John's homebrew pages
Transverter for 23cm (1.3GHz - with 432MHz IF)
This is a more complicated project so sections will be added as they are completed. I have left them more or less in their original form, so the story gradually unfolds.
- Oven Controlled Crystal Oscillator (OCXO)
- Local oscillator multiplier chain
- Mixer board
- Receive board
- Transmit board
- Modifications and testing
- New driver amplifier
- First QSO
This is the first of a series of projects aimed at a move up in frequency. It will be a key component in a transverter for 23cm.
As usual there are lots of designs on the web, and descriptions in the RSGB/ARRL Microwave Handbook. I'm using a basic oven control circuit as used in the G8ACE designs.
An oven is a box that you heat up, in this case something to keep a crystal at a constant temperature using a thermostat. I have these nice chunks of copper bar bought for heatsinks, so a piece of that is ideal for the oven.
Here are the main components: the copper bar, power transistor to heat it up, crystal to provide the frequency reference, and negative temperature coefficient thermistor to provide feedback for the thermostat circuit.
It didn't take too long to drill out three holes and file them out using needle files into an appropriate slot to hold the crystal.
Then a hole next to the crystal is drilled to take the thermistor, and a hole drilled and tapped to mount the power transistor. The holes at the end are to mount the block on the circuit board.
Just checking that the crystal fits nicely ...
Electronics - the thermostat
The oven control circuit was assembled first. The crystal is not yet fitted; below the slot for the crystal is the thermistor embedded in heatsink compound in its hole, and below that the power transistor. To the right of that is the 78L09 regulator, and to the left the LM358 (only one op-amp used), with a trimmer pot to adjust the setting for the thermistor voltage.
Below is the circuit under test. It works really well; on start up, a current of 600mA heats up the copper; this reduces once the working temperature is reached, and even on the bench it stabilised at about 7mA. Blowing air over it caused a pretty quick increase in the current, as it cooled slightly and put the heat on again in response before stabilising within 10-15 seconds. It should work considerably more efficiently inside an insulated box.
Here is the heater circuit:
The next job was to get the crystal oscillator circuit working. This is a Butler oscillator, and was initially constructed as a "dead bug" prototype as seen below. Most bits are from the junkbox or salvaged - I decided that I could sacrifice the single channel 2m FM set I built in 1978 which is where the transistors and coil former came from.
The oscillator appears to work well, and generates lots of nice harmonics; using the FT-817 I could easily detect the fourth harmonic at 96MHz (which is the required one) and even that at 432 MHz.
It didn't take long to transfer the prototype onto the oven board. Just a few tweaks were needed: the oscillator LC circuit needed a smaller C (68pF instead of 100pF) to enable tuning without the slug poking right out of the former, and the output circuit had to be designed properly. The coil was calculated using RFSim99, then constructed, measured (!) and re-calculated to be sure all would be well. It peaks up on the 96MHz nicely (grey trimmer). The 432 MHz harmonic was used to get zero beat using the FT-817 and tweaking the crystal load capacitor trimmer (green). The yellow wire is an antenna to get a bit more signal out so the FT-817 can be a lot further away - it had a similar antenna poked into its BNC input. Note the crystal is stuffed into the oven and held in with heat conducting paste.
The circuit needs to be tested with the oven now in a thermally insulated box, but all seems to be well, so I can start on the multipliers.
Here's the oscillator circuit in its original form:
The circuit seemed to be working well, so I started on the multipliers.
Note that eventually I made changes to this circuit - see further down under "Modifications to and testing of ..."
Following construction of the OCXO, this is the next key component in a transverter for 23cm.
I decided to use an IF of 432MHz (to 1296MHz) so that I have the larger frequency range at 70cm to play with to allow me to use more of the 23cm band. That has the advantage that the IF is further from the wanted frequency too.
