How about blueprints? https://www.formulastudent.de/pr/news/details/article/pats-corner-design-reports/ --- Post updated --- There is a lot of data from these cars, even more blueprints etc are available, just look for formula student and it will be ok as long as engine is combustion engine. --- Post updated --- About minimum weight, you kinda can go really low, but if it is not drivable and does not handle, that will loose points again, I think that will add possibility to go different routes to best, necessarily going quite bit over minimum weight is not bad at all. Also I think that if thing explodes in crash, then that is a fail, like if real formula student would be unsafe it is flagged out of competition, so instead of setting weight to some range, I would like to set goals and it is up to modder how he/she gets to that goal. That would then allow creativity and different strategies, which in my mind at least should stimulate creative minds to test their true skills. But say if this sounds insane plan, it is not set to stone yet --- Post updated --- Tested how badly track would lag, I got a good 26fps, 150 physics objects might be bit much, so cones will be static objects, only few might be physics objects.
That's a good point Some people would be very good at just tuning a basic car, but may not have any real modding ability, so a basic vehicle that could just be tuned would probably be a good thing to have for something like this. It would be a bit of a limit for them, but perhaps they could have their own section in this competition (unless they end up being just as competitive as the fully custom ones, then in that case you might as well lump everyone together) So that's what a normal car looks like... how far would be be allowed to stray from that design (dimensions, suspension, tire sizes, so on and so forth) No limit on weight ehh? You might end up regretting that if I can get it as low as I used to in the "Low Weight Competition" we were holding on the RoR forums Um... could you just use the new course creator to make a course? Might save you some time... and a lot of CPU usage
It really is up to you how you design your vehicle, those uni designs are quite close to optimal though, but you can try whatever you figure out. Has to be open wheeler though, following FSAE rules, but not too tight as some things are pretty hard to pull of in the game. If you can find some way to get weight really low, using pigeon engine, vanilla tires or tires provided to you (that is one point that needs testing too) no rockets, thrusters, while making car be fast around the track, then go for it, but that 200kg lower limit probably stays, might increase that bit higher though, but 200kg already is quite unobtainable with BeamNG wheels and pigeon engine, you can't use FLT_MAX either as that will have quite bit of penalty on points This is bit of trouble right now as rules are not 100% formed yet, but I think main lines are pretty much there as people are already working on vehicles. Blijo and I can go pretty low on node weights (even down to 2.5 grams), but going lower than 400kg with wheels and engine, while having deformation starts to be bit of a puzzle already. I wish Blender would not use dot for new objects, now I have 269 renaming to do Anyways cones are placed somewhat, I just need to rename them and add to gridmap, also need to apply scale (as I forgot do that early) and location to each. Dimensions should be quite accurate, scaled cone to US standard cone dimension, it is 1mm less though, but I thought it would do. Kinda painful with this mouse wrist, but such is part of the artists, entertainment trough blood and tears for your enjoyment Inflammation of nerves and all that: Oh and about template vehicle. Those FSAE competitions are such that vehicles are required to be made by students themselves, they can source or manufacture parts, but it has to be made by themselves, so idea is kinda to have you to make vehicle from scratch, technically sourcing suspension bits, steering wheels etc. from other vehicles should be ok. Suspensions are quite hard to make, so that might help a bit. However this is point under consideration at the moment. Techinally using hopper frame, slapping radiator etc. in, could be within rules, but probably it will not win against better performing scratch made vehicles. More effort and skill put is giving more points obviously. Me for example can't really model, can't do any textures, building car from a scratch is very challenging for me and certainly will not be able to do best design, I think that I can tune car to go pretty good around corners though. Obviously not everyone has equal chance to win, those that are better, tends to be better, but like in running competition there is a race among other position than 1st. I have no chance winning, especially when I'm making rules and probably do part of judging, my vehicle is not really allowed to enter or at least not to be in result list, it would be unfair as need to ensure everyone has a fair result. Perfect car should be enough challenging to challenge even best of the best, that would give them some motivation to participate too, I think it would be okay if none can make best possible, at least on first try, then field will have different level of modders. By allowing use of vanilla suspensions etc. it should help to get more people into making car from scratch, but I am bit against template car for reasons mentioned. I am not against of running other competition about car setup tuning, we could use winner of this fsae car for example there, but I think if we make template cars it kinda voids the point of this competition of creating a race car. However as jbeam builders might not be able to model, they might team up with modeler, also such team can include setup maker and driver too, it would be really cool if people would come to working together such way, so more people can take part and maybe it even makes built cars even better! Also that would easily teach people new skills as working together usually shares knowledge.
