Month: September 2026

All drilled out and all put back again, but at last my Savannah’s new screen and cabin roof plastic are in. All that’s left is some finishing off – white silicone sealer around the base of the outside of the screen and a small amount of black acrylic at its foot at the back on the inside. It wasn’t all plain sailing as you’d expect with the rivets having been drilled out for at least the second time in the life of the aircraft but there’s little to show for that and it’s all gone back pretty well. Only I will notice the blemishes.

I’ve also got some new plastic trim to go along the top edge of the instrument panel. The rubber trim that’s on there was only supposed to have been temporary but it’s lasted for ten years. It’s also had the annoying habit of coming off and dropping down at the most inconvenient moments while in flight and hopefully what I’ve got to replace it will be a lot more ‘grippy’. I’ll put it on at the same time as I do the screen sealing which may be tomorrow, I don’t know, as I’ve got an important appointment in Perigueux so it and the sealing may have to wait.

I’m so pleased and relieved to have come to the end of the hard work. Compared to the past few weeks I’m now in the position of having resolved all my computer problems and once I’ve finished off the Savannah’s screen I’ll be able to turn my attention back to the outside of my house and my work in my garden. And who knows, I might even be able to get a few flights in.

I topped the Savannah’s tanks up in March at what turned out to be the end of a long spell of calm weather. The weather then unexpectedly changed the next day and then we had the heatwave during which not only was flying during the day impossible but also in the mornings and evenings due to the high temperatures. We’re still suffering from it to a certain extent. When I left Malbec at 6.00 pm my car temperature gauge was still showing 29 degrees C in the shade.

Below are some shots that I took of my Savannah after I’d cleared up and vacuumed the aircraft out again. Now its exterior needs yet another good wash and I’ll also have to find time, as I did before, to put a dab of white paint on every new pop rivet. And also as the first shot shows, I think that I the last time and the person before me incorrectly fitted the wing root trims. I’ve now fixed them under the front edge of the roof panel which I think is correct because the inner edges of both trims have a sharp corner which I think should be hidden. This is supported by the corners of the roof panel, which are now exposed, being rounded. I’ll have to tidy them up now though as they’re a bit scuffed and dirty looking.

As Mike Tyson said, as did Joe Louis before him, ‘Everyone has a plan until you get hit in the face’. My plan for today was to finish fitting my Savannah’s new screen but that went by the board. For the past few days I’ve been arguing the toss with the supplier of my Isward lawn robot. As I mentioned previously, I’m now on my third machine after a mechanical failure with machine 1 and a software failure with machine 2 (the latter is now in my tool shed for ‘spare parts’ with a complete set of brand new accessories – charging station, charger and GPS beacon) and the current machine has been dogged by an irritating problem.

It has been refusing to dock as it should either mid-task in order to recharge its battery or at the end when it has finished the job. Instead it has returned to a position in front of the charging station and then aimlessly started slowly spinning around before finally coming to a complete halt. It has then been dropping off-line and if left it has then remained stationary until its battery has become totally depleted. This problem became my top priority because if I can’t get the robot working reliably again my lawns, which is how I laughingly describe them, will get out of control again, especially if we start to get more rain as the Autumn progresses, in which case I’ll have to start mowing them again with my ride-on instead of using the time to do more productive things.

I’ve shared photos and videos with the supplier who, if you ask me, has tried to take the easy way out ie by blaming me. He spotted that I have built a canopy over the charging station to protect it and the machine from the weather and he has suggested that it is preventing the machine’s camera from getting a clear view of the docking station’s position. I said no, it didn’t stop the other two machines. He then suggested that part of the canopy’s top cover was obscuring the ‘target’ on the rear of the charging station that shows the machine precisely where to dock but I said no it wasn’t, because neither of the other two machines were affected by it. So he then said that probably it was to do with the ambient light level. I replied funny that, because an evening or two before, the machine was finishing a mowing task in semi-darkness with its headlights on and had succeeded in docking just once having just prior turned its lights off.