Given an OCXO output frequency of 96MHz, I need two triplers, the first to 288MHz and the second to 864MHz; 864 + 432 = 1296 as required. I found various circuits, and used S53MV Matjaz Vidmar's design as a basis. The triplers each have two coupled tuned circuits which should give reasonable selectivity of the correct harmonics.
The inductors are flat line on the single-sided board; for 288MHz with 12pF centre on the trimmer, 25.4nH is needed; for 864MHz and 3pf, 11.3nH is needed. I decided on 3mm wide tracks, which makes them 34.8mm and 18.6mm long respectively. There's a handy online calculator for this.
The circuit is to operate at frequencies higher than I'm used to, and though there are many designs around using older components, most modern designs use surface mount components. These are a lot cheaper than the older types, and also easier to obtain, being made in huge quantities for modern consumer electronics, which is much higher tech (e.g. mobile phones) than it used to be.
However, for one-off circuit boards I'm still (a) poor (being retired I can't afford commercial pcb production) and (b) old fashioned in that I like making things myself as far as possible. So here's the circuit layout drawn out as 2:1 enlargement on 2mm squared paper, with the (single-sided) board drawn up, the tracks drawn in with indian ink.
The last board I made I used a permanent marker pen for the etch resist, which was a bit marginal - it was not thick enough in places. The indian ink is much better, though next time I will remember to degrease the board before I start drawing on it. I found the hot water used to rinse after etching also removed the indian ink quite easily, so I may yet revert to my old method of using enamel paint.
Here are the first few surface mount components added. I bought new semiconductors and capacitors (since I need some small values, and can't yet be bothered to salvage and test unmarked capacitors from old boards), but the resistors are salvaged. Note the shaggy edges to the tracks caused by using a simple inking process.
I must admit I've had a lot of fun with the surface mount technology though - I'm going to enjoy this microwave stuff!
Here's the local oscillator multiplier circuit:
Testing and alignment
The circuit takes less than 20mA in total and the transistors seem to be biased correctly. I have a little problem to begin with - since I'm operating at frequencies much higher than I have used before, I don't have a good enough RF detector to be sure the alignment is right. So I think a little sub-project to build a suitable RF sniffer was in order. However, with the FT-817 tuned to 432MHz (a harmonic of the OCXO frequency) I could detect the effect of adjusting the multiplier tuned circuits, so something was happening!
Having built the RF sniffer, I had another look at alignment. The sniffer is very sensitive, and the small loop picks up a big signal from the strip inductors in the triplers. In fact, it goes off scale with the loop very close to the inductors - that's more than 10dBm in the loop, so I'll probably need an attenuator in the path to the mixer. I'll need to try measuring the output directly with the sniffer in circuit (after an attenuator).
Knowing the expected value of the trimmer capacitors, I think it might be set up on the correct harmonics, but I'll need to have the transverter receiver finished to be sure. Another diversion could be to build a microwave prescaler for my ancient frequency counter, but I'll try without that first (though it's a very useful project for another day).
Looking at the measured current taken by the multipliers (it's more like 24mA now after adjustment), it appears that the circuit peaks in output with a dip in the collector current when adjusting the trimmer on the collector inductor, but (maybe?) a peak when adjusting the trimmer on the coupling inductor. Hmmmm. More later ...
The transverter mixer, receive RF amplifier and transmit RF amplifier are really on one board, but I divided it to allow me to test it a bit at a time before being committed to the final etched board.
DesignThe design is basically swiped from Sam Jewell G4DDK as given in the International Microwave Handbook. I've made a few changes; the IF is 432MHz not 144MHz, the IF is relay switched, and the oscillator chain has space for an extra MMIC amplifier on the mixer board in case it's needed, but otherwise I used most of the DDK design.
As this project progresses I'm getting more into surface mount stuff. Being a cautious person I decided to split the transverter board into the mixer, receive RF and transmit RF stages so I can test a bit at a time. I had done the first draft of the mixer board on squared paper and then found PCB, an open source printed circuit board editor. It took me a couple of hours to get it installed (from the source code), needing lots of bits and pieces that I didn't already have set up on my Linux machine, but it's basically OK now, though I haven't got image export set up properly. However it does .pdf export so no problems. It looks to be a great tool, and will be superb for when I finally start producing pc boards optically.
