Monday, 15 August 2016

EME2016

This last few weeks have really all been about getting ready for EME2016, later this week.
The International moonbounce Conference is held biannually. The last one was in France and the previous one in Cambridge, UK. This year's Conference is being held near Venice.
As has become usual for me, I am offering a selection of my EME VLNA preamps, both kits and semi assembled  kits. It has been the preparation of these preamps that has been occupying my time. In addition to the VLNA preamps I have been building a number of PGA432 70cm preamps for evaluation by a number of other amateurs as well as taking one or two to EME2016 to be independently measured.
And.....the new Anglian3 transverter boards arrived last week and so I've been busy building up some of these for testing. I'm pleasd to say that they work as well as the Anglian 3 that they replace, and with some added features. Unfortunately I hve not had time to get Anglian 3 kits redy for EME2016, although I will have the first ones on show.
I've also been finishing off my presentation on failures in LNAs (too much RF - they don't like it up 'em)
The nub of the presentation is just how much abuse a HEMT LNA can take before it fails, and what happens along the way! I'm scheduld tobe the second talk. It will be nice to get mine out of the way early on, although as an ex-lecturer I don't mind standing up in front of such a bunch of established academics, enthusiasts and EME  experts.....much!
As well as all these preparations I have been trying out the new digital mode of QRA64. This is in alpha release for testing and will form part of the new release of WSJT-X. Possibly after EME2016. Joe Taylor, K1JT, and Nico Palermo are presenting at the Conference.
My testing has been with a well-known EME op, but on HF. Honestly!
QRA64 is a weak signal mode, well suited to EME, and usable at HF. It answers many of those critisisms of digital, aimed at how some of the computer processing has been considered 'cheating'. Troglodites..........
QRA64 and the new fast mode of MSK144 could quickly supplant some of the existing digital modes like JT65 and FSK441. We will see what happens over the next few years.

Sam



Sunday, 24 July 2016

Anglian 3 update

For those of you who don't subscribe to Twitter you will not have seen my progress with the Anglian 3 transverter. I have been tweeting regular updates (for anyone interested!)
The second set of Prototype PCBs were deliver last week and a second unit built up to prove the design works as it should. I'm pleased to report it all worked as planned, although there were still a few very small errors on the PCB. Version 3.1 PCBs have now been ordered and are expected in a few weeks. 
The new Anglian 3 is designed as a direct replacement for the Anglian 2, but with a few small improvements. These include transmit enable delay, better transmit amplifier gain distribution, programmable transmit IF gain (for use with ICOM rigs, with their low transverter output level), better transmit IMD, but retaining the same excellent receive performance, lockable local oscillator, separate IF receive and transmit connections and all SMD construction as the 2.
Unlike the Anglian 2 I have no plans to offer the Anglian 3 as a fully assembled PCB, as this has proven to be uneconomic for small production runs of boards with specialist RF parts.
However, I may offer partially assembled boards, depending on demand.
Also, depending on how Brexit affects component purchase (many parts are sourced from overseas), I anticipate the cost being lower than the Anglian 2.

Sam



Sunday, 17 July 2016

70cm low noise preamp for Contesting and Satellite mode J


I have had a 432MHz version of my PGA144 preamp on the stocks for some time. I never bothered toadvertise  it because I knew it needed more development. I took the time over the last few days to look at the filters again and then put one together. As always some tweaking of values was required compared to what the design software said, die to stray capacity of pads etc.
The results are quire pleasing.
I've not yet measured the input two tone third order (IIP3), but it ought to be pretty good.
Of particular interest is the deep notch I have been able to place on 144-146MHz to suppress signals in this range. Similarly the frequency response is well down in the Upper digital TV band, GSM band and at 23cm.
The noise figure measures under 1dB ( 0.8 to 0.85dB on the prototype) on an open board, which is not the ideal way to measure noise figure due to 433MHz SRD interference.
As the accompanying plots show, the input match is exceptionally good, so that a filter could be placed at the input if required and its frequency response would not be badly affected. try that with many common 432MHz designs!
Noise figure and gain (can be varied if required) at 435MHz

Yes, I know that attenuators is spelt wrong......


