I started building up a small collection of cufflinks (I also have a collection of salt and old 1950/60s Gillette razors but that’s another story). Cufflinks are ridiculously overpriced in NZ (what isn’t?) so I tend to order mine from overseas. You get better quality ones at better prices if you hunt about.

Favourite cufflinks.

Above are some of my favourites. Yes, I am something of an Anglophile and a train-spotter. The Dalek ones are very cool and I always had it in my mind that the TARDIS would make a nice cufflink. It is possible to buy some but I never got around to it. The London Underground ones a friend of mine got for me when he and his wife visited London from their home in New York and I then had to wait till they visited New Zealand to bring them to me. Those are well travelled cufflinks!

One of my other (many) projects is an N scale model railway I am building to go into a dining table. That is a 1930s British themed layout and when looking around at what buildings I could get for it I discovered you can buy a tiny N scale Police Box.

Hornby Police Box.

This is one of the Hornby Lyddle End range of resin models. I bought one, as well as some buildings, online since they don’t seem available in NZ easily. Oh, if you’re a retailer in NZ (and elsewhere actually) and have an online shop here is a hint – listing products you don’t actually have in stock and not telling people until they get to the checkout is not good!

The Police Box arrived and it is very cute. Perfect for my railway but also perfect to make some cufflinks from. The model is about 22mm high.

Lyddle End Police Box model.

It looks just like another TARDIS I happen to have here at home! I didn’t want to cut up my railway one so I ordered another from the UK. I got mine from Celestial Games and Books on eBay who were excellent. It’s interesting that ordering things from the UK is much easier than ordering from the US and shipping takes far less time than getting things sent from America.

I also bought some brass cufflink bases from a local online jewellery supply shop called Beadaholic. They, by the way, have an excellent online shop!

Model and brass cufflink base.

Next I used a modelling razor saw to carefully cut the model in half. This unfortunately results in the loss of the lamp from the top of the box but that is easily fixed later. After cutting I flattened the back of each half by rubbing them on some wet and dry paper.

Model sawn in half.

To attach the model to the bases I made some small brass backing plates. These were cut and filed from 0.9mm thick brass. I first made one then used that as a template to make the second. They are a little rough but at this scale they don’t need to be perfect and I like the hand crafted look. A quick clean up with some wet and dry and a final clean with steel wool finished them off.

Brass backing plates.

I was then able to solder the cufflink bases to the backing plates and later I could glue the model onto them.

Bases soldered on.

Next I repainted the model halves. I used small tins of Humbrol enamel paint – matt number 25. I love the smell of that stuff, reminds me of childhood and making models. Unfortunately I couldn’t recreate the fine lettering that was on the original models. These were actually stickers stuck to the model. Again the loss of detail doesn’t matter too much as these will be seen from a distance usually. I didn’t have a brush fine enough so I used a toothpick to put on tiny dabs of paint. White for the “Police Box” signs and white for the “Pull to Open” sign with some black lettering.

Painting.

I also made a replacement lamp using a tiny piece of styrene plastic glued on with superglue.

Replacement lamp.

Finally I superglued the painted model to the brass backing and gave the whole lot a coating of satin varnish, Humbrol 135, and I was done!

Little TARDIS in front of big TARDIS.

TARDIS cufflinks completed and ready to wear.

Right, where was I? I stuck the templates for the letters and the direction arrow onto the remains of the steel case.

Templates glued to the steel.

The letters and arrow were then cut out using the jigsaw and a hacksaw then filed and sanded smooth.

I then worked out how to make the pivot. I have no idea how real weather vanes do it. Looking online all I could find were silly toy ones. I found people online who make and sell real weather vanes but none explained how the pivots work. Perhaps it’s a trade secret? So I invented my own.

I am using a tube sealed at the bottom with a ball bearing at the base as a pivot and a bush near the top to stop the vane flapping about. The vane itself is attached to a straight steel shaft that slides into the sealed tube part. The whole thing is full of oil so it shouldn’t rust or wear out and it should provide a smooth rotation without being too sensitive due to the damping effect of the oil. Pictures should make it clearer.

Pivot parts.

The main body is simply a piece of steel tube.  Into the bottom of this is brazed a bolt, the head of which was turned on the lathe to make it fit into the tube. I put a small countersunk hole into the head of the bolt to locate a small ball bearing to work as the pivot. The ball bearing I got from inside an old spray paint can. These are what you hear rattling when you shake the cans. Now some cans have steel balls and some glass marbles. How do you tell which has which sort? Stick a magnet to the outside of the spray can and gently tip the can so the ball inside rolls past the magnet. If it is a steel bearing it will stop when it reaches the magnet whereas a glass marble will just roll past. Actually it makes no difference to me, I cut them ALL open and rescue what’s inside before putting the steel can out to be recycled! I have a jar full of the damn things.

Anyway, the bush at the top of the tube is just a nut I turned down to fit. The main shaft of the vane was some rusty old 1/2 inch steel rod I found in my old garage. The bush is pushed into the tube then held in place via a small hole in the tube I then plug welded through to hold it all in place. The picture above shows the relative positions. The whole tube is eventually filled with oil so the bush is actually totally submerged meaning it won’t rust or wear out in a hurry.

The pivot ball bearing.

You can see above how the ball bearing sits on the bolts at the base of the tube. The bolt is brazed into the bottom of the tube as shown. Later that bolt thread is used to screw the pivot into the main mounting bracket of the weather vane and it allows it to be turned to accurately position the direction markers.

To this pivot base (which doesn’t turn) the direction indicators are attached. I brazed the letters to lengths of 6mm diamater steel rod then welded these to the base.

Direction letters.

Once the base was done I brazed the direction arrow and locomotive to the main 1/2 inch steel shaft. I changed the rear of the arrow from my original design to one with more surface area based on this original LNER sign from Briglington I found online. I also added a small water shield made from a piece of tube and a washer. This gets welded to the weather vane shaft and it forms a cap that goes over the top of the pivot base to stop rain getting into the tube. This needs to be welded on very carefully so it is concentric to the main shaft so it won’t rub on the pivot base as the vane turns. I also added some balance weights to the front of the arrow head to balance the main vane.

A weather vane needs to be balanced with equal weight each side of the main shaft, so it spins easily, but has to have more surface area on the rear half than the front so it points into the wind.  I didn’t get this quite right (hey – it’s my first try!) so I needed to add weight to balance things out.