Here's the mixer (and DC control) part of the board as a screenshot of the PCB drawing window; the eagle eyed will recognise the basic outline as nicked from G4DDK.
The last couple of boards made using indian ink were OK, though the indian ink tended to detatch late in the etch process. For this board I reverted to my old method of using enamel paint; below is the board part painted. My art skills coming into use here!
Here's the painted board together with one of the pages from the .pdf output of PCB. It can group various layers as required, do mirror images, show the board silk (print) layer and so on.
The etched board (below) is now ready for drilling then I can get on with building it.
The first step was to add basic components, just the mixer and helical filter for 23cm and a few capacitors. As expected, I wasn't able to detect anything!
The next step step in this project is the receive part of the board. The transverter mixer, receive RF amplifier and transmit RF amplifier are really on one board, but I divided it to allow me to test it a bit at a time before being committed to the final etched board; having finished the mixer part I needed the receive part to enable me to test it properly.
DesignAs noted before the design is basically swiped from Sam Jewell G4DDK as given in the International Microwave Handbook. I've made a few changes; the IF is 432MHz not 144MHz, the IF is relay switched, and the oscillator chain has an extra MMIC amplifier on the mixer board in case it's needed, but otherwise I used most of the DDK design.
I used PCB (an open source printed circuit board editor) again to do the layout for the receive section. This circuit is lifted directly from Sam Jewell's design, although the initial build has some capacitor and resistor values changed a bit as I didn't have (or couldn't find on old boards being used for salvage!) those specified.
Here's the board painted and ready for etching, alongside the printout from PCB.
Here it is built up and attached to the mixer board - the whole thing is gradually being put back together. Note there are currently wire bridges that will be replaced by components once the transmit side is built. Note the two very tiny MMIC amplifiers.
For bench testing the mixer board has the basic components for a 5V supply added. Here's everything built so far, the OCXO board providing 96MHz into the LO multiplier, which goes up in two stages to 288MHz then 864MHz to give the 432MHz IF.
Bench testing wasn't easy with my almost nonexistent test equipment; I could do little more than make an attempt to get the LO chain aligned using the RF sniffer and my knowledge of the design capacitance of the trimmers in the multiplier tuned circuits. I really must have a go at making a decent UHF/SHF wavemeter, and a prescaler for the frequency counter.
I built a small bi-quad antenna (design by DL5NEG in the International Microwave Handbook again) but indoors wasn't able to detect any beacons. However, there were clearly some signals being picked up that didn't seem to be simply 70cm feedthrough, so I was hopeful things might be working.
Luckily there's a very local beacon - GB3EDN - and by taking the bits in the car to the car park on Blackford Hill where I used to work I could get the beacon as a line of sight test source. Woo hoo - I was absolutely delighted to tune to the beacon frequency and find it there without further adjustment - I had indeed managed to align the LO chain using nothing more than common sense and my microwave RF sniffer. Here's a photo of the setup taken on my mobile phone - I was so excited I had to record it. Note the fairly insubstantial RF screen used to make the system work reasonably well; I can now go about ordering some boxes.
I even took a very short video clip on the phone; the picture quality is dire but the sound is OK.
Now the basic receiver is working the next step is to get it all into screened boxes and start on the transmit side. One step at a time though - first into the boxes. The next photo shows the board reassembled, waiting for its transmit components, but also the switching components for Rx/Tx switching.
The OCXO board now also needs to go into its box, but this will have insulation to try to maintain a stable temperature. I used some plastic packing foam which cuts easily with a sharp knife. The box base is screwed to a piece of aluminium to make mounting in the final transverter box (which will contain all these little screened boxes) an easier job.
