Somewhere between Premiere and UA-cam I lost some audio between 6:02 and 6:15. Sorry about that everyone! It's a short explanation and rant about how difficult TPU is to print.
Tpu and abs and petg are all ambient temp dependant. Having an enclosure, even just a cardboard box, around the print bed is advised. Yes extra heat will kill your hotend and fan, and your power supply if it is anywhere near the print bed/enclosure, but it is the best way to prevent the edges from peeling off the bed (and thus warping) as well as keeping layer adhesion consistent. I learned this the hard way in a temp controlled room. Enclosures are key.
I went to school for engineering and I have to say I would have much rather just gotten a shop and started tinkering around and started learning with hands on experience instead. All you do in engineering school is book work and equations and very rarely anything hands on so when you finally go to do your project that inspired you to go to engineering school for, you would in actuality be a thousand times more skilled at that project if you had just been building projects like that for 4 years straight instead of reading books about how the modulus of elasticity and laminar flow. Or maybe you could meet half way, there are cool tech schools that do 2 year programs and have some good hands on stuff.
Find a school with an FSAE team, if you like this kind of stuff you will not regret joining, I’m currently a freshman and joined my schools team and we are doing many of the things he did in the video. :)
Consider getting employed in a Machine shop along the way as you attend school. You'll likely start with menial task's, but you'll also be starting to practical app allot of stuff as well.
Welcome back, Wesley! What a wonderful series. I'm so excited to see the Kart take shape. This thing is one to pass on to your kids; it's that great of a project. Also, love the Synthwave background music.
So happy to see you back at this project, Wesley. I wouldn't be terribly disappointed to see an entire video dedicated to the equal length exhaust so we can hear the V12 scream... just saying
Oh it's going to sound good- I'm debating on whether I'm going all mandrel bent on the stainless or if going to weld it together. Cost will probably be the question there.
@@WesleyKagan don't forget the time component. Every time I fire up the tig, it feels like the sun is about to come back up as I finish the job. I liked the steering wheel bit too btw, it's cool to see the design in how communication will be handled and what data will be read based on what condition is active on the wheel.
ALso you should strongly consider using a Teensy 4.1 and Canbus for all comms between systems on the car. I also have my ECU directly interfaced to the Teensy4.1 datalogger as well through CANbus. cuts cables down hugely and getting different systems talking to each other stops being a new problem each time.
I was just gonna comment this. The CAN-bus is very robust, quite fast and it is easy to add stuff later on. Also as you say, with a Teensy or something similar it is very easy to implement!
I'm excited to see this trilogy or however many episodes it will have. I enjoy the way you share your process of testing and building. The iterations on the steering wheel add lightness!
I originally got suckered in by your mechanical logic gates. While I’m disappointed in the lack of further episodes of that this far, I’m more than impressed by the effort that you give to us on your channel and each new video you release is exciting anew! I’m a graduate mechanical engineer. Great to get inspired here!
If you would like to use the i2c bus in higher noise environments, or environments with longer runs, I would look into a bus accelerator. Analog devices sells the LTC4311 which I have seen used to fortify i2c busses on cat 5 cable. Signals competently and reliably can be transmitted for distances over 30m. I know you've found a workaround at this point, but in case you feel like it, you could look into this. Also: I would point out that the arduino uno/micro uses an ancient 8 bit microcontroller with extremely low speed and memory resources. I highly recommend looking at the arduino zero or nano, which use a much more modern 32 bit arm microprocessor and have tons more space in them.
Alternatively move to CAN. There are some simple/cheap CAN transceivers like MCP2515, with corresponding Arduino library. Similarly to I2C, you can put a (large) number of devices/uCs on the network, which consists of only 2 wires, but is much more robust.
Bus accelerators do work, but they're an EMI nightmare - if you're running that I2C parallel to any other busses or analogue signals crosstalk can become a huge problem. I second using a modern 32 bit micro. I spend a lot of time working with the pi pico and I'm a huge fan.
