Sunday 27 November 2016

A WiFi-Controlled VFO

Now that I have placed a DDS on a board with Wireless connectivity, it is only obvious to set it up as an AP Web Server and allow remote control via a 'soft' user interface...

When I first started to play with the AD9850 DDS, I made some beacons and wrote the Kanga VFO demonstrator code (as application code to support the Kanga / m0xpd DDS shield). Well, now I've produced this new internet savvy AD9834 DDS board, I've followed the same pattern; I've done the beacon thing and now - in this post - I'm covering the VFO. Only, of course, this VFO won't have physical knobs and buttons like its predecessor. Understand, it could have all the physical controls if you wanted it to - but that's not the point. Instead...

When  the new code fires up, the ESP8266 sets itself up as an access point and offers a new wireless network:

which you can join from any phone, tablet, computer or similar wireless enabled device. This will provide the interface to the VFO. The network name 'm0xpd Kanga DDS 94TD' is formed of a generic part ('m0xpd Kanga DDS') and a four character index associated with the particular board (such that two VFOs in the same area could operate independently).

Opening a browser and going to the VFO's 'web page' will open the simple control interface seen below, which reports the frequency at which the oscillator is running, offers 'buttons' to adjust the frequency and 'buttons' to change band:

The picture above is a photo of my iPad mini screen, controlling the VFO. I've added the red annotations to make it clearer for you.

The web page is generated entirely by the m0xpd / Kanga ESP8266 - AD9834 board - the iPad is just interpreting it (as HTML).

Clicking on any of the 'Adjust Freq' hyperlinks will cause the VFO frequency to change according to the label. Clicking on any of the 'Select Band' hyperlinks has the obvious effect.

All this is rather dynamic and needs a video to demonstrate (which I haven't provided) - so here are some rather dull 'stills' of the system on 40 metres

and 80 metres...

And here's a 'sniff' of the requests received from the web page of sequential increments and decrements in frequency whilst on 80m, along with the resulting frequencies...

This code will be released as an application 'demo' for the new m0xpd / Kanga ESP8266 - AD9834 board (along with a multi-mode beacon).

Of course, the demo code above is intended only as an illustration of what is possible. The rather dry web page could be replaced by an application, written specifically to control a different piece of VFO code, etc.. This is just a start point - but it sure gets me thinking...

There are other exciting opportunities to be exploited via this access point - watch this space!

...-.- de m0xpd

Friday 25 November 2016

Occam Going Dutch

Kees, pa5cw, has produced a new variant of my 'Occam's Dirk' software for a multi-band CW transceiver.

'Occam's Dirk' was a multi-band development of the original 'Occam's Microcontroller' concept, adding automated 'CQ' calls and one or two other little refinements (like RiT, VFO A/B, etc). The original 'Dirk' was presented with an Si5351 in the RF generating role (as I wanted to try using the then new Kanga / m0xpd Si5351 shield).

Kees has decided to revert to the AD9850 DDS for his version of the code, which makes the architecture similar to that introduced in my 'Kanga VFO' demo software (and elsewhere).

Kees has also added some new functionality: the means to change keyer speed using a voltage input to A0, most conveniently generated by a potentiometer. The original code had speed change as a software function under the menu system - which works, but isn't immediately accessible for a quick change.

I've always built keyers into my software (and provided both a paddle and a straight key input) but - if I'm honest - I've seldom used these keyers in anger. Here at the shack, I usually drive all my rigs (via their straight key input) from the old faithful "Funky Keyer", which has its own physical speed control. Thus, menu-based speed adjustment of the software was never a big handicap for me. But Kees' approach certainly is convenient (at the expense of one knob).

Any of you interested in trying Kees' code can find it here on the Occam's Software page. I haven't tried running it - but I have confirmed that it compiles correctly.

Many thanks to Kees for sharing his work,
...-.- de m0xpd

Saturday 19 November 2016

SNA Junior

I have (finally) got round to building my own instance of DuWayne, KV4QB's prize-winning Scalar Network Analyser Jr:

and a great little instrument it is too!

DuWayne and I have been corresponding for a couple of years, sharing mutual interests. I was pleased to be able to give his work a shout in both the printed and  'spoken' version of my talk at this year's Four Days in May event in Dayton and - more importantly - to catch up with the man in person for a quick eyeball QSO. I also got a PCB for SNA Jr, which has been sitting on the bench for months - until last week.

The SNA board finally bubbled up to the top of the pile and I looked around for the bits I needed to complete it. Perhaps I should explain (to those of you who don't know) what's involved...

DuWayne's baby uses an AD9850 in one of our familiar modules to generate RF, under the control of an Arduino NANO. You can read on DuWayne's blog how the SNA Jr is the descendant of earlier experiments in which an Si5351 was used as the signal source.

