Roman

After a few iterations my bellmouth turned out pretty sweet! Had heaps of trouble trying to get any of the dissolvable filament options working though, so I still havent made second part of the pipe. The next thing to do is buy an all alloy radiator so I can chop it up and have a front facing air intake, so no 90 degree bend just before the throttle. In unrelated news I found this 90 degree oil filter adapter which is an absolute bloody gem: https://www.summitracing.com/int/parts/fms-m-6880-a50/overview/make/ford My remote oil filter setup has been an absolute dick punch right from day 1 and I've always battled leaks with it. This uses same threadpitch etc as the Toyota filter so works with 3S or 4AGE or similar. Lush! This will now clear my swaybar so I can get rid of the AN lines and so on. I've been making a bit more progress on my digidash thingy, 3d printed a rear housing that accepts the factory loom plugs: And I thought about using 2 screens side by side like this but in practice all of the writing is too small and it sucks looking at just numbers instead of gauges etc. But either way I'm learning a lot about electronics so I'll keep at it. I've nearly finished writing the code for my cruise control based closed loop ignition trim computer which this all ties in with. Will be interesting to see if it works. Once I've got my mish mash of parts together I'll design a circuit board that incorporates everything and then get a company to etch it or whatever they do. I've been lazy and havent had a wof or reg on this car since summer time. I signed up to a trackday in another 2 months so I've been putting in some work to fix up some shit to get it all sorted again. Cant wait!

Hey cheers. Yeah one problem I found was that even on the least bright screen setting, my screen was still too bright for night time. So darker colours needed for everything. Also at a real life viewing distance, everything was just too small. I was thinking I had heaps of screen real estate and I'd struggle to fill two screens, but in real life I probably need half as much stuff on there. I know what you mean about distracting elements/colors, will take a lot of trial and error to get the right balance of not looking too bland but also being easy to read. In other news. So far I've only been putting all of my code into a single Class, in the "loop" void. However it's become such a lumbering beast that I've now split it all out to seperate functions. Now my main loop is just down to only a few lines of code again, and more represents the logic of the order of things rather than the nuts and bolts. Part of the reason for this, Is that sending and receiving can values works insanely fast, so I will do say 20 read/writes of can data for every 1 write to the screen.

Oooohhh I love a good bar graph. Histograms are great

Theres nothing actually wrong with the factory dash it works very well. But I want to finish this one way or the other as everything im learning is very relevant to work. And it will be the first thing ill have made in a while that combines a lot of different disciplines. (Cad, 3d printing, tuning, canbus, electronics, programming, gui design) so will be satisfying to finish it even if i do end up preferring the standard instruments. The Nextion screen can run fast enough to have analog type gauges that are useful. So ill see what i can cobble together.

Okay so my idea sucks. Reading digital digits while driving is distracting and.... shitty. Back to the drawing board.

