Loopy

 It's arrived. Even if Waters and Stanton confused me... I placed the order for an "ICOM" AL-705 loop antenna, and this large box arrived from Alpha antennas USA. No paper work so it made me think a bit. I called W&S they confirmed it was mine. Whoopee.

Now built the antenna

And "borrowed" my wife's camera tripod, got an iPhone clip mounting screwed on top and wired it up. 

First trial I found the tuning very sharp, but I found a way to do it. Display the frequency you want on the SDR specturm display (I use an ELAD FDM-DUO TRX). You can then peak the noise. I found I could get an SWR of 2-4. So I put my MAT-125E auto ATU inline and this was able to tune it down to an SWR of 1.4 or better!

Noise levels are much lower (S4-5) than the small 60cm loop (S8-9) I used before, and signals are up 2-3 S points.

I checked it out also with my NanoVNA, and oh boy is the tuning sharp, and the knob very sensitive.

I have now had a few weeks to use the loop. And this is what I have found

1. It does work from 40-10meters (Alpha antenna can supply a 2 extra turn loop to expand to 80m)

2. It is a bit floppy and wuld not stand up in any windy conditions

3. It is sensitive to the shape of the loop and position of the coupling loop

4. It is incredibly sharp tuning as the frequency goes up, and to get a low SWR you must be exactly on tune.

I have used my NanoVNA to look at the SWR, and also to tune the loop This is the SWR on 17m (18.100MHz for FT8)

As you can see the SWR is around 2.3 at the best point. The tuning was very very sharp and difficult to target exactly the right frequency. The problem is that touching the tuning knob affects the resonant frequency, so you have to tune a little bit low so that when not touching it the resonance goes up 5kHz or so. I am starting to think about a motor driven system which I can remotely control back and forwards. Very low RPM (3rpm?) motor or a stepper motor...?

Looking at the Smith chart shows

Here you can see the impedence is not pure resistance at 50R which we want. This is the measured characteristics

An impedance of 61R plus a reactance of -j65 (Capacitive).

I use the loop via my auto0tuner MAT-125e. This brings the SWR to my transceiver down to 1.22 when I add a little inductance.

I use  small antenna switch to 

change over from Operate to Tune

I have started to be impressed with this loop. I have it indoors and even so it is reasonably sensitive, especially on 30m and above. I have taken up tuning it with my NanoVNA to whatever frequency (+/-20kHz at best) that I want to use. Then connecting it via my MAT-125E auto ATU and optmising the SWR again at the freuency of transmission. Normally I am able to get the SWR from 1.2-1.8.

RF Combiners (or Splitters)

One of the tangled things for which there seems to nothing very clear on the web is about combiners. specifically a way to combine two RF signals with zero phase shift and equal gain on each channel.

These are typically used in an SDR to combine to two outputs of the RF mixers which have 0/90 phase RF and 0/90 AF inputs.  So here is what I have found, three solutions:

The single transformer solution, 50R in 25R out

Port A = 50R

Port B = 50R

Port S = 25R

Rint = A+B = 100R

Obviously this does not have both 50R input and output, the two ports on the right are 50R, but the combined one on the left is 25R.

The two transformer solution with matching 2:1 to give 50R inputs and output

J2  = J3 = 50R

L1 tap mat 0.7 to give 2;1 ratio, J1 = 50R, tap = 25R. 10:7 turns FT37-43

L2 Bifilar 10t+10t FT37-43

R = J2 + J3 = 100R

C for HF compensation.

A single transformer with an impedance ratio of 1.4:1, to give 50R inputs and output

R = 50||50 = 25R (51+51R parallel)

10t / 7t+7t bifilar FT37-43

Starting on a new SSB DIGITAL EXCITER

PHASING SSB_EXCITER

Note: all files and sketches on my M0IFA site

I have fiddled about before to use a Teensy 3.2 processor to create 0/90deg audio outputs, using Hilbert filters, and then mixing these with RF 0/90deg from a Si5351 synthesiser (inspired by ZL2CTM). I got fairly good results using a Teensy 3.2 processor. But I could not improve the filters which used just 100 taps in the filters. This caused the +/-45deg phase shifts to fall off at lower frequencies, which in turn lead to unwanted opposite sideband generation at lower audio frequencies.

The solution to this I think is to increase the number of Hilbert taps, to 200-400 say. I have now mapped out a new design:

1. A Teensy 4.0 will both generate the 0/90 deg audio, but also scan the rotary encoder and drive a small OLED display - showing frequency, tuning step and sideband selection LSB/USB.

2. Use commercial mixers, a pair of ADE-1+'s and a combiner (see blog post above). This will drive an output amplifier which will be an MMIC such as the MAR-6 or ERA-1.

