QDX arrives - tomorrows Digital Transceiver, today

 QDX Digital Transceiver

I have blogged before, just down a bit, about the remarkable QRP Labs design for a digital transceiver for 80-40-30 & 20m. This design is a conventional SDR receiver, including digital sythesised frequency control, IQ audio to digital I2S codec, software filtering & decoding of USB signals and finally a "sound card" USB interface to your PC or Mac. [more below]

It is the transmit side which is very innovative. The digital USB audio from your PC has it's frequency measured, very fast and very accurately, this digital timing then sets the output frequency square wave of the synthesiser. The resulting RF frequency is amplified digitally and in a class "D" output amplifier. Giving an output of up to 5W across HF bands.

Also included is a CAT serial port implemented on the same USB socket, so your PC sees two ports, audio in/out and serial in/out. These can be directly selected in WSJT-X for FT8, WSPR or other digital modes.

My kit has arrived

QRP Labs were very kind to put me on the early release shipment list for the next batch of QDX kits and mine arrived yesterday. This is it

QDX kit, that I have started to build.

It is sooo small. Fits in the palm of your hand

The build

1. I read the documentation twice. Very detailed very clear. Seems most of the build is winding toroids. Anyway I have started as instructed to mount the very small capacitors. And I have had a lot of difficulty reading the tiny values printed on them. I mounted 7 and have now taken a break.

2. Finally finished the caps, by a process of elimination as I simply could not read the writing on final value. Hope OK...

3. Moved on and mounted the RFCs & diodes - one diode was cracked at the end where the wire comes out I hope it performs OK.  I then mounted the PA transistors and bolted them down, easy.

4. Now the coils. I have wound a few toroids in my time, but I hate it and feel like a novice every time I start. I have chosen the 12V operating power supply so wound the PA transformer 3:2. And I lost count so many times when winding the 19-30-36-41 turn input BPF inductor. But finally arrived here:

First components fitted and

started on the coils

5. So on with more and more toroids. I do wonder if QRP Labs could find a source of pre-wound toroids as this is the most tedious part of the kit construction. But then may be it would be too simple. Seven toroids to go... six for the transmit LPFs 80-40-30/20m and one for the balanced RF input

6. At last got all those pesky LPF coils wound and fitted. I am always scared at this point that I have got bad soldered joints and/or wrong number of turns on the coils. 

LPF coils mounted, did I get the turns right?

7. Seems very little 0.33mm wire left for the last transformer T2. Just starting measuring it now. Well I did not by a long way have enough 0.33mm (AWG 28) wire left, 3 x 20cm... so I had to use some 0.28mm (AWG 32) which I had in my components draw. What a faf to twist these three wires, then wind 10 turns, then a faf to identify the correct connections. At last to insert them, finally to solder them. They are very very close tother components so great care is needed. But I must praise the documentation for its calm and clear instructions, without which  I would have gone mad.

8. And now to finish up. USB, 12V power and Antenna sockets mounted. Finished. And here it is

I am proud of myself. A neat tiny 5W transceiver!

9. Now to try it out. Think I will box it first as this is easy and straightforward, and I can easily unmount it if anything is wrong, So I did

Boxed QDX, case ground connection not shown 

here (see comment below)

Testing

First let's get it connected. Plug in the antenna (an AL-705 loop and Mat-125 ATU). Whoops, the op manual says, "DCPower: use a regulated well- smoothed DC supply of not more than 12V". My shack is all running on a 13.8V supply but I think I have a home grown 12V supply somewhere in a dark pullout draw, I look it out. Yes, here it is, check the voltage, OK 11.8V. Last use this fancy new, posh, white  USB B to USB C cable that I bought for the purpose to connect the QDX to my MacBook Air. 

Software - MacOS

No drivers to install as the Mac has drivers for all main uses, digital audio and serial comms included. I have the latest version of WSJT-X installed which I normally use for FT8 with my ELAD FDM-DUO transceiver.

WSJT-X - preferences

1. Radio setup: I chose Kenwood TS-480 for the correct CAT commands and the QDX serial port (mine was called "/dev/tty.USBmodem101") should be set at at 8-1-N, the baud rate can be left at 9600 as the USB virtual port does not need this to be set to a particular value. PTT should be set to use CAT commands.

2. Audio setup: For input and output I chose the "QDX Transceiver"

I set the WSJT-X PWR slider to max as the QDX converts any input to square waves and the output is not dependant on the audio input level, only the power supply of 12V to the switching Class D PA.

Now the big moment, connect the 12V supply. Docs say the front red LED will

1. Quick flash for 5 sec on power up. Then go steady red and be in receive mode.

2. When in receive mode, steady red, for one second or

     a) 3 quick flashes - transmit mode, 

     b) 2 quick flashes - computer sending audio to QDX but volume too low to detect

     c) 1 quick flash - PTT activated but no audio detected.

Now to operate. I have RF output on 20 & 30m, but none on 40m. Oh. So I guess I have a poor connection somewhere in the 40m LPF, or may be it is that broken 1N4007 "pin"diode I found during construction... have to look into this.

Later: fixed the problem, poor soldered joints on LPF coils. Great care must be taken to burn off the coating on the wire used to wind them, do as docs say, hold your iron on for the full 10 sec!