So I've been busy working on my dash some more. It was a battle getting the canbus working on the Teensy but figured it out. It's funny how just one simple thing can bloat your code so much. I've been drawing up the layout for my digidash in MSpaint, because, thug life. And you get this super crisp looking font on it that's great. But then when I blank out the text, and put the same sized font into the display editor so the numbers can change in realtime. It looks yuck at pixelated. Well, it turns out that the screens processor isnt smart or grunty enough to anti alias the text to smooth the edges against the background, which is what most PC programs do. Neither the 4Dsystems or Nextion screens can do this. Shitballs. So I was pretty frustrated here as it makes my screen look like shit when I already know it can look so much better with a smoother font.. However since I'm only really changing numbers on the screen, rather than words. (I can just make the titles part of the background image) I thought I'd give it a try by making a small image for each digit 0-9, and then so a 4 digit number would be 4 seperate "Write" commands to the screen, that tells it to bring up image number 9 to show 9. and so on. Previously my code for writing a 3 digit number to the screen was as simple as this: (Where IAT is Intake Air Temperature variable) genie.WriteStr (9, IAT) ; Which is essentially just "Write IAT value to String number 9 on your screen" But because I'm now sending so many more write commands, they are by far the most time consuming thing in my loop of code. I cant write every single value that comes in because it will slow things to a crawl and I lose the advantage of my fast processor. So I need to economize the amount of writes that I am doing to the screen as much as possible. So for starters I am comparing the existing value to the last one, if it's the same then dont bother doing anything else. Then it measures the difference in each digit by subtracting the new number from the old number, decides if it needs to write a new value and then writes that digit. Using the modulo command which is what all of the %10 bits are. All of this is great but what was previously one line of code has now bloated into this for every variable I write to the screen: if (IAT != IAT_old); { IAT_Diff = (IAT - IAT_old); if ( ( (IAT_Diff / 100) % 10) != 0); {genie.WriteObject (GENIE_OBJ_USERIMAGES, 9, ( (IAT / 100) % 10) ) ;} if ( ( (IAT_Diff / 10) % 10) != 0); {genie.WriteObject (GENIE_OBJ_USERIMAGES, 8, ( (IAT / 10) % 10) ) ;} if ( ( (IAT_Diff / 1) % 10) != 0); {genie.WriteObject (GENIE_OBJ_USERIMAGES, 7, ( (IAT / 1) % 10) ) ;} } IAT_old = IAT; Now works pretty sweet! The write to screen time, even when optimized is still about 1000 times longer than anything else in my code though. So might try a Nextion screen as apparently they're faster in every conceivable way. They use a cortex m3 which is the same as the brains of an Arduino Due. Dont know what the 4dsystems screen uses but dont think its very grunty.

+1 Black and Silver. Body coloured engine parts just look out of place.

Hey thanks - That's cool but the Teensy is way gruntier It can have 32 bit variables instead of 8, and it's got a floating point processor so it doesnt take a zillion years to calculate a float. And processor is many times faster. Canbus itself is usually 2 chips (a transciever which converts canbus frames to SPI or similar, and one that acts as a filter for messages) My main dash unit / ignition trim computer thingy will be trying to do a lot of stuff at once, so faster the better.

Okay soooooo Arduino Due. Its canbus can work between two Arduino Dues, but you need a transceiver chip to make it work with a "proper" canbus. So bought some transceivers, which apon arrival I realised they are the wrong ones. D'oh. So rethinking it some more, although I like the fast speed of the Due, and the fact that it's got a native canbus thats easier to write for. Eventually I want to add some other canbus devices in addition to ECU and dash, in which case I dont need or want the bulk of a Mega sized arduino board. So thats the next problem, if I write a bunch of code for a Due, and then want to add a smaller arduino device I'd need to rewrite all the code rather than just chop snips out of it to make it work. So from here. I found out about the Teensy. Teensy V3.6 is 180mhz 32 bit and basically trumps the Due in every conceivable way while being exceptionally tiny! And it's arduino sketch compatible. https://www.sparkfun.com/products/14057 It's also got 2 canbus lines and can use this exceptionally small double transducer. https://www.tindie.com/products/Fusion/dual-can-bus-adapter-for-teensy-35-36/ So since it's small enough and not incredibly cost prohibitive I'll probably make everything on my bus use a teensy + that transceiver for sake of keeping the code easy. Then 3d print some housings to suit. Now that I understand things a bit better I dont need to use any of the shields I had. Should be nice and compact for whatever they end up doing.

Agree with that 100% Look up virtual dyno. It's great for quantifying differences from road tuning if you can minimize variables. I had a situation where I gained 20hp and heeeaaps of midrange power and the car felt slower.

You definitely don't need a dyno especially for a low power car. As you can just nang it up the open road and get some good results because you aren't accelerating too fast. Also it's really easy to find the timing plateau. Also, it's immensely satisfying and heaps of fun. In my dyno session I gained nothing over my road tune.

8% more power is another way to think about it... Which isnt bad when you stop to think that it pays for itself over time with the economy difference.