Input match (in yellow) marker 1 at 435MHz and 145MHz with frequency response in blue withe same markers

Tuesday, 12 July 2016

Anglian 3 update

As I have run out of the current Anglian 144MHz transverter kits I decided that rather than order more boards to the current design, for reasons that will become obvious as you read the following, I needed to design a replacement. The replacement is the Anglian 3.
It is the same size as the Anglian 2. 

The redesign gave me the opportunity to make improvements, especially on the transmit side. The original Anglian 2 was designed to work with HF rigs like the K3/S or TS590 that have typically 0dBm 28MHz  transvert output. I found that I could get it to work with the typical -20dBm output from some ICOM HF rigs by slightly modifying the transmit IF input attenuator. Whilst the output would not quite reach 100mW (+20dBm) when driven at this level, it has proven adequate to drive a Mitsubishi module to full output. However, I never quite felt comfortable driving the mixer that far below its nominal IF input level where the full IF to RF isolation was achieved. At this drive level the transmit amplifier chain needs the full gain of the two transmit amplifier stages.

Some users have reported a short burst of transmit output when switching from receive to transmit. This appeared from my tests to be dependent on what was connected to the receiver input or transmitter output. The gain from receiver input to Anglian 2 transmitter output could momentarily be 70dB! It didn't take much additional gain in a masthead preamp or PA and an RF burst could appear. The changeover from receive to transmit originally had a small overlap of a few ms where both were on. Reducing regulator reservoir capacitor values speeded up switching since  there was then less charge to keep the receive amplifier working for those few ms.
However, the proper solution was two-fold. Decrease transmit amplifier gain and introduce a delay between the receiver switching off and the transmitter switching on (sequencing).

An improved output amplifier, with lower gain and better IMD, has solved the first problem. A MOSFET switch instead of a bipolar switch improves switching speed and allows the introduction of a simple but effective sequencing delay to solve the second problem.

Of course the reduction in transmit gain now meant that a low IF transmit input could become a real problem. This was solved by the introduction of an IF transmit (28MHz) amplifier stage. If not required it can simply be left out or a link arrangement allows it to be bypassed.
With the new TX IF amplifier the Anglian 3 now easily produces more than 100mW output at 144MHz for less than -20dBm input at 28MHz. Harmonics are less that -60dBc at 100mW output.

I had also been asked if it was possible to include a 28MHz band pass filter in the 28MHz transmit IF input, which I have now done. The TX IF input now passes through a single stage bandpass filter centred on 29MHz.

The LO chain is largely unchanged except that it now works entirely on 5V and runs from a single 5V board voltage regulator. The Anglian 2 had three voltage regulators! The LO still produces +20dBm output and can be locked to an external 116MHz input, if higher stability is required.

I have not changed the receive converter side as it already has about optimum performance for the design and components used.

My plan is for kits to be available later this year. I need to check the revised PCB and if satisfactory, will order a batch of production boards in August. I will probably initially revert fo full kit version with assembled SMD boards coming later. Small runs of assembled RF boards are not economically efficient! 

The aim is to keep the price (component costs permitting) similar to the Anglian 2, or maybe a bit less. We'll see how the pound fairs later this year.

Sam





Tuesday, 28 June 2016

Yet more IC7300

I have taken delivery of an Alpha Delta 4 way coax switch to direct the output from the 7300 to the 4/6m antenna, HF antenna, dummy load or to the Anglian 2m transverter.
I have opted for the simple approach go interfacing to the Anglian and Gemini 2 PA.
I use just a 20dB power attenuator after the switch into the Anglian. I have the 7300 28MHz band power set to 8W. The reasoning is that if I was to accidently forget to turn the 28MHz power down (it is set per band) then the max that would hit the Anglian would be 1W and transverters tend to be more tolerant of overlod due to the usual input attenuator before the mixer. The 20dB attenuator is a25W unit which will stand a degree of overload before burn out.
My tests using the spectrum analyser on max hold etc indicated no significant spiking of output power when set to low power. I have seen some reports of ALC overshoot on SSB and CW. I have not seen this.