I chose first 30m as I often use this band. I connect an antenna and my auto ATU, I hit WSJT-X  "Tune" to tune the ATU (<1.5 SWR indicated), then switch back to receive and FT8 signals magically appear. First contact!

First contact

Round up

Whoopee. By the way I get 3-4W output 20-80m, not a full 5W. I also noticed that the case QRP Labs supply is not connected to Ground of the circuits! I have screwed a small solder tab on the inside back panel and wired this to the ground side of the 12V connector.

So what is this QDX wonder?

It is an example, very well conceived, of the transformation of amateur radio to the digital world. So little of this transceiver is analog. At the front end there are BPFs which are digitally switched (no relays here). After that a base band detector, where the incoming sideband signals are shifted, or mixed, down to I & Q audio frequencies (actually 12kHz away from zero to allow the digital decoding to happen at an audio of 12-15kHz providing better signal filtering). And then we go to digital audio, with a high end analog/digital sound convertor, the output of which is standard PCM stereo audio signals along a standard I2S bus to the heart of the system, a microcomputer and DSP.

Microcomputer

The microcomputer handles a lot of things. It DSP filters and decodes the I & Q audio I2S inputs to mix them down to cancel one sideband and output Upper Side Band signals. It generates USB output digital audio signals to send to your PC, where applications such as WSJT-X decodes the spectrum and interpret the received audio tones of FT8 or other "digital" modes. So far so good, and fairly conventional as SDR receivers go. But innovative to use a dedicated software in a microcomputer to do this.

Transmit

The heart of the QDX innovation lies in the transmit chain. Applications such as WSJT generate audio tones, for example for FT8 a series of pulses of audio frequencies over a 15sec or so time to encode data in FSK. And these are sent out by your PC over the USB connection as digital audio. It is what the QDX does with these signals that is innovative. The incoming audio to the microcomputer is a digitised sine wave, and the period of this is accurately measured. This then gives the frequency of the tone.

But what we need is to transmit a Upper Side Band RF signal with this audio as modulation, in other word translate the audio back up to RF. This is done in the QDX by an Si5351 digital signal generator capable of generating RF square wave outputs up to 200MHz from any of its 3 outputs. If we chose to be on, say, the 40m band then this synthesiser will be programmed to output the FT8 frequency of 7074kHz plus the instantaneous audio tone frequency we have detected. Thus generating an FSK transmission, with no suppressed carrier and no wrong sidebands at all.

The huge advantage of this digitalisation, rather than analog mixing, is that two of the Si5351 outputs  can be programmed to output the digital signals 0/180deg out of phase. Why? Because these can be used to drive a push-pull power amplifier stage in class D operation. This has very high efficiency and thus very low heat dissipation. The PA generates 5W of RF from four small TO92 MOSFET transistors mounted on the PCB which is sufficient to provide a heat sink! Of course square waves have harmonic content, though by using push-pull amplifiers there is little 2nd harmonic and the first one to filter out is the 3rd at a much higher frequency. So conventional 5 element Low Pass Filters are all that is needed to meet legal harmonic output levels, and in practice much better.

But what about changing bands? As we have seen on the receive input digital CMOS switches change to different BPFs, and here again on the output RF side the switching of the LPFs is done digitally by "pin diodes" under control of the microcomputer.

Lastly, CAT

The final task of the microcomputer is to provide a digital serial USB port which uses the same cable connection as the audio, to provide CAT control. The program emulates the Kenwood TS440 or 480 transceiver - using ASCII text strings for communication - a much better solution than the complex ICOM hex data coding...

Thus applications such as WSJT-X can set the band and provide CAT transmit/receive PTT control. And there is but a single USB cable from transceiver to PC carrying both digital audio and serial data.

Wonderful

Thus we have an almost entirely digital transceiver which fits in an incredibly small box that you can hold in the palm of your hand. The receive chain has an option to be used as a simple IQ audio detector and have the microcomputer send these IQ signals directly to your PC, where applications such as Cubic SDR or HDSDR on a PC can be used to display a spectrum and waterfall of the received signals. But of course it cannot transmit analog speech SSB as it is only a digital transmit system. 

Do we need a transceiver - or just a TX

 How about just a transmitter?

Let's think about this. Many people working 160-80-40 & 20m very often receive using WEB SDRs which are extremely popular considering the huge increase in noise at most QTHs reaching S7-9 levels.

So there has to be a case to separate TX & RX not to build transceivers at all. The Ham shack would then be a PC & Internet connection, browser and web SDR plus your own transmitter which you tune to the same frequency as reception.

So what should be the spec for a transmitter only

- Audio/MIC input with EQ & compression 

- digital; audio input for digital transmissions, and/or audio line input, no EQ or Comp

- Key input, straight and paddle

- SSB & CW 1.8-30MHz, probably SDR as much easier to get wide coverage, use ready made modules?

- 100W say, PA with ALC 5-100W selection, LPFs for all bands, relay switched

- a low level 0dBm output for transverter?

- Auto ATU with 50R coax, balanced and long wire outputs

- built-in power supply - SMPS for low dissipation.

- Front panel or remote operation. PTT & CAT commands

That's quite a list but it is do-able, ain't it? Got to define a way to switch TR between transmitter and reception on a PC...