The local oscillator multiplier board was designed before I know about standard size tinplate boxes, but fortunately it fits in one dimension. Note the capacitor feedthrough for the +12V supply. The input and output RF connections will be coax poked through holes in the box sides; it would be a bit extravagant to connect everything together with expensive SMA connectors (maybe later when I need to interchange stuff). The board characteristics will of course have changed now it's in the box so it will need retuning.
No photo of the main board in its box yet, but they have all been connected together and taken out again to listen to GB3EDN. Things are a bit more stable now, though I had to do some tweaking of the tuning, especially the multiplier chain. The next step is to complete, align and test the transmit side.
The transverter mixer, receive RF amplifier and transmit RF amplifier are really on one board, but I divided it to allow me to test it a bit at a time before being committed to the final etched board; this step adds the transmit part and integrates the components (transverter board, OCXO, multipliers, control) into a single box.
I used PCB (an open source printed circuit board editor) again to do the layout for the transmit section. I haven't put yet another PCB printout and photo of the board - you will have the idea by now!
Here's the complete transverter board in its tinplate box; the MMICs are indicated by the red circles.
When first put together in the tinplate screening boxes, everything had stopped working. My first thought was that the local oscillator multiplier chain had gone off tune (resonance affected by the box?) and I spent some time messing about with that. I could get the system to work and receive the GB3EDN repeater but only with the lid part way off the multiplier box! - something not quite right there. So I had another think. I decided that the multiplier board was probably at fault, since it seemed a bit unstable; I added quite a lot of extra decoupling (realising that the power lines down to the transistors were probably resonating) and if you compare the photo below with an earlier one of this board, you will see the extra capacitors, and an extra bit of board to make sure the leg of copper near the output was not resonant and properly grounded.
This made a big difference - the adjustment of the trimmer capacitors on the resonant lines of the mutiplier (visible above, green and blue trimmers) was much smoother. I peaked up the output using the RF probe. Oh dear - on another visit to listen to the beacon it still didn't work (I still can't receive it from home, there's a hill and a lot of nearby stone built houses between me and the beacon).
More thinking - and a bit of calculation - showed that the trimmers could pick out the wrong harmonic quite easily. More careful adjustment, looking for the harmonic I needed with the trimmers set to roughly where I expected them to be using the RF probe suggested this idea was right. Time to put the lids on the screening boxes again and try it once more.
This time - success! The beacon signal was clear and strong even without adjusting anything (in fact I couldn't make it any better by twiddling). The beacon is so strong there that it's just about audible using the transverter with its main box lid on and no antenna connected! Here's the setup in the car boot whilst I was playing with it - you can see lots of bars on the FT-817 S meter.
The transmit side is also now aligned but not yet tested - I need to go outdoors with my small antenna, and the weather has taken a turn for the worse. Watch this space!
With the main transverter board finished, and things seeming to work, it was time to do some proper testing; this part describes what happened and how I dealt with the issues raised.
Thinking that everything was working and all hunky dory, I thought I'd breeze ahead with the final piece needed to make a working system - the box containing a small test amplifier to give a bit more output power, and also containing the antenna switchover relay.
Here's the completed circuit board for the test amplifier. The design (by GI0GDP) comes from the RSGB VHF/UHF Handbook and uses a BFG235 transistor - I had bought a few of these from Farnell before they become obsolete.
I installed it in the box (below) with the relay board (overall circuit will follow in due course), and did some tests.
Unfortunately it turned out that what I thought was working wasn't - the RF coming from the output when transmit was switched on didn't go away when I removed the drive - the output MMICs in the transverter were oscillating. I cured this - only by replacing one of the MMICs with a lower gain device, my layout is clearly not as good as the original - but still didn't have a sensible measureable output on transmit. This was very annoying, especially as the receive side seemed to be OK!
So it didn't work ...
