40m dipole up

Mayday brought the combination of a day off, plus sunny weather, plus no wind.

Go back two weeks:

I had a first go at hoisting up the 40m dipole. This was initially frustrated by discovering that the heavy winds of the last couple of months, although they had proven that my experimental mast was fairly safe against the wind by itself, had managed to lift the rope off the top pulley. I hadn't realised there was enough room for the rope to slip down the side of the pulley, but the wind had managed it. I had to wait for a hand to get the pole down again. I added a couple of cable ties to the pulley arrangement to stop the rope coming off again, but I can see now that this pulley is not really reliably enough. Note for the future: get a better pulley, probably some sort of marine block and tackle.

I had my 21m length of house wiring wire (2.5 square mm intended for lighting circuits I think, I will check next time it is in reach) and originally planned on making a delta match into 450 ohm feeder. I made a few inquiries, including putting a question on http://www.antennex.com/. After some suggestions from Tim (VK3IM) on this forum I decided instead to cut the dipole at the centre and simply connect the 450 ohm feeder. I gather that this will eventually need tuning at the far end of the feeder, but I understand this could be the basis of a multiband antenna, rather than only working on 40m.

Anyway, I hoisted the centre point, to some alarming bending from the top 2m of 12.7mm (half inch) tubing, but it got to the top okay. I tied off the hoisting rope, and then tied some more rope to each dipole end. The uphill side was easy to tie off to a tree, nearly horizontally due to the slope. I don't have an obvious place to secure the "downhill" end of the dipole, so for now I simply extended out the rope from this dipole end and tied it to a concrete terrace chair. This results in the dipole dropping at about 30 degrees to the horizontal. I will need another pole arrangement for this end eventually.

Then the wind began to build over the next day or two, and the top 2m of the pole was thrashing around a lot, so I lowered the dipole until I came up with a plan.

It occurred to me that the problem with a hoist is that the downward force is twice the weight hoisted, due to the pulley idea. I could reduce this with a multiple pulley arrangement so the tension on the rope is reduced at the expense of a lot more rope! Also I needed some more guy ropes for the top section of the mast.

Back to yesterday. The fine weather tempted me out, and I hoisted the dipole again. The top pulley continued to play up, as although the rope could not actually slip off, any sideways displacement of the weight bearing rope caused to jam, presumably against the cable ties that stopped it coming off the pulley, but I couldn't really see. Luckily I could rotate the main pole to keep the pulley angled directly so the load rope could run up straight.

Once I got the dipole to the top, I was able to use the tension in the lift rope to double as a guy rope by tying it off to the western guy anchor. I put some bend on the top 2m of pole - the prevailing wind is from the west so this should anchor it well. The dipole itself is almost north-south, and I angled the feed line off the house in the east. I realise that I should have tied off the feed wire with some string to relieve the strain on the copper cables. My guess is that this is going to be the weak point.

The photo is taken from the uphill side (north east) so is looking down at the support pole, and shows the south dipole wire tied off to a chair.

Daytime tests into my HF225 receiver showed the usual range of strong Spanish stations, and not much else. Sunset brought in a strong Japanese station in conversation with a number of Spanish amateurs. I swapped to the softrock and couldn't get this signal until I eventually substantially reduced the gain on the A-D converters, which suddenly cleaned up the band. This is odd, as level meters on the converter did not show any peaks, but I suspect these level indicators. At some stage I will experiment more with converters, I am trying out a few 192k converters I have around.

Today we have strong winds again, this unusually from the east. I will watch with interest how my skyhook behaves...

Completed 80/40

A few weeks ago I finished the build of the 80m/40m PA. A quick test of the RX with a short wire aerial into Rocky software (including the best part of a day messing with soundcards  before I got a dual soundcard system to work) revealed a lot of interesting whistles, but no signals. Not having a transceiver to generate a signal, and running out of time, a few weeks have passed while I've been away a bit on business. 

Thinking about how to test further I decided to build a one transistor LC oscillator to generate a test signal. I found a BC109 in the junk box and wound a small transformer on an FT137-43 core (8 turns in collector circuit and 2 turns for feedback to the base). After getting the feedback phase right and dropping the gain with a suitable emitter RC network I got vaguely sine wave oscillation but with only stray capacitance it ran at under 6MHz. I soldered in about 100p of capacitor across the collector turns and got it to run at about 3.5MHz, giving me a harmonic around 7MHz. My AOR scanner gave me a nice signal around this frequency and a fun time playing it as a "Theremin" as waving hands near it scanned it across the 7MHz band. Reverting to the Softrock I now got a good spectrum display as the oscillator zoomed around the band, so I guessed that the receiver was actually okay. The correct display seemed to require the I and Q selection of Rocky to be reversed from what I expected. Still not sure why.

A quick test of TX into a dummy load and detected on my scanner showed that TX seemed to be working. I had been dreading this, expecting smoke, but all seemed in order.