All in all, it could save us a lot of money to have such a product. It might also stimulate better and more advanced web SDRs to be developed.

Computer and Radio interfacing - chaos

 The mess that is computer control of your rig. 

Connections!

On our computers we have almost universally today just USB input/outputs. These can handle digital audio input/output and/or serial data communications. Often two USB A connection are used for audio and serial separately, on others a single USB C handles both audio and serial. And this works, with the coveat that some OSs demand careful driver and levels setups. I am using a MacBook Air and frankly it just works, no driver installations, no fiddling around with sound panels!

Interface computer and radio

We need to send audio in and get it out, which needs an A to D and D to A convertors, running at any of the standard PCM audio bit rates from 8kHz to 192kHz! 48kHz is common.

The biggest issue is the serial data control data from and to our transceivers. There are two main camps, one using sequences of hexadecimal numbers (ICOM, Yaesu not the same), the other using much simpler ASCII character strings (Kenwood). Add to this the signal for PTT, which can be a special sequence of serial CAT signals, or a hardware serial interface RTS or DTR lines, of the right polarity.

And to interface the computer USB serial data to separate TX and RX digital lines, which can be 5V or 3V3 volt levels and speeds of anywhere from 9600 to 38400 baud (possibly lower or higher), the logic must be correct for example positive logic where +5V = logic ‘1’ and 0V = logic ‘0’, or it could be negative logic with +5V = logic ‘0’ and 0V = logic ‘1’. The serial data must also have the correct format, for example one start bit, 8 data bits and one stop bit, a total of 10 binary bits.

For ICOM radios, to make things worse, these signals are carried on a single wire C-IV, not two separate connections TX & RX, requiring special I/O circuitry between the computer and the radio.

Then comes the problem of the rig audio levels. What does it need for input and output? These could be standard line levels (about 1V rms) or most likely the direct Microphone (10-100mV) and LS output (0-5V). With various input impedances of maybe 100R to 10k and rig output impedance suitable for driving an 8R loudspeaker.

Finally some radios have moved completely away from audio and data interfaces to use WLAN and communicate with your computer over LAN and via your home router. Giving the advantage of remote operation. But here again we have a new set of communication protocols and signalling.

Xieigu G1M example

A couple of blog posts back I describe the fiddling around I had to do to get my Xiegu G1M to work FT8

Better solution in the ELAD FDM-DUO

On the other hand my FDM-DUO is simplicity itself. It has separte USB B connectors for CAT and TX/RX digital audio (standard printer cables can be used). These can be connected through a simple 2 x USB A to USB C Hub to my MacBook Air where they appear as a digital "device" and an audio "device" which can be selected directly in WSJT-X. Which also has a built-in "hamlib" control configuration for the FDM-DUO. And it all just works. 

QRP Labs is the best

Even better solution from QRP Labs on their Digital Transceiver QDX. A single USB connection carrying both the CAT serial data and the digital audio from a built-in "sound card". One wire, just plug it in. Excellent.

Follow me

The rest of the radio industry needs to follow ELAD and QRP Labs QDX and provide a simple USB connection direct to our computers, with implemented CAT signalling - preferably using ASCII text commands.

Plugs confusion

 Plugs, plugs, plugs

How many of us have this problem? Main stream amateur equipment has no standard plug. The choices seem to be PL259, N Type, BNC or SMA. OK they all have their advantages and disadvantages. The tendency is to use PL259 for higher powers up to 1GHz, though the N Type is often preferred. Then BNC are frequent for smaller equipments and lower powers, and finally SMA which are used for all frequencies as "mount on the PCB" interfaces like most board SDRs.

But this confusion does not help in the shack, does it. And you end up with tons of convertors like this

PL259s, SMAs, BNCs joiners and convertors

And if you play around a lot with the configurations in your shack, as I do, you find yourself forever searching for cables and convertors and screwing and unscrewing SMAs.

So enough is enough. I have decided to standardise on BCN, because its quick and easy to plug and unplug. I have bought the right cables and/or convertors for my QRP Labs U3S I use for WSPR, my ELAD FDM-DUO the main SSB/DIgital modes transceiver, and my new (not so sure about) Xiegu G1M transceiver. Then of course there is my NanoVNA with SMA which I use to tune up my loop antenna. I have included an SMA output from my antenna switcher just for this and use an SMA-SMA cable... was this wise, I don't know.

Antenna switcher

This is a very simple switch, far from coaxial! Just a 4-way rotary in a small box.

Antenna out and transceivers in BNC

SMA out to NanoVNA

End up

Now I need to buy a bunch more BNC/BNC cables. On the common web sites these come is very limited lengths, so take care. 

Checking for RFI in the shack

Are you sitting in a puddle - of RF?

One thing you can do when suspecting RFI in the shack is to just see how much RF is around. I do this with a sensitive RF meter based on an Arduino, OLED display and a AD8307 log detector module. See this previous blog posting for details.

The RF METER reading from -80 to +20 dBm input

See this blog September 2018 for design details

The circuit is very simple. It is an Arduino Nano with connections

* A4 & A5 (I2C bus) to the OLED screen

* A3 as a DC input from the AD8307 module

* AREF for the A/D converter from a 3V3 voltage regulator

Power is from a 12V input feeding the Nano and the 3V3 regulator. Nano 5V output to AD8307 & OLED display.

