Homebrew

Welcome to my ‘Home-brew’ page

From time to time, it’s nice to create a piece of equipment that you’ve made yourself. Below are a few projects that I’ve recently been working on. I hope to add more as time goes by and I’ll also include some antenna designs that are simple, effective and easy to construct and cost only a few pounds. There will also be some simple low-cost solutions on antennas that are easy and straightforward to implement.

Click on the links below to go directly to the item

HF and (Bonus VHF!!) – A Real QRO Extra Heavy Duty Antenna Switch (Listed directly below)
HF/VHF/UHF 300w 6-Port Antenna Switch using a MIL-SPEC relay
How to ‘Wideband’ a 3 Element Yagi
Constructing and Matching a Single Element Mono-Band Delta Loop

HF and (Bonus VHF!!) – A Real QRO Extra Heavy Duty Antenna Switch

Here’s a really nifty heavy duty antenna switch that’s better than most you’ll find.

Homebrew 10k HF/VHF Antenna Switch

The ‘Relay Box’

When we prototyped the ‘Vortex V-RAT’ 6-port antenna switch, we found ourselves with a couple of spare chassis bases with holes already drilled for SO239 or N-Type connectors. Rather than just put them in a junk box, I thought they could be used for some home-brew projects. Given that also I didn’t want to cannibalize a good V-RAT box, I set about building a home-brew switch which would be a little different. As I planned for high power, Amphenol and Rosenberger PTFE SO239 sockets were used.

HF/VHF Antenna Switch Underside

The switch is based around the Russian B2B-1B W2W-1W (15A 4kv) single pole single throw vacuum relay which is quite tall (115mm), so I had a friend construct quite a tall housing from 2mm aluminium sheet which was then folded and welded on the sides to make it watertight. Each tube is supplied with the factory spec sheet plus the unique serial number. Here is ‘Unit 1606’ which was manufactured in the Svetlana factory, St Petersburg in 1992.

B2B-1B W2W-1W Vacuum Relay

In the Vortex workshop I cut a 12mm section of PTFE sheet to use as a base to mount the relays. These were then drilled with a hole saw to accommodate the relays. M8 threaded spacers were used to separate and also hold the base to the chassis. Now we have the base and two chassis sections to work with, so let’s look more at the relays.

The relays were built solely for military use and as such are very conservatively rated. I managed to get 6 NOS (New Old Stock) tubes on EBay which were manufactured around 1992 for around £30 each from Bulgaria. Given the amount of stock around, this was a pretty good price especially for new pieces. Expect to pay around £30 used and £40 new each as of 2019.

The published rating is 5kw key down at 30MHz, so lower frequencies will have a higher rating especially 40, 80 and 160m. Additional info from QRO Parts in Russia gave them a rating of 25A at 32MHz and 35A at 3 MHz. Here’s a direct link to purchase 8 brand new ones plus all the technical data you need (and you get a couple of spares for other projects).

I didn’t want to push it too far as destroying the project was not on the agenda. Originally the relays were set up and the switching was tested together with a VNA curve to get some idea of frequency response. As per the published data, the relays saw no issues up to around 33MHz. Going higher in frequency caused quite a step rise in SWR which was to be expected as they are only rated for HF. The relays are spec’d to work at 6-16v DC and work well when switched with a 13.8v dc supply and also seem quite happy all the way down to around 2v although we wouldn’t recommend using voltages this low.

I fitted each relay with an 0.01uf RF bypass capacitor across the coil and a flyback diode (1N4xxx) series. Since an inductor (the relay coil) cannot change its current instantly, the diode provides a path for the current when the coil is switched off. Otherwise, a voltage spike will occur causing arcing on switch contacts. Since I have hundreds of 1N4007 rectifier diodes in the workshop, these were put to good use.

Relay Mount - Underside

Following on from the original build, I thought it would be a good idea to try and get the system to operate on 6m. Testing with a VNA gave quite a bad SWR on 6m (about 1.6 to 1 – usable but not great) and on 4m the SWR was up to 1.9 to 1. I tested the 2m band but 144MHz was totally unusable at around 3 to 1. After searching on-line, I found another US based home brewer (W6PQL) using capacitors to remove the unwanted reactance created the further up the band you go. He suggested using 5pf on the input and 5pf on each output. However, the project shown was using relays that were much friendlier at the lower VHF frequencies than my chunky Russian comrades.