I then made the base which is simply a section of tube welded to a steel base. PArt way down the inside of the tube is a nut that matches the thread on the base of the pivot. This nut is welded inside the tube. I drilled a hole in the tube at the point the nut needs to sit then just plug welded the nut in place from the outside of the tube. For some reason I didn’t photograph this! The base was just some scrap 3mm steel plate I had that I cut to what I felt was a pleasing shape. The tube is welded to this.

Mounting plate.

This mounting plate is in turn screwed to a wooden base that is bolted to the eaves of the garage. The wooden base was made from some scrap marine ply I had about the place.  I needed to know what angle to cut the wood but since I didn’t have a ladder tall enough to get up to measure the angles I made a little measuring jig. This is just a long piece of wood with a T piece nailed to to top arranged so, with some force, it can pivot. To the pivot I hung a plumb line. I then took this contraption to the front of the garage and held it up making sure it was vertical using the plumb line. I then pushed it up onto the eaves. This moved the T piece to the correct angle.

Contraption for measuring roof angle.

I then transferred that angle (turns out it was 20 degrees) onto the wood. Simple! The steel mount is simply screwed to the wooden base which is in turn bolted to the garage eaves. I turned up a pointed plug from a piece of dowel to fit to the bottom of the mounting bracket just to finish it off. The steel base was primed with a zinc rich primer.

Wooden base for steel mounting bracket.

Another detail I didn’t get a picture of was the spacer used between the base and the mount. It goes on the threaded part of the base and is carefully cut to the right length so that when the base it turned fully into the nut in the base the direction pointers point in the right direction. I used a pice of 1/2 inch copper water pipe. Using copper means the spacer has some give in it. You can snug the base up very tight into the mount, the copper distorting a little for the last part of the rotation so you can get it in the exact right place. I got the idea from the crushable spacer used in an MGBs rear differential to set the differential pinion gear.

Oh, about now I discovered I could buy online some really nice locomotive cufflinks that happened to be almost the same as my weather vane. So I bought some! $30NZ shipped from the UK, arrived in 4 days! Cheaper than buying locally.

Locomotive cufflinks!

With everything done it was just a matter of painting it all. I painted the steel with black satin paint. Obviously I used the right stuff since the spray can has a picture of a weather cock on the front! Who wants a giant cock on their roofs?!? I painted the wooden base white and I used carriage bolts (appropriate for a railway themed vane) to bolt the whole thing to the eaves. So the vane sits into the base which is filled with oil to cover the ball bearing and steel bush. The base is in turn screwed into the mount via a spacer and a into nut welded inside the base shaft and the whole thing is bolted to the garage.

Finally painted.

Finally with everything painted and finished I just needed to bolt the whole lot to my garage roof. Luckily my father had a ladder long enough and he came over to help me by holding the ladder while I clambered up to bolt it all in place. Once done we stepped back to inspect it all and this is how it came out!

Finished weather vane.

Now if you wondering why the N-S letters are backwards in this view there is a good reason for this. As my mate Dave pointed out to me weather vanes will have one correct orientation. From other views the letters will be backwards. I was to busy making sure I got the N-S-E-W order the right way round to notice! Mine is the right way round when viewed from inside the house!

Yep, that was totally planned, not lucky chance at all…..

View from indoors.

Finally, to prove it works, a little film on YouTube.

Works at night too!

So in light of all that I decided to start a new project – a railway themed weather vane to put up on my spare garage. Currently my house and spare garage (I have two – doesn’t everyone?) are being painted and they are looking really good. The peak of the spare garage roof is the perfect place to put a nice weather vane which should be a fun little project to build.

Spare garage soon to be painted.

I seem to be going through a railway phase just at the moment. Reading everything I can about old British railways, locomotives, stations, etc, etc. I’ve always been a bit of a trainspotter I guess. I decided to make this weather vane with a railway flavour. I knew I needed some suitable steel and looking through my junk cupboard I found an ancient PC whose case I could cut up. It is a 486SX-25 so slightly past it’s best. The top cover of the case is nice thick steel – 1.2mm thick to be exact.

Nice steel case.

I decided to reuse this steel to make the bulk of the weather vane. I hunted online for a suitably representative locomotive that would look good as a sillouette. In the end I decided on a LNER J83 tank loco. The Hornby model railway provided me with a nice clear side view of one so I copied the image and put it into a graphics program, turned it into a black image and printed it out. I used that image to take measurements from which I then scaled up 3x to transfer onto the steel cover.

Outline drawn onto the cover.

I did tweak the design a little to make it clearer and easier to cut out. I just drew the design onto the steel with a Sharpie, some of it freehand, some with a ruler. The wheels I made slightly oversized. A screw jar lid made a useful template to trace around for those!

Once drawn out I drilled holes in strategic places then, using a metal cutting blade in my jig saw, I started cutting it out.

Holes drilled to allow the pattern to be cut out.

It took a lot of careful cutting but eventually I removed most of the bits that aren’t locomotive.

Shape roughly cut out.

I then started filing, grinding and cutting the edges to clean up the shape. I found that was well as using the bench grinder and hand files I was also able to use my joggler, a panelbeaters tool used for bending and punncing holes in steel for plug welding through. I used it a lot when restoring my MGB. With it I was able to nibble close to the line in some tricky spots which then minimised the amount of hand filing needed.

Joggler tool.

After about an hour and a half of mainly hand filing the main cutout was done. I gave each side a quick once over with the random orbit sander to smooth off the rough edges.

Finished cutout.

About now I was felling pretty pleased with myself when I suddenly realised ‘Bugger, the damn thing looks a lot like Thomas the Tank engine‘. Bah!

<trainspottervoice> Oh well, everyone is sure to realise my locomotive is based on the LNER J83 where of course Thomas was based on an LB&SCR E2. Obviously completely different locomotives! </trainspottervoice>

There is just no way I am going to paint the thing blue and red now though!

Next I need to make a direction arrow in the style of the old station Way Out signs and make the letetrs for the direction pointer. In preperation for that I used an online font sampler to print out the cardinal directions. The font is Gill Sans which is the font used by the LNER railway back in the day for all their signage. Since I am using their loco it seems fair I should also use their lettering! I will glue the letter onto the steel then cut those out.