Off topic to the above, but. My cars exhaust noise is really sensitive to ignition timing. Like if I dont have enough timing its way louder and tinny sounding. And also the oxy sensor readings dont seem right when the engine is cold. Like it will say 16:1 when it is chokingly rich at the exhaust pipe. So for ages its always been much louder when its cold, but I put it down to either an exhaust leak that closes up once hot, or... something. But so thought I'd experiment by adding some ignition timing when cold. Annnnddd car is now way quieter when cold! And the air fuel ratios showed closer to what seems accurate, now shows at about 13:1 So I reduced my cold fuel enrichments until I was at my goal AFR again, and now my fuel usage when cold has decreased: I'm gonna call that a win! I think its because when the engine is cold the fuel and air just hasnt mixed very well at all... Lots of the fuel will still be droplets that dont start evaporating until ignition has started. So it takes way longer to burn properly (relatively speaking) Which explains both why it was louder and why oxy sensor readings werent right. As there is still uncombusted air and fuel exiting the exhaust port if you dont have enough timing. I've always had on my mind that you need extra fuel for cold conditions but never thought about ignition at all.

Success! Edges need trimming but you get the idea. Pva glue over plaster was the good combo. Then a little wax for good measure. Carbon sleeve is the best! The inside is super smooth and tgecoutside feels like crocodile skin or something. The biaxial weave is really cool. Gonna order a bunch of different sizes I think. This one is probably a tad flimsy with only one layer of sleeve but two will be sweet with a 6k weave I think. Or maybe just go 12k single. But it's probably lighter than the bolts that would hold it in place. Only about 25 more prototypes and I'll have a usable part haha

I think pva glue might be the answer for sealing the mold. Its water soluble, fairly thick so doesnt massively soak in, doesn't strengthen the plaster too much, super smooth finish, non toxic and cheaaap. And can wax / pva release agent over the top too. Will give it a go.

Also my latest CF related abomination I thought it would be an interesting idea to try make some molds for a destructable inner core that I could easily make a bunch of. So firstly drew up a trumpet shape that I wanted to make (omitted the return curves for simplicity for starters) Then made a 2 piece mold following the inner surface shape and added some locating tabs 3d printed it like so and filled it up with plaster It took a few goes to figure out how to avoid bubbles in the finish and blah blah, but got a nice result eventually. So the great thing about the plaster is that just a quick whisk of sand paper and you've gotten rid of any layer lines from the 3d printing. Get a super smooth finish which is awesome. But the big problem is that its massively porous. So I thought giving it a coat of epoxy would give it a nice glossy finish that the final part would seperate nicely from. So did that, left it for a day and it ended up with a really nice glossy finish. So PVA'd over this (No wax which was a mistake. mistakes were made. I'm stupid. etc) Then bunged some CF sleeve over the top and some normal matting at the bottom to cover over where it's thin. Then vacuum bagged this, with wet layup. Had some problems with the bag not being able to tighten fully up to the part, as I didnt allow enough bag size essentially. So I had to fiddle with a it a bit which ruined the outer finish and creased part of it. Bums. Annnnddd then my second problem. Coating the plaster with epoxy to seal it creates this rediculously hard outer layer that is almost impossible to smash haha. Damnit. Underneath the parts I've managed to break/pick off, it's got a super nice finish where it was up against the glossy epoxy finish on the plaster. The PVA seems to be doing its job to some extent as when you pick some of it off there's a clean seperation. But it just seems mechnically locked in place since theres still a curve all of the way around sticking to the part, and both the part and the plaster are stupidly rigid. But I'm really happy with what the internal finish looks like where it was up against the mold. I dont have the patience to get this entire part unstuck, and there's no point because the outer of it looks yuck anyway. So I'll try find a better way of making that center core part and try again. I've ordered some PVA filament which is dissolvable, so I could 3d print a core that just washes out after. But it takes ages to print, need to sand the layer lines every time, and is expensive. So if I can make plaster or something similar work (any ideas for another something similar to use?) I'll stick with that. But if the plaster needs epoxy or similar every time, PVA might just be more cost effective as well as a lot less messy. (It's stupidly good fun pouring plaster and then seeing something cool come out of a mold though) I think my next iteration of this piece will be good. My resin infusion is working awesomely now so I'll do that again for the next iteration. It's much easier to sort out any of the bagging issues when the bag doesnt stick to the part because it's already wet with epoxy. TL;DR: Lots of time and materials, have produced nothing useful since previous posts, still have hope for non shit parts in future

Haha this cracked me up http://www.motoiq.com/MagazineArticles/ID/1704/PageID/2400/Project-Miatabusa-Part-43--The-Intake-Manifold.aspx

Shit these things are cool.