After an hour or so of trial and error with some small junk-box capacitors, I had settled on 10pf on the each output and 10pf on the input. The return loss on each port was excellent and on two ports the SWR at 144MHZ was below 1.4 to 1. Not perfect but usable. On 70MHz all ports were below 1.2 to 1 and at 6m no reactance was present with a nice flat 1 to 1 SWR. Looking on-line again, I found someone from Bulgaria selling some pretty hefty 10pf doorknob capacitors. These again were NOS Russian military items rated at 10kv. I purchased 8 for around £23.

10uf 10kv Russian Doorknob Capacitors

When the caps arrived I duly fitted one on the input, just after the SO239 socket and one on each output, just before the SO239 sockets. Now fingers crossed to see what the return loss was and to check if there was any difference. It was great to see that the figures with the Russian caps were pretty well much the same when compared to the junk box test caps.

Whilst testing a super QRO Vortex commercial antenna for a customer, we had the chance to use a 20kw high-power wideband HF/VHF amplifier so making good use of the time available to us; we put the switch into action for some testing during lunch.

On 20m the relays were handling 8kw key down carrier and at 30MHz 6kw key down carrier without any issues. I decide to play on the safe side at VHF but the system didn’t break a sweat with 5kw key down at 50 and 70MHz. Even at 2m on the two best ports, 3kw was passed through without issues.

Homebrew 10k HF/VHF Antenna Switch Relays and Capacitors

The Shack ‘Switch Box’

The switching control for the box was constructed using a simple ‘Black Box’ enclosure purchased from ‘Rapid Electronics’. The box was the drilled on the base to accommodate four rubber feet purchased from Farnell which we also use for the VRAT boxes.

A six way rotary switch also from Farnell (I used a hefty 2.5A version) and knob was fitted together with six green LED’s (one for each port) and a red LED for the power on/off indicator. A small heavy duty 5A toggle switch turned on the power which fed via a small 5A DC connector on the rear.

Shack 6-way switch box

To prevent a mistake of connecting the device to the wrong DC terminal, a hefty polarity protection diode (SB340 – Schottky Rectifier) was used in series with the main positive DC feed. There’s a minor price to pay in a small voltage drop (about 0.3v) but this is better than frying everything in your new project.

Small quarter watt current limiting resistors were used in line with each LED. The resistance will vary depending on the value of resistor used but I used 470k ohm resistors when using a supply of 13.8v and 2.2volt LED’s. You can work out other values here. After a few days, I added extra resistance to the on/off LED as it was a little too bright. Power was supplied via 13.8v DC power supply.

The main switching connector used was a ‘SP17’ series 7-Way by Weipu. Don’t use the SP13 smaller connectors as soldering these is really fiddly. These are Chinese connectors but are very robust. Being rated at IP68, they are fully waterproof which makes for a good outdoor solution. There’s plenty on EBay to find and a plug/socket set will set you back around £6.00 (You’ll need two, one for the relay box and one for the switch box). Originally I had only intended making something fairly rough of the switch box so I’ll have to get the interior smartened up – but as it worked first time, for the moment I’ve left it as it was.

Shack 6-way switch box interior

Although I have been busy over the winter, it was nice to make time for something different but it did make a great winter project. If you add up the cost of the components and time, it’s not a cheap and cheerful project. It uses military grade hardware throughout and takes quite a while to construct but it’s worth it. In all I spent around £500 on components, hardware and fabrication and a few hours every now and then over a three week period to get everything as I wanted it. I also found that with the additional capacitors that it also works well on VHF.

As a PS, N-Type connectors would make a better solution certainly at VHF as the good old SO239 can easily give sloppy impedance readings the further up the band you go. Of course putting this kind of power through SO239’s for any length of time is not recommended but as I’m only using UK legal limit:) – then there shouldn’t be any issues.

Although my shack is currently undergoing some re-decorating and a full revamp, I’ll be putting the switch to good use during the course of the year.

Sorry – before you ask, no we are not manufacturing this – it was a home brew special project only but our great
Vortex 5kw V-RAT is available for next day dispatch.

More items will be added in due course….