My idea about the template car was more like a box on wheels, but with all the necessary parts like powertrain, wheels, maybe a basic double wishbone suspension. Even if you build the structure from scratch, you'd have the jbeam syntax ready and a working car that's easy to start from. It would be possible to make very lightweight and rigid but unrealistic structures. Some minimum weight would encourage more realistic designs, I think. Just the engine, smallest race wheels and 75 kg for driver already weigh about 180 kg. My car ended up around 400 kg total and seems challenging enough to build and drive.
It would be nice to have that, especially for me as I've never modded before and only know the basics of skinning... At least I have my last exam tomorrow and then I'm on holidays until February so I guess I'll just learn myself by reading the wiki and watching blender tutorials. I'm determined to do this though so hopefully it turns out well. Also are people going to create a 'frame' mesh (like the CRD) and then create a 'sheet-metal' mesh to go over the top so it looks realistic?? That's what I want to do, I just hope it's not too hard for a beginner
Ah, this is how differently we work I believe. How I make jbeam such file structure would not help much, I do pretty much all in Blender, scripts handle everything, I just need to model the jbeam, then set parameters, groups etc: Plugin which I have funded with around 50$ I believe, can be found https://www.beamng.com/threads/blender-addon.56981/ Then I just add some parts in Notepad++ I don't technically have flexbody made yet, I just did simple export of jbeam coltris to act as flexbody. Developer of plugin is looking a way to get node names preserved, which currently is greatest limitation. I think that at least engine we should be able to provide, but not sure about files as it varies quite a bit between modders. 400kg is about same I and Blijo have ended up in our tests, so probably upping minimum weight to that like Blijo suggested, unless some magic about getting rid of wheel weight is found. I did test then set tire pressure of 0psi and tires did carry weight still (smallest tire on smallest 4bolt wheel), so I think we need to craft new tire based somewhat on Pigeon tire, but need bit more testing. I'm soon running out of time for my deadline as even real life puts few tasks I need to handle before Sunday Also when you have found a name for your team, I could add those to first post. --- Post updated --- Hmm, I think that if someone is willing to make such, I can put it on first post, it just is quite different from way I build, so I might probably do something wrong. For example hydros/steering I did copy from 200BX, but then rewrote a lot as pretty much nothing matches, so new node names, parameters, part names, slots etc. Still not done and very buggy. You make new parts by copying a block, placing it in any .jbeam file, renaming bolded parts, make sure slot matches to slots you define in slots section of main file. You can then copy from any car and alter nodes, names etc. It does not matter which file is in as long as file is .jbeam and in folder of vehicle. Code: "fsae_fufsgfen_steering": { "information":{ "authors":"BeamNG, Fufsgfen", "name":"Steering", "value":500, } "slotType" : "fsae_fufsgfen_steering", "rails": { "steeringrack":{ "links:":["r76", "l75"], "broken:":[], "looped":false, "capped":true } }, "slidenodes": [ ["id:", "railName", "attached", "fixToRail", "tolerance", "spring", "strength", "capStrength"], ["r86", "steeringrack", true, true, 0.0, 5001000, "FLT_MAX", "FLT_MAX"], ["l85", "steeringrack", true, true, 0.0, 5001000, "FLT_MAX", "FLT_MAX"], ], "beams": [ ["id1:","id2:"], {"beamPrecompression":1.0, "beamType":"|NORMAL", "beamLongBound":1, "beamShortBound":1}, {"beamSpring":4001000,"beamDamp":150,"beamDeform":49500,"beamStrength":92500}, ["r76","l75"], //Alignement adjustables ["l21","l68"{"beamPrecompression":"$camber_FR","beamPrecompressionTime":0.5}], ["r76","r86"{"beamPrecompression":"$=$toe_FR+$steer_center_F","beamPrecompressionTime":0.5}], ["l85","l75"{"beamPrecompression":"$=$toe_FR-$steer_center_F","beamPrecompressionTime":0.5}], ] "hydros": [ ["id1:","id2:"], {"beamPrecompression":1.0, "beamType":"|NORMAL", "beamLongBound":1, "beamShortBound":1}, {"beamSpring":4001000,"beamDamp":40,"beamDeform":"FLT_MAX","beamStrength":92500}, ["r76","l85",{"factor":-0.26,"steeringWheelLock":450, "inRate":1.25,"outRate":1.25}], ["l75","r86",{"factor":0.26,"steeringWheelLock":450, "inRate":1.25,"outRate":1.25}], {"beamPrecompression":1.0, "beamType":"|NORMAL", "beamLongBound":1, "beamShortBound":1}, ], "input": { "FFBcoef":13, }, }
I just read through the whole thread. It was very interesting and helped me understand a little more how everything works The plugin seems extremely useful I was just wondering if it ever got updated with the improvements that were mentioned and if progress is still ongoing?