@@huzeff CAN would be the logical choice here, since it's pretty much designed for this purpose... Also later in the video he refers to an MS3Pro by which I presume he means MegaSquirt, which supports CAN for clusters...
Oooohh! some CanBus folks!!! Hey ah, can someone tell me why we even use CAN? I dunno just doesn’t make sense with mini and micro PCs, GPUs, etc. One would think all that processing speed would be beneficial to engine performance. And then watching movies, ya see kids racing with friggin laptops in the passenger seat. Yea I’d rather ask the question than look it up and tumble down the research rabbit hole - kinda old school, but not old school enough that I wouldn’t mind a few 100 core processors (or whatever we’re up to now) on-board. Na don’t want a Tesla! Rather breathe car exhaust than mine the Earth for all she’s worth. (can’t wait for them to leave for Mars. I hope they can hold their breath!). Anyway, I ask because I own a Tech II. Boy! it’s painful to use, especially since GM only let’s you use a IBM Thinkpad from the ‘90’s and RS-232 serial connects (pre- USB). I got it new from China (can’t get a letter to my neighbor within a week, yet get pkgs from China next-day. I just don’t get it). Yea, don’t ding me for mining silicon - I got nothin’…😊
@@pistonwristpin1 CAN was designed from the outset with automotive (and similarly electrically nasty environments) in mind, it's got very high noise immunity without needing a bunch of shielding and such. It's also resilient enough that it'll keep working even if you break one of the two CAN lines (though at diminished performance). It's also *relatively* simple to implement in hardware. Also being an addressed, bus-based system it means you can stick all the stuff in the vehicle onto the same (or more often a couple of separate) CAN network(s) and have them talk to one another without having to build and wire dedicated point-to-point communications interfaces between each system and all the stuff that system needs to talk to. Means you can just lego stuff together, for example you can stick something like a Haltech Nexus ECU, several Haltech PD16's, a pair of Wideband controllers, an electronic dash and whatever other gubbins into a system and just use the same two wires to connect the lot together rather than having to individually every circuit through the car separately.
Very nice. Wow that is a lot of wires to wire through a steering wheel connector though. I did try something similar once (ended up parking it for now) but used a 6-pin mini-din in the inside of the steering shaft. Lined up so that when the steering wheel was in the correct place it would just slot in. Only used 4 pins though, gnd, +12v, CAN L and CAN H. Used a custom circuit board to use CAN with an arduino. I'd link pictures but youtube won't let me.
Anyone with a V45 Interceptor in their garage is cool. I had the 2nd one in the state of NC in 1983. Now I have someone else to follow and learn from, great channel.
I wish I were even a fraction as driven and as talented as you to be able to do something like this, but it makes me very happy that people like you can concoct things like this. Great work.
Duuuude, good to see you back. I was thinking the other day I had not seen a video of yours for awhile. It's interesting to see you are 6'2". Just a shade taller than me, so I can about how I would fit in that car. I like how you are designing it for your size.
Glad to be back! Thanks! Yeah, It bothered me for a long time because the roll hoop on the Lotus was 31" and the original kart was 36, so even at unsafe height it was still taller. I gave in and made it to fit the one who is driving it hah. With the adjustable pedals, ideally anyone should be able to drive.
Roll hoop: check out the Crosslé 90F, it might be a happy medium between your two versions. They've been using them at a driving school for 5 or 6 years and nobody died yet.
I am up for this as well. Or at least in the case, you have 2 microcontrollers communicating with each other. The other option would be CAN, but most CAN modules for small microcontrollers use SPI, which will be busy controlling the display. The RS485 is basically uart, so it uses uart module/block inside the uCPU. If you use RS485, it is handy to implement/use some messaging protocol to ensure message transmition and validity. This though is actually quite easy.
Switch to using CAN rather than I2C. There is an arduino library available and there are plenty of shields. You can keep using the CAT5/6 cables (in fact doing so will be good if you use a pair that's twisted for CAN. It'll make the electronics side more difficult but all your data transfers will just use the serial port and it's still 2 wires. If you're building a custom board for the steering wheel you can include the CAN chips directly there anyway
Because I'm already expanding the circuit board, I agree. I originally was expirimenting with wireless protocols (bluetooth, exc) and couldn't get a reliable connection. I2C worked well for this, but as I expand, it really needs more headroom. I want to use a board for all the connections on the wheel instead of all those soldered wire connections anyway.