In the SNA Jr, the output from the DDS is fed to the device under test and the returned signal is observed in a detector system. DuWayne has 'history' in using simple diode detectors in this role (and I was praising kv4qb for this minimalist approach in my talk at Dayton) - again, you can read about this lineage. However, the SNA Jr now replaces the earlier simple diode detector with a fancy AD8307 detector, in the well-known Wes Hayward, w7zoi, circuit. This gives superior performance in terms of dynamic range and 'linearity'. Also, with the availability of cheap AD8307s (of dubious parentage) from China, this option is also becoming attractive for cheapskates like me! [I have some Chinese AD8307s on order and will report back on performance when the slow boat docks.]

You can't see the detector in the photo above, because it lurks under the screen - so here's another shot (with apologies for my wayward handling of some of the SMT devices):

The detector is supposed to be enclosed in a screening can, which I haven't made yet - so final performance will be better than I'm going to show you below.

The DDS RF source and measurement of the RF level returned from the device under test are all under the control of the little Arduino NANO, which runs a sketch provided by DuWayne. This sketch compiled for me under Arduino 1.6.12. The user interface is provided through just a rotary encoder and the 1.8 inch TFT screen.

I found I had everything needed to build SNA Jr in the 'junk box' - except the screen and a spare NANO, so these were quickly ordered through usual suppliers.

The result, as you see above, was simple to put together and works very well.

DuWayne's software offers a number of options, including a 'signal generator' mode, in which the output of the DDS module is set at a single frequency, whilst the returned RF amplitude is displayed numerically and on a bar display (useful when the numerical display is flicking between two values). This mode is illustrated in the graphic below, which shows the system driving a simple switched attenuator, seen in the graphic, with and without 20dB of attenuation switched in (two 10 dB stages).

I'm sure the (in)accuracy of the -20dB step is down to my cheapskate attenuator (with its low tolerance resistors, lack of screening etc.), rather than SNA Jr.

Another, more important series of modes sets the DDS module generating RF sweeps, which result in graphical displays. These are illustrated below, in which I've contrived a test of the low pass filter which has been conditioning the output of the 'connected beacon' (Blogs passim) on 30m.

SNA Jr can also be used with various 'attachments' such as a 40dB tap (with which DuWayne's software allows it to function as an RF Power Meter) or a Return Loss Bridge, with which it can perform SWR Scans.

Here's a scan looking into my (g5rv) antenna, with (right) and without (left) the Made-from-Junk 'Deluxe' Versa Tuner II switching in tuning appropriate for operation at the CW end of 40m...

As you see from my additional labeling in the graphic above, the scan was set up to run from 6 to 8 MHz. I reported an equivalent measurement on my own system (originally reported here) in one of my slides at FDIM (although it was presented in terms of Reflection Coefficient, rather than SWR)...

There's even more to SNA Jr - it can even locate minima to impersonate a dip meter (but I haven't been able to try this yet).

I said at the beginning of this post that SNA Jr is 'prize winning'. DuWayne won the 'Best in Show' award at  the homebrew competition at FDIM (which is no mean feat, given the very high standard of the submissions I saw there, in several widely different categories). The prize was well-deserved.

The project was written up in QRP Quarterly (vol 57(3)  pp 22:25, July 2016) :

which is nice to read. But the best thing to do is to get the information from DuWayne's blog and build an SNA Jr for yourself  - or take the inspiration to build something similar.

Great fun - thanks, DuWayne.

...-.- de m0xpd

Saturday 12 November 2016

ESP8266 Production PCB

The first sample of the production version of the PCB for the ESP8266 / DDS system arrived a couple of weeks ago.

As you see, it follows the plans established in the earlier prototypes, with the surface mount components already fitted (such that tyros don't need to face the challenge of dealing with these pesky little things).

I've been detained by work and by a short break up in M-land, where I stayed on the banks of the River Nith, playing at being mm0xpd/p. However, now I'm home, I've stuffed the new board - this time with the intended 1/8th Watt resistors:

I set her up for a quick test yesterday morning, programming the ESP8266 with my beacon code and plugging a Kanga / m0xpd TX shield on top of the new PCB. Here are the WSPR spots accumulated on 30m in the first thirty minutes (from 09:00 GMT):

Looks like things are working FB.

I had a nice time working CW, PSK-31 and even a little SSB on my TS-480 yesterday, so the station was out of commission for beacon operation for most of the day. But I turned the beacon back on in the evening and let it run overnight.

Here are the accumulated 24 hours of spots on 30m ...

(It looks even better as I write, 'cos I'm also down to Stewart, w4mo in Venice, FL - but you've got to draw the line somewhere).

I hope it won't be long before Kanga can offer the new board as a kit to anybody interested in playing similar games.

...-.- de m0xpd