I spent about 2 hours on a car last night... after screwing around with some other stuff in Beam. I used the stock Pigeon wheels along with the 600cc engine. Basically just copied the files over to my cars folder and adjusted the connecting nodes. I figured those were good bases for a car this small (and the fact that you all are using the Pigeon engine anyways) I rigged it up extremely basic right now... it's basically a frame with some hubs that act as suspension. Mostly just hashing out a file system first (like Miura was talking about) as that makes changing things down the road MUCH easier once you get everything sorted out. So it does run and drive... no tuning yet though as there really isn't much structure to the thing yet. That being said though... it told me that I was at about 236lbs (107kg) with tires, wheels, suspension, steering, frame, engine, and transmission... with completely unaltered Pigeon parts... so I am gonna have to add some weight back in, but I don't have any wings or anything like that yet, so that will add some weight too. Then again, it is so light right now that with a lack of sway bars, if one of the tires falls over, the whole car tips over not incredibly helpful really... I didn't think to add weight in for a driver though... How much weight should I add in for the driver? I see Miura suggested to use 75kgs, so is that what we are going with? OR... here's a thought... should we have standardized JBeams for parts that we MUST use to compete? (aka, a driver, wheels, engine) I already have wheel and engine files made up (as do all of you) but if we were to standardize it... it might help others a bit... Like I said... I don't know how much weight to add for a driver.
It has been updated after that thread and I have exchanged messages with Khalibloo frequently, retaining node naming is something that is being looked into as it is not the easiest thing to do, but there has been updates, you can see commits here: https://bitbucket.org/Khalibloo/beamng.drive-blender-addon/commits/all Yeah, I need to add driver weight to rules too, 75kg is pretty much what standard weight of human is everywhere, afaik. Blijo probably knows or can get lat G of such vehicle on skidpad, I have seen 1.3G to 2.5G first might of been figure 8, maybe it is possible to modify Pigeon tire to work with high grip and cornering. There is some information though, but will not work for 400kg vehicles I guess as these are for 200kg vehicle, but you can get pretty good data from there and if knowing how, I guess it should be possible to create tire that works, for that I have to rely on some of you engineer guys or randomly poke Pigeon tires which quite often does result something ok: http://www.avonmotorsport.com/resource-centre/tyre-applications/sae-formula-student
Im currently getting up to 1.2G while cornering with an overweight, unoptimised and simple vehicle I don't really think that I can do much more since the new quartile at uni just started. I think that my evenings will mostly consist of differential equations, sending love letters to LaPlace and battling with MATLAB and simulink instead of chilling with beams and reading about vehicle setups. That said, I love challenges like this but I just don't have time to compete. I might join next competition and run my current wonky-car for fun now(development time: 7-8 hours) Maybe someone with decent skill wants to team up with me? I have a base and some knowledge but don't have time to optimise, research and develop stuff right now. Or, I help teams by answering questions and uploading a basic frame with engine, gearbox and fueltank.
UPDATED 16.11.2018 Start: Finish: It is very simple and crude, but you can choose if you want to collisions on or off. You go to world editor and add new object autocross_layout from art/shapes/autocross/cones, also autocross_path from art/shapes/autocross Don't move camera after you add 1st object, it is then easier to keep them lined up, but it is not too difficult to line up later, just don't rotate or scale as that messes things up easily. Path has no texture, but I kinda liked that orange so left out materials.cs from that. If you take out files and place to some other map folder structure, you can add them to any map. I'm not sure if it is gridmap or my object, but it's not 100% flat, I did set all vertices to same height and set to 0 with s, z, 0, enter + typing 0 to z transform, making sure global was selected, but still ends go into gridmap ground. I will add more cones and some other stuff maybe too, but so that you have something to test, there is now simple cone track.
Pigeon 600 uses the clunky stock Pigeon exhaust, which reduces power from 102 to 81 hp. I guess we can make a custom exhaust part without torqueModExhaust, or just have no exhaust to keep it simple? CVTs seem to be fairly common in FSAE cars. Would be interesting to have free transmission choice. Some kind of higher speed event would be fun too. My car currently does well over 200 km/h (no aero yet), but the autocross course has something like 50 km/h average speed.