Here's a pretty poor hand drawn circuit I will try out - this is in no way definitive, just to give you an idea of my thinking...

USB/LSB switching is in the Teensy audio chain.

The Si5351 two outputs CLK0 & CLK1 have a 0/90 deg phase shift, programmed in software and good across the HF bands, though initially I target to test them out at 7.1MHz. These feed through a couple of pi attenuators to set the drive and impedance to the mixers (target +7dBm). The audio inputs to the mixers are via a couple of diplexers. I am looking into combiner designs for the two outputs of the mixers - see next blog post, but will try out first a single transformer with windings of 7+7/10turns.

The Teensy connections are

 

Teensy 4.0 and Audio board

The Teensy is very flexible with the function of the pins and great care has to be taken to use the right ones. Several are dedicated to communication with the Digital Audio interface board. The Rotary encoder is on pins 2 CLK, 3 DT, 4 SW as usual. And the OLED shares the Audio board I2C serial bus in pins 17 SDA, 18 SCL. The Teensy will be supplied by a 5V regulated supply and the Encoder and OLED run from the 3V3 output. Pin 15 is brought out by the audio board as a "vol" pin.

The first effort software uses the Arduino IDE with the Teensyduino (on PRJC web site) add-on installed. And,

1. The Teensy Audio library, installed when you install the Arduino Teensyduino app above

2. The digital audio development web site, here to define the architecture of the DSP processing

3. Hilbert filter design software, here

3. My OLED header, Oled.h defines the graphic library used (u8g2lib) and a range of display init & functions

4. The actual code is SSB_EXCITER. This includes the Hilbert filter taps (HT_45_200.h).

Hilbert filter taps are generated by the filter design program, and the floating point values are cut and pasted to a spreadsheet, this is output as a CSV file to comma delimit the values. These are then pasted into the sketches HT_45_200m or p. The outputs on the serial monitor of these sketches is code that is put into a single "HT_45_200.h" text file for use in the SSB_EXCITER sketch. Bit comploicated but it works.

The DSP organisation is simple, mixer1 has a gain of +/-1 to switch sidebands.

And the Audio System design Tool outputs the headers and wiring as a header you can paste into the SSB_EXCITER.ino sketch.

I will update my progress as I get to grips with the hardware build. 

And progress is being made, here's the PCB on which I am building the SSB Exciter, lots of wiring and soldering to go... going to be a forest of wire, but hey ho...

Progress - decided not to mount transformer in the middle as gets in the way of wires, will find another spot for them

Close up of board, massive bus board needed! For SCL, SDA, GND, 3V3 & 5V

And it works, it tunes, it displays it selects sidebands & Si5351 outputs signals

Working board, so far

Output at 7.1MHz

And now it has been tested with audio MIC input and generates 0/90 audio outputs. (Note if you are following that new code is now on my web site here, as the first code had wrong DSP signal paths.

When running the Teensy with the SSB_EXCITER.ino sketch and an audio input to the MIC connection it gives these audio outputs with a phase shift of 0/90 deg for LSB or USB transmission, the BW is around 100Hz to 3kHz.

DISASTER: made a huge mistake when wiring board: connected PSU 5V to 3.3V rail. Blew up everything. So project stopped for now... and may never start again... expensive to buy another Teensy...

5W PA

 I have been working for some time to fathom the secrets of power amplifiers. I have been targeting 7MHz and an output of 5-10W. The output stage needs a load impedance according to this

This is the circuit I have come up with, sorry poor photo, click to enlarge,

5W 7MHz PA

Transistors 2N3904 - 2N5109 - IRF530

The gain is adjustable by the trimmer on the emitter of the first stage. With the supply at 13.8V an input of -2dBm gives a gain with outputs at each stage of +12dB, +20dB and +37dBm. The board copper is used as a heat sink and does not get excessively warm. An external 40m LPF is used, but could easily be included on the board - may need a little more space.

5W PA board, 

the emitter resistor is made of 3 x 3R0s as I don't have any 2W 1R resistors.

Power output, via a 40m LPF,  is an easy 5W @ 12V, with quite a bit more at 13.8V

Power output 50R

Some home grown measurements

Not perhaps very accurate perhaps, but a few small boxes of tricks can make home brewing and measurements possible.

Oscilloscope - who would be without one? I use a LabNation USB interface scope with two channels, with software running on my MacBook, seems to be enough for my audio to HF needs.

