low volt drop regulator to regulate the incoming 12V down to 5V directly to supply the FPGA board and hence the Receiver board, but it got hot, not good for reliability. So I considered using just a chinese DC/DC converter to drop the 12V supply directly down to 5V. However having had these boards fail in the past and put the full input voltage at the output I devised an alternative. I used the Chinese LM2596  DC/DC converter set to produce 8V output then used an LT1084 Low volt drop regulator to produce the 5V. This way if the DC/DC failed and produced full output the LDO regulator would protect the board. This was the method eventually used.



Sunday, 22 December 2024

Low cost SDR (1)

Occasionally interesting projects do get shown on the BATC QO100 net which runs on Thursday evenings.

One such occasion in November 2024 was when GI4DOH mentioned he was using a cheap A to D board in front of a commercial FPGA board to do cw skimming and rtty skimming instead of the expensive Red Pittaya board

I had been using my ANAN 10 to do my skimming which is overkill . It is limited to 4 receivers. His system could do 16 so time to Investigate!

Full information was available at https://pavel-demin.github.io/qmtech-xc7z020-notes/ who was the designer of the  A to D board. All the necessary files were available to get the boards assembled in China. Luckily I found someone who was getting a batch made so I ordered one from him. The cost (shipped) was £58

The FPGA board QMTECH ZYNQ7000 XC7Z020 Development Board was available from Aliexpress  for around £80. It came with a 5 Volt Power supply

The only other items needed were a 2x25 pin header and socket to mate the boards,  a heatsink to keep the FPGA from overheating and an edge mount SMA socket. I found the 2x25 pin connectors recomended in the links to be too tight a fit, needing a screwdriver to seperate them (risking damaging the board) so I ended up using some single row  headers and sockets, cut to length and superglued together

The A/D PCB has a reminder to remove R14 and R15 on the FPGA board before mating the two boards and to connect 3.3V wires between the 2 boards (the positive lead on the RX board goes furthest away from the edge of the board), It didnt take long to assemble.



 Software was easy. Format an SDHC card then copy the contents of qmtech-adc-20240614.zip to it. I also copy the start.sh from sdr_receiver_hpsdr_77_76 directory to the root directory so that on power up the receiver can be accessed by HPSDR

After power up I found the IP address assigned by the router and pointed HPSDR running on my PC . It connected and I could see signals


Overnight I compared the 600m WSPR decodes on the new receiver with the 600m WSPR decodes on the ANAN-10. They were within 1dB of each other, the new receiver has good sensitivity!

Now to make the receiver protected from accidents from falling items!

 

Sunday, 9 June 2024

IARU TV Contest 2024 Day 2

 An early start was made from Membury services after coffee and pastries. I then drove to Combe Gibbet, IO91GI61 from where I could see Gareth G4XAT/P on top of Walbury. 


He was worked on 5GHz, 10GHz and 24 GHz easily, it didnt really matter which way the antenna was pointed

Noel, having driven from Devon appeared from Stockbridge and was worked from IO91GC on 5GHz, 10GHz 24 GHz 47 GHz and 76GHz 



Dave G8GKQ/p appeared from Lane end, IO91JA and was worked on 5GHz, 10GHz and 24 GHz

Noel then moved to Morestead Road, IO91IB and was worked on 5GHz, 10GHz 24 GHz 47 GHz and 76GHz

While Noel moved to Butser Hill IO90 MX. I moved to Combe Masts which had a good takeoff in that direction.  On the way past I visited Gareth


G8GTZ/P was worked on 5GHz, 10GHz 24 GHz 47 GHz and 76GHz. 76GHz was fairly easy and set a new UK distance record of 53km. 

76 GHz

47GHz was harder, taking 30 minutes, mainly due to Noels synthesised LO not locking up


47 GHz

Having worked all I could on the TV contest I had a quick tune around 144MHz and found a 2m QRP contest was happening and was surprised to work GW1YBB/P using the IC705 and vertical whip on the car roof. I think he was equally surprised by the locator of G4FRE/P!!

Saturday, 8 June 2024

IARU TV Contest 2024 Day 1

 Noel G8GTZ decided he would be starting the contest at Dunkery beacon IO81FD so the obvious place to start the contest was Cleeve Common IO81XW, at the roadside position which always completes the path. The normal road to the car park was closed for 6 weeks for repairs so a convoluted route was taken, which was full of sponsored walkers. However Noel was worked on 5GHz, 10GHz and 24 GHz easily as usual

10GHz towards Dunkery

Gareth G4XAT had meanwhile installed himself at Walbury hill IO91GI. I packed the gear into the man portable fishing trolley and moved to the trig point and worked him on 5 GHz 10 GHz and 24 GHz

5GHz man portable towards Walbury

Next I drove to Notgrove IO91BV and again worked Gareth G4XAT/P on 5GHz, 10GHz and 24 GHz easily

IO91BV 5 GHz

I then drove to Membury services on the M4 and stayed the night there at the Premier Inn hotel, which would be convenient for the day 2 activities. 

Wednesday, 15 May 2024

A visit to Dunkery Beacon IO81FD

 Having booked a 7 night walking holiday at the HF Holidays location at Selworthy, Devon I realised it was just down the road from the famous Dunkery beacon IO81FD which I had worked many times from Cleeve Common IO81XW. It was decided to take some narrowband microwave equipment to see how the site really was.