The new Arduino sketch code for this project SIMPLE_WATT is here (note you must load the Oled.h header and the U8g2 library for the screen display (download libraries.zip and put them in your Arduino libraries folder). Unplug the 12V supply, connect the Nano to the IDE on your PC and upload the sketch, then plug in the 12V to establish the 3V3 A/D reference.

An extending aerial is connected to the SMA (50R) input connector. 

Note: if a 2K2 resistor is placed in series with the input then the meter will read -30dB down, or up to 20W (= 20mW) input across a 50R antenna or dummy load.

WSJT/FT8 with Xiegu G1M

Let's get the Xiegu G1M on FT8

I think I have made some progress in getting WSJT-X to issue CAT commands my new Xiegu G1M transceiver. I am using a MacBook Air M1 computer. The G1M is connected by the supplied blue USB/3.5mm Jack cable. It presented a serial port and an audio device on the Mac. Here's what I have done so far:

1. Loaded the Mac version of Flrig

2. Configured it like this

Configure/Rcvr to Xiegu G90

Chose serial port

set 19200 baud

INIT to connect (green)

Note: Serial port number may be different as dynamically issued. Just chose the one that pops up. Then connect.

Configure/PTT Generic

'x' only in PTT via CAT box

Back to main screen

Vol does nothing, AUT (squelch) does nothing, Pwr does nothing, ATT does nothing

PRE switches in/out the RX preamplifier

PTT tests the PTT

Mode can be selected in the box (USB etc)

Freq can be set for VFO A(left) or B(right)

3. WSJT-X set up. Chose Rig as Flrig and PTT as CAT. Press Test to check it is working, you can also check PTT...

WSJT-X Preferences. 

Rig FLrig FLrig, PTT CAT, USB None

TEST CAT should turn green

I have yet to make up a couple of audio cables to connect the radio to a small USB audio dongle I have here. I need a mono cable tip to tip for dongle HP out to radio MIC input (the MIC ring is used for manual PTT by the hand held microphone), the audio output from the LS can use a standard 3.5mm stereo cable.

4. Switched to DigiRig Mobile 1.5. I have changed from the "Blue" USB cable that is supplied with the G1M to use a DigiRig interface. 

The Digirig

USB C - Serial TX/RX - Audio IN/OUT

The Digirig 1.5 has a single USB C connection to your computer (neat solution for my new MacBook Air M1, leaving the second USB C for charging...), and provides both serial data TX/RX and audio in/out for your rig in the small box. The connection are:

* Serial, simply use a normal stereo audio 3.5mm cable to connect the Digirig SERIAL to Xiegu COMM

* Audio, needs a specially made cable. I used a 3.5mm-3.5mm with one end a right angle plug. This is then cut in half and spliced to a cable for the 3.5mm Digirig plug.

Note Digirig 1.5 is used here, this has different connections to the version 1.6 which uses four way jack plugs to include RTS/DTR (PTT) and 3V3 outputs. So chose the right one!

Digirig TIP (AF_OUT) <- G1M LS TIP - output from rig

Digirig RING (AF_IN) -> G1M MIC TIP - input to rig

Digirig schematic for Audio in/out

Digirig cable splitter for audio connections (red)

RING (RIG_AFIN) to MIC TIP and 

TIP (AF_OUT) to LS TIP

Flrig must be configured to talk to the Digirig Serial USB device, then INIT to connect

Flrig configuration

The WSJT preferences must chose the Digirig audio device "USB PnP Audio Device" for input and output. The audio level output can be set by either WSJT "power" slider, and for input by the Audio MIDI app on the Mac. The audio level input can be set by the G1M volume control, and in the Audio MIDI app on the Mac.

The Xiegu G1M then looks like this, with the right angle 3.5mm plug at the front to make it neater

Xiegu G1M, audio MIC input 3.5mm right angle plug

This gives FT8 and other digital modes, on 80, 40, 20 and 15m bands.

Too much audio

I have found that there is simply too much audio, the output of the LS jack is way higher than the USB MIC input of the A/D dongle can handle so it is distorting, and the output of the HP output of the dongle is way too high for the MIC input of the G1M. So I need an attenuator in both audio channels. So far I have tested one simple plug-in 40dB attenuator in the LS connection (which is stereo) but Ihave yet to build an attenuator for the MIC connection, tip-to-tip connection only not stereo as rin gis used for PTT.

Alternative WSJT-Z

 I guess it had to happen! 

WSJT-X has been seriously forked. The fork is called WSJT-Z. It has a number of complicated additions but the one thing they have done, which the original specifically didn't do, is to have "Auto CQ". That is it calls CQ continuously and restarts automatically after any QSO. Kind of unattended operation. This is a bit dangerous as people will be tempted to just leave it running, against licence conditions...

WSJT-Z in Auto CQ mode

Here is it running, and you can see the auto QSO's I made in the log file.

I have tried it, I don't like it, I will go back to WSJT-X.

Here we are back to WSJT-X

QRP Labs does it again - a revolutionary digital SDR transceiver - the QDX

 QRP Labs have done it again. And designed a revolutionary SDR Transceiver for digital QSOs.