I had an interesting problem when set up. The Gemini 2 would fail to SWR when driving bove 100W output. This was eventually traced to the bulkhead lightning arrester on the antenna feed starting to trigger at this

level. The arrester has been replaced by an N bulkhead adapter. All is well now.

This evening saw several new locators added to my 6m total in the NAC contest. Now to try and leave  the system as it is now instead of yet another rearrangement.........
WTS

Sam






Saturday, 18 June 2016

More IC7300 tests

Since I want to try my 2m system with the IC7300 I was interested to know how well the power control on the new Icom works. In particular, does it spike on switching between receive and transmit?
I set the rig to 28.2MHz and (carefully) connected it through a 150W/40dB attenuator to my HP8592 spectrum analyser. The analyser was set to 20mS sweep and 5MHz span. With the IC7300 set for full output, the reference level was set to 0dB on screen. The mode was SSB. The analyser was then set to Max Hold in order to capture any spikes. The IC7300 mike gain was set to zero.
Keying the IC7300 transmitter on and off produced a small spike, caught by the Max Hold on the analyser. The level of the spike was around -55dBc
The IC7300 was keyed in different modes with similar results.
Winding the power down to '0%' resulted in the spike amplitude being similarly reduced.
Power output at 0% was 22dB down on maximum output, or about 700mW. This is below the  2W claimed.

It would appear that the IC7300 does not power spike to any significant level.

Connecting the IC7300 to my 2m Anglian transverter with DG8 mast head preamp, though the 40dB attenuator, produced a barely audible increase in noise.  No surprise there!
Reducing the in-line attenuator value to 20dB result in a greatly increased noise output and lots of beacons appearing on the 7300 spectrm display.
16dB is about optimum for my system and so it was with an MCL 15dB 2W attenuator that the system seems to work well and achieve about as good a dynamic range as I could expect with this set up.

Using a 15/20dB power attenuator is required to further reduce the power into my Anglian. I don't like to dissipate more 28MHz power in the transverter than absolutely necessary.

Although the IC7300 does not appear to spike, this is still not proof against forgetting to turn the power down.............

A 20dB, several watt, IF attenuator will give some protection and then fuse if severely overdriven. Hopefully you will have noticed a problem by then and there could even be time for an overdrive protection circuit to operate. I do not advocate this approach in order to drive an LDMOS SSPA as it would probably fail a whole lot faster than any protection circuit unless carefully designed to prevent any drive reaching the PA before checking its level.

Sam



Monday, 30 May 2016

Cold Bank Holiday

So far it has been a very dull and chilly Bank Holiday Monday, so I've been using the time to catch up on some writing. I have two Conference papers to write with both of their deadlines looming in the next week.

In addition to those I have an equipment review to write for Radcom, so what am I doing writing this blog as well?  Good question, but whilst the words were flowing seemed like a good time to put something onto the blog about the new rig.

The new IC7300 has been working out well, with a number of new locators and DXCC worked on both 6 amd 4m since Saturday. I've also been thinking about how to add a sutable transverter interface to the 7300 so that I can use it to access other bands.

The touch screen spectrum display is a pleasure to use. That doesn't mean it's perfect, by any means, but it is a big improvement over some other implementations and I think in ICOM's next SDR transceiver it will be even better.

One of the shortcomings of the touch display is that when you select the signal you want to listen to it brings up a static, magnified, view of the spectrum around the signal. You then need to touch the screen again to select the signal. I find I cannot touch the right spot to tune accurately to the signal without then having to use the tuning knob to tune it in accurately. I am sure that with the power of modern DSP processors it should be possible to do this automatically. Of course you would then probably need a 'defeat' facility in the event you wanted to select a weak signal next to a strong one!
One other observation is that there is no 70MHz band 'button' necessitating a frequency input, which means several button presss to go to, e.g. 70.2 input
You quickly get used to this, but a 4m band button would be nice, please, Mr ICOM
More anon.
Sam