@@WesleyKagan Yeah, you do NOT want to do wireless in this kind of environment and I2C is great for small amounts of data, but it's very slow, which causes problems when you want more than a few frames per second.
Good to see you back in the garage with some new content, Wes! Suggestion on the steering wheel rotary control knobs: (I know this, because I am also building my own steering wheel for my street car, and I am trying to print circuit boards to retrofit steering wheel controls from a ~2018 Hyundai into a 1998 Trans Am) the way that the factory avoids the spaghetti, is this--Instead of running each resistor to a separate wire out (a good idea at the start), each resistor runs into the next adjacent resistor, similar to how a rotary phone works. Select item 1, and the signal goes through one resistor, and then out. Select item 2, it goes through that resistor, followed by the resistor from item 1 for a total if 2 resistors, and a different resistance value. As expected, item 9 runs through all 9 resistors before exiting the steering wheel. In this way, only 1 wire need go in, and one back out again. Hope that helps.
Instead of TPU, have you considered printing molds and pouring up some silicone? I've done some of this, including making little silicone one-way valves like they have on the full-face snorkels. It's worked pretty well.
Hey man! Here are some tips for printing TPU: n.1 (most effective) - use a direct drive setup. It makes printing tpu an absolute dream as it pulls the you more than pushing it through Bowden. N.2 - print slowly, usually printing at about 40-45mm/s works best, possibly slower too. N.3 - not often mentioned by high acceleration and constant flow is vital for getting good tpu prints, as the less pressure change in the nozzle with flexible filaments the better the end result. High acceleration means that there's less time spent changing extrusion amount and there is less flow / pressure change. N.4 - use high flow nozzles, they allow for more consistent extrusion of tpu as the bigger melt zone can prevent funky tpu moments. N.5 - bigger nozzles print better. A combination of a 0.8mm nozzle and a high flow hotend (phaetus rapido for example) means that you can get more plastic down quicker at the same amount of pressure, and smaller pressure changes don't effect the end extrusion amount too much. These things have helped me print flexible filaments the most, though I would definitely say that the extruder matters the most. Look up printable extruders capable of printing 70-85a TPU and they will be more than enough to print any material reliably and at speed
I don't know what shots anymore, the fact that this channel only has 165k subscribership, or the fact that you do NOT look like a Gearhead and can do things that are pretty much considered an art! Subscribed!
@2:47 The normal way to do this is to use the same value resistor linking adjacent legs so that your voltage divider has even steps, then put 0v on one end of the chain and 5v on the other end. For example: 0v to pin 1 1k resistor between pins 1 and 2 1k resistor between pins 2 and 3 1k resistor between pins 3 and 4 1k resistor between pins 4 and 5 1k resistor between pins 5 and 6 5v to pin 6 Signal from sweeper pin. This should give 0v at position 1 1v at position 2 2v at position 3 3v at position 4 4v at position 5 5v at position 6 If you want some error banding then you can also put a resistor inline on the 0v and 5v lines to condense the steps further. This will allow you to program an error function if either 0v or 5v is read on that input. 0v to pin 1 with 1k inline resistor 1k resistor between pins 1 and 2 1k resistor between pins 2 and 3 1k resistor between pins 3 and 4 1k resistor between pins 4 and 5 1k resistor between pins 5 and 6 5v to pin 6 with 1k inline resistor Signal from sweeper pin. This should give 0.71v at position 1 1.42v at position 2 2.14v at position 3 2.86v at position 4 3.57v at position 5 4.28v at position 6
What really helped me print TPU was getting a filament dryer, TPU wicks up moisture really quickly from the air and creates artifacts and other print issues. I just got a cheap food dehydrator off Amazon and cut the racks until I could fit a roll of filament in there. I normally put the filament in for a couple hours before I print
Absolutely awesome steering wheel build. Glad someone showed that a modern looking F1 style wheel can be made at home. I think another good route that’s less time intensive, but also less cool, would be to mount an AiM dash to a steering wheel
Pumped to see this revived! Steering wheel looks good though it's so close to the rest of the dash would be easy to accidentally hit switches and stuff while turning, I know all that's gonna change anyways but yeah, something to plan, a wheel looks good in it's initial position but you have to think of ergonomics mid-turn too and not hitting anything else with your hands.