The QDX

It uses SDR techniques for reception generating audio IQ signals and DSP techniques for RX, and an amazing new digital signal generation technique for TX.

Design highlights

1. Si5351 VFO with a very stable TXCO, generating directly RF IQ signals for the detectors.

2. Normal Tayloe IQ detectors, feeding an on board DSP processor for USB reception

3. Digitally switched bandpass filters for four bands, 80, 40, 30 and 20m on RX, and electrically switched LPFs for TX -  no relays!

4. > 5W output from 9 or 12V supply

5. A new transmit chain, that does not just take the audio output of the PC and send it to an SSB transmitter, as most designs today. But detects the input audio frequency and programs the Si5351 synthesiser to generate the RF tuned frequency PLUS this audio frequency for direct transmission. With an output stage which is pure digital switching. Simplicity though superb software proprietary coding. 

6. CAT emulation for TS480 

7. Integrated sound card and single Audio + CAT interface to your PC. (All in the single microcomputer chip)

This is how it is done

Block diagram

And all this in a box for $80. Remarkable, of course with the current shortage of semiconductors it is not available. But I guess I could show willing by placing my order... no I can't ordering is not open!

Can you build a 100W Auto ATU for £40? Yes

 Amazing, yet again, these Chinese suppliers can deliver an ATU kit here for £25. Just connect the display OLED (or as shown here a larger one) and apply 12V and here it is

Have not tried it out yet but if it is built onto a £9 case with a 12V socket and a pushbutton that's it Here's the display of Power, SWR and L/C values used

My Shack

 Here is a photo of my shack.

Left to right. 

1. A home built GPS clock with TFT display (code on M0IFA.me here)

2. QRP-Labs WSPR 200mW TX for 10MHz (30m).

3. Xiegu G1M SSB/CW four band TCVR

4. ELAD FDM-DUO all mode, all band TCVR

5. Antenna switching unit (four way)

6. Auto ATU, MAT-125E

7. MacBook Air M1 used for Digital with the FDM-DUO, mainly FT8 across the bands

8 An older Macbook (x86) using Apple Bootcamp to run Windows 10 Pro for the SDR software for the ELAD FDM-DUO.

9. On the table a NanoVNA used mainly for tuning my AL-705 loop antenna and checking the SWR

So far so good, all working.

WSPR 30m / 200mW IO92 / M0IFA and antenna update

 My Al-705 loop antenna has the capability of one, two or three loops. I will make a series of tests of WSPR reception on 10MHz (30m) with these options, with a transmit power of 200mW. 

Comparison

Now I am making a careful comparison, it will take 3 weeks! I want to see if the number of loops on my antenna makes a difference. First here is five days with 3 coils.

10MHz 5 days, 3 coils

I each case I have decided not to connect the loop directly to the QRP-Labs 200mW transmitter, but to connect via an ATU. My MAT-125E  auto ATU needs a minimum of 1W TX to tune, and so I have switched to my ELAD FDM-DUO at the same frequency 10,140,200Hz, and output a tuning signal of around 1.3W. Then switch back to the QRP-Labs 200mW output. The FDM-DUO showed an SWR of  1.5:1 on tune (down from 2.6:1 without the ATU. I checked the tuning also with my NanoVNA to see the SWR minimum at the correct frequency.

10MHz 5days , 2 coils

There is practically no difference between 3 and 2 coils on the loop antenna. May be just a little lower S/N ratio and fewer spots. 

Seems there is little difference between ,3 2 or 1 coil. Here's the one coil data

10MHz 5days , 1 coil

But more interesting might be any correlation, as you might expect, between sun spot numbers and the results. For example, the last day of one coil above shows a very poor day for 10MHz propagation. And sure enough the SN was a low 31. 

A fast way to check the propagation is a new app from N0NBH called "Solar". This runs on the iPhone but also on the new MacBook Air M1. Here is todays info:

SUN

And if you want the latest picture of the sun click here

Club QO-100 transmitter

An uplink converter (here) for transmitting to the QC-100 satellite has been built by numerous amateurs

This is the layout with individual "ebay`' items fixed to a mounting board, interconnected by SMA cables.

The Banbury Amateur Radio Society is building one of these along with the TX/RX antenna and LNB. So far we have connected up the units free across the table, and are feeding it a 432MHz input at up to 10W through a 20dB attenuator. Awaiting PSU module for +8 and +5V. No results yet, but will document as the project goes ahead.

So here's the project going ahead - it is NOT a shambles, promise

Some of the modules scattered around

And proposed box which will go outside near the dish

Just a couple of new photos as we have a club presentation about the Ofcom EMF checking we all have to do: spreadsheets and all that. I think I am exempt with max 5W TX and indoor loop antenna! Or also with my 200mW WSPR TX.

The task goes on... the antenna is proving difficult to bild as the plates keep warping when heated to solder them, tried twice so far but still a problem. This is the state of play, I'm sure something will rise out of the jungle...

We have a receiver going with a commercial LNB and dish, but it drifts quite a lot and we must xtal stabilise it. We are using an SDRplay with SDRUno software to tune to the LNB output on 739MHz.

