Assembled 20m WSPR transmitter

Here's a picture of the top and bottom side of my 20m synthesised WSPR transmitter. Measured results were 1W pep and reference spurs 42dB below this, or less than 100uW.

Now to put it in a box and really smoke test it.

WSPR - 20m Success

Wow! After 24 hours continuous operation I'm amazed how well the prototype transmitter my son put together for me worked. Despite the output of 100mW it has been heard far and wide.

So a big thank you to Alex for soldering up a prototype and Georgia, my daughter, for helping me get a quarterwave groundplane  strung up in a gum tree.

The second really pleasing aspect is the synthesiser has exhibited no drift. Despite running in a shed with no lid and over 10 degrees of ambient temperature variation it has not suffered from frequency drift eg

Timestamp	Call	MHz	SNR	Drift	Grid	Pwr	Reporter	RGrid	km	az
2016-05-17 00:42	VK6TT	14.097069	-18	0	OF78	0.1	VK7ZL	QE37pd	3104	123
2016-05-17 00:36	VK6TT	14.097069	-10	0	OF78	0.1	VK7ZL	QE37pd	3104	123
2016-05-17 00:30	VK6TT	14.097069	-17	0	OF78	0.1	VK7ZL	QE37pd	3104	123
2016-05-17 00:24	VK6TT	14.097070	-16	0	OF78	0.1	VK7ZL	QE37pd	3104	123
2016-05-17 00:18	VK6TT	14.097071	-16	0	OF78	0.1	VK7ZL	QE37pd	3104	123
2016-05-17 00:12	VK6TT	14.097072	-22	0	OF78	0.1	VK7ZL	QE37pd	3104	123
2016-05-17 00:06	VK6TT	14.097073	-12	0	OF78	0.1	VK7ZL	QE37pd	3104	123
2016-05-17 00:00	VK6TT	14.097074	-15	-1	OF78	0.1	VK7ZL	QE37pd	3104	123
2016-05-16 23:54	VK6TT	14.097075	-16	1	OF78	0.1	VK7ZL	QE37pd	3104	123
2016-05-16 23:48	VK6TT	14.097076	-20	0	OF78	0.1	VK7ZL	QE37pd	3104	123
2016-05-16 23:42	VK6TT	14.097077	-12	0	OF78	0.1	VK7ZL	QE37pd	3104	123
2016-05-16 23:36	VK6TT	14.097078	-14	0	OF78	0.1	VK7ZL	QE37pd	3104	123
2016-05-16 23:30	VK6TT	14.097079	-13	0	OF78	0.1	VK7ZL	QE37pd	3104	123
2016-05-16 23:24	VK6TT	14.097079	-9	0	OF78	0.1	VK7ZL	QE37pd	3104	123
2016-05-16 23:18	VK6TT	14.097079	-18	0	OF78	0.1	VK7ZL	QE37pd	3104	123
2016-05-16 23:12	VK6TT	14.097080	-26	0	OF78	0.1	VK7ZL	QE37pd	3104	123
2016-05-16 23:06	VK6TT	14.097081	-22	0	OF78	0.1	VK7ZL	QE37pd	3104	123
2016-05-16 23:00	VK6TT	14.097082	-17	0	OF78	0.1	VK7ZL	QE37pd	3104	123
2016-05-16 22:54	VK6TT	14.097083	-19	0	OF78	0.1	VK7ZL	QE37pd	3104	123
2016-05-16 22:48	VK6TT	14.097084	-21	0	OF78	0.1	VK7ZL	QE37pd	3104	123
2016-05-16 22:42	VK6TT	14.097084	-19	0	OF78	0.1	VK7ZL	QE37pd	3104	123
2016-05-16 22:00	VK6TT	14.097082	-23	0	OF78	0.1	VK7ZL	QE37pd	3104	123

So I'm calling this a success. I've made a board for a final prototype and once that is completed to confirm there were no layout errors I will be ordering commercial boards for a kit  for those that wish to duplicate this.

Cheers
Richard VK6TT

WSPR Transmitter - Amost There

While I only ran the 40m prototype for a few nights before thunderstorms saw me disconnect everything. I had a ball seeing how far the 100mW could go. More thunderstorms forecast for tomorrow evening but after that I will put it back on air for a few nights.

I will build a 3rd prototype to confirm everything is as expected. While I wait for parts to arrive my focus will be on getting the second prototype on air on 20m.

It's been a long road but I have almost arrived. 

73's
Richard

WSPR Transmitter - update

It has been an interesting few days with lots happening keeping me away from the workshop. In that time I did some reading and discovered a flaw in my logic in relation to the output power of this project. So with a slight circuit and layout revisions I have made some progress.

Firstly, the intended output power was 500mW pep. Since this is a DSB transmitter I incorrectly assumed I would have 250mW in each sideband. This is wrong. I would in fact have 125mW in each sideband. The reason is that when the voltage of both signals is in phase the output voltage doubles, giving a fourfold increase in power.

So I revisited the circuit and concluded that 1Wpep was as far as I could stretch the class A broadband final stage before matching networks would be needed. Since the long term goal was to use this board for other bands I didn't want to move away from the concept of a broadband transmitter. It appears that the 2W of heat the transistor is dissipating is manageable and the device is still operating in it's safe operating region. So back to 250mW of usable output.

I did look at using a broadband push-pull arrangement, and even parallel transistors as in the JBOT (just a bunch of transistors) concept. But the first problem was to source ferrite material at reasonable cost. I have ordered some material that may be suitable but it will take weeks to arrive and I want to finalise the boards. So despite having great success in the past using ferrite tubes removed from EMI filters on VGA cables, something I will write up for you one day, I will just push on. The only problem with junk-box ferrite is each transformer is bespoke because the properties of the ferrite were unknown. As much as I love experimenting I strive for reproducible results at all times.

One step forward was to revisit the layout. I now have the layout dimensions as  110mm x 60mm so it will slide into a commercial extruded aluminium case.While my intended housing is an old VHF commercial radio case it is good to have a nice case like this as an option, albeit more expensive than what I will be using.

That's it from me for today.
Richard

WSPR Transmitter

So after testing I was excited at how well the 20m WSPR transmitter my son soldered up for me worked. It performed flawlessly on the bench with a power output of 0.5W, just as intended. If it hadn't been such inclement weather this week it would have been on air by now. Hopefully if fines up in a few days so we can hoist up a dedicated antenna for this.

I have attached some shots of the top and bottom PCB layers. Pushing stuff around so the silkscreen would be legible took a lot longer than I expected. But it is almost done now and I will route the board again soon and get some fabricated in China.

Naturally I'm pretty chuffed about how well this project has come along. If you'd like to emulate it please let me know.

Regards
Richard VK6TT

WSPR Transmitter

It was an interesting week. I had my 11yo son put together my second prototype WSPR transmitter. Watching someone build this was both instructive and rewarding. I saw first hand how important it was for consistency in labelling parts on the circuit diagram and parts placement diagram.  And rewarding to see that one of your creations can be replicated.

I lost a lot of time testing the synthesiser though. It took me several attempts before I realised the reason the phase lock loop (PLL) was not locking up was not related to the loop filter which I thought was the issue. Instead, it was the oscillator was too low level for the PLL counters to work properly all the time.

As the frequency changed the oscillator level also changed. And it turns out the levels were right on the limit for reliable counting. So it would appear to be working properly when the frequency into the PLL was too low, but as the frequency increased it started counting in error so the error signal into the loop filter was still saying the oscillator was too low.

All is good now though. After some more testing I should be in a position to release further details shortly.

73's
Richard