Transplanting the Softrock from the workbench to the shack bench I connected my roof antenna and - amazing - lots of stations across the 40m band. The spectrum display and DSP detection was amazing to play with, quite an eye-opener. 

More about some weird issues with mutliple images and testing the TX further in the next update...

Storms Passing By

Last week's storms were a good test of my sky hook, even though no aerial wire is yet attached. It was bending quite a lot, but the guy ropes held it in place, and It seems to me that bending is better than breaking.  I believe we had wind gusts forecasted at 50 - 70 kph. I don't know if this is what we actually got, but it felt like a sample of the worst we normally get around here.

This weekend the weather was clearing, but I thought I'd like to get further with the Softrock before trying any new aerial ideas.

In the end I decided just to press on and get the main board assembly finished. I decided not to stop for all the tests at each stage as it really breaks the flow of the soldering, so to speak. Just stopping at each stage to test the power rails to make sure of no shorts seemed enough for now. I can run through a full signal test when it is all assembled.

Some observations:

The build instructions advise fixing the PCB when working on it. I don't find this necessary. In fact being able to spin the board around quickly to get to it from all angles seems a great benefit. The board is heavy enough to sit well on the surface. The four corner screws no doubt helped here.

Winding the coils: I found it useful to start the single windings from the middle. It is easy to estimate the middle of the recommended length of wire, then doing half the turns in one direction then the other half from the middle going round the other way seemed to quite convenient. It also means half the length of wire to pull through each time. 

I noticed that the bandpass filters have an undocumented phase reverse in the way the PCB is laid out. The circuit diagram shows the top of the T100 secondary winding connected to pin 3, but it actually goes to pin 2, and conversely the bottom goes to pin 3 instead of pin2. However the circuit does not show any phase "dots" identifying the matching ends of primary and secondary, so I assume it is not important. Apart that is from the secondary pair needing to be correctly connected to get the centre tap, I assume that phase reverse between primary and secondary is not a problem.

Power testing: For some reason the current scale on my multimeter seems to have stopped working. It is probably the built-in fuse but I don't have a replacement. In the end I skipped the first test of each stage with the 100 ohm series resistor, and settled for a 10 ohm in series. This seemed to me to represent a safety measure for any shorted power rails. A quick test for voltage drop would show up a short quickly enough to remove power. It also meant I could measure current consumption by the voltage drop of about 1 volt across this, and I could leave it in circuit for the actual rail tests and other voltage test points as it did not drop enough to interfere with normal operation. This sped up the power test before moving on at each stage.

The story so far:

The underneath:

Anyway, this weekend I got the main board assembly finished, a couple of bandpass filters done, and started on the 80/40m PA stage. I'm hoping next weekend to finish the PA and to get the PC connected up and Rocky software installed  for some signal tests.

Soldering the Chips

Last weekend I got as far as getting the clock chip soldered on and working, which felt like a bit of an achievement...

Soldering on the small chips is definitely not easy. In the end I was thankful that I have a microscope, and I did the soldering under it. You get used to zeroing in with the soldering iron and tweezers into the field without melting anything on the way, and it geat to see exactly how the solder is flowing. Maybe my multicore is too big, but I found it very difficult to get anything less than a lot of solder onto the joint. Still, in the end it seemed to flow okay, I got almost no spills, and tests so far seem to indicate that it works, so I think it will do. The picture attempts to show how the solder steeps up against the edge of the pins on the clock chip. I can only guess whether it runs underneath ok.

To check this, I got out my very old and cranky oscilloscope, and got this picture of the quadrature signals at 7MHz, so it looks like the oscillator is okay, and the PIC is programming it okay.

I had one doubt here, as I want to control the oscillator from the add-on USB module when it arrives, and the connection pins seem to be underneath the 8-pin switch module, obviously designed as an alternative to the 8-pin PIC. In the end I mounted the switch above the board, as high as its pins would allow. This will make it easy to cut off if I need to remove it later. On the other hand, the I2C programming pics can be soldered onto underneath the board, so simply removing the PIC should be enough to free up the oscillator chip programing pins later. 

That was all I had time for last weekend, but it felt like I was getting somewhere.

Softrock Arrived

The Softrock xtall-v6.3 transceiver kit arrived last week, though I am still waiting for the 10m PA kit and the USB interface kit. Still, now the main board is here I can get started assembling, as neither of those is essential.

Fortunately I have had to work on surface mount boards before, though only to change odd components when problems with prototypes have been found during software development. I've never had to build one from scratch though. 

Managing the components is a challenge. It was good advice in wb5rvz's build notes (http://wb5rvz.com/sdr/RXTX_V6_3) to check all the components against the inventory before starting. This helps identify and distinguish similar parts. I have quite a good collection of small antistatic plastic bags into which I put the separated parts to keep them organised. 