I parked at the layby previously used by Noel G8GTZ on the North side of the Beacon. First QSO was on 24 GHz SSB with G4UVZ IO80KX near Taunton.

G8CUB/P had been persuaded to go to Cleeve Common IO81XW and was a huge signal on 24 GHz NBFM, He was loud enough to work on 47GHz SSB for my best ever DX on the band at 136km. No signals were hear either way on 76 GHz.

The takeoff over the  Bristol Channel

24GHz towards IO81XW from the layby


47GHz (a dish was needed for the QSO)

Meanwhile G4LDR, G8GKQ and G8GTZ had gone to Bullbarrow Hill, IO80UU (96km)  I worked G8GKQ and G8GTZ on 24 GHz NB and DATV and G4LDR/P on 24 Ghz SSB and 47GHz CW
The setup to G4LDR/P on 47GHz CW

I also tried working G8IKP/P IO80RQ near Hardy monument and M0GHZ/P at Cold Ashton IO81KP (obstructed path the plots later showed) but no success was to be had

After packing away the gear I did walk to the Top of Dunkery Beacon to see if it was worth the extra effort of going man portable. I was not convinced!


The summit of Dunkery Beacon

The takeoff towards IO81XW from the summit











Saturday, 17 June 2023

IARU TV Contest 2023 Day 1

 This year I decided to go somewhere closer to home. Titterstone Clee seemed a good choice  but the usual spot would be a bad choice as the path North was blocked and G8GTZ was threatening to be QRV on Meryton Low. There were a couple of spots on Hoar edge that looked useable. Mid morning I was on my way with 23cm 13cm 9cm 6cm 3cm and 24GHz,

Arriving on site after an easy journey I had a look and found a site in IO82

I worked G3VKV two way over a 65km on 13cm 9cm and 3cm On  6cm band Graham could receive me but I had no chance of receiving him through the local widehand WIFI QRM.

I had been monitoring the skies all afternoon as the forecast had been for storms. They were seen approaching from the South from the Hereford Direction, but they seemed to go either side of the location without dumping rain!

Noel then appeared from IO83AD On 13cm and 9cm I could receive him over the 92km path with a MER of 20dB. However he could not receive me. Turned out he had a receive issue. We tried narrowband on 10GHz which just about worked. I called Noel on 2m Talkback with no response, so after a 2m qso with a station who took pity on me calling G8GTZ for so long I packed up and headed home.

The homeward trip was challenging as there was localised flooding and tree branches on the road, Glad it missed the hill


Takeoff towards Titterstone Clee


9cm towards G3VKV

View South

View towards G8GTZ/P









The journey

Saturday, 5 November 2022

DATV-Easy

Having listened to the presentation by PA3FBX during the BATC Convention 2022 Part 2 about the F1EJP DATV-Easy 2.08 software, I thought I would give it a try.

The installation procedure is well described in "DATV-Easy-2.08-En.pdf"

First I needed to install OBS. This would be a learning experience having never used it before! I installed OBS 2.8 which meant I would need to install Droidcam for OBS as the virtual camera. 

I launched OBS and was met with complaints that Virtualcam for OBS could not be opened which was a mystery as I had installed Droidcam for OBS, After a lot of experimentation i discovered the solution was to delete Droidcam!

Next I installed DATV-Easy 2.08 which was easy. I then installed the Pluto driver. My Pluto (with F5OEO software was found. The callsign etc settings were changed. On starting to transmit, the left LED in the software lit, but the right one did not nor did the transport rate box change from zero. I eventually found out that DATV-Easy Encoder defaults to NVIDIA (which i didnt have)  so it needed to be changed to Soft(ware) With this change the second LED lit up, the transport rate got close to the expected value and a DVB-S2 signal was decoded

I confirmed that the maximum rate that could be supported by the software encoder is 333kHz. Next I need to investigate hardware encoders

This system is going to be fine for fixed station useage, but forget it when going portable, as it is very complicated and needs much hardware




Friday, 17 June 2022

Portable on Rosedale Moor IO94MJ

Before returning South I wanted to try another path with Rob M0DTS/P The slightly obstructed path was 115km between IO85XF20 (Winters Gibbett) and IO94MJ97 (Rosedale Moor) 

 Two way QSOs were had on all bands 2GHz and 24GHz on 333ks DATV with MER ranging from 24dB on 9cm to 5dB on 24GHz(rob forgot to pack the 23cm antennas!) 

 A Two way QSO was also had on 6cm FMTV, the previous intermittent issue with carrier and no video on transmit being traced to a poor quality Chinese BNC plug 

My FMTV signal was also received by G1LPS in IO94EQ at 55km after he did some careful arranging of components through the shack window

9cm equipment

24 GHz equipment



6cm FMTV equipment


24GHz signal received


13cm signal received (Note Gibbett at Robs site!)

The following day I went to Pickering and caught the NYMR train to Whitby, which was much busier than the visit the previous Monday. I was staying at the White horse farm in at Rosedale abbey. 

On Sunday I intended to get up to see the sunrise but my GPS was on GMT so i missed it by an hour, so contented myself with a walk on the moors, getting back for breakfast at 0930




It took 4.5 hours to get from Rosedale Abbey to Malvern. 




In total, 560 miles were driven in the 10 days!