Gill Sans font for the cardinal point markers.

More updates soon.

Part two is now available here.

Click on the pictures to see larger versions.

A film is available on YouTube showing the pistol and how well it compares to a standard Nerf Maverick.

The main body is simply a length of PVC pipe.  The barrel assembly and the two centering rings are screwed into one end.

The main spring goes between the piston and the end cap. To assemble this you simply remove the ring from the end of the plunger and push the end cap over the end of the rod then reattach the pull ring.

With the end cap screwed into place the weapon can be cocked. As you pull back the plunger the telescoping rod pulls out until the internal stop hits the back of the catch. Then as you pull further the rod and piston move backwards compressing the spring. When pulled all the way back the catch protrudes through the end cap. The catch wire is bent so that the plunger pushes it aside as it passes. The catch wire can then move back into the notch in the end of the plunger preventing it from moving forwards again. Once the catch is locked into the notch the telescoping section can be pushed forwards. This is done to avoid having the plunger rod having to extend past the rear of the pistol. To release the piston to allow it to spring forward all you do is push the wire out of the notch.

The wire is moved by pulling the trigger which rides in a trigger housing. The trigger itself is made from a solid piece of aluminium cut and filed into shape.  This is attached to a flat aluminium plate by two small screws. One screw also holds a small brass tab used to attach the trigger return spring. At the other end of the plate is attached another piece of aluminium with an angled ramp cut on one end.  This piece is the part that actually pushes the catch wire aside to fire the pistol.

The trigger housing is made from a piece of 22mm wide aluminium extrusion. The housing is what the pistol grip attaches to as well as holding the trigger and trigger plate. Holes in the housing correspond to the holes in the end plate and barrel spacers so that one of the three screws holding each of these in place is also holding the trigger housing to the main body. The front screw also holds the trigger guard in place. At the rear is a small rectangular cut out the catch wire passes through. And there is also a small screw used as an attachment point for the catch spring.

The trigger and plate slot into the housing and are a nice, sliding fit. The bolt that holds the pistol grip in place passes though the trigger plate. The spacer mentioned earlier is what allows you to tighten the bolt to secure the grip but still allow the trigger plate to move freely back and forth without allowing it to move up and down. With the bolt done up tight the grip is held securely to the trigger housing. The acrylic side plates are then attached with the 3mm bolts. The top of the plates have a step filed into them to fit snugly over the sides of the trigger housing. This prevents both the grip from rotating on the housing and also prevents the side plates rotating on the grip.

With the trigger guard in place a long screw passes through the housing and guard to attach the housing to the gun body. This screw is also what the trigger return spring hooks onto. The spring pulls the trigger plate forwards. When the trigger plate is forwards the ramp at the rear of the housing is also forward, away from the catch wire.

When you squeeze the trigger the plate slides backwards. The ramp on the end of the trigger plate also moves back and the ramp pushes the catch wire to the right, out of the notch in the end of the plunger and the pistol fires. The spring then returns the trigger forwards.

With the trigger housing attached to the body and the catch spring in place you can see above how this works.

Above you can see how the NERF dart fits into the barrel.

And finally some pictures of the (mostly) finished article.

Finally a shot showing the advantage of the telescoping plunger. With a straight plunger the pistol would be much longer.

What’s left to do? Well, it can obviously be cleaned up and painted. I also want to remake the ramp piece on the trigger plate from steel rather than aluminium as this wears too quickly. I could also make a small cover to go over the rear of the gun although this would be purely for looks rather than any functional reason. I would hesitate to paint it black as then you start getting into ‘real looking gun’ terriority and people start getting twitchy and I don’t really want twitchy armed offenders squad members on my doorstep!

The range, well that’s tricky to say since the darts don’t fly that well as they are fairly light. Around 8 – 10 metres or so (26 – 33 feet) the darts are fairly accurate and travel fairly flat. Any more than that and you need to start lifting the barrel to get more range.

In the film on YouTube all shots were fired from 7 metres (23 feet) away. The standard Nerf Maverick can barely make that distance.

Click on the pictures to see larger versions.

A film is available on YouTube showing the pistol and how well it compares to a standard Nerf Maverick.

Homemade pistol with NERF Maverick.

Now, this isn’t new. In fact there are forums for doing just this as well as some very impressive homemade NERF guns around online. The site that inspired me most was www.boltsniper.com where Evan Neblett has documented his very impressive and remarkably sophisticated homemade NERF guns. Even more impressive is these have been made by hand. I have the luxury of having a small lathe!

My effort is not quite so impressive. I just wanted to see if I could make a simple pistol that would fire a NERF dart better than the original toys. This is how I did it. Generally the photographs should explain it so I will try to keep the text to a minimum.

The pistol is made from 32mm ID PVC pipe, some aluminium extrusion and aluminium pipe, wood for the grip and various pieces of metal and plastic. The main spring is one I got from Bunnings warehouse. This gun should be able to take a cut down AR-15 spring as used in the Boltsniper weapons but these aren’t that easy to come by in New Zealand. I believe I am legally allowed to buy one but by the time I found that out I already had the Bunnings spring. I may still get one to play with. The idea of a toy gun using a part from a real one appeals to me for some odd reason…

The spring seems to be this one from Century springs. It is 11/16ths (0.687) OD, 6 inches long and the wire is 0.072 diameter with 29 coils.

Pistol stripped down to component parts.

I will go through the parts in turn.

Here we see the aluminium trigger guard and the screws used to hold the pistol together. I used stainless steel screws. The two small springs are the trigger return spring and the release catch spring.

Next is a bolt and insert that holds the pistol grip to the trigger housing. The insert is a 16mm diameter piece of steel with a 6mm hole threaded thought it for the bolt. On either side is a 3mm threaded hole. The bolt has a small spacer under the head the purpose of which should be clear later on.

The pistol grip is a piece of wood 18mm thick. A 16mm diameter hole is drilled through the grip to hold the threaded insert. A long hole is drilled vertically though the grip to allow the 6mm bolt to pass through it into the insert. The grip is bulked out with an acrylic plate each side. The plates are filed and sanded to provide a comfortable shape. Each plate is held with a 3mm screw that screws into the metal insert in the handle. Each acrylic plate extends past the top of the wooden grip so that when the grip is attached to the trigger housing with the single screw it won’t rotate.