HF/VHF/UHF 300w 6-Port Antenna Switch

A Cheap and cheerful but professional MIL-SPEC antenna switch that’s good for 300w at HF/VHF and UHF using a professional switching relay which is good up to 18GHz and costs only around £65 to build.

As I’m doing quite a lot of work outside on the tower and antennas, one recent project was to provide switching capability for a variety of VHF and UHF antennas. Higher frequency switches tend to be more expensive because the higher up the frequency you go, the more margin of error from relays and from split feed lines. Everything in the tx loop wants to see 50 ohms but splitting your coax for whatever reason causes a ‘bump’ in the 50 ohm feed line. Sometimes a split of a few centimetres will cause a 600 ohm+ bump. It’s not so critical at HF but start drifting over 100MHz or so and the bump can cause quite an impedance shift. At 70 cms it can be quite dramatic.

Transco 146C70600

After a search online, I saw a few guys from the USA using ex-military/commercial ‘Transco’ switches which although not super high power, seemed to be perfectly designed with VHF and UHF in mind. An EBay search revealed a variety of different ‘Transco (Dowkey) Switches’. One in particular was a 6 port switch with SMA connectors which was an ‘MO’ type relay switching six outputs (The model was 146C70600). The unit was MIL-Spec and rated at 300w key down so certainly adequate for what I required. It had a great low return loss with the VSWR curve going right into the GHZ range. In fact it’s rated so well, even at 18GHz the maximum SWR was 1.5 to 1. The relay worked off a 20-30v DC supply and with superb isolation (better than 100db below 10MHz and better than 80db below 500MHz this seemed to fit the bill nicely. All this for only £25.00 and in great condition.

Transco 146C70600 connectors

First item on the agenda was to obtain an enclosure. On this occasion I chose a rugged plastic ABS enclosure from Rapid Electronics. I had also decided that with six outputs, I’d use a combination of N-type connectors (on four of them) and SO-239 sockets on the remaining two. The input was also an N-Type. It’s worth noting that at VHF and above, many SO239 sockets are not 50ohms and struggle to retain a good level 50ohm impedance. I’ve heard stories of quite hefty impedance bumps using SO239’s which doesn’t seem to be the case when using ‘N-type’ connectors.

HF/VHF/UHF Enclosure

To power the relay, I sourced a 24v DC 3.75amp laptop power supply again on EBay for £12.00. With the relay wired purely for testing, I applied 24v DC to each terminal and everything switched without any fuss. I used a small 50ohm SMA dummy load which was good up to 5GHz and a wideband scan with the spectrum analyser backed up the manufacturer’s claims. Pin 7 on the relay is the common negative feed and each port is positively switched so pretty straightforward stuff so far.

OK time to do some drilling. Six holes were drilled on one side of the enclosure and one for each end for the N-type socket input feed and for the SP17 Weipu connector. I then soldered the terminal on the relay to the SP17 connector and left it loose in order to fully tighten later.

HF/VHF/UHF Enclosure

Now the fiddly bit. I purchase some small lengths of RG142 MIL spec coax which had SMA connectors pre-fitted at each end. Each one was measured and spliced and then soldered to the appropriate output socket. It took quite a bit of time to get things right and to also tighten up each SMA connector accordingly. Even fitting very small spanners around this tiny area proved a somewhat complex affair as each one had to be done in the correct order, otherwise the one next to or underneath could foul.

HF/VHF/UHF Enclosure

As I had wired the switching on this unit exactly the same as the QRO switch, this means I could also use the very same 6 way shack switch box that I had made for the HF/VHF switch. Well at least test the functionality before making a duplicate. I then put the analyser back on and took readings. The splicing of the coax had caused some slight impedance bumps but nothing to worry about below about 600MHz. So from 70cms downwards everything looked pretty well flat. Even up to the 23cms band the SWR was pretty reasonable (below 1.5 to 1) but if you plan to use this on 23cms then you’d need to look closely at any coax splices and be very precise in connecting them to remove any potential impedance bumps.

HF/VHF/UHF Enclosure

All-in-all for about £65, I had a quite neat HF/VHF/UHF 6 port antenna switch. OK it won’t do high power, but at 300w it’s a good all-round unit especially for the great frequency range it covers. More on how it works in the field in the coming months

How to ‘Wideband’ a 3 Element Yagi

From time to time, it’s nice to have a directional antenna that covers most or all of the band you are operating on. Bands such as 17m are very narrow band and there is little point in designing an antenna with a good bandwidth, it’s best to design purely with performance in mind. With a computerized model, that will give you in region of 7.9dbi forward gain and over 25db front to back. Rear side lobes are also pretty well suppressed. You can see what a plot looks like below.