I suggested that there were actually two more plausible explanations for the machine’s behaviour. The first was that when the machine successfully docked it had not long been recharged and was returning with a still highly charged battery as it was finishing off mowing just the last small area of grass. I’d noticed than when pausing to recharge mid task, this machine was doing so at a lower battery level than the other two machines and that it might therefore be running too long and returning to the charging station with a battery that was too low for it to then dock. The supplier denied that this could be the case. I therefore proposed the second explanation, which was down to ‘mapping’.

Old generation lawn robots use a physical wire either pegged to the ground or buried just below the surface in order to define the boundaries of the area to be mowed. RTK GPS machines like mine don’t do that and instead use precise GPS coordinates which are obtained by ‘driving’ the machine around the boundaries in record mode and saving a ‘map’ in the machine’s memory. I did this for machine 1 and although the supplier suggested copying the maps over from machine 1 into machine 2 when I received the latter in fact I did complete remaps from scratch for my two areas in back and front. But I didn’t do that for machine 3, which is giving all the problems. Instead as instructed by the supplier I’d copied the maps over.

I suggested that in my opinion the problem was that although the machine knew where the area boundaries were and followed them exactly it just had no idea where the charging station was. The supplier poo-pooed this but said that all I should do to make sure was enter mapping mode and reset the charging station’s position. I did this several times and after one of them the machine went berserk, missed the charging station by a mile and began climbing up the slope towards the rear of my house. The supplier told me to move the charging station. I said not likely as it had taken me long enough to create the present position and anyway, it worked fine for the other machines. I said that no, when I had the time I’d do a total remap from scratch of the boundaries, charging station position and no-go areas (trees and septic tank upstands in the lawn) of the larger lawn at the rear of my house and see what the machine did then.

And today was the day. And guess what? The machine performed impeccably, even docking mid-task to recharge with a battery level significantly lower than for either of the other machines. I told the supplier that copying maps from one machine to another using their established procedure obviously doesn’t work. I also told them, a bit tongue in cheek, that apparently I’m beginning to understand their machines better than they do but have not yet received a reply. I’ll be interested to see what they say when they do. So my original plan for today, to finish fitting my Savannah’s new screen, crashed and burned but hopefully I’ll be able to do that tomorrow. Unless something else turns up to disrupt it before then. 😕

One of my problems is that because my To-Do list is so long I invariably end up trying to do too much. Today I had to spend much of the morning arguing over an issue with my lawn robot with the Chinese manufacturer then I had to pick up a new computer disk drive from Mondial Relay and then I had to start cloning an existing hard disk onto it. In between I had to grab a quick bite for lunch and all the while my Savannah cabin top panel was waiting in my car to be taken to Malbec and refitted. Oh, and before that I had to pop rivet into it a retaining cross bar for the new transparent plastic panel that I cut and sealed in a day or so ago.

So it wasn’t until well into the afternoon that I was able to get over to Malbec and begin the work on the Savannah, starting with giving it a thorough vacuuming out to remove all of the pieces of pop rivets and aluminium swarf that had fallen into it and the accumulated dirt and grit on its floors. After that I was at last able to get onto the serious job of refitting the cabin roof panel. It’s not as easy as it sounds. There are lots of pop rivet holes to be exactly aligned so it’s not just a matter of banging pop- rivets into them. And things also have to be done in order.

Later on in the afternoon I was at the point when up to four layers of material needed to be lined up before they could be pop riveted. This is more tricky than it sounds and any mistake could lead to pop rivets needing to be drilled out again and even damage to the items that were being riveted. So I thought that slow and steady is the way to go and and this would be a good time to stop. Tomorrow I’ll be able to return afresh and work my way through but for now I’d rather have everything all good and waiting for that so I can finish the job off without any hitches. I’ve burned myself in the past and don’t want to do so again.

My Savannah’s new screen is officially in. The screen itself is perfectly in position and all of its securing pop rivets are in. And that includes the two troublesome ones up in the wing root cavity. It’s hard to see in the fist image below, the one on the left which was difficult to get out because the pop rivet pin snapped in the rivet head did exactly the same again, which is annoying, but at least hopefully I won’t be the next person who’ll need to drill it out next time. The one on the right was actually the more difficult to get in but as shown in the second image it ended up looking the better of the two.

All that’s left now for the screen is to apply some white silicone sealer around its base on the outside and a dab of black sealer in its centre between the two vertical tubes on the inside. So tomorrow I can get cracking on the cabin top panel. I’ve got plenty of both white silicone and acrylic sealer for the screen and top transparent panel but I’ll probably use silicone as I did last time as it’s lasted about 10 years. So it would be nice if I could get the top panel totally done and refitted tomorrow. However I tend to procrastinate even more than usual on Sundays so I doubt that I will. 😉

The parcel that I expected was delivered around lunch time so although I wasn’t able to get over to Malbec until this afternoon I managed to get a reasonable amount done on my Savannah’s new screen. The screen is now basically fully installed. The interior and exterior side trims are also now fitted except for those annoying pop rivets at the very top inside the wing root cavities and they would have been except the cut down 3mm drill bit I was using in the angle drive broke.

I’ll need to cut down another for tomorrow but this time I’ll make it a 3.2mm. So when I deal with those pop rivets tomorrow it’ll then be time to start on the roof top panel and replace the transparent polycarbonate window in it. It’s not that bad but as its weather sealing is broken as a result of removing the panel it would be silly not to put new plastic in to make it as clean and bright as the main screen, which as shown in the third image below is now excellent once more.

It’s not even the beginning of the end. But it is the end of the beginning, to paraphrase a very famous gentleman in history. I’m talking about my Savannah’s new screen. But before I get onto that, something else that’s been winding me up. I didn’t get to bed until 01.30 am after being preoccupied with yet another ongoing computer problem. Right now the cost of memory chips has gone through the roof, the reason being apparently, that supply cannot meet the demand of AI businesses and datacentres. Whereas in the past computer memory costs have fallen even as they have become faster now the opposite is happening.

The cost of the solid state disk drive that I installed in my current machine that I built in 2022 has doubled from just under 200€ to just under 400€ which is putting such drives out of the reach of many amateur builders like myself. The Chinese have been quick to notice this and all of a sudden, not because of philanthropy on their part I would imagine, many low cost solid state disk drives and RAM sticks have been arriving on the market. Knowing that I could return them for full refunds, in recent weeks I have purchased three inexpensive solid state disk drives. Two of them have failed immediately on loading them with data and the first failed a few days ago after having been installed about a month ago.

All have been refunded to me in full but my experience can’t be coincidence and the Chinese supplier(s) must know what’s going on. So why do they do it? Do they just calculate that not every customer who purchases the junk will send them back? That’s the only reason I can think of. I’m awaiting delivery of yet another more expensive SSD from China and I’ll be interested to see how that fares, but if it fails it’ll also be sent straight back.

So what kept me up last night? Because of my experience with the three SSD’s that I’ve mentioned, because I’m now in desperate need of computer back up storage I decided to go for a more conventional external USB disk drive. It arrived yesterday and during the afternoon I began copying data onto it, a lot of data because that’s how you find out if it will fall over or not. And it did. I persevered until 01.30 am because I knew that I had to get restarted on the Savannah’s screen today and needed to arrange the return of the disk drive before doing so. I managed to do that but it meant I didn’t get to Malbec until after lunch.

I did OK in the time available but wasn’t able to finish the job off. It’s something that you can’t rush because that’s how you mess it up. I started by removing the protective film that was still on the outside of the screen and placing it in position to get a feel for its fit, which wasn’t bad. So then I was able to fit it much more firmly using two wood clamps, one in the doorway on each side. The clamps are the perfect tool for that and once the screen was firmly in place I could see that the problem I had with crazing of the old screen around the two pop rivets that had secured it to the two vertical front struts would this time be academic.

The reason for this was that although it can’t be shown in photographs, the screen was standing some way off the struts and as the dimensions of this screen are not much different from the old one, it’s no wonder that the stress of riveting it to the struts eventually led to stress damage. This time I won’t be doing it, so I plugged the holes with pop rivets and gave the struts a refreshing light spray with matt black paint as now they’ll be more in view from both outside and inside the cabin.

So after that I pop riveted the screen lightly in position and called it a day. I found afterwards that although the passenger side of the screen is perfectly in place, the captain’s side will need a bit of adjustment to pull it down. That’ll involve drilling out a couple of pop rivets but it won’t be difficult and any spare holes will be hidden under cover plates anyway. But that’s for tomorrow, hopefully. I’ve received notice that a parcel I ordered is being delivered tomorrow and I just hope that I won’t be waiting in all day for it and wasting what would otherwise be yet more valuable time. This being France, I just have a feeling… 😕

Tell you what though… take a look at the final photograph below. It wasn’t until I’d taken the old screen out that I realised just how brown and cloudy it had become. The new screen will be almost like having nothing there. 🙂

For the first time in several weeks last night I went to bed a happy man. No, not for that reason… The reason was that I at last managed to crack the many computer problems that have been dogging me. In a previous post I mentioned that they all started when one of the secondary disk drives in my main computer crashed and I lost a considerable mount of important data, including the engine hour records for my Savannah. I managed to get just that one file back but much other data was still lost.

Then on top of all that, while I was trying to deal with the lost data problem the new CPU that I fitted in my computer by way of an upgrade began to run incredibly hot – dangerously so, at 100 degrees C which is its maximum before it shuts down before damage occurs, so that was another problem that had to be solved. I surmised that the root cause was a failure of the CPU water cooling system that I’d installed when I built the computer back in 2022 so went ahead and purchased a replacement from Amazon.

Fitting it meant dismantling virtually the whole machine as I had to extract the motherboard in order to fit a new cooler mounting plate and when I’d finished and had the machine back in one piece it failed to restart. It was totally and completely dead. Now Amazon have a great little scheme for various products, especially for many computer components, whereby they offer ‘second hand’ or ‘good as new’ items at reduced prices. If you read between the lines these are often brand new, unused items whose packaging has been damaged in the warehouse and can represent really good bargains.

On the other hand, they could be items that have been returned by previous buyers who have damaged them and sent them back for a refund without declaring the damage and often there can be no practical way for Amazon to test them before offering them for resale at a reduced price. My replacement cooler was one such item which left me in a complete quandary. My initial reaction for my computer’s failing to restart was to put the problem down to the cooler. It’s wiring is quite complicated and I thought that maybe a previous buyer had screwed it up and shorted the cooler out, but how could I test it?

But it got worse. I then tried replacing the new cooler with the old one which I knew was probably faulty as it was allowing the computer’s CPU to heat up, but it did at least allow the machine to boot to a running system while it was doing so. The result was a disaster. Even with the old cooler installed my machine now remained totally dead and I began to wonder whether I had unknowingly done something stupid while installing the new cooler which had possibly shorted something out and destroyed the motherboard. But, once again, how could I test it?

I’d already packed the new cooler up and requested a return and refund when I had a last desperate idea. I have what was a brand new motherboard which I think I foolishly damaged but which I knew powered up but did not fully boot into a running system. I wondered what would happen if I swapped the motherboards and installed the new cooler on the faulty motherboard. Now it just so happened that late yesterday evening I’d been watching a video on Youtube by someone who had purchased a totally dead computer on Facebook Marketplace for peanuts and got it running within minutes as a result of one simple measure, one that you learn about when you first start messing around with computers but which I’d forgotten and haven’t applied for donkeys years. It’s called resetting the CMOS.

CMOS stands for Complementary Metal-Oxide-Semiconductor and in computer terms it’s a small memory chip on the motherboard which is powered by an on-board battery. The chip contains the motherboard settings which are stored in the factory but which can be subsequently modified either intentionally by the user or inadvertently by events. This hit me like a kick in the teeth. Resetting the CMOS is done by shorting across two pins on the motherboard and only takes 10-15 seconds. And I hadn’t done it.

So at around midnight last night I stripped apart my complete machine yet again, reset the CMOS on the ‘damaged’ motherboard which I knew would probably power up but not run and installed that with the new cooler that I’d had to unpack again. And bingo! Sure enough it and the new cooler powered up but sadly, the machine would still not fully boot. However, it told me that the new cooler was perfectly OK, so what about if I reset the CMOS on my previously working motherboard which had been running but now appeared to be dead and would not now start and reinstalled it with the new cooler?

And know what, it sprang into life. So there I was at about 1.00 am putting everything back together and then putting the machine through its paces. It was brilliant. It ran like a Swiss watch, totally silent and with its new CPU and an additional stick of RAM (memory) faster than I’ve ever known it. How crazy is that, after all the problems I’ve had. The images below show the machine running. The cooler fans change colour, which is the latest fad among ‘gamers’ but most importantly, the temperature of the CPU is being kept incredibly low by the new cooler, at not much above room temperature, which is amazing.

But on top of that I have more good news which is the cherry on the cake. While all this has been going on I’ve also been putting my ‘backup’ machine together and trying to recover the data I lost using two pieces of specialised software, Disk Drill and Recuva. The good news is that I’ve managed to recover ALL of the key data from the two disk drives involved, one that was internally damaged and the other which I must have inadvertently wiped during the data recovery process.

And there’s also a possibility that I may be able to recover even more, although I’m now not too bothered as what’s left is only old data that I was ‘hoarding’ for old-times sake. So now I can get on with reassembling the recovered data on the two machines, my main computer and the ‘backup’, so I can get back onto other pressing matters. Like fitting the new screen in my Savannah and the work in my garden. Happy days at last!

Today I was supposed to be getting straight back onto fitting the Savannah’s new screen but first I had to finish a task that I started yesterday, namely an attempt to recover the data lost on the computer disk drive that had crashed. I was using special software that crawls very slowly over a drive seeking each little piece of lost data and reassembling them file by file. It takes forever and cannot work on a disk drive, or parts of a drive, that have been physically damaged but it’s worth a try.

Initial results looked very encouraging because the software could ‘see’ almost everything that had been lost, but seeing it and being able to lift it off the drive are two different things. In the event the only file of real value that I managed to recover was the Excel spreadsheet of the Savannah’s hours. Don’t get me wrong. If I’d had to choose just one file this was it, but I’d hoped that I’d have got more. Nevertheless, I feel blessed and relieved that I’ve managed to recover this one.

I had some other things to do and by the time I’d done those I have to admit that I wasn’t motivated to then go to Malbec to work on the Savannah’s screen. Instead, later in the afternoon I thought I’d go and start my excavator up and raise its boom which slowly sinks as the hydraulic oil in the system passes the seals. That’s quite normal but previously the boom dropped far enough for the end of the boom to contact the tarp wall covering of the excavator’s garage and if I hadn’t noticed it would have ended up splitting it, so now I raise the boom on a regular basis to prevent that happening.

And then I had an idea. I thought that as the excavator’s engine needed to be run anyway I’d tackle something that I’ve been thinking about for a little while. Looking from my house down to the bottom of my back garden the ground slopes downwards to where the front of my tool store/workshop will eventually be and then it immediately begins to slowly slope upwards again. This means that the ground at what will be the back of the structure will be higher than at the front and as the concrete base on which it will be built will be flat the ground on either side will have to be progressively shaved down from front to back and on the same level from side to side at the back.

It’s difficult to see in photographs but is easiest with the excavator garage in the picture. As both the dumper and the excavator needed to be started up and run and I need the practice, I decided to get them both out of the garage and tackle the side further from the garage. I had to be careful because the distance between the hole I’ve already dug out and the fence is pretty narrow and the ground is dry and soft without much support for the excavator’s weight. I’d swapped the buckets over to the wider, smooth edged one weeks (months?) ago before the heatwave but hadn’t used it for grading the ground. I also I have nothing like the skills of a pro but I’m quite satisfied with the results of half an hour’s work, as the photographs below show.

I’ll need to get to the back via the side closer to the garage but I feel pretty confident that it won’t take too much to get the ground around the hole level enough. Before pouring concrete the hole will receive a sub-base of stone and ‘remblais’ that will be pounded down with a compactor and wooden shuttering will then be put in place. Once the concrete has been poured any gap between the concrete and the ground after the shuttering has been removed will then be back-filled but before then I have to complete the work on my Savannah and also prepare the ground around my house for some other works that are planned. So much to do.