Oscilloscope

Signal generator - I would not be without my general purpose signal generator built with an Arduino Nano driving an Si5351 synth chip (and it has a RTC built-in too). Programmable!!! can be used for many applications. A VFO from <1MHz to > 150MHz, with the possibility of up to 3 outputs. Another use is for generating RF signals with 0/90deg phase shift for experimenting with phase method SSB RX & TX designs. And then again it can generate CW for beacons or training and QRSS, WSPR transmissions. The only thing to recognise is that the outputs are square wave, not pure sine wave. The output is +12dBm (860mV , 15mW into a 50R load).

Signal Generator

Power or RF voltage meter - Based around an AD8307 chip this RF meter has a high impedance or a 50R dummy load input. As the AD8307 is a log amplifier it has a  huge range of -40 to +40dBm (1uW to 10W!). 

Three little boxes - a 7MHz low pass filter, a 20-10-5dB attenuator and a 30dB tap with 50R dummy load built in. 

Top :  LPF, Centre: 20-10-5dB attenuator, Bottom: 30dB tap with 50R dummy load

The 30db tap is particularly useful for checking > 10W power inputs to the dummy load, giving a -30db output to a second RF Meter I have with a direct AD8307 input capable of  -70 to +20dBm with a 50R input impedance. It is also useful the other way round.

Power amplifier - one thing I have missing is a small power amplifier, with say up to 10W +40dBm output from a 0dBm input, input and output 50R. I am working on this design right now as a project to better understand medium power PA stages and preamplifiers.

MY SHACK

13.8V PSU, 

GPS clock & QCX+, 

paddle  antenna switch

and ELAD FDM_DUO transceiver

The RF Signal amplifier

 I was talking to another amateur the other day, about a preamp stage for a xtal filter and a following amplifier stage. The xtal filter needed to see a source impedance of 250R and a load of the same.

Now the basic transistor, common emitter, transistor and equivalent circuit looks like this

Where the collector is a current source/sink. A complete circuit would look like this

Transistor AC gain is B

Where Rs is the input source impedance, RBIAS is the bias resistor(s), re is as above, Rc is the collector load (which determines the output impedance) and RL is the load (for maximum power transfer Rc = RL).

Using this circuit we have

Gain Av = Vout/Vin = Rc * ic / re * ie, 

  and since ic ~ ie, Av = Rc/re

Input impedence Zin = re*ie/ib = re*B (when RBIAS >> B*re and B >> 1)

Zout = Rc

So an amplifier with 50R source impedance and 250R output impedance (for the xtal filter source impedance) would look like this

Amplifier

Assuming

B = 100

Ic = 10mA

re = 26/10 = 2.6R

Transistor Zint = 26/10 * 100 = 260R

Amplifier Zin = 50R with a 1:3 transformer

Amplifier Zout = 250R

The transistor bias can be arranged however you want provided the bias resistors are large compared to Zint.  

The output of the xtal filter also needs to see a load impedance of 250R, but this about the same as our amplifier input impedance (260R), so we could have this simple design

Xtal filter post-amplfier

The output side of the post-filter amplifier can be whatever is needed for the following circuits.

But the input impedance is still critical on the collector/emitter current and on the AC Beta (gain) of the transistor. Chosing Ic = 10mW has determined the input impedance.

Feedback

So the next thing to look into is building a feedback amplifier where it might be possible to determine the input/output impedances using some sort of negative feedback across the transistor. And so it is.

This is a simple amplifier (sorry poor screen grap from youtube...)

You can see

1. A 1k feedback resistor, meaning with 50R input source, the gain is Rf/Rs = 1000/50 = 20.

2. The bias is set with the Ve = 3V, and bias is set by the 470R resistor, which does not affect the source resistance much

3. The amplifier gain has to equal the the feedback gain, i.e. 20. This is determined by the RL/Re. RL is 200R (chosen so that a 1:2 transformer drives a 50R output load). And so Re has to be 200/20 = 10R.

Two radios need an antenna switch

Foolish I know, but if you have two radios then you need either two antenna or an antenna switch! I was fed up with unplugging one and plugging the other so I made this

The antenna switch and

my QCX+ and FDMDUO TRXs

The unselected radio is terminated by a 51R resistor, just in case you accidentally put it into TX mode!

Morse?

I am using the QCX+ at the moment to learn morse code - with the help of a random letter sender based on an Arduino Nano and a Si5351 sythesiser, transmitting on nominally 7023kHz (CW_LEARN code here). The QCX+ is tuned to 7023.0 after calibration using a GPS NMEA & PPS input (Menu 8.11 & 12).

By the way the "clock" (see pic above) under the QCX+ is my GPS clock, used also to provide NMEA and PPS signals to the QCX+. The GPS operates at 3V3, so I use a 3V3 <-> 5V converter to drive the QCX+.

For fun

I am playing also with a Pixie TRX, also tuned to 7023kHz and picking up the Arduino morse tutor.

The Pixie $5 TRX!