Update: the main components are mounted on the board and we have an output at 2.4GHz from a transceiver and 20dB attenuator at 430MHz. Now to add the filters and power amplifier.

So this is what it now looks like, but building has stopped as one of the power supplies blew up and we are waiting for a replacement.

RX

There is talk of moving to an RTL dongle and I would like us to try GQRX software... as it runs on the Mac!!!

QO-100 frequencies

The tuning and TXRX frequencies are

Field day 4 Sept 21

Saturday 4th Sept was Field day and Banbury Amateur Radio Society (BARS) went to Balscote Mill to put up a /P station.

I also went along to put up my ELAD FDM-DUO and an AL-705 loop antenna.

SSB

I have updated my old MacBook 12" to run Windows 10 Pro and install the ELAD FDM-SW2 SDR suite. Looking to make any contacts on 40-30-20m.

FT8

I switched to my new MacBook Air and WSJT-X and try for FT8 contacts.

Most of the day was a disaster. We were located in a deep dip with little take off. I was heard in Croatia but no contact confirmed. The main station's G5RV  antenna broke in the centre and was irreparable!

Buying an EV?? Our needs for EV charging by 2030 (9 years away!!)

 Chucking out the peanuts

Day after day our UK government is announcing this and that million pounds for peanut initiatives for the upheaval change we must make to move to Electric Vehicles. It's a new Transport System guys...

What about 2030, just 9 years away? Will there be 6m or 18m EVs on our roads? It varies depending which part of government you ask, BEIS/Dft or Ofgem. So what? It's just chicken and egg. And here most definitely the chargers come first, for without them along our strategic road network EVs will not come, people will not buy them.

We are faced by seemingly random government actions. "190,000 home/work chargers of 7kW power by ???" "25,000 chargers, unspecified, and without declared funding, by ???" Magnificent £1.3bn declared over 4 years, for what? Not defined, but turning out to be handouts to cowboys operations putting chargers in the wrong places like McDonalds or Chester Zoo, not along our roads. Lots of talk about "£950m for rapid chargers on our roads, 6000 by 2035" (far too few). Plus a miserable £20m for Local Authorities to divide their car parks and prevent fossil fuelled cars equal access to parking, oh, and some unspecified "street" charging in 2021 & maybe 2022...

Like I said, chaos.

Pools, chargers and points

And on top of this Boris's stated bold angelic vision of charger pools "every 30 miles along our Strategic Road Network". Which, when you work it out, means 2,300 miles of motorways needing 80 charge pools, at which at least 20 pools with 10 chargers each or 9,300 points.  And even then 15m EVs will need up to 15m home/work fast charge points. We have to deliver the power, a mean of 1kW per EV by 2030 is the EU estimate for electricity suppliers.

But this is not the direction things are going. Is it? The government firmly believes in the approach of "the market will provide" (read: it's not our job, we are only the government). But the market  will not. Today few of the chargers are finding their way to motorway pools and very few, if any, to "A" roads. They are being installed at local destinations, supermarkets, fast food outlets, in town car parks. The result of which will simply be huge traffic jams as people drive in and out to put energy in their cars. And as EV's fight ICEs for a parking place.

And again just think, EVs need a 45-60 minute charge (it is getting faster but not until better battery technology comes along) versus fossil fuels cars needing just a 10 minute stop. Or 6 times as long. And the range of the EV is, not as makers claim 200-250 miles, but a realistic 150 miles, or lower in cold weather, versus a fossil fuelled car of 500-600 miles. Or four times smaller. This means many more EV stops and many more chargers. It is obviously foolish to think of just replacing petrol stations with EV charging stations. A new architecture of a revolutionary Travel System is needed, only achieved by good top down thinking and planning.

This is where our government is failing us.

It seems the only company that knows this is Tesla, with its manufacture of batteries, cars, of  navigation systems that route you to/via chargers and of an adequate national charger network.

Bad place

So where are we today? In a bad place. The useless national data base of chargers, on line as Zap Map, shows puddles of chargers. Except when you drill down to find one you need right now, right here, for example in East Anglia desert. There are more vast desserts when none can be found. Worse still are the fifty different "networks" run by 50 different cowboys, all given handouts by government to put chargers anywhere they want. All branding no substance. 

No plan, no hope.

The government financing model of handouts to stimulate private initiatives will not provide a pan-national charging infrastructure. What it promotes is seed money promoting sharing of costs with destinations. You know, "come to McDonalds and eat as we provide Rapid EV charging"... In other words the focus of the private investment is shared with destinations. No destination, no chargers. And worse, the paltry sums handed out to Local Authorities most of whom have no idea what to do with it, who employ "consultants", more cowboys, then sprinkle their patch with useless EV only parking. Some even refusing to provide on-street points for those without a private drive.

