Mr Bravo, bit of brain picking if you dont mind, any good contacts you could suggest/recomend to laser cut these locking tab washers? the origional steel measures .7mm but .8mm would be fine, finish could be plain/zintec/galv doesnt really matter, the scan should be actual size
Moss Motors has various sizes for various cars. But, not sure if I have asked before, why not just use locktite? Tabs were used before chemical technology was advanced enough to hold bolts in place. Or alternately individual washers, the yoke is there for ease of manufacturing and installation.
Mr Bravo, bit of brain picking if you dont mind, any good contacts you could suggest/recomend to laser cut these locking tab washers? the origional steel measures .7mm but .8mm would be fine, finish could be plain/zintec/galv doesnt really matter, the scan should be actual size Have you got a .dxf or .dwg drawing of it? Most laser cutting places won’t even quote without it. Also,laser cutting is about volume. You ask for five of those,you are going to be hit with a minimum order charge of about £60. If you aren’t in too much of a hurry,I can add them to my next order. If you don’t have a dxf drawing,I can sort that for you too…
Unfortunately I dont have the capability to provide a dxf or dwg but if you could do something when time allows that would be much appreciated, I can send you an origional part if that helps? PM incoming
“Can you make us some Hetzer lights? They have glass lenses. “ How hard can it be? 🤣 So started on the easy bit first. Turned a tool for the housings Then got the housings metal spun out of 1.2mm steel sheet Milled the little threaded inserts that get brazed to the inside of the housings, as they were too small to get lasercut
Then it was off into the great unknown, as all I know about glass is beer is stored in it….
So moving on, I managed to borrow an original light to take measurements from, and made a mold out of stainless steel Then I managed to find a dental kiln, spares and repair but with the added bonus of a spare element Ripped out all the obsolete controls and rebuilt it with a WiFi enabled programmable controller with live data logging. More on that later.
Then seeing I am environmentally friendly, I didn’t go out and buy an EV, but rather I did something constructive. Everyone talks about these dangerous greenhouse gases, well , that’s simple…let’s destroy greenhouses and recycle the glass
Then it was either learn how to cut glass or improvise….
Wasn’t about to take glasscutting lessons, that’s proper witchcraft and goes against my ethos of if you cannot do it on a milling machine or a lathe, it’s probably not worth doing. So, quick removal of greenhouse foliage from the glass and onto the milling machine Diamond hole saw makes quick work of the task. Prying out the blank is not for the faint of heart, but having cut about ten blanks and not breaking any,it’s not as dodgy as it feels 🤣 Once you get over the initial trepidation,it’s plain sailing and you realise glass isn’t as fragile as you sometimes think it is
In the meantime I cobbled up a new controller. Most of which came from the “ Might come in useful one day cupboard Even had some really nice Industrial 16A connectors courtesy of a company that is constantly finding new ways to kill its existing client base… So it was all ready for a test run. Mold had been sprayed with Boron Nitride in the meantime. Tool only just fits in the little kiln🤣 Initial programming done on the IPad via Wi-Fi Repurposed case was from a dosing pump ,was still waiting for a stainless steel plate to cover up where a keypad had been previously Typical, the blank off plate turned up after the job was finished 🤣
So, let the games begin. To start off with it was a simple program with no glass in the kiln, just to check if the controller was working.Initial test program on the iPhone but it soon became apparent for proper programming a larger format, Ipad ,was going to be a much better option. A very basic program, ramp up from ambient temperature to 300c in 30 minutes and hold temp there for an hour. This already started to highlight issues All along I was looking at the massive overshoot in temperature, this was going to be a serious issue. The controller works basically on fuzzy logic, with PID style control. It should have been able to keep this temperature within +-2c of target, and yet it was nearly 65c hotter at 300c, in a seemingly state of thermal run away. It did try and reign it in after the overshoot, but it was fighting a losing battle. The strange massive drops on the graph near the end of the run also confused me at the time. I didn’t realise just how significant this data log was actually going to become later,and not for the data that I initially was very concerned with. This data log will return later. It was the beginning of the Easter weekend, so I would not be able to liaise with the controllers manufacturer, so I did what any normal person would do, switched it all off rather disappointed and went and drank a beer or three. This definitely wasn’t the start I had envisioned. I was seriously missing something here 😕
Glass is funny stuff. It’s still liquid even when it’s solid. Seem like a bit of an odd statement? Easy enough to prove if your inner nerd is willing 🤣. If you ever come across an old single pane window( non double or triple glazed,but they would show the same characteristics,)try this out. Mark which side was the top. Actually, an old greenhouse is ideal for this,and you have easy access to multiple panes, to prove it’s not a fluke. Take the glass out and measure the thickness. Gaurenteed as God made little green apples, the side of the glass that was at the bottom will always be marginally thicker. This phenomenon is called slumping, and it’s the principle upon which I will form the glass into the shape I need.
So, a bit of a science lesson. Glass doesn’t like shock. Physical or thermal. Both cause it to shatter. So, when trying to make something out of it, you need to be mindful of both. It’s a lot more robust than you think, and once you stop pampering it with kid gloves, you seem to break less 🤣
So, to be able to reform this skanky greenhouse glass into a thing of beauty, there will be a couple of distinct phases it will have to go through.Skip to the pictures now if your bored…oops no pictures yet 🤣🤣
First phase is to warm the glass up to a temperature around 500c in a controlled manner. At this temperature it’s pretty inert and you are kind of out the danger zone at the minute. The rate at which you heat it,is important . Too slow and you waste electricity. Too fast and it breaks. Very conflicting information on the internet here, many saying do not heat up at over 150c per hour, others saying that’s a load of tosh, don’t exceed 250c per hour. Everyone’s experience is different. In hindsight I can see both arguments.