Hahahaha awesome. I'm at the stage of "Why the fuck did I put all those houses there and what do I do from here" then look at my code for 10 mins and go do something else hahaha. Need to start again with the canbus code for the arduino Due which is waaayyy simpler. Too many 1/10th finished projects at the moment haha.

Yeah it seems counter intuitive but with epoxy it starts going off in the pot way quicker than when its on a part. Because it builds up heat which causes the epoxy around it to build up more heat and it gets in an exponential cycle. Can start smoking and melt your pot if its plastic! but when laid out in a thin layer the part its on helps cool it and it stops this process. Like your pot might last 15 mins or less but your pieces of cloth could sit there with epoxy brushed on for hours before even starting to cure. So if you get the epoxy onto the cuts of cloth asap and then there isnt really any time pressure anymore. And it gives the epoxy a bit of time to let off some of the gas that can cause bubbles in your surface finish.

Which ever side is up against the mold is what comes out nice. So if you lay up against the light directly you'll get a nice looking inside finish that no one will ever see haha. +1 to just waxing / PVA the shit out of your headlight lense and then make a mold, then layup CF into this. There are a few videos posted earlier in this thread on Youtube, by Easy Composites I think its called. Watch all their stuff its awesome! I've found the best way to avoid bubbles when wet layup is I bought some big shallow plastic bins, and I'll wet layup the sheets (already cut to shape) and then let them sit for 5-10 mins. Then drape them onto the part. And get CF with a fine weave or it frays like crazy and it's bloody hard to do anything with.

Well I havent made any more useful parts but I've made a lot of test pieces that have ended up in the bin. I tried another resin infusion but using a larger diameter feed hose and with a resin that looks a bit thinner. This flowed across the part a little quicker but then still had poor absorbtion into the part. So tried again using peel ply instead of the perforated plastic sheet between the CF and the flow mesh. (This was probably half my issue) I read a few posts where people had success by heating the resin as its viscosity massively drops with some heat. But, then how do you avoid making your pot life too short... Answer is to heat the part and keep the feed pot cool. So got my part up to maybe 35 degrees by sitting it on the heated bed of my 3d printer with a heat lamp on top too. And then holy shit, the speed at which the resin flows through the part was amazing. Was almost like water. And, the resin penetrated right through 4 layers of tight weave CF! Awesome. Apart from one corner which was a bit dry because it wasnt very well in the flow path. Because I was just using a hose in, hose out. Rather than spiral tube or similar to disperse the entry and exit points a bit. So I'm confident that my next part will be fully saturated with resin if I use spiral tube. The next part I try will be hopefully be a "keeper" so I'll let it cure properly before unbagging haha. But need to make a more rigid mold now that I'll be using vacuum bagging rather than my initial plan. Which was laying up around a dissolvable core which I havent been able to print successfully. Also, the spray adhesive stuff is an absolute lifesaver for trying to dry lay up parts in a meaningful fasion. Jesus. Night and day difference. Hopefully soon I'll have something worthwhile to show for my efforts. Oh, and using PVA instead of the wax and the surface finish no longer looks white-ish anymore. Awesome.

Keep at it man! Have always wanted a 2002 so I'm living vicariously through your progress.

Pump goes down to about 8kpa. Yeah ive got a big degassing chamber in line with the pump so resin wont reach it. Thanks ill try again with more resin. I wasnt using peel ply but the perforated blue plastic stuff thats supposed to let resin through. Then that flow mesh stuff on top of that. The pump pulls down to 8kpa with bag on it when the resin line is shut off.

Actually here's a fun fact about the nomenclanture of the M series engines, and why the numbers arent sequential as you'd expect. While still in prototype phase as a planned obsolescence engine, Toyota Engineers described each iteration based on how many critical flaws in engine design there are. 4Mistakes 5Mistakes 6Mistakes 7Mistakes They later dropped the idea and went back to making reliable engines.