The barrel is made from 19mm diameter aluminium tube. It is held inside the PVC body by two machined rings made from plastic chopping board.  Each ring is held into the body by three screws. Two circlip grooves are machined onto the outside of the barrel and the plastic rings slide up against the circlips to accurately position the barrel which then can’t move either forward nor backward. I used an o-ring under one clip to ensure the barrel was a tight fit and would not rattle or rotate in the housing. A standard NERF dart is 1/2 inch diameter, much smaller than the barrel. I used a piece of black irrigation pipe that is a 1/2 inch internal diameter which the NERF darts are a nice, tight fit into. I machined the outside of the pipe to make it a very tight push fit into the aluminium barrel.

The end cap is also made from machined chopping board. This is at the rear of the pistol and the main spring pushes against this when the pistol is cocked. The hole is to allow the plunger attached to a piston to pass through the rear of the gun. A fender washer is attached to the plastic cap with two screws to provide a flat, hardwearing surface for a music wire catch to slide against. The wire is pivoted through the end cap. A small brass piece was pushed onto the end of the wire to stop it falling out. The inside face of the cap is chamfered so the piston plunger will more easily pass though the hole as the piston and rod are pulled back.

The piston is again plastic chopping board. I machined an o-ring groove into it to provide a good seal inside the pistol body. The main piston rod is made from brass tube. Another o-ring is under the head of the piston rod to seal that against the piston. The plunger is designed to be telescoping.  The telescoping section is made from steel and is the most complicated part of the pistol. I used a 1/8 inch steed rod to make the moving section. On the far end of this rod is a metal piece to attach the pull ring to. On the other is a small stop. These ends are silver soldered to the rod. In the middle, but free to slide,  is the main catch piece.  This is machined so that it is a tight fit into the brass tube of the plunger rod and it is soldered in place. The 1/8 rod can now slide in and out of the brass piston tube but it won’t pull out of it. The catch piece has a chamfer on the end so it can easily pull through the hole in the end cap of the pistol. It also has a groove machined into it. It is this groove that the music wire catch hooks into to hold the piston back in the cocked position. Because the stop on the end of the 1/8 rod that is inside the brass tube is a reasonably good fit I drilled holes in the brass tube to allow air to escape from it as the telescoping section moves inside it. I should have made the internal stop on the end of the 1/8 rod smaller so this wouldn’t have been necessary.

Continue onto part 2 here: http://www.asciimation.co.nz/bb/2011/01/19/building-a-better-nerf-gun-part-2

The driver would turn the little knobs to scroll through a little list of directions to their destination. In car navigation before GPS!

It occurred to me (and others going by the comments in the link above) that you can easily make a list of directions like that using Google maps. So I gave that a try. I selected a route and plugged it into Google maps and produced a list. A little editing in something like Openoffice and you can print out a nice shopping list of directions. All I needed was something to use them in.

I went out to my shed and rummaged in my box of brass bits and pieces and threw this together.

Google Maps Wristlet Navigator.

The body is all hand soldered from brass modeling tube. There are two cylinders, each with a slot cut along one edge. They are soldered to a flat back plate that the strap is threaded though. The ends are made from more small pieces of brass. On the front are some flat pieces and rods that the glass window press up against. The window is made from a glass microscope slide I cut down and glued in place with some windscreen sealant. The strap was just a scrap of green leather I had with a buckle attached to one end. I made the strap large so the watch can be worn over your sleeves so you can see it when driving.

The far end of the cylinders are plugged with small turned pieces I machined up on my mini lathe. The scrolls are made from long 1/8 inch brass machine screws. The ends were machined on the lathe and they are a nice, smooth fit into the cylinders. They are soldered onto the screw leaving a about 1/8 inch of thread on the far end for the knobs to be screwed onto.

The knobs on the end were made from acorn nuts and a piece from a binding post screw. Those are used to bind loose pages together into a booklet. I drilled out the centre of the binding post screw and pushed that down over the top of the acorn nut and soldered it in place. This gave me a little knurled knob with a domed end that can be screwed onto the end of the scrolls.

The directions are printed out onto a 35mm wide strip of paper and then simple taped and wound onto the scrolls. The scrolls need to be would into an ‘S’ shape. This is so that you can advance the list by turning the top knob clockwise and you reverse the list by turning the bottom knob clockwise also. This is very important! Because the knobs both use a standard thread you turn them clockwise to tighten them.  If you turn them anti-clockwise the knobs unscrew! There is enough friction in the scroll and paper that turning the knobs backwards will cause them to unscrew.

To load the device you unscrew the knobs and insert the scroll into the watch ensuring the paper slides nicely through the slot in the left hand end. The threaded ends of the scrolls will poke out the right hand side of the watch. You then screw on the knobs which then prevent the scrolls from sliding back out.

Knobs to you!

The glass window, an improvement on the original device as far as I can tell, stops the paper getting wet in the rain when you are driving your convertible sports car with the top down.

The amazing thing is it actually bloody works! You can easily fit the length of an A4 sheet of paper into it which will provide around 25 different directions when printed at the size shown. Of course you could simply shrink the size of the text and fit many more.

Here are more pictures of it in action. Oh and the beauty of this device – it is silent!

Update: 22 August 2010

Just a quick note for all those interested in this. It isn’t Steampunk! At least I don’t call it that. It’s just made of brass. Also the car is an MGB.

Finally some results! But lets start with what I have done so far. Since my initial experiments with my scanner were somewhat successful and I had my new QFH antenna made and ready to go I just needed my Minikits weather satellite receiver to arrive and then I could try getting some decent images.

The kit arrived after just over a week and ended up costing me $NZ125 including shipping. This is far cheaper than any other kit I could find online, all of which you have to pay in Euros or US dollars.

The Minikits kit.

The kit is rather nice. It comes with a tinned double sided plate through hole circuit board. Several little bags of components including all the coils and ICs you need. It also came with an F connector to solder onto the board but I have chosen to use BNC connectors instead. It also comes with the instructions and the description of how the circuit works.

The kit uses an MC13135P receiver chip which is a complete FM receiver on a chip basically. There is also a TL072 op-amp and an MC34119 audio amplifier. All of the components are standard leaded types except for a surface mount NTC thermistor used as part of the automatic frequency control circuit to account for temperature variations.