So I did make some updates to 1st post, updated some rules, also I did go trough this thread and attempted to list people participating or wanting to participate but looking for a team, yell at me if there are some mistakes. Also if you can come up with some team names, I could list them there too. Wanted to add engine in, but figured out that I need more time to verify that, also I had a thought of compensating extra weight with a small turbo setup, but that is something that would take time to craft again. I could upload engine now, but it would need to be changed very soon again, so I rather would not go to update hell route with that. Very good points! This is what I have got so far for the engine, took out everything from the slots, to keep it simple, no torquemod of any kind should happen, all intake, exhaust etc. would have just visual appearance to keep it simple and power level same for everyone. { "formula_BeamNG_engine_600_YOURTEAMNAME": { "information":{ "authors":"BeamNG", "name":"YOURTEAMNAME 600cc I3 Engine", "value":650, } "slotType" : "YOURTEAMNAME_engine", "slots": [ ["type", "default", "description"] ], "powertrain": [ ["type", "name", "inputName", "inputIndex"], ["combustionEngine", "mainEngine", "dummy", 0], ], I would love to have CVT and other transmission types allowed, but for that we would need to determine plausible ratios for a such vehicle, which would need research a bit, but if you have stumbled upon ratios they use, I could add that too. Also need to add rear diff acceptable ratio. For sequential I want to have 600cc sport bike ratios, few different to represent different brands, which from one has to be chosen to be used. That is as IRL these are not changing gearsets between events, afaik. For higher speed track test I have been thinking about Automation test track, it has also few nasty turns that put suspension setup to test quite well. Maybe Blijo can relate to this thing I stumbled upon? I'm not sure which category I should list you in first post? Jbeamer looking for a team perhaps? Oh and if someone wants to use track layout of cones I posted earlier and make some proper map, with decal roads and race track textures for the race line etc. be free to do so! Also it is easy to make other layouts based on that as each cone is own object, so when imported to Blender you can arrange them to new tracks etc. I'm not very good at visual aspects. Cones are from Beam vanilla cones.zip just resized to be 710mm tall and to work as static object. To have collisions with cones, in world editor setting collision type of cone objects to visual mesh final enables them, but collisions are quite nasty with such pointy objects. --- Post updated --- Added few points about diff and gearbox ratios to first post, if some other ratios are used, I would like to hear about them as defining realistic range can be off at the moment.
eh, not really, more like a lot of fun stuff and not enough time in a week I think you can put me under general advice and looking for a team?
That's nice, then it is not quite horrible bad there Find fun in work and life has not a dull spot, or something close to that. I did add you in first post, if you want adjustment to that then yell I try to craft endurance layout of Lincoln too, as from that there is really good data available, how well my wrist can just cope, can't make it quite the pace I would like to. I saw 1.4G as highest here, latG and skidpad is at latter part of the video:
give your wrist a break and extend the deadline, nobody will have visually/technically optimized vehicles finished within a month anyway. maybe the end of february, which gives plenty of time (if used wisely) to make the formula cars "dev quality" if that's the goal.
I think that main run of the competition will happen from beginning of February, until end of this year we run pre-competition, but all work done to vehicles this year should help for next season as indeed not many will have vehicles fully done to end of the year and that way we can have really good next year's season as we (or at least I) need bit experience from pre-run to flush out bugs that are not quite apparent yet. Put drift steering to street tuned Miramar and it is kinda ok on this track. I think that I did cut at least 2 cones, would need a lua that adds time from each touched cone: You can also see how I improved path a little and added most of the cones that were missing, so it is now much easier to go around the track. 416 cones now, thankfully I found mass renaming script for Blender, which saved already something like 2000 clicks. I will update track download shortly.
Looks nice! I like the real sharp hairpin halfway through the track... and how you actually bothered to put in a cooldown area... nice touch.
I still need to add cones for skidpad area and acceleration test area, but it should be pretty close to real thing as I used drawings from real event to place cones, I might try with higher resolution shots to see if some cones need to be adjusted as there was bit of blur issue, but there should be deviation of no more than 2 cone width at most, but with me anything is possible. Download and instructions are there https://www.beamng.com/threads/form...petition-for-modders.59114/page-3#post-959423 You can compare to this run and make complains, then need to check if it is my crafting or their deviation from plans, more likely my crafting: If someone manages to make 1:1 realistic car, then it should result pretty close to same time. I wonder, how would Vendetta compare to it's real life partner? You don't happen to have some 500 cones at garage and huge parking lot to spread them over? Anyway by importing those cones to Blender, setting new places them and exporting to same cones folder, anyone can make new autocross layouts with relative ease, you just need to set 2d plane to be exact size of your drawing and then drag cones to correct places. Making full map is then bit more work and requires skills that I don't have that much, but placed cones might make it bit easier for map makers to build nice parking lot and surroundings. --- Post updated --- 546 cones placed pretty accurately, using Lincoln 2015 map that was posted to FSAE board : Also skidpad and acceleration test is made, I did actually add few extra cones for skidpad so it works better. Now would just need terrain with road textures that has two white lines one at both edge and placing those should make it quite ok performing map. Also decal roads should make AI possibilities but I don't know anything about that, at least not yet. If I would figure out how to make white line meshes to follow bezier curves in Blender, I could make them to same dae as cones are, that would allow one map to be platform for numerous of autocross layouts, however that would not support AI then, so I guess it would be better to have pretty much own maps, but I'm not much of a map maker, so I don't know if there are better solutions for having numerous 'configs' of the map. But at least we can now have test drives on real autocross layout.