Yeah, I didn't have room on the build now for gauges or dash switches or anything, so that's why it all moved to the wheel. Makes things easier for sure!
i'm an embedded software engineer, and the display is pretty dope. i've been designing a similar display for an F1 style go kart i have (made in the 70's). if you are still having memory problems, there are some tricks you can do. you can store some (or all) of the static image data like the lines in the program space using the keyword PROGMEM. if you haven't used it before, read the arduino page on it. you also want to live in the stack as much as possible, and use as few variable declarations as possible. in some cases, you can use the video memory to do calculations. if you have a chunk of video memory you know is all 0's or all 1's, you can write some math stuff in there, and then just put the 0's or 1's back when you are done (John Carmack perfected this method with Commander Keen years ago). Basically, try to avoid any alloc calls if you can, and reuse variable as much as possible. anything static, stick it in progmem. you can also usually compress your graphics a TON. i noticed you have a lot of lines , and instead of storing a bunch of 1's in memory, you can store where the line goes and how long it is. typically storing compressed data in progmem is the way i save the most space when it comes to fancy screens. you also might write your version of Wire.h that does the bit banging for I2C. you can often squeeze out some more space by doing this because the libraries have to be general purpose and often include stuff you don't need for this particular project. this goes double for screen libraries, which often have a TON of complicated and bulky code for screens that can rotate like a phone. The adafruit screen library has a huge amount of code that is useless if you don't plan on rotating your screen at all.
Took a peek at your code. IIRC, you don't need to redraw the whole screen every loop, you can update just the parts you want to change. In other words, draw it all at the start, or when you change modes, then just update the relevant bits. Hope this helps.
also, the pi pico is mostly compatible with Arduino libraries and is quite a bit faster, so you could look into that on top of loading images from the sd card to improve display speed.
I can’t wait for this car build. I wanted to create my own open wheel/f1 style car for the longest time, but make it the other way: instead of going back to 60s, I wanted to use as much banned tech as possible and make it modern. As I learned a few videos back, we’re the same age, so looking at your progress, I should get on it 😅
This is insane and I love every second of this video. I can do the mechanical work but I’d definitely need to learn the electrical side of this but it’s absolutely stunning. Best of luck I love this idea!
Great to see this project back on the front burner. I truly hope you reconsider the Top Fuel dragster roll cage and go with something a little more period-correct, just taller and braced better.
The really crazy part about all this is in the end you likely could have just built a second car and kept the Boxster racer. Evolved instead so survival of the fittest I suppose! ;) Awesome video as always lots of insights.
For good results with tpu, the best thing I’ve heard to do is get the straightest and shortest length direct drive extruder as possible. Things like ender 3’s are especially bad at tpu because of their Bowden tube setup where the filament has room to squish. Hope this reaches you and is of help 👍🏻
Hey man, just wanted to say you're one of my favourite home build engineering creators. Your videos are so, so great. Up there with the best, and I can't wait till your channel blows up.
you can use momentary switch on the shifter for the gearbox to send signal to the arduino. my bike (honda supra x 125) use some sort of rotary switch on the shifter, which is pretty much the same
I remember seeing a Honda V-12 F1 car at Watkins Glen in 1968 or 69. What I recall is that it was, by far, the best sounding car on the track. It had those insane blending harmonics at 9 or 10 thousand RPM. Yeah, I loves me a V-12 sound.. LOL
You can get good results with tpu. I have a Bowden setup and adding a PTFE guide inside the extruder stepper to constrain the filament made the biggest difference. I also dropped the retract setting down to absolute minimum.