After a bit more thinking, I decided that the 96MHz output from the OCXO was at too low a level to drive the multiplier stages properly, so I decided to add a little amplifier to beef it up a bit before being fed into the two triplers. I tried a little tuned amplifier using a ZTX3866 (like the 2N3866) but with not much success. I then decided that I need to rethink the local oscillator chain just to get on air. I was thinking about replacing the lot, starting with a computer-type oscillator module, but a lot of work had gone into the OCXO so it would be good to use it if possible!
Jon Joyce GM4JTJ had made some helpful suggestions which really got me thinking more about the LO problem, and I decided that I simply didn't have anywhere near enough drive. I eventually decided to give up trying to get 96MHz out of the OCXO box, and replaced the output stage with a switching buffer, that provides a nice harmonicy 24MHz at about 16mW. I fitted an SMA socket on it and decided not to open the box again until I (eventually) want to put the frequency adjustment spot on! Here's the modified circuit of the oscillator:
With that level of drive set up, I read up about multipliers in the Microwave Handbook and set about the job properly. The test replacement was all built rat's nest style (though a neatish rat did the building) on a groundplane. This made the 96MHz a nice solid signal, producing milliamps on the wavemeter, and the next tripler also gave a nice big 288MHz signal - I discovered that my old VHF wavemeter JUST (or nearly) reaches that frequency, so I was able to detect resonance with both the microwave RF sniffer and the VHF wavemeter. Here's the board with the 96MHz coils and the tuned lines for 288MHz.
Eventually I got the final multiplier done (864MHz) and it seemed to be working OK - the RF probe showed a nice smooth peak as I tuned each section of the filter. It just needed the final output buffer, then I could put the signal into the wavemeter. This was much better than the first design - it gave a much bigger signal at each stage (the RF probe showed 2-3V at the output of each buffer; the buffers are emitter followers). I included an emitter follower buffer between each stage of multiplication. The buffer for the 288MHz stage was (with the multiplier) an oscillator until I put some ferrite chokes on both the collector lead, and also at the base in the high bias resistor lead.
I had never used SMD components in "dead bug" construction before - there are a few capacitors supporting other components - and the BFR92 transistors are suspended in space above the groundplane. It's a bit like a surface mount pcb but with an air substrate! Here's the board just before adding the final buffer.
Unfortunately I found I couldn't get the third harmonic from the final multiplier. Second (576) and fourth (1152) yes, but not the wanted 864. The wavemeter I had built as a result of Jon GM4JTJ's advice was of course a tremendous help in checking this out. The circuit also had a tendency to oscillate - clearly there was too much gain, using those buffers.
New LO multiplier chain
Having had a think and a look at others' designs, I decided that everyone else seemed to go for doublers in the later stages of the multiplier chain. So I did a redesign - triple up to 72, triple up to 216, then double to 432, and again to 864. I also decided to try without the buffers. A photo of the result is below. And yes - it produced the right harmonics (checked on the wavemeter) and finally, I could (just!) hear the GB3EDN beacon indoors on my bi-quad antenna, even through much of the house. The drive into the mixer was probably a bit low, but I could increase that since I put an attenuator between the MMIC on the transverter board which ups the LO level and the mixer itself; there were several volts of signal (on the RF probe) from the MMIC into the attenuator. With an apparently working local oscillator multiplier chain, I built screening round it using tinplate from some "Hula Hoop" potato ring tins left over from the Christmas festivities!
Here's the circuit of the low frequency half (left hand side in the photo above) of this multiplier configuration:
And here's the circuit of the high frequency half (right hand side in the photo above):
With this success, I had been thinking about how on earth I had been hearing the GB3EDN beacon before. I now think I know - from only a couple of km away, line of sight, the beacon is extremely strong. So what was happening, I think, is that the beacon was getting through and acting as the LO into the mixer, and the very feeble 864MHz harmonic that was in the transverter produced the required IF. It's the only thing I can think of to explain it.
Once the multiplier chain was screened I redesigned the arrangement in the main diecast box, and this is shown in the next photo. It is a working 23cm receive transverter, with the transmit side still to be properly tested.