This is a 10m band high gain version which has a 2 to 1 SWR bandwidth of around 1400khz running from 28.000MHz to 29.400MHz. That’s pretty good in most cases, but the 1.5 to 1 SWR edges only give a bandwidth of 850KHz from 28.300Mhz to 29.150Mhz making it in reality quite narrow banded.

10m Band 3 element High Gain Yagi

As soon as the SWR moves above 1.5 to 1, many of today’s modern transceivers will start to complain of a mismatch and begin shutting down, lowering the power output. Of course an external ATU will help the situation but you can design your antenna to be a little more forgiving especially on bands with a wider bandwidth.

So, how do we get our antenna to tune over a wider bandwidth? Let’s take our 3 element Yagi. We are not going to change anything regarding the length of the boom, it’s a reasonably long boom at 3.4m in length, but by moving the driven element forward by 30cms (typical for the 10m band – other bands will differ) towards the director and retuning the reflector element (adding 5 cms on each side) and then a minor adjustment on the driven element (+2 cms on each side), we have an antenna with nice gain (albeit slightly lower than the high-gain version), but with great front to back which covers a bandwidth of over 2.0MHz.

The 2 to 1 SWR limits are now 27.500MHz up to 29.600MHz – that’s 2.1MHz. The usable 1.5 to 1 SWR limits (so you don’t have to use an ATU) are now 28.000MHz to 29.300MHz- that’s 1.3MHz which is great for the CW and SSB portion of the band. If you are an avid 10m FM operator you could even re-tune the driver (take 3cms off each side) which would give you the SSB section and the 10m FM section. In either case, the whole band is covered with an SWR of under 2 to 1.

10m Band 3 element Wideband Yagi

The antenna is presumed to be matched with a hairpin match and the natural impedance of the high gain version is around 33 ohms and the wide-banded model 45 ohms.

One other way to get more bandwidth is to apply the ‘OWA’ (Optimized Wideband Antenna) design which includes an extra element to the array. Adding an extra element close to the driver and in front of it helps to be an impedance controller of the whole array. With the correct design you can build an antenna with a bandwidth exceeding the normal Yagi ‘Wideband’ design by quite a margin.

To give you an idea of the SWR, here’s a run-down of an antenna we built for an 11m operator who was also a ham. He wanted to operate on both 10 and 11m. So we built him a 6 element Yagi that has great gain (over 11dbi) and over 20db front to back and an SWR that barely moves from 26.500MHz all the way to 28.500MHz – that’s a great 2.0Mhz bandwidth
of below 1.3 to 1. The only downside is that there is a small reduction in forward gain when comparing to a standard Yagi design.

So as you can see, there’s more than one way to crack a nut and there’s no ‘right’ or ‘wrong’ way in antenna design – just different methods. Of course a good PC antenna simulation program (such as EZNEC or MMANA-GAL) is a must as you don’t want to just build your own design testing it by trial and error. Computer designs are really the only way to guarantee a design will perform as close to what has been modelled.

Constructing and Matching a Single Element Mono-Band Delta Loop

Matching a single element delta loop is quite a straightforward affair. The user has a couple of options. You can use a gamma match which is really a glorified capacitor to do the matching for you. However, you need to get the correct tube length and use an antenna analyser to get it set up correctly as the internal rod and shorting bar will need setting at the correct positions. It’s a good all-round solution.

You can however make a slightly cheaper matching system by using a ¼ wave length of 75 ohm feeder from the feed point and then join it to any length of 50 ohm feeder which then goes back to the shack. It’s a really simple solution but the Achilles heal is working out the correct length of the 75 ohm quarter-wave section. As its coax, the velocity factor of the ¼ wave line comes into play.

You can second guess as most manufacturers quote the velocity factor (or V/F) of the coax but you will need to know this figure, but beware; coax velocity factors can differ widely between different manufacturers. Foam coax tends to have higher velocity factors (normally above 0.80), as does MIL grade coax like those with a PTFE dielectric. Whether you have an antenna analyser or not, you can work out the correct length quite easily and we’ll show you how below.