Since the pictures were taken I've got to the end of step 1 and I'm about to test the power regulators before moving on. Since I'm nervous about my soldering I've been testing for shorts after every component.

I was confused for a few minutes when it came to placing the first tiny 5-pin 3.3v regulator IC as the pads were hiding between a couple of capacitors. The lack of silk screen component ref on this side of the board made it difficult to identify which caps were which. I finally found the site though, as in this picture:

The size of the pads was challenging. Luckily I have the luxury of a microscope so I could check that my "first tack" had managed to get it well within its pads, so I completed soldering it under the microscope. I hope I did not overheat it as one pad was slow to flow.

Now on to the first test then the next assembly step...

Sky Hook

My next problem is how to get an antenna up a bit higher than just lying on the roof.

Now I am not a natural mountaineer, and I am not going to go clambering on the roof or shinning up chimneys; I have a bit too much metalwork in my legs for doing that. But anyway, at the moment I am thinking about simple wire antennas so all I really need is a convenient sky hook to attach wires to.

I found some 2m lengths of steel tube at the local hardware shop, which was useful as they would fit in the car. They had a number of diameters so 4 rods of 25mm, 22mm, 18mm, 12mm could be fitted inside each other, sliding about 20 cm of the smaller tube inside the larger, and drilling through for a couple of bolts. This way I had a pole of about 7 meters. I attached a pulley to the small end and put a rope through it. I am hoping this will enable me to haul up some wires when I know what I want to try.

Next, I got a 1m length of 30mm tube and with the aid of a few brackets and bolts screwed this to the wall of an out-house. Now I could drop the 7m pole in (with the aid of a friend), tighten it up and admire my new skyhook.

I took the precaution of adding a few guy strings as it can get pretty windy here, as there is not much between us and the Atlantic ocean 60km or so to the west, apart from some medium sized hills.

I am also worrying about lightning, so decided to attach a sort of "lazy" earth. I ran an earth wire from the bottom of the pole along the ground for about 10m (without insulation) and buried it a little. This is mainly to provide a leakage path for static electricity to ground and might help prevent lightning strikes. I did not want to wire it to the outdoor mains earth as the wiring is rather low grade and I think a real strike back into the house wiring might not be so good.

This is still pretty much an experiment to see how this works out. The guy ropes are basically plastic string of unknown material, so I expect the sun will have a disastrous effect. Meanwhile the general springiness of the construction should help it withstand the weather for a few months by flexing rather than breaking. After that, I will have a better idea of how to make something more permanent, and whether this location for the upright is going to be okay, both cosmetically and performance-wise. This corner of our terrace is at the edge of big drop to the west, maybe 20m, so this should help with take-off angle even though the maximum height with this arrangement is only about 9m above the ground level on the terrace.

Adding a balun

I was not happy with wiring coax directly to the centre of the dipole, so decided to make a variable balun as an experiment. I expected I would be using some twin feeder for future antennas so thought that a set of impedances might be useful.

I decided to provide 50 ohm, 200 ohm, 450 ohm and 600 ohm on the antenna side, with a 50 ohm winding for the unbalanced coax side. This is what I believe is known as a "voltage" balun. I decided on a separate winding for the radio side mainly because all the stuff I've read doesn't, so I thought I'd find out the disadvantages.

I got a ferrite 140-43 toroid off ebay (GBP 2.95) and decided on 8 turns for the 50 ohm "tap". Since I wanted a separate unbalanced winding I actually ended up with 5 windings twisted together to make a 5-filar (quintifilar?) wound transmormer. One of the winding stopped halfway at 4 turns. By wiring the antenna sides in series I got turns ratio of 1:1, 2:1, 3:1 and 3.5:1, giving impedances of 50, 200, 450 and ~600 ohms. These I wired to some 4mm sockets (does anyone call them banana plugs and sockects any more?) so I could easily try different antenna connections. The other winding went to the coax SO239 socket.

I put it in a plastic box to make the assembly easy and to make it vaguely water resistant as I intend it to be outdoors, though semi-under-cover. I decided that I was likely to get RF radiation from the twin feeder connection anyway, so hope the extra radiation from the wiring will not be significant. I am worrying a bit about what might happen if I start putting significant power down it, but since my first experiments will be QRP I am not too worried. I can rebuild it into a metal box if it ever seems necessary.

Anyway, I wired a short length of wire to the dipole (a sort of random impedance feeder), plugged it into the 50 ohm antenna side and checked out the rx. I was amazed to find that reception appeared to be much better. Difficult to quantify but I'd guess about 6dB better (1 S point). I tried different taps and got little difference, and I settled on the 200 ohm setting. I am pretty sure the QRM from the shack PC got better too, but this might just be because that's what I've been told to expect when the coax outer is not feeding interference back up to the antenna.

All in all, it was an interesting first build item. Well worth it. And I am ready for a better antenna, as I can use twin feeder to the balun outside the shack, run it through the balun and into the house on RG58 coax.