Once up to 500c you can turn the wick up. Serious ramp up to the next temperature, pretty much throw as many electrons at it as you can and get it to the next temperature as quick as you can. For my application it was going to be somewhere between 650-675c to get the glass to slump into the mold. Determined by trial and error. You then hold this temperature for a certain time,to allow the glass to slump. Again, trail and error.
Once that’s done you need a steady ramp down to 550c. It needs to soak at this temperature for awhile. This is akin to stress relieving or normalising a piece of metal. It’s to get rid of the stresses you have just induced into the part by subjecting it to what you have. If you don’t do this stage correctly, your part will shatter as it cools down
Next stage is ramping down again. This time to around 425c. Again holding it for a time. This anneals the glass and stops it being brittle. Again, if not done right, your part shatters as it cools.
Then it’s the last stage, cooling down to ambient temperature. You have to resist the temptation to open the kiln at anywhere under 400c because the thermal shock of the cold air rushing into the kiln is enough to shatter your part .
So , if you followed all the steps correctly, you should land up with a program with a fair few features, with three distinct plateau’s
Drank more beer and analysed the initial data some more…
Impressive work there, hope you don’t mind if I offer a few thoughts. Those oscillations look to me like a derivative error in the PID values, the huge dip looks like derivative kick, did you alter the set point at that time?
If you can change the P, I and D values independently in the software that would help. If the controller has the ability to self-learn as suggested it might even tune itself eventually but you may need to disable the derivative function when you make a set point change. I seem to recall something to do with loop back. You probably know this but those oscillations should be reducing in amplitude by about 50% on every swing, the graph looks like it just went back to proportional control before the big drop. Plus you have an offset from the target which is I & D combined but mostly I.
It’s a while since I did PID tuning but that’s what my thought was when I saw the graphs. I spent the first 10 years of my working life in instrumentation and control for the C.E.G.B on power stations, this brought back fond memories of spending a whole night shift watching a chart recorder!
Discussion is always most welcome, so don’t be shy 🤣 It’s been a steep learning curve, but have learnt a lot. Not being able to get a definitive answer straight away from the manufacturer was probably a blessing in disguise. I’ve written most of the next post already, and the issues you brought up are touched on. Nice to know from someone with real life industry experience that I was on the right track🤣🤣. Electrics and electronics are not my forte, would classify myself as a semi skilled bodger when it comes to electrons 🤣🤣
Glad to help. I’m a consultant now in a slightly different field (IT/AV) but I always stand by the statement that ‘you don’t know what you don’t know’. PID control systems is something that I got drilled into me for five years of my apprenticeship. On the plus side I had an expense account, free accommodation, travel allowance and generous wages to compensate so it wasn’t that bad. The source of all knowledge is Instrumentation Systems by E.B. Jones, and Instrument Technology also by E.B. Jones. These are known as the ‘Jones Books’ and cover everything you would ever need. They were published in the 80’s but are still valid today. I’m a bit rusty on all of this but if you get stuck post on here and I’ll see if I can help out. PID tuning is part maths and part intuition, your kiln is a very simple closed loop system so should be mostly maths as you don’t have too many variables to deal with.
So the famous data log and what I took it to mean… Having thought about it a bit, I interpreted that data log as follows:
Given the stupid ramp up time of well over 600c per hour, it’s highly unlikely there would be any parameters that could be applied by the controller. Coupled to it being a tiny kiln, with a massive heating element with exceptionally good insulation, no controller was going to hold this in check. So the overshoot highlighted with the yellow arrow was attributable to poor programming rather than a controller poorly configured.
Given the tiny kiln/ huge heating element/ exceptional insulation, the is no way there was temperature fluctuations of the magnitude outlined by the red circle. I subsequently put this down to a temporary loss of Wi-Fi signal in the garage and it being a glitch in the live data log. Having put a Wi-Fi range extender thingymabob in the garage since, no further anomalies of this nature have been recorded.
Having bought this controller as a programmable kiln controller specifically for glass and pottery firing kilns, I also surmised that the PID settings undoubtedly would have been tweaked for an infinitely bigger kiln.( I subsequently found out it’s configured for a 16l kiln and calibration is done at 800c) .I decided there was no point in changing any of the PID settings yet until I had a couple of full data logs from full firing runs, because at this stage,the kiln had not been over 400c.
So taking a bit of care I started off with this program For the eagle eyed, you can see that contrary to what I originally said that you ramp up as fast as you can from 550c, I toned it down a bit in the hope of the controller would have a sporting chance of attaining the slumping temperature. The slumping temperature and time was deliberately kept on the low side to see what would happen.
Kiln loaded and programme started And now we wait….
Nope, I couldn’t wait either 🤣🤣 So it had worked, it wasn’t fully formed but it wasn’t broken( when I took it out the kiln. It did crack about ten seconds later though,but that was to be expected as the kiln was still over 300c when I opened it) So buoyed on by initial success I added another 10c to the slumping temperature and 15minutes more soak time , and tried again. The controller was within 4c over the programmed temp and dropping 8c below target. I put this down to it being calibrated at 125c higher and it being in a kiln about a fifth of the size of what it was calibrated for. Whilst not ideal, I don’t think it was going to be overly detrimental.