There's some great ideas in this reboot, the trick is to finish it rather than constantly iterating individual parts. Thanks for sharing Wesley, I'm looking forward to seeing the build progress!
Why aren't you using deutsch connectors? Also, EEPROM for storing your graphics off-chip. Cheap and easy (SPI/I2C/Serial IIRC). Also also, why not use a CAN bus? That tinsy you got supports it IIRC.
Nice to see this project going again. Don't think I've ever seen an open wheeler with a dragster style roll cage. Aesthetically it looks a bit odd, but I suppose safety was a higher priority than looks here. Keep up the good work!
Another potentially useful habit to have with I2C communication, or most embedded programming, is to stay away from transmitting text whenever possible. It looks like you are sending a raw concatenated string of values that you then unpack into your various numerical values. If instead of the ASCII text you are sending as a string, you just send the value as a byte you will see something like a 2-4x decrease in the data sent. Just convert it to a string when you need to print it. For example the example string you showed on screen is 14 bytes. But if you sent the values it would be 6 bytes. (each value besides tach is a 8 bit integer value, tach could fit into a 16bit integer). Your code seems mostly fine, a lot of optimizations are not worth the hit to readability. But trading ASCII chars for integer bytes usually involves little change to the code, and often removes bottlenecks like you mentioned running into around 32 characters, so it's usually a good trade to consider.
That's a really good point- I didn't explore that, as I really am not sending _that_ much data- with everything I can't see much more than 50 characters. Still substantial, and I'll do some testing with switching them over. This was easier to read during testing.
@@WesleyKagan yeah that is one advantage of sending ASCII, is it can be easy to read in some situations. I fall back on that for a lot of web requests and such, especially since in those settings I am not limited by compute capacity usually. Usually you can do a print command that will print a integer/float/ect value as text directly. For example printf(“d”, value) in C which Arduino essentially is. It adds an extra step, but is simple enough that it can become habit. I have found that I often work best by writing the base fragments of my program first, such as a function to handle packing and unpacking whatever data I am sending, then a function that takes the packed data and sends/receives it, then whatever functions actually use that in the logic of whatever I am trying to do. I test each as I go, and once I am convinced they work I treat them loosely as a “black box” that just works. That being said, that is just what I find works well for me. YMMV
love the project, did you think about using CAN for the dash ? ms3 supports it, and you wont be limited by bandwidth or cable length (plus its only 2 wires from ecu to dash)
So cool to see you do a project similar to what I've been working on, as far as the arduino and screen go, Been working on an oil temp and pressure on a similar Nextion 320x480 screen. Not sure if their GUI works with the Adafruit, but the Nextion Editor program is super helpful on creating those graphical backgrounds. Looking forward to the next couple videos of this!
@@WesleyKagan Oh this is lovely on code, it gets loaded on that SD slot and there's not much code to send the data to each box. You can even change designs and not touch the code as long as you keep the box ID's the same.
@@WesleyKagan I was going to recommend the nextion as well. Besides the editor the other benefit is there is an existing library of dashboards that the sim racing community has created. It should just be a matter of tweaking the identifiers in the editor to those the Arduino is using... at least in theory.
Quick recommendation, to get around the printing challenges with that material, you could print yourself molds and then cast your forms with material from Smooth-On. There's also expanding foams on the market now with many density options. You could always wrap the handles with something nice like Alcantara. 👍 cool project 😎
For printing TPU, my experience has been that you want a direct drive extruder, print relatively slowly and very importantly calibrate your extruder steps. I need to up the flow by 30% compared to PLA because the TPU gets squished on the extruder gear giving it a smaller apparent diameter. The bed doesn't need to be heated, cooling is dependent on what you're trying to print. I've only ever used NinjaTek NinjaFlex on a modified direct drive Ender 3 Pro for reference. Also, if you want a part you can compress check the infill types, they have different properties in different directions. In cura there is a plugin that gives a lot of explanation about all settings including infill, including what their strengths and weