The actual greenhouse construction has been going on for some time now (myself again more in the assistant roles). I must admit that there are times when I am getting tired of the entire project, even if I am not even carrying the main burden. There are several issues in the Juliana greenhouse kit that make things more complex, confusing and difficult than they need to be. The instructions leaflet is one thing: it might be that all the aluminum parts are indeed painted black, but it is not optimal to print everything as black in the instructions with small details and bars with complex profiles. There is too much guesswork in the construction now. There are mysterious gaps in the instructions, where you just need to make a guess how to get from phase A to C, and trial and error is not good in system that is put together with soft, aluminium nuts and bolts – fastening and loosening them just a couple of times can lead to threads of nuts breaking and bolts getting stuck. Also, if there is a gap between parts in one part of the aluminium framework, it is often very hard to figure out where the source of problem is, when there are dozens of parts that connect to each other like a giant ‘himmeli’. – But, we are making slow progress, pictures below.
The foundations for the actual greenhouse are now ready, I think. The last steps in this phase included e.g.: using fine gravel and sand (“kivituhka”) to create a top layer that was at right angle and height to serve as the basis for concrete tiles that form the final top layer, supporting the steel plinth. My old heavy wooden ram (“juntta”) was again in use, manually banged down to compress the earth layers. (My academic hand muscles are not perhaps best suited for this kind of work, btw, I noticed the following day…) Before installation of the greenhouse plinth (the steel base), there was one important extra item to take care for, however. The greenhouse was being constructed near – almost underneath – a large old birch, which had already started to rot and drop branches. It was now the very last moment to say goodbye to her (cutting down a large tree later, with delicate glass house underneath, would had been much harder, or impossible). We hired a couple of professinal loggers come and cut the tree into pieces, and there was much work left also to us in sorting out branches (some the size and thickness of regular trees), twigs and leaves. After a full day of that lumberjack work, there was finally time to cast the concrete (oh yes, and a mixture of sand and mortar was brushed into gaps between tiles and dampened to fill them, as a final touch), but before that, the last round of measurements and fine-tuning was done to adjust and fix the steel plinth into the right positions. We had not managed to get all plastic pipes into same exact depth, and a Dremel tool was used to cut few millimeters away from the top of three of the protruding cast pipes. Bubble level was also much in use, and I also applied a couple of spanners to open all the corners of steel plinth, apply small wooden wedges underneath the plinth in a couple of places, and then fix them again, hopefully sealing everything into correct width and length measurements, while also keeping the entire construction at even level, using the bubble level tool. Then it was time to mix some concrete – a small power tool was used, but mostly manual blending (we did not have a proper concrete mixer). It was getting really dark at this point, but finally all the pipes were filled to the brim with concrete (there was beautiful full moon rising, I noticed, and the fireworks from Tampere Venetian Festivals were making popping noices somewhere far away). But: while the doing the control measurements we saw that while the X and Y directions were correct, the steel pinth was not direct – the thin steel bars were tilted (in Z, upwards direction)! It was pretty hard to solve the problem at this point, in dark and middle of the night. But the trick that did it was to loosen again all the nuts and bolts in the corners, move away the concrete tiles that we had used to support and fix the plinth into the right position, as some of them pressed the inner edges of the steel bars inwards so that all of them became tilted. Fixing all that required some fast action, as concrete was already slowly solidifying while all this was going on. While we managed to do all that, there was one measurement that I forgot to take. In the morning, I wanted to kick myself: the front part of Juliana’s steel plinth consists of three separate elements, and I had forgot to check that they were fixed so that the width of plinth in front was the same as the width of plinth at the back (there was room for adjusting the bolts in their elongated holes). We had carefully measured that the diagonals from corner to corner were identical, but forgot to measure that width of back and front were identical. And now all six steel pegs were fixed deep into concrete (with their tips bended, to prevent them from slipping ever out of concrete). Luckily, there was only c. 2 millimeter error between front and back, so the situation was finally not so bad. Hopefully the actual Juliana greenhouse construction has some adjustment room when the aluminium and glass parts will build and fixed on top of the steel plinth.
At this point, we could almost take a little break. But there was still some chainsaw work to do and log piles to build, before thunderstorms arrived to Tampere in the Sunday afternoon. So, maybe a bit later.
The method of building the foundation for our greenhouse has proved to be a bit more demanding than perhaps some others – hopefully the final outcome will be worth the extra labour. The tricky part is casting concrete into long pipes, that go down into 80 cm depth – i.e. beneath the soil frost line. These concrete pipes, or pillars, need to be cast so that the the steel plinth (base) that will provide the basis for the actual greenhouse will be exactly square shaped in horizontal X & Y directions, while simultaneously also exactly at right even level (Z) so that there will be no tensions into the glasses of the finished greenhouse. The actual freezing protection will be provided by several layers of gravel and sand, separated by filter fabric layers, plus a couple of layers of Finnfoam (extruded polystyrene foam – XPS – thermal insulation). The tricky part is that those pipes I mentioned need to go through all of those other layers, and they need to be made and fixed in pretty much their correct, final positions first, before any other elements of this foundation have yet been built. From the photos underneath you might get an idea how our ambitions have met with the reality so far.
Oh, and did I forget to mention that there of course need to be also underground drain pipes, so that all rain water will not turn that big hole in the clay earth into a swimming pool? And that those pipes need to be installed at the right angle of bank.