You do need a few other things so I added the speaker, the signal meter, a AFC switch (the kit comes with a jumper) and the audio out connector. I also used full sized pots for the volume and squelch control since I will mount them off the board. The kit comes with trimmers to use on the board though. The kit runs off a 12 volt supply which is regulated down to 5 volts on the board for most of the circuit to use. I made a small power supply board with a 12 volt regulator on it and powered that from an old power supply from a set of old computer speakers. The used speaker I am recycling came from the same set of old speakers. The signal meter was a nice find. The first place I look for odd electronic stuff, Surplustronics, didn’t have anything so I headed up the road to the top of Queen Street to a real electronics junk shop. I have no idea what it’s called but it’s been there for years now. It’s cram packed with junk. Most of it is useless but the odd interesting thing here and there. Nothing is sorted. It’s all just in piles. I dug about and came up with this nice signal meter with a 0-5 scale on it. He wanted $3 for it. I had $2.20 in change (good thing I hadn’t bought a Mars bar that day) so he said close enough. It was unused and still in the bag!

Assembled kit set. The red thing is a plastic tuning tool.

Building the kit is easy. The instructions say it takes about four hours but I did it in two. All those years of working as an electronics tech to pay for my way through university came in handy! Once you have built it you need to set it up.

The way the receiver works is the incoming signal, which is around 137MHz, first goes though three bandpass filters. These eliminate any extra signals from getting through and just pass the weather satellite transmission frequencies that we want to receive. This is important in some places where there are strong pager frequencies nearby (people still use pagers – probably more reliable that Telecom’s XT network).

We then want to take the incoming signal and convert it down to a much lower frequency. This is called an intermediate frequency.  An intermediate frequency, or IF, is used because it is easier to work with and  filter on the lower frequency and because if a receiver converts all the frequencies you want to receive down to the same IF then all your circuitry from that point on can be optimised for just that IF frequency. This receiver uses the same IF most FM receivers use, 10.7MHz.

To get to the IF the receiver chip uses a mixer. The mixer takes the received signal and mixes it with another signal that we generate on the board. This signal needs to be 10.7MHz higher than the signal we wish to receiver. The mixer takes our signal and subtracts from it the received signal to get the 10.7MHz IF signal. This process is called Hetrodyning (which is probably what my friend Fing tells his friends he is doing when he comes to visit Auckland and we all go to the pub for lunch). The on board frequency is generated by a thing called a voltage controlled oscillator or VCO.

In this circuit the VCO is controlled by a coil, which gives a rough setting, and the tuning resistors, to fine tune to for each frequency.  The kit uses 25 turn trimmers for tuning. Each channel uses one trimmer. The kit only comes with one but you can simply add more and use a multi-pole switch to switch in the appropriate ones. For now I just have the one which I adjust for each different frequency when I need to.

The signal passes though various stages in the chip and eventually gets spat out as audio. There are two audio paths. One is to the audio amplifier chip which is controlled by the squelch circuit. The other is a line output to the PC sound card.

To set the receiver up you need to adjust the VCO to be 10.7MHz above the frequency you want to receive. The receiver needs to receive from 137MHz to 138MHz approximately. The tuning resistor needs to tune across that range.  If you set the tuning resistor to half way in it’s travel that would correspond to 137.500MHz. We therefore want the VCO to be oscillating at 148.200MHZ (148.2 – 10.7 = 137.5).

To set the VCO frequency you need to adjust one of the coils on the board. These coils have adjustable ferrite slugs in them that you screw up and down to set the frequency. They are fairly delicate and also metal near them will upset the frequency so you need to use a red plastic tool for adjusting these. This was available as an extra from Minikits cheaply.

The best way to know you have the right frequency is with a frequency counter. Unfortunately I don’t have one. So the next best thing is to use my scanner again. Because the coil is oscillating it is also radiating radio waves. If you hold the scanner close to the circuit board set to 148.200MHz when you adjust the coil and it hits that frequency you’ll hear it on the scanner.

With the VCO set this way you know the receiver should be receiving across the 137 to 138MHz frequencies over the 25 turn range of the trimmer tuning resistors.

Next the instructions say to attach a RF signal generator set to 137.500MHz to the antenna input and adjust the filters and output demodulator for the cleanest signal. Again, I don’t have a signal generator so instead I had to wait until I knew a satellite was overhead. With the antenna sitting out on the front deck railing I used the WXtoImg software to know when to listen and on which frequency. I used my scanner to be able to hear when the signal was there.  I then connected the antenna to the receiver and tuned the receiver to the right frequency and quickly adjusted the rest of the circuit while I still had a signal. This can take some time given the low number of passes per day and the short time the passes last. You have about 15 minutes in which to tune things.

To do this tuning I used my ocsilloscope on the audio output to see how clean the signal looked as well as by watching the signal meter and adjust it to give the highest signal as suggested in the instructions. This worked very well.

Initial setup and testing.

One interesting part of the circuit is the automatic frequency control, or AFC. This is quite clever and it will allow the receiver to account for the Doppler shift in the satellite as it passes towards then away from the receiver by adjusting the VCO.

You will see that the EeePC is displaying an image here but I found that it can’t actually decode the pictures properly. I think this is because I was using the mic input instead a a line in (which the EeePC doesn’t have) and so I wasn’t getting a proper signal.

The next thing to do was transfer everything over to my main PC and to put the antenna in a better position (on top of the guttering on the edge of my roof for now). With everything set up like that I finally got my first real picture.

The first image!

And this is what I got. The noise and blank part at the bottom of the image was me trying to tune things in as the satellite passed (from south to north). The horizontal lines were when I was adjusting things. In between though you can see a good image.

With more tweaking and tuning I am now getting nicer images. I am just using the basic settings in the software for now. I think there are a number of different settings to show you different things but I haven’t played with them must so far. These images are using MCIR map colour IR and an Equidistant Cylindrical  settings.

First good image.

I still need to manually tune hence the missing bottom to the image but this was a imagine received from NOAA 18 on 137.9125 MHz (that’s why so much of the bottom is missing – I had to turn that 25 turn trimmer a lot) at about 2pm.

The same satellite did another pass about 1 hour 40 minutes later and I got the image below.

Second pass by the same satellite showing Australia.

You can see how the satellite follows different paths on different orbits so you end up receiving differing pictures. Using some (free) software such as WXTrack you can actually see the path the satellite will take overhead. If the satellite is very low on the horizon you will get noisy images. The antenna needs a clear line to the satellite so obstructions in the way will degrade the signal.

So, what’s next? Well, I am still only using the one tuning resistor so for each different satellite frequency I need to re-tune the receiver. This won’t do as a long term solution. Instead of the idea that is given in the kit of using 5 tuning resistors all switched in via a rotary switch I want to try something a bit more automatic.

The kit does provide connection points for adding a phase locked loop or PLL for short. This would allow scanning across the whole frequency range then have it stop once it detects the appropriate signals. You can buy suitable experimenters PLLs that are micro controlled but these are rather expensive and I am trying to do this on the cheap!

I considered using a micro controller to switch between five or so pre-tuned frequencies. Because there are only a few satellites transmitting on fixed frequencies (the status of the satellites and information about them is available on the NOAA POES status site) you don’t actually need to scan the whole range. You can just step through fixed channels looking for a signal.  I could use a micro to switch each channel in for a brief period, say 2 or 3 seconds, and if a suitable signal was detected stay on that channel until the signal drops out again.

Switching in a set of preset tuning resistors can be done using a bilateral switching chip. A micro could control the switching. I then though that’s perhaps over complicating things so perhaps I can do it entirely using normal ICs.

My current plan is to use a 555 timer providing a slow clock pulse. These I feed into a 4017 decade counter (one of the first ICs I ever experimented with so a favourite). The outputs of the counter feed into a 4066 bilateral switch to switch in each tuning resistor in turn. They also turn on various LEDs to show what channel is currently selected. When a signal is detected the counter stops on that signal until the whole signal is received. This part I haven’t worked out yet. I was thinking of using the squelch signal but I don’t think this will work too well. Instead I might make a tone detector to listen for the tone you hear when the signal is coming in. The APT system used to transmit the images used a 2400Hz sub-carrier it should be possible to detect easily.

Progress so far.

So far I have got the clock, counter and switching working. You can see those on the breadboard in the picture above. I need to start on the tone detector. Perhaps using a LM567. My backup plan if I can’t get that working is to use an Arduino perhaps. Everyone loves Arduino!

Also I need a more permanent antenna installation. Perhaps a small mast mounted on the roof of my garage so I can put it there and get a nice, clear view of the sky.

And of course once done everything needs to be put into some nice housing.

Will update this with progress soon.

Now I will describe how I made a much neater QFH antenna to receive the satellite signals. My wooden one worked fine but it was not waterproof or particularly pretty or accurately made. So I decided to make a waterproof, prettier and more accurate one.

Again I followed the directions given here: http://www.g4ilo.com/qfh.html

I started off with the raw materials which I got from one of my most frequented places – Bunnings Warehouse. There are probably cheaper places to buy this stuff but Bunnings is convenient. I bought a 1m length of 32mm diamater PVC pipe and three lengths of black 15mm diameter Enduroflex pipe. I have no idea what that stuff is used for. Water piping I think. The PVC pipe cost NZ$15.51 and the black pipe was $11.16 for the three lengths although I only ended up using two of them. That black pipe incidentally is the perfect size for the barrel for a home made Nerf gun firing suction darts (a project for another day)!

I also bought some RG-58 (50 ohm) co-axial cable from Jaycar rather than the RG-6U I used on the wooden antenna. The The problem I had with the RG-6U is that the braiding isn’t copper but rather aluminium so it was impossible for me to solder the connection easily. The RG-58 is somewhat thinner diameter than the RG-6U which does mean the antenna is more easily bent out of shape but as my antenna will only be used occasionally it should be fine for now. The 15 metres of co-ax was $28.50.I cut the PVC down to about 800mm long and marked it up ready to drill the holes for the cross arms. I was very careful with the dimensions on this antenna. An easy way to mark around the pipe is to wrap a piece of printer paper around it ensuring the paper is square to the pipe by making it overlap itself perfectly as it wraps over itself. You can then mark off the place the sheet overlaps. You unroll the paper and the distance form the edge of the paper to the mark you a made is exactly the circumference of the pipe. It is then easy to mark off the quarter and half way points on the flat paper. Then you wrap the paper around the pipe once more to transfer those marks to it.

I then put the pipe on a nice flat surface and use the ended of a long, straight piece of wood to transfer the lines down the entire length of the pipe.

After carefully marking the pipe and measuring it several times I drilled the holes to hold the cross members. I first drilled a pilot hole with a 3mm bit then I used the spade bit to drill out a larger hole (sorry for the badly focused photo).

To make the holes a nice, tight fit I drilled slightly undersized then used a tapered reamer to enlarge the holes to the exact right diameter.

Once done you end up with mast finished. The positioning of the middle set of holes holes isn’t very critical as those arms just hold the ends of the cables in the correct place. Obviously you want them close to the mid point though.

I next carefully cut the six cross arms from the black pipe. Note the two different sizes for the two different sized loops in the antenna. I trimmed the ends square in the lathe (because I could, not for any technical reason) then carefully marked and drilled the holes in the ends. Looking at the plans you can see these holes need to be a fixed width apart to ensure the loops of the antenna are the correct size. They also need to be drilled at 45 degree angles to the mast so the cable has the correct spiral to it.

To drill the holes so they are at 45 degrees to the mast when installed you actually need to drill the holes in each cross member 90 degrees apart, one at either end of the support. Since the drill press drills down vertically all I did was drill one hole then push a piece of aluminium tube through it temporarily. Then, turning the tube around I simply adjusted it in the vice until the aluminium tube was parallel to my drill table. With it parallel when I drilled the vertical hole I could be sure the two were 90 degrees apart. When you push the crossmember through the mast you simply twist it so that one of the holes is at 45 degrees to the mast and the other will be aligned at 45 degrees too. You do have to twist it the right way though so that the cable is spiraling down anti-clockwise when viewed from above.

Because I might want to experiment with this antenna a little I didn’t want to glue the crossmembers fully in place yet so I am temporarily using some pieces of split ribbed tubing to stop the crossmembers moving in the mast.