How I got WSJT working under MacOS Big Sur

First download the latest version of WSJT-X.app from 

https://physics.princeton.edu/pulsar/k1jt/wsjtx.html 

Then execute the ".dmg" file, this will open the install page. Drag the WSJT app to your apps folder. Then open the Terminal and do this: 

1. List long the file supplied on the download volume

ls -l /Volumes/WSJT-X/com.wsjtx.sysctl.plist

-rw-r--r--@  1 antonywatts  staff  543 28 Mar 12:53 /Volumes/WSJT-X/com.wsjtx.sysctl.plist

As you can see it has the "@" marker indicating special attributes, and has the wrong owner and group. The file as needed in /Library/Daemons is

-rw-r--r--  1 root  wheel  543 21 Aug 15:18 com.wsjtx.sysctl.plist

2. Copy the file to the correct location

sudo  cp  /Volumes/WSJT-X/com.wsjtx.sysctl.plist  /Library/LaunchDaemons

it has special permissions set - as seen by a '@' after the mode, as well as the wrong owner and group

-rw-r--r--@ 1 antonywatts staff ...

3. After copying the file to /Library/LaunchDaemons both issues must be solved. First change your directory

cd /Library/LaunchDaemons

and list the files using

ls -l

where you will see the things that are wrong with the .plist: the "@" and the wrong owner and group

4. Remove the special permissions "@"

sudo xattr -c com.wsjtx.sysctl.plist

and check that is has gone with another

ls -l ...

5. Set the correct owner and group

sudo chown root:wheel com.wsjtx.sysctl.plist

The result should be

-rw-r--r--  1 root  wheel  543 21 Aug 15:18 com.wsjtx.sysctl.plist

6. Now restart your Mac and check the settings with

sysctl -a | grep sysv.shm

If shmmax is not shown as 52428800 WSJT-X will fail to load with an error message: "Unable to create shared memory segment".  And you have to fiddle around a bit more.

UPDATE

I just bought a new Apple MacBook Air (M1 chip). On this the installation went more smoothly. I copied the com.wsjtx.sysctl.plist file across to /Library/LaunchDaemons (see #2 above). It arrived with right owner and group and no special permissions, and then I restarted the MacBook Air and the sysv.shm value had correctly been  updated.

So WSJT is right at least for new MacBooks.

Antenna switching unit

Ok so I have one decent antenna - an indoor AL-705 loop. But I have three things I want to connect it to

* My ELAD FDM-DUO SDR transceiver, possibly via my MAT-125E auto ATU

* My QPR-Labs QCX+ WSPR transmitter

* My NanoVNA, for tuning and checking the SWR to the loop

So I am building a switching unit. And for fun I am putting in it a stupidly simple resistor SWR bridge, with LED indication for high/low SWR which will also help with loop tuning (in addition to the power meter described below which still is, for now, the best way of optimising the matching)

Here it is. Without the three way switch which will mount on the right & I have ordered but which has not yet arrived

The red switch bypasses the small SWR bridge. I chose this green LED after trying a bunch from my scrap box as the brightest of the bunch, there was a notable difference between LEDs. I have standardised, good or bad, on BNC connectors, they are so quick and easy to plug/unplug when fiddling about.

Project finished, works great, now can switch antenna from ELAD for FT8, U3S for WSPR and NanoVNA for tuning.

More outputs

A later modification expanded the outputs to four, for my QRP-Labs WSPR TX, my Xiegu G1M, my ELAD FDM-DUO and the NanoVNA. Very convienient

Update to Power Meter (AD8307), EMF detector, loop tuner

My power meter has a direct input to an AD8307 Log Amp module. This gives a range of -80 to +20dBm sensitivity, with reliable measurements from -70 to +10dBm. 

AD8307 characteristics

To use it at higher powers attenuators can be used. These can be made with an high wattage input resistor of 50R, to act as the dummy load, a series resistor into the 50R input impedance of the AD8307 module in a resistor divider. The maximum input to the AD8307 is +10dBm or 10mW for linearity (it can tolerate 20dBm, 100mW).

The aim is for 10W in to give a reading of 10mW on the meter, this needs a series resistor of 2200+100R. I have checked the readings by comparison with the power output indication of my ELAD FDM-DUO SDR.

This is the screen layout, the small signal showing is from a short antenna connected to the input picking up local RF noise... The bar has a range of 100dB

Power meter display, Bar is -80 to +20dBm

The new Arduino code for the power meter called SIMPLE_WATT is here.

LOOP TUNING

A new use for the now sensitive power meter is to tune my AL-705 loop. I connect a short antenna to the power meter and place the unit close to the loop. On transmit (either by my FDM-DUO 5-8W, or my new U3S WSPR TX 250-200mW) it deflects and allows me to tune the loop using my new remote motorised capacitor controller (see below)

Great stuff.

So how is the WSPR/PSK going?

 I have started on a more longterm project to look for myself into my propagation and reach. This is somewhat limited as I cannot mount outside aerials and I use only an indoor AL-705 loop.

These are the long term graphs I have right now for spots on WSPR, from my U3S 200mW 30m station and from my ELAD FDM-DUO 5W FT8 station

WSPR 200mW

WSPR

PSK FT8 5W

Thanks to Paul Marks down in VK land for the development of an excellent Mac app "WSPR Watch" for these graphs. More to come as I leave the WSPR on for a week or so (occupying the loop so no FT8..)

SWR meter, motorised loop antenna tuning, QRP-LABS U3S for WSPR

 I have some new projects in the pipeline. 

LOOP TUNER

A longer term idea is to fit a stepper motor to the loop tuning capacitor to remotely step up/down and eventually auto tune the AL-705 loop.