The building project of this summer has been to make a greenhouse – for my chilli peppers, as well as other vegetables for the family. Construction of proper foundation for a rather sensitive small building that will mostly consist of sheets of glass and thin bars of aluminium is important; however, July was mostly rainy and we also were travelling a lot, so most of the digging was still to be done in early August. Now, in mid-August, the hole in the ground is almost deep enough (in Finland earth can freeze in quite deep during winters, and our garden is on top of several meters of clay, which expands when it freezes; thus – a lot of showel-work). Since our backyard is a bit on the small side, there was a narrow spot where the entire construction had to fit into. But the hole is now there. Have I mentioned that going down into solid clay is somewhat heavy digging?
In other news, our pick for the greenhouse kit manufacturer was Juliana: http://juliana.com/en/products/juliana/
I have also already got the watering system (that I have also already used during our travels for my chillies), from Blumat.
Next steps will probably involve some concrete and a lot of gravel. And a shovel and a wheelbarrow.
(To be continued.)
Continuing to troubleshoot our persistent home networking problems: while we have got a high pile of various routers, last several years the heart of the network has been Asus RT-N56U dual-band model, which was awarded as the fastest router available in 2011. It was a slim device, and after I updated the firmware to the Padawan version, there was also more than enough room for tweaking. However, the constant connection failures and speed dropping finally pointed towards the life-cycle of our router coming to an end. The router has been located in very narrow space, without cooling, so it should not have come as a surprise that its components have started failing after a few years.
The selection of a router for a home where there are a fair number of connected devices (smartphones, tablets and computers are just one part of them) is a tricky business. I wanted to have a 802.11ac model, but otherwise I kept on reading and comparing various options. According to specs, speed and configuration options, the current top model of Asus, RT-AC87U, was for a long time my number one choice. However, the actual user reports were a rather mixed bag: there seems to have been various bugs and issues with both the software and hardware of this, 4×4 antenna configuration, dual band ac model. And I have come to learn that I have less and less time and patience for tweaking tech — or at least, I want the router and network infrastructure to “just work”, so that I can use the Internet while tweaking, testing and playing with something more interesting.
The conclusion was to get yet another Apple product, this time AirPort Time Capsule (2 TB model). It does not reach quite as extreme speeds as the RT-AC87U, but then again, there is limited support for hardware that is capable of reaching its theoretical 1,3 Gbps top speeds. I am increasingly relying on my Macbook Pro Retina also when at home, and we are actively using several iPads and other Apple devices, so having the full Apple compatibility, while not a “must”, was a nice bonus. The user reports about the new AirPort Time Capsule have been overwhelmingly positive, emphasising its robust reliability, so I am interested to see whether this router lives up to its reputation also as the backbone of our household. So far, so good. All our devices have succesfully got online, and the speeds are close to the 100/10 Mbps maximum, also in Wifi, when close to the AirPort. And the Macbook is now making its automatic backups in the background, which is nice.
My Raspberry Pi had arrived while I was at the Microsoft Research Faculty Summit, and I got finally some hours to test drive it. As far as contemporary PC hardware goes, RPi is of course seriously underpowered little plaything. On the other hand, when you compare it with to some other devices (like smartphones, embedded systems), it does not look so bad. The principal reason for its development should also be taken into account (promoting computer literacy, encouraging tinkering with hardware and software tools, helping kids learn to code). I have been looking for some time for an affordable and functional HTPC system for serving media in our living room, and thus my first test drive involved setting up RPi as a media center PC. The Raspian “wheezy” distro that they recommend on the Raspberry Pi Foundation website was too slow and unresponsive for my taste to do anything. I tried also Raspbmc version of XBMC media center, but I could not get it to install any addons at all. So finally I did find a place that instructed how to install OpenElec, an embedded operating system that has been built to run XBMC – from a Windows PC (http://www.squirrelhosting.co.uk/hosting-blog/hosting-blog-info.php?id=9). Now XBMC was getting online, updating itself and installing addons nicely. It also booted up decently in c. 20-40 seconds.
It turned out that the major issue for me finally was a network infrastructure related one: we did not have a LAN socket in the corner where our TV set is situated. I tried to learn about WiFi USB dongles that could run out of the box, plug-and-play style with the OpenElec/XBMC, but it would had been necessary to know the exact version of chipset and firmware to make sure whether the USB dongle in question would work, so I decided to stay with the wired Internet/Ethernet connection instead, and added another layer to the (rather instesting) network topology of our home by setting up a Powerline Ethernet bridge (using two Zyxel PLA4215 units). While I was at it, I also got a powered USB 2.0 hub (a basic Belkin thing) and wireless keyboard-touchpad combo for comfortable sofa-based media surfing. The latter was a Logitech Wireless Touch Keyboard K400, which is a rattling, plastic thing, but has two important benefits for me: (a) it is cheap, (b) it has an inconspicuous power switch hidden on the side. Anyone with one or two (or, indeed, three) hyperactive toddlers in the house can witness why these are good things. I have already e.g. a broken Logitech diNovo Edge lying around somewhere. Surprisingly, everything seemed to work after a couple of system reboots.
As to the actual use of the OpenElec/XBMC/Raspberry Pi system, I have not yet much experience to share. I can say that the software is still buggy and occasionally rather slow. It is difficult to say what the system is doing when the playback or a menu does not open immediately, whether it is buffering data or whatever is going on. Attempting to stop the playback of a HD video file can suddenly jam the whole system to a complete halt. But yes, I can play music, videos and watch photos in a full HD screen from multiple sources, from both local network and from various online services in a more or less satisfactory manner. There seems to be much potential and room to explore further in this surprising little system. One can only hope that the energy of the community does not die out, but the development of software continues far beyond this early stage. It is, after all, really early in the evolution of Raspberry Pi ecosystem, as some developers have not yet even received the unit they are waiting for. Much of the OS distributions and applications are thus more at ‘alpha’ rather than even ‘beta’ stage at this point. But taken that, this is really entertaining little playground to experiment with, and to fool around.