In the top and bottom sets of holes I needed to file slots so that the cable could pass through into the mast. The holes for the arms at the ends of the mast I drilled so that the centre of the co-ax cable would be sitting at the exact right measurement given in the plans with the cable running along the outside of the crossmember , not inside it. Running the cable inside the crossmembers would of course be neater. But there is then no way to make sure the cable lies flat in the correct place. Does this make much difference? I have no idea! But the dimensions are given down to fractions of a mm accuracy so I decided to get things as close as possible.

I started putting the crossmembers in place and also soldering the connections. Shown here is the bottom most connection. That is the feed wire soldered to the shield of the bottom of one of the loops. I insulated the connection with waterproof tape then carefully pushed it into the mast and then pushed the crossmember fully home.

Similarly at the top I carefully soldered the connection inside the top of the mast once the loops and the crossmembers were all in place.  Notice the anti-clockwise direction of the spirals. I used lots of cable ties to hold the co-ax down flat across the crossmembers. Eventually I will remove the split tube and just glue the crossmembers in place once I know everything is correctly positioned.

To finish off the antenna I put a PVC cap over the top of the mast and I also made a quick and dirty stand from a scrap piece of wood and another off-cut of pipe that the mast happens to slide into perfectly. Eventually my plan is to hang it from a line suspended above my garage roof where it will have a clear view of the sky. The antenna might be connected to a line going around a pulley so I can hook the antenna onto the line and pull it up into position but still bring it inside if the weather is too bad. Either that or I will make a third, more robust version to permanently mount on my house roof.

With the antenna on it’s temporary stand I was able to move it about outside to see how well it would work. It works well when it has a clear line to the satellite down to very low on the horizon. The signal is very easily blocked though so building and even trees will cause you to lose the signal.

Now I just need my receiver kit to arrive so I can build that and really see how this works!

Update: The kit arrived! Read about it in Part 3.

I should point out that I am not a radio amateur, or HAM, as they are known. I am just an electronics tinkerer. Anyone with good, general electronics skills can do this.

Bender showing off my scanner.

One thing you can do is pick up pager frequencies. I pick up 157.950MHz easily at home which is a Telecom pager frequency. It sounds like a random collection of bleeps and bloops – R2D2 with his knickers in a twist. Apparently with a simple mod you can break out the unfiltered audio (called the discriminator output) in the scanner and then, using some software called PDW, decode the pager signals.

Even though the mod is easy to do (if you are electronically inclined) looking at other peoples pager messages strikes me as being terribly uninteresting, kind of like having to overhear other peoples one sided phone calls on the bus, so I haven’t tried this modification myself.

One other thing I could try (after doing the discriminator mod above) is receiving AIS signals. These are the identification signals used by ships. Since I work in the city and it’s not too far down to the port I should really try seeing if I can receive them. Perhaps the Northern Steamship Company pub might be a good place to try!

The frequencies used are 161.9750 for AIS1 and 162.0250 for AIS2. Apparently you can then feed the signals into a sound card on a PC and using some software like Shipplotter you can track the movements of shipping in your area. Something to try another day.

One thing I did want to try, being interested in all things to do with space and space technology, is receiving weather satellite transmissions. Now, before I start I should say that you cannot actually decode the weather satellite transmission using this particular scanner. You can HEAR the signals but you cannot generate any kind of useful pictures with them. Well, I have been unable to. But given that you can hear them was enough to start me off on this project.

Lets start with some background first.  Up there, in space (well, low earth orbit 850km up), there are some weather satellites. Both the Americans and the Russians have them up there, the NOAA and the METEOR satellites respectively. There are others of course but the NOAA are the ones you can easily receive signals from so they are the ones I am listening to. These satellites are continuously transmitting signals down to earth using a system called APT. The Wikipedia Automatic Picture Transmission article gives you all the background you need but here are the basics.

The NOAA satellites are in a polar orbit around the earth and as they orbit they are continuously scanning a line beneath them 3000km wide and transmitting this line back to earth. Being in a polar orbit means each satellite will pass over the entire planet as the earth rotates beneath them. As a satellite comes up over the horizon and passes overhead a receiver on the ground can start receiving these lines, decode them in sequence and build up a picture. Due to their low altitude these satellites orbit rapidly and each satellite will make several passes over your location each day.

The signals are transmitted on a group of frequencies around 137MHz.  The actual picture received  contains two views of the earth below the satellite – one in visible wavelengths and another in infrared. A line is scanned twice every second and each line contains the image data as well as non image data. This system has been used since the 1960s so it is quite remarkable it is still in use today. You can get up to date information about the NOAA weather satellites from this status page.

To receive the pictures you need three things: a receiver, an antenna and a PC (with a sound card and some software to do the decoding).

First, the receiver. As I mentioned above my particular scanner (or indeed most scanners) isn’t suited for doing more than hearing the signals but it is enough to let me work on the second piece, the antenna.

The antenna obviously needs to pick up the 137MHz signal the satellites transmit on. It also needs a wide receiving pattern as it has to be able to pick up the signal from horizon to horizon. The signal from the satellites is circularly polarised so the antenna must be designed in a particular way to receive the signal. Two suitable antenna designs are the turnstile antenna or the quadrifilar helix antenna (QFH).

Looking around online I came across several references to the QFH antennas with very good instructions on how to make them specifically for receiving weather satellite pictures. The design I decided to use was from here: http://www.g4ilo.com/qfh.html

Because I didn’t know if I was going to be able to receive anything I made my first QFH using materials I had on hand, namely wooden dowels. I made an antenna using the instructions and dimensions given above. I used RG-6U coaxial cable as that was easy for me to get quickly in order to try things out. I first tried using the antenna indoors until a friend pointed out that the metal roof would block the signal (a duh moment) so instead I simply took the antenna outside and hand held it in order to try it out. To my amazement it worked! I could hear, albeit with a lot of noise and static, the satellites transmitting their signals.

It’s worth mentioning that because the satellites are constantly orbiting you need to know exactly when a satellite will be passing over. This brings us neatly to the third thing you require, a PC with a sound card and some software running on it.

The software I am using is called WXtoImg and it is available to download and use for free (you can also pay to register it for additional features). Interestingly it seems to be written here in New Zealand. This software does two things. It decodes the received audio signal and formats that into usable images and it also tells you when (and where) the next satellite will appear.

Another useful piece of free software is WXTrack by David Taylor. This software lets you see exactly where satellites are over the earth at any given time.

Both of these programs require you to download up to date satellite tracking information called Keplers but this is well explained and easy to set up.

Since I could now hear the signal and I had downloaded and installed the appropriate software I rigged up a temporary support to hold the antenna up outside my window were it would have a (reasonably) clear view of the sky and attempted to feed the signals into my PC.

My quick and dirty wooden QFH antenna and dodgy mounting method.

The results were not good (it wooden work)! This I expected and the problem isn’t the antenna itself but the limitations of my scanner.

There are various problems with using scanners for receiving weather satellites. Other weather satellite sites explain the problems better than me but to summarise the main problem is due to bandwidth. Now bandwidth, as I understand it (and I could be wrong!), when talking about receiving weather satellite transmissions is the difference between the lowest frequency and the highest frequency that you need to be receiving signals over to get the full signal. A scanner like mine is designed for receiving voice signals. Voice signals don’t need a very wide frequency range so we say the scanner has a narrow bandwidth. The bandwidth of a scanner like mine is about 15kHz so the scanner will filter out any signal outside that range around the main frequency.

The satellites transmit their signal over a much wider range spread around the main frequency, i.e. is has a wider bandwidth. To accurately receive the entire weather satellite signal you need a receiver with a 30kHz to 50 kHz bandwidth. So in effect my scanner is blocking out part of the signal. We basically lose information from that lost signal so the software can’t build a good picture from it.

Other issues with a normal scanner are sensitivity, how strongly it picks up the weak satellite signal, and selectivity, how well it picks up only the frequency you want. Another, more subtle problem, is the doppler shift caused by the movement of the satellite as it passes overhead. This causes the frequency to shift as the satellite passes overhead.

All these issues mean my little scanner isn’t up to the job. To really demonstrate this here is what a properly recorded signal of a weather satellite should sound like: http://n8imo.com/APT/images/N14.WAV (from http://n8imo.com/wefax4.html).

What I get is this: http://www.asciimation.co.nz/misc/06250521.wav\

Image produced by the scanner.

My friend Dave wanted to see a picture of what the scanner produced. This is the best that I ever managed to get. Mainly noise. With some vague picture in the middle as the satellite passed right overhead.

So when people say a normal scanner won’t work, they are right!

The next thing to do is sort out a proper receiver. There are a number of options here. You can buy a scanner that does have the correct bandwidth but these are very expensive. Or you can buy a ready made receiver just for receiving weather satellites. These are also expensive. Finally you have kit receivers you buy and assemble yourself. Being an electronics tinkerer this is what I was after. There are various kits on offer but I chose the Minikits weather satellite receiver.

This receiver is very reasonably priced, should do everything I want and the company is more or less local (well, Australia) so I was able to pay in Australian dollars rather than US dollars or Euros which tend to work against someone earning NZ dollars. Shipping was also cheaper and faster.

I ordered the kit last week and it should arrive any day now so I will be building that as soon as possible.

In the mean time, having proved that a QFH antenna would indeed work I set about making a more accurate, more weather proof version.

Will describe that in part 2 here.

http://www.asciimation.co.nz/bb/2009/10/18/a-rocking-horse-for-sophie

With time moving on I had to really move to get the rocking horse done. After protecting the foam with the tin foil I applied layers of fibreglass and polyester resin. I built up a nice, thick skin of glass all over the horse.

Layers of fibreglass to make a shell.

After the fibreglass shell had hardened I gave it  a quick sanding just to remove any sharp bits then set about adding in Horseome Paul’s super secret feature – a hidden hiding place where Sophie can keep things. The cavity was made form a piece of PVC piping. A hole was cut in the horse and the pipe gently inserted.

Gentle insertion of secret hiding place.

Once the pipe was fully inserted the excess was trimmed off with a cutting disc on a Dremel.

Cutting off the excess.

Once the cavity was done I then covered the body in several layers of 5mm thick upholstery foam. This was glued on using Ados F2 spray contact adhesive. I put extra layers on the head for some extra padding in case of face plants by the young rider!

Covered in foam padding.

Next I had to figure out how to cover him in fur. A visit to Spotlight (a material and craft super shop – kind of like Bunnings for girls) provided me with some great fake fur and some pink frilly stuff for the mane and tail.  I roped in a friend to help with the sewing (at the cost of making her dinner). I wrapped the horse in plastic from an old plastic bag then wrapped that in layers of brown packing tape. The idea is to make up a tape shell that you can then cut off with scissors and cut to form flat pattern pieces that will exactly match the three dimensional shape you are trying to cover. I was going to buy a sewing machine and learn to do all this myself but with only a few weeks before Christmas I ran out of time.

Wrapped in plastic and tape.

So, with time running short my friend Dani came over to help with the pattern making and sewing. One look at Horseome Paul though and we decided it was going to be trickier than we though. Dani makes quilts which are flat. A three dimensional horse was a much trickier problem. We decided I should try to find someone with a bit more experience to help. Still made her dinner though

As it happened there is an alterations place near to where I work (errr, used to work as I was made redundant just before Christmas). I popped in there and talked to a woman who didn’t speak much English but after must gesturing and explaining she realised what I wanted to do and said she could help. I took the horse and material along to her and she sewed on the covering for me in a very short time and for an extremely reasonable price! She also made a bridle and sewed on buttons for eyes and even gave him a little charm necklace. He has a zip running down his back so the fur can be removed if needed.

Horsey head!

In the mean time I completed the wooded base. I originally painted it pink but in the end repainted it white and the pink become overwhelming! I attached some cardboard stickers I got from Spotlight to give some detail and to put Sophie’s name on the base. They were supposed to be self adhesive but I found I needed glue to get them to stick. Finally the body was simply bolted down to the base and held with large fender washers and Nylock nuts.

So here is the final product!

Horseome Paul is complete!

I ran out of time to make a saddle but I think he doesn’t really need one anyway. Christmas day is today so we shall see what young Sophie things in a few hours or so! Hopefully it doesn’t scare her!

Oh, that secret hiding place. Well, you can get to it by simply unzipping the zip under the tail a little. I am not sure my sister is going to approve…

And finally, 6 months later, an actual picture of Sophie on her horse. Yes, it is still bigger than she is!

Sophie aged 2 years 5 months (or so). Horse aged 6 months.