This is the state of the development with stepper and Arduino. Button drive it up/down and pot adjusts speed. Code is here.

Some progress has been made, my son has kindly designed a mount for the loop tuning capacitor in its box and the stepper motor. This he has 3D printed. We have a few problems with it. One is that the capacitor on the Alpha Antenna AL-705 tuner is NOT in the centre, but about 1mm offset to one side, meaning the stepper does not line up with the capacitor shaft! Thank you Alpha... will have to re-design and re-print...

Another issue is to find a flexible coupler from the stepper shaft to the capacitor shaft, the capacitor seems to have a strange diameter, may be American "Imperial" not metric? It measures just <8mm, but is loose on the 6-8 coupler I bought on eBay... my son has an engineerng company and he says he can drill out a 5-6 coupling to fit.

The other issue is whether the stepper has enough torque to drive the capacitor which is quite stiff. This can be check only when I can get the coupler problem solved.

I have tested it. Here are some photos

Stepper to loop connector, book is there to align shafts!

SWR on NanoVNA

Arduino and controls, two direction and speed

The stepper driver

There is an expected problem, the stepper driver is a switching signal of 1A, this generate a S9+ interference on reception. So the stepper driver must be switched on only to tune the loop, not all the time. This is implemented in the latest ANT_TUNER Arduino sketch, here

Mount 3D printing

Tuner and Arduino mounted

What it will look like, now to wire it up

There is wiring to do and a control box to build. And a software problem to solve - to inhibit the driver while not moving, so to stop the switching RF interference, hopefully.

Jungle wired

I have mounted all the electronics and wired it up, bit of a rats nest, but... And I have built a controller box, pot for speed control and a `left and `right buttons for tuning.

Control box

IT WORKS superbly, fast is just right to scan across the bands, and slow is just right to nudge across a band, while watching the NanoVNA resonance frequency and SWR.

I feel clever.

WSPR

Second I want to TX on WSPR on an ongoing way, i.e permanently on. I have built a QCX+ CW transceiver which has, it claims WSPR programming, but Ihave been unable to get it to work. I think the problem is that the class-E PA is not running in class-E due to some inductances being wrong - I get only about 2W output on 40m*. This in turn generates quite a lot of heat in the PCB substrate which is used as the PA heatsink. This heat in turn is heating the Si5351 xtal and causing it to drift over the 100+sec TX period. I can see this drift quite easily on my base station RX (ELAD FDM-DUO). It is also not decoding with WSJT software as a consequence. Anyway I don't like the QCX+ software UI for Beacon use...

So I will build the QRP-Labs U3S WSPR TX which has a GPS calibrated frequency control or an oven controlled xtal oscillator circuit to stabilise the frequency. This kit was ordered back in Dec 2020, but QRP-Labs have had long delays in getting a new GPS module built and tested and so my order was completed only at the beginning of June...

* since found my power meter is reading about 6dB LOW! So this is probably 4-5W in reality...

QRP-LABS U3S built. 

Quite a tricky kit as very tiny boards and lots of link options for different operating uses. I have the GPS board, but cannot fit it as QRP-LABS do not seem to supply the nylon standoffs for this and I have had to order some from eBay. Also had to order new push buttons as one is faulty. I have bought a small 12V to 5V regulator board for the power supply. I think it is wrong that QRP provide a 1/4" jack plug for 5V input as this size is traditionally used for 12V and could cause an accident. Could have used a USB B socket? To start with I have not built the OCX temperature controlled synthesiser module and will try with the standard Si5351 module. Here's the result on 30m when feeding 200 mW to my loop antenna:

I struggled with the menu system. But it is ingenious and when I had it sussed out it was easily done. Initially when I programmed in the frequency of 10,140,100 the U3S output was 10,140, 316 which is outside of the WSPR band (Only +/-100Hz wide). But after I left the system running for 24 hours and self GPS calibrating after each transmission, the frequency had been reduced to 10,140,206 which is closer to the band centre, probably as close as it goes... We shall see if it moves further. Very impressed with self calibrate feature.

By the way I calibrated mu ELAD FDM-DUO at 10MHz by reprogramming a small Ublox GPS module to output 10MHz in place of the usual 1pps. See down this blog...

U3S transmitting WSPR on tuned frequency 10,140, 200Hz

from location IO92 at 23dBM (200MW)

The output spectrum has tons of harmonic side lobes (phase noise?) not a single neat carrier. This is typical of the Si5351 and I have seen it in many other projects I have built. Locally decoded this brings 3-5 decode lines, the main central frequency and 3-4 other side lobes, only 10-20dB down. Not much to be done about this, unless there is a better way to program the Si5351 to remove these side lobes... 

Here's the output power into 50R load, as seen on my home made power meter (described further down this blog as part of BARSICLE project)

I have tried it driving a small loop (Wonder loop) and the result are disapppointing, pity I was planning to use this a a permanent installation for propagation monitoring. But connecting it to my new AL-705 loop brings results.

By the way there is a new version, an update, of the excellent MAC and iPhone app "WSPR Watch" which you can use to see your spots and display some stats and trends.

Update to U3S. Have installed all 3 BS170 MOSFETs, power output unchanged, as need to use higher Vcc to get more. May try 12V line to PA..