No analyser? – working out the length of the ¼ matching section using 75 ohm coax with 0.66 velocity factor (v/f) for the 15m band:
First, find the value of a ‘Full wavelength’ on the frequency you are operating on. Let’s say we are operating on the 15m band (21.230MHz) and require a matching section. We use the following formula.

Wavelength = 299,792,458 (speed of radio waves in air) divided by the frequency. So 299,792,458 divided by 21.230 equals 14.12 meters.
Then multiply 14.12 by 0.66 and you get 9.3324m. Divide by 4 (gives you a quarter) and your total length is 2.333 meters.
You’ll need to know the approximate velocity factor of your coax then you’ll be close. Maybe not ‘spot-on’ but close enough.

No analyser? – working out the length of the ¼ matching section using 75 ohm coax with 0.84 velocity factor for the 20m band:
First, find the value of a ‘Full wavelength’ on the frequency you are operating on. Let’s say we are operating on the 20m band (14.230MHz) and require a matching section. We use the following formula.

Wavelength = 299,792,458 (speed of radio waves in air) divided by the frequency. So 299,792,458 divided by 14.230 equals 21.06 meters.
Then multiply 21.06 by 0.84 and you get 17.696m. Divide by 4 (gives you a quarter) and your total length is 4.424 meters.
You’ll need to know the approximate velocity factor of your coax then you’ll be close. Maybe not ‘spot-on’ but close enough.

If you have access to an antenna analyser (such as an MFJ) – then follow these instructions. The info below shows how to cut stubs of various lengths. We require a ¼ wave stub – so the far end is not connected and left ‘Open’

THE ‘X FACTOR’ – X = ‘Zero’

Here we show you a section of 75 ohm coax hooked up to the analyser. The far end is open (although it does have a PL259 on it). It was trimmed for a quarter-wave at 17m (18.130MHz). You can see that X = ‘0’. You don’t need to take any other of the readings into account. This actual section of coax measured 3.21m which made the velocity factor (v/f) 0.78. The manufacturer claim on their technical data sheet was 0.80 so it was fairly close.

MFJ Analyser - THE X FACTOR

Here’s how to measure using your MFJ Analyser:

Select the first (or opening) measurement mode in the MAIN menu.
Connect the stub under test to the “ANTENNA” connector of the analyser.

Note: The line must be open circuited at the far end for odd multiples of 1/4 wave stubs (i.e. 1/4, 3/4, 1-1/4, etc.) and short circuited for all half-wave stub multiples (like 1/2, 1, 1-1/2, etc.). For our experiment, the far end of you coax should be ‘OPEN’ (i.e. nothing connected to it). Put an PL259 on the other end and connect it to the analyser.
Use the ANTENNA connector’s shield for one lead and its center pin for the other.

Coaxial lines can be in a pile or coil on the floor, it makes no difference. Internal or external power can be used, and the analyser can be placed on or near large metallic objects with no ill effects. Coaxial lines connect normally, with the shield grounded.
When tuning stubs, gradually trim the stub to frequency. Adjust the feed line or stub using the following method:

1.) Determine the desired frequency and theoretical length of the feed line or stub.
2.) Cut the stub about 15 percent longer than calculated, and short the far end for a half-wave (or multiple of a half wave) stub or feed line. Leave the far end open for feed lines or stubs that are 1/4 wavelength or odd multiples of wavelength long.
3.) Measure frequency of lowest resistance and reactance, or lowest impedance. For fine tuning look only at the “X=?” display. Adjust for X=0, or as close as X=0 as possible. The frequency should be about 15% below the desired frequency as we have cut the stub ‘long’ so it can be trimmed.
4.) Take no notice of the SWR or ‘Resistance’ readings. Begin to trim 1-2cms off the coax far end and re-take the reading. You will notice that the ‘X’ (reactance has gone down). Keep trimming until the reactance is as close to ‘0’. When you get close to ‘0’ fine trim at 1cm a time.

You now have a quarter-wavelength stub cut to the velocity factor of the coax at the frequency you are operating at.
Now you have successfully made your quarter-wave matching transformer, check out the diagram below which shows you how to get your loop up and running.
Constructing a mono-band Delta Loop and matching it