However, the reasonable drive into the mixer was enough to produce some transmit output. I guess my layout wasn't as good as the original DDK one which was the cause of the Tx amplifiers oscillating. That was only cured by replacing a 21dB gain device with what I had - one that gave me 15-16dB gain. So instead of 50mW out I would already be down to 12mW. In fact it was definitely worse than that - I was only getting a couple of mW or so out of it. However, that was good enough to allow me to use the output signal level to align the output 2-stage helical filter, and also the 3-stage filter next to the mixer used for both Tx and Rx. That meant that the GB3EDN beacon was very clear (in fact it could even be heard at home without an antenna attached - the box being open), and I could also just hear GB3ANG on 23cm (as well as the 70cm GB3ANG beacon breaking through on its own frequency). This still on the bi-quad indoors!
The output therefore seems to be OK apart from the miserly power; it's the right frequency (good old wavemeter) and SSB input makes the RF sniffer LED bars go up and down satisfactorily as I speak.
Back to the driver amplifier
With the BFG235 amplifier in a box already, I thought I might put a little MMIC onto a board to boost the signal and hope it would coexist happily in the output amp box. However I tried the BFG235 amplifier on its own first.
I had convinced myself that the BFG235 amp was giving a few dB gain, in that the RF probe measured more into a 50 ohm load after the amp than before it - so I decided to build a little preamp. I realised that something like the MAR-6 was not really suitable as its maximum output is only 3dBm; I had bought several of the Agilent MGA71543 MMICs for the transverter design (it only needed one) and one of these can put out maybe 14dBm at 30mA power consumption, according to the data sheet. So I knocked together a pcb (using vinyl tape - this gives a lovely result, thanks for the tip from Jon GM4JTJ) and built it up. I was supremely disappointed to discover that it was clearly a wonderful oscillator, in spite of me following the suggested layout. However, the RF probe showed lots of RF near the "ground" pins of the device (unless you have a negative supply, it needs a bias resistor, and the ground pins grounded through capacitors!!), and the data sheet said good decoupling of these is vital for stability. So I drilled a little hole beside each pin, and stuffed a 100pF 0603 capacitor in each hole, and soldered them directly between the device pins and the groundplane. Magic - the oscillation stopped and it's an amplifier. That means I now have maybe 20mW output from that device on transmit (and it seems to saturate at that level).
I had another look at the BFG235 amplifier and was not convinced it's working - it was behaving as an attenuator not an amplifier. The trimmer capacitors on the input and output lines tuned up to maximise the output (though I had to add a few more pF to each to make this happen), but it just didn't seem to amplify - the measured output into 50 ohms (or the antenna) was lower than that coming out of the MMIC. I decided I would need to play around with amplifiers a bit more at this frequency to start to understand what's going on - that circuit ought to amplify!
I had seen the use of the "puff" microwave design software in the RSGB VHF/UHF Handbook and the RSGB/ARRL International Microwave Handbook, and thought that at some stage it would be good to get hold of a copy and look at it. Then, as I looked on the web for ideas about the driver amplifier problems, I discovered that there is now an open source version of puff for Linux - just what I needed. I downloaded it and compiled it (very easy, it has a makefile so it's not complicated to install) and found that it is going to provide me with hours of fun. I couldn't resist playing with some designs, and decided I'd try a simple single stage amplifier using the BFG591.
After an hour or two I had a reasonably matched setup looking like this:
I thought hard about it and added a few little tweaks (e.g. allowing for the fact that the input line isn't directly connected to the device - there's a very short section of narrower line around the transistor base connection, which makes quite a difference). I used λ/4 chokes on the base and collector lines to feed the DC. It's not perfect, but it looked quite good, so I decided to build it. The circuit is as follows:
I used the pvc tape masking method to produce the circuit board, which gave very nice neat transmission lines - much better than painting or indian ink - thanks again Jon GM4JTJ.
Here's the board on the bench ready for testing. The pads near the top were for playing around with the bias circuit, and were needed. My simple design of a regulated bias didn't work - there was clearly the beginnings of thermal runaway. So I built up a standard feedback circuit to control the collector current from the voltage measured across the collector load resistor, and that proved to be very stable indeed, though the transistor is a bit hot to touch at 70mA - so I'm running it at 60 mA (I have the manufacturer's S parameters for every 10mA, and it makes very little difference - the data sheet suggests IMD performance is best around 60mA).
Bench testing looked pretty good, in that it was definitely an amplifier. However, with pigtails of coax with BNC connectors on the ends for input and output it didn't perform quite so well into a 50 ohm dummy load. Still, I decided to get it in the box and try it.
Installed in the PA / preamp / relay box it looks like this; you can see the little MGA71543 MMIC amplifier mounted on the side wall of the box. (I still have to include details for that.) It turned out to work much better with the much shorter input and output leads, and gave an output that looked to be over 400mW, which was very pleasing - it tuned up very smoothly.
This amplifier apparently working, we were ready for a real test.
Having an apparently working amplifier (even low power - less than half a watt) I thought it was definitely time to try out the system. Jon GM4JTJ had been tremendously encouraging in the final stages of the build and was happy to fix up a sked to try out the system. He lives about 90km away so I needed to get to a reasonable site to have a good chance of a contact.
The chosen site was on the lower slopes of the Pentland Hills to the south of Edinburgh - easy to get to and park near the artificial ski slope, without having to carry everything too far for the first test. We arranged to talk back on 2m and set up a time for the tests.
Here's the site with everything set up and ready to go. The 2m antenna (5 element Yagi described on the antennas page) was lowered since (a) I'd forgotten to take the guys for the pole, so it was attached to the tripod instead, and (b) every so often there was a strong gust of wind which tried to knock over the 2m antenna. The 23cm antenna was set up on an old tripod.
Here are the antennas pointing towards Jon's location further up the east coast of Scotland near Arbroath.
On time, we made contact on 2m, then I listened on 23cm whilst Jon transmitted. Previously I had only ever heard beacons on the receive transverter, so it was gratifying to hear Jon's voice coming through as a very clear signal.
After Jon's next over I changed to transmit on 23cm, and was more than pleased to be told by Jon that he could hear me. Jon made a recording which plays my end of the QSO, unfortunately not his, as the transmission at his end overloaded the recording device. I've edited it down to the first two overs.
Here's the shack for the tests; only two batteries were used, one for the FT-817 and one for the transverter boxes. I used the bi-quad antenna briefly but found the signal from Jon greatly down compared with that when using the Yagi. I'll need to look at the matching of the bi-quad.
So finally, it's really working. Of course, I now know what I want to do to improve it; I'm going to do a complete redesign using 2m IF rather than 70cm IF, and build a new local oscillator chain. Then I'll be able to build the PA properly for the new system; I have some transistors waiting which should give me up to 10W output, which will be more than enough for /P operation. As usual, watch this space! - that will be a new project.
Further tests - UHF UKAC contest 15 June 2010
The day after the first QSO was an RSGB UHF UK Activity Contest day - so I set off up Allermuir Hill (after alerting it on SOTA in case any keen SOTA chasers with 23cm wanted a contact!) with all the gear.
This was a successful trip - only 5 contacts, but I was well satisfied with that for a first try. They were all in the first half hour anyway. I struggled to hear a station in Ellon - better conditions would probably have allowed a contact there. The 400mW to the 15 element Yagi was giving me 59+ reports from central Scotland, and a 59 report from Jon GM4JTJ 95km away. All very satisfying. There are some photos on my Flickr photostream.
I subsequently discovered that Andy MM0FMF had been out with a 23cm portable hoping to give me a summit to summit contact, but I didn't know he was going to be there and wasn't checking the SOTA spots. A pity - that would have been a great extra, and a test of the transverter on 23cm FM.
So I think that's this project pretty well done. On to the next!