Replacing a rotator at GM4FVM

In brief, I tilted over the mast, took off the Yaesu G-450 which I had been using, installed a new G-1000 model, screwed everything back together, raised the mast and after calibration and installing a new USB controller, everything worked. 

OR you can read the long winded version (or just look at the pictures) next:--- 

Here at GM4FVM we have two masts and thus two rotators. On the main mast is (was) a Yaesu G-450 rotator. 

At present on that mast are three antennas. First is the two band (two feed) DUAL four element on 50MHz and five element on 70MHz. Above that comes an 18 element I0JXX yagi for 432MHz and above that is an eight element I0JXX for 144MHz. Since I recently moved the 70cm antenna to a position between the other two antennas I have become a bit concerned that this may have added too much load for the G-450, especially in this windy area.

The G-450 looking tired

Another concern I had was that the G-450 actually comes from an unknown background. I bought it in a silent key sale along with various other things (like the IC-7100 which I still use daily). I had no plan to use it, thinking it might give up parts. However, when my own G-450 failed I just put up this one and gave it no more thought. I do not know how old it is, or what work it has been doing before I got it.

And finally, I wanted to get more accuracy when it comes to directing the antennas. The G-450 has been operated using a Hy-gain YRC-3 controller. This gives me antenna control from the computer screen, and allows automatic moon tracking and so forth. The Hy-gain controller is fine for use on the lower bands, but it was not really accurate enough for the narrow beam width of the 70cm antenna. It would do for terrestrial work, but for moon bounce it cannot reach the accuracy I need (or what I think I need).

I have had computer control for a long time, originally using a controller kit from EA4TX and a Yaesu G-600 rotator. That arrangement went back about 20 years and the G-450s and the Hy-gain controller arrived about ten years ago. In fact the G-450 had been a step backwards in specification compared to the old G-600 and recently I began to worry if the G-450 was no longer up to the job of turning more, larger, antennas I now have.

So I began to think of a new stronger rotator. In the end I decided on the Yaesu G-1000. Compared to the G-450 this offered 33% more braking torque, up to nearly 100% more turning torque (depending on rotation speed) and 100% more weight capacity. The K factor, the overall measure of load performance, is 130% greater and the wind loading is 120% greater than the G-450. Overall this looked like a sensible step up to take.

The G-450 alongside the G-1000

The G-1000 has a DC motor while my older G-450 uses an AC motor. Hy-gain YRC-3 is switchable between AC and DC so it would drive the G-1000, but I wanted to step up to more accurate positioning while retaining computer control. The G-1000 controller has a six pin DIN socket on the back which allows the use of a Yaesu outboard box to connect to a computer via a USB socket.

The snag with the Yaesu USB box is that it costs £400+. After looking at a couple of alternatives I chose an ERC-Duo USB box which is often available in the UK for about £120. As it was out of stock when I needed one I ordered one from Wimo in Germany. The ERC box would allow me to interface between the G-1000 controller and PST-Rotator running on my computer and allow me full control and tracking.

When a suitably still day arrived to change over the rotators I hatched a plan to replace them without taking the antennas off their supporting pole. This would be tricky, given the weight involved but I hoped to support the pole from underneath. This would take the pressure off the rotator when the mast was tilted over, and allow me to change over. In the event it worked well. It would have worked even better if I had managed to get every aspect of the plan correct.

Mast tilted with pole supported and G-450 uncoupled from the pole

Once I had got everything in the right place I could uncouple the G-450 and remove it. Then by undoing the bolts on the upper bearing I could move the pole up the rotator cage and leave enough room to get the G-1000 into the space.

Along with the rotator in the SK sale came a spare Yaesu rotator cable. This had been cut and mended in three places but has proved to be of much use down the years. I used it to test and calibrate the G-1000 in the shack before the day when the change over was to be done. Whilst the calibration might be slightly different with my normal cable it would be close enough. In fact it proved to be within about two degrees. What went wrong here was that I set the rotator to 110 degrees when I always set my antennas to 100 degrees. Good plan which worked in almost all ways save one silly mistake.

I was able to slot the new rotator into the gap and lower the pole into the top bracket. Then there was the tricky job of tightening the bolts on the rotator which Yaesu insist has to be done in strict order in quarter turns at a time. This is good advice as I have snapped alloy fitting before with these rotators. Do not over-tighten!

G-1000 installed

That done I could lower the pole with the rotator (and antennas!) attached, do up the bolts on the base and start re-aligning the bearing at the top of the rotator cage. I was aware that this bearing had become slightly out of kilter over the years and so it needed to be done anyway.

With all that complete there were still a few jobs to do. Yaesu recommend testing and I did this first using the Hi-gain controller and the spare cable. I took this out to the mast and did various tests ensuring the everything was rotating freely. This was done with the mast still tilted but with the support removed, which meant that the antennas could rotate through 360 degrees.

Next task was to do some final tests from the shack using the G-1000 controller. These proved that only minor calibration corrections were needed, but everything was 8 degrees off target. That was because I had set it up in advance at 110 degrees. My rule of thumb is that antenna booms parallel to the ground equals 100 degrees. Doh! But it would have been 2 degrees off on the other side anyway so that was not so much of a problem.

This time I did a careful job, checking the "parallel to the ground" theory with a spirit level. I set each antenna carefully and then raised the mast vertically. Outside job done.

Careful twiddling got the G-1000 controller spot on. The G-1000 controller is much more accurate in its alignment than the G-450 was as it allows five different settings to be juggled. Once this was working OK, it was time to align the ERC-Duo USB box.

ERC-Duo box with 50 pence piece for scale

In contrast with the big heavy Yaesu controller, the ERC-Duo box is tiny and draws its current from the USB cable. This is basically another controller which gives its instructions to the Yaesu controller. Thus it needs to be calibrated separately. It comes with a link to download and install a "Service tool" and a "Rotator Control" piece of software. I used the service tool to do the alignment and save the data onto the computer. I will not be using the rotator control as the ERC-Duo can take instructions from PST-Rotator.

Final result.

All done now. I managed quite easily to get the Yaesu controller and the ERC-Duo to agree on the bearings and for that to line up with my key points viewed from outside. That needed an eight degree correction, easily done by moving the Yaesu pointer and setting an eight degree offset on PST-Rotator. 

I find it more accurate to use this "parallel to the ground = 100 degrees" alignment than trying to find North - eccentric I know, but practical. Obviously this angle will be unique to every tilting mast, but once you have worked it out it will stay true. On my other mast it is 98 degrees so the original and extension walls of our house are not exactly parallel ....

The only slight issue just now is that when I transmit on 2m the pointer on PSTRotator moves about four degrees clockwise. Just the pointer, the antennas and the pointer on the G-1000 controller do not move. This is RF is getting into the ERC-Duo leads somewhere, and I am busy trying to solve that minor problem. 

The new rotator moves quietly and quite quickly - the rotation speed is variable. It certainly behaves much better than the G-450 ever did. I am happier that this can take the strain of those three antennas, even though I am already wondering if the 432MHz beam would not be better on the other mast !!!! 

But why, oh why, do Yaesu keep selling rotators with huge expensive mechanical pointers when a simple digital display would do just as well? And why do they charge £400+ for a USB box when all rotator controllers should have USB connections as standard. SPID rotators have their failings, but at least they have no moving parts to break down and they connect easily to a computer.

73 Jim

GM4FVM

The lazy man's way to update Icom clocks

Why do I bother to set up the clocks on Icom radios? Once the battery runs down, why not just leave them as they are? 

HINT:- Icom make a software app that does all this for me!

Well, this goes beyond just looking at the radio and telling the time. Updating the clock (and the calendar at the same time) puts a date stamp on everything you record onto the SD card. 

For instance it dates and times recordings of various tests recently with Pom, DG7AC. Pom responded to my request for Ionoscatter/Q65 tests and we did some for several days. During these tests I recorded what I was receiving using my IC-7300 and IC-9700. If the clock is correct then the timestamp on the recordings allows me to work out afterwards which recording relates to which test. Without that everything is recorded with the wrong date and time.

No doubt there are all sorts of DStar things which need the time set too, but I do not know what those are (of course). 

When the rigs arrive they come with factory fitted lithium batteries to hold the time and date in the memory. These batteries are soldered in and are located in a pretty inaccessible place.  Over a period of a couple of years these batteries become exhausted and the clocks revert to zero every time the radio is turned off and back on again.

I see one of radio supplier in the UK is offering to replace the battery with a battery holder for a standard CR2023 battery holder which is located in an easily reached location under the rig's covers. Inclusive of transport for the radio down to them this costs £125.00. That would be three radios for me (IC7300, 9700, 7100) at that price, £375. Downside, you still need to go in to change the battery and reset the time when it does run out and this is not an Icom approved alteration. Hmmm.

£375, that is more than I paid for a good HF transceiver [That was 50 years ago Jim] .

On the other hand, I could just install Icom's free ST-4003W software tool which sets the rig time to your computer's time whenever you want. Sure, you lose the setting every time you turn the radio off, but it is easy to set it when you start everything up each day.

So I downloaded the software tool which you can find here. The instructions are simple. If like me you use data modes like FT8 you already have the rigs linked up to USB sockets on your computer. This software tool works with Windows and a wide range of recent Icom radios.

My IC-7100 taking a little longer to update its calendar and clock

Once I had downloaded it and installed it I could update the IC-9700 and IC-7300 in seconds. The IC-7100 takes up to a minute while it "sets the seconds". For some reason I do not understand the IC-7100 then holds the date and time for several days.

GB3NGI beacon switching from JT4 to CW seen from GM4FVM - note the time

The file in this case is recorded on the SD card as 20260823_143100, giving you date and time information.

This blog does not seem to upload audio recordings but I would share these too if I can find a way to do it.

I simply saved three tabs to my taskbar, one named for each rig. Then I just click on them in turn and all three clocks set themselves. You cannot do this when data mode software is using the USB connection, but I do it first when I turn on the computer. Once the clock is updated the Icom software releases the USB link and you can start WSJT or whatever you use. If you only have one rig then this is even simpler, just pin that Icom shortcut to the task bar, or run it from the start screen.

So, you get a nice correct clock on your rig. It will show your local computer time, not UTC or something else, but that is easy for me to correct as I convert between local and UTC all the time. 

It is a fiddle to do every day when I start the rigs. For me the warm feeling of having saved £375 quickly deflects my mind from that problem.

EDIT: Thanks to Carlo, IU1KGS, for emailing me with "An even lazier approach to update Icom clocks".

Carlo points out that it is possible for the Icom IC-9700, among other modern Icom rigs, to directly update their clocks via an internet time server. It seems this is covered in the advanced manual. My own 9700 is not networked so I need to try this out before I comment further but this seems even  more straight forward. However, for my older IC-7100 and 7300 I will need to stick to the Icom software (though the IC-7300 Mark 2 is covered by this).  

I appreciate that news Carlo and I hope to be posting various things to do with the IC-9700 later. I also agree with your comment that using soldered batteries in radios is crazy when every PC maker can fit a button battery holder and get longer life with a CR-2032. So why are we short changed in this way? On the other hand, it gives us something to think about. 

73 Jim

GM4FVM 

23cm Aircraft Scatter - Beaming at aircraft rather than along a direct path and the benefits of Q65.

The basic concept of aircraft scatter (AS) is similar to that of radar. However, in the case of aircraft scatter we are generally relying on signals being scattered forwards to a distant receiving station, rather than back towards the originating station.

This posting is about a series of tests carried out to establish the best way of targeting the distant stations using narrow beamwidth 23cm antennas and which data (or other) mode to use for this purpose. It is an expanded and updated version of my article which appeared in the GMDX group's Digest (gmdx.org.uk).

As usual, click on images to enlarge them if necessary. 

--------------------

In practice radio amateurs can use AS on the VHF and UHF bands to achieve QSOs up to a distance of about 600km. The exact distance will depend on the frequency, the height of the aircraft and the curvature of the Earth. This range can be exceeded by tropo ducts, but these are fairly rare. In the absence of a duct aircraft scatter is the most reliable way of reaching stations in the 250-600km range. The larger (or “heavier”) the aircraft, the better chance you have.

I recently started a series of trials on 1296MHz with Gordon, GM4OAS. Our stations are broadly similar, with Gordon running 250W and two 36 element yagis. I have 150W and one 36 element yagi. The path is 257km in length which might make us think that a tropo contact in normal flat conditions would be possible, but the terrain here means this would not be easy. The path crosses close to Ben Nevis and many other Highland peaks in the Munro Tops category. It starts close to sea level at each end and reaches over 750m several times on the way.

The mountainous path from GM4FVM to GM4OAS makes UHF tropo contacts difficult (AirScout) 

Despite many attempts we have failed to make a tropo contact between us and we are certain that any signals between us rely on aircraft scatter. 

AirScout

Starting out on using AS the immediate problem is not knowing where the aircraft are, and thus not knowing when to call in the hope that an aircraft is passing between you and the station you wish to work. To solve this dilemma there are several sites on the internet offering maps of overflying aircraft using their transmissions on the 1090MHz band to capture the height, direction and speed of the aircraft plus a host of other details.

Many amateurs use AirScout software by DL2ALF. AirScout uses reception of the 1090MHz transmissions to produce a map showing the aircraft. It integrates this with the location details of the station which you wish to work. Thus, when you enter the station you wish to contact it adjusts continuously to present a diagram of the path involved, the terrain, and any aircraft along the path which are at the desired height. (See reference [1] for a comprehensive description of the technique and the software). I have been using Airscout for several years.

After several AS contacts, initially using a single 36el yagi at each end, Gordon achieved dramatically improved results by deviating his yagi direction away from the strict direct path between us and towards specific aircraft instead. This involved using an ability AirScout has to directly communicate with rotator control software.

Up to this point I had used AirScout in standalone mode, setting the beam direction to the distant station manually. It seemed to me to be impossible to follow a fast-moving aircraft with the narrow beamwidth of a microwave yagi. I was not aware that AirScout has the capacity to communicate with the antenna control software in my computer to allow tracking individual planes - planes which could potentially be some distance away from the direct path.

Aircraft Tracking
Tracking as a concept will be familiar to many VHF+ enthusiasts. Many of us use a similar set-up to automatically follow the Moon across the sky while attempting “moonbounce” QSOs. However, the idea of tracking individual aircraft was new to me. Obviously, tracking the plane makes more sense than pointing the antenna direct at the distant station. It is the plane which reflects the signal, and it should be the obvious target. However, planes travel more quickly across the sky than the Moon does! Gordon explained to me the steps I would need to take to set up aircraft tracking and after a few false starts I got it working. I had not known that I had to upgrade PST Rotator (Software for Antenna Rotators) to the VHF/UHF version first. See reference [2].

My rotator for the 23cm antenna is a SPID RAU. The control unit for this rotator either has a USB or a parallel socket (in my case, parallel) to allow connection to a computer. The software I use in the computer to drive the SPID is “PST Rotator”. When tracking is engaged PST Rotator looks for a dump file in AirScout which contains the latest instructions from AirScout to point in a particular direction. AirScout knows which plane to follow because the user identifies it by clicking on targets from amongst the options presented. 

AirScout turns the aircraft image red if it appears to be likely to cross the direct path – otherwise they are grey. In the past I only considered red targets, and only when they were near the direct path (Airscout also indicates when they are very close to the path). Once we started targeting specific aircraft we mostly selected images which are grey and not very close to the path, the opposite to what we did before.

Without having to consider how all this works via all the different interactions etc, I can now click on the image of a plane on the map, and the software continues to find out where it is and points the antenna towards it as it moves relative to me. Just to assure me that AirScout is keeping everything under control, it continuously shows a compass bearing towards the plane which it is following, and I can see the bearing changing regularly on the rotator controller too. I can also hear the SPID turning through the shack wall (!). Eventually the plane goes out of range and the tracking stops.

AirScout showing an aircraft North of the direct path to GM4OAS which was suitable with tracking. Note that none of the planes are near enough to the direct path to be coloured red and they would not have been considered candidates without tracking.

The result of changing from beaming along the direct path to tracking the aircraft was remarkable. QSOs were longer and signals stronger. We had three or four contacts for each plane, allowing repeated QSOs to check the outcome. 

It has to be said that this path to GM4OAS possibly favours this approach. On long paths, perhaps over 400km, the difference between the beam heading for a usable plane and the direct path will be smaller. For long paths planes in some directions would be below the radio horizon if far off the direct path.

While it was not easy to calculate the exact distance away from the direct path we were able to use, we estimated it as in the range of 25 to 40km. This brought many more planes into consideration. Once we were both using tracking, we had a series of near-continuous contacts over a period of two hours, selecting likely planes as they arrived on the screen.

Somewhat chaotic screenshot taken during the tests. Two aircraft following each other off path provided time for repeated QSOs. The two traces can be seen on the screen.

Why use tracking? 

It would of course be possible to estimate the position of the aircraft and simply point the beam manually in that direction. The plane could then be followed by dead reckoning using the map and constantly updating the rotator via the controller. This would require a lot of guesswork and antenna repositioning. As a halfway house, the on-screen compass bearing could be used to follow the plane whilst controlling the rotator by hand. However, as AirScout could do the whole thing automatically that seemed like the best way to proceed given that aircraft scatter contacts require full concentration on their own.

Compass shown on AirScout while tracking an aircraft

We used tracking not just because it takes all the hard work out of keeping the beam pointed at the aircraft (but it does that nicely). It also finds the correct angle far more accurately than would be possible chasing the plane manually. I do not yet have antenna elevation, but it would set that automatically too. And finally it solves the biggest problem - it can keep pace with the fast moving planes.

Which mode to use? 

For the tests we started by using FT8, which failed. We then switched to Q65 version 15B. Q65 is particularly good at handling Doppler shift created by the motion of the aircraft. A 15 second tx period is short enough to complete a QSO during brief flyovers. We did not just use Q65 because I like it (though I do like it). We then tested Q65 against FT8, FT4 and CW, using aircraft tracking. 

We became familiar with the plane passes which worked with Q65. Over a number of these passes FT8 was a failure.

Typical failure to make contact using FT8, with zero decodes reported at either end of the path. Note the exaggerated Doppler shift common to all these tests.

FT8 failed for several reasons. It is less sensitive than Q65 - its coding makes it less suitable for scatter use as it cannot deal with what Doppler there is - and its longer 30 second periods allows for twice as much Doppler shift in the first place. 

We then tried a comparison with FT4, hoping that its shorter period might help. However, FT4 is even less sensitive that FT8, so the benefits of the shorter period were largely lost. Still, over a period FT4 give roughly one decode for every five decodes using Q65 on comparable plane passes. 

This is not an ideal way to rank these three modes as it did not directly compare them at the same time but instead measured decodes per plane pass. However, with a fairly large sample size I think it gave a representative result. Broadly we found the FT8 did not work at all whilst FT4 worked to some extent but not as well as 15 second Q65. These results were what you would expect given that Q65 is designed for scatter propagation and FT8 and FT4 are not. 

As an experiment we also exchanged callsigns on CW.  Later we tried to complete a QSO on CW while using tracking and this was successful. The signals were quite strong so this should succeed but the effect of Doppler was rather off-putting. At some stages the Doppler shift meant that the CW station fell outside the receiver's CW filter as the QSO progressed. 

We did not try SSB, which is often used for AS contacts during contests. Whilst tracking increased the duration and strength of reception, it also increased the potential for Doppler shift. Whether this would cause problems for SSB contacts is something which needs to be investigated. It may be that at 432MHz, where the shift would be smaller, there could be useful gains to be had.

Doppler shifts of 400 to 700hz over four minutes were common and this is difficult to deal with in most modes except Q65 and similar protocols specifically designed for the purpose.

Summary
Tracking the aircraft rather than beaming at the distant station produced improved results in the tests. This can be done automatically, which was something I had not considered. It is a method which I will be keen to try more often in future. Provided AirScout can access a reliable feed of 1090MHz plane data (something which it sometimes fails to do) this approach seems to have merit.

Q65 seems to offer benefits over other data modes and CW.

Thanks to Gordon GM4OAS for guidance and encouragement with this experiment.

References: 
[1] http://airscout.eu/index.php. The latest issue of the AirScout Aircraft Scatter Prediction Manual may be downloaded from here: http://airscout.eu/downloads/AirScout_V1_4_0_0_EN.pdf
[2] Information about PstRotator – Software for Antenna Rotators may be found here: https://www.qsl.net/yo3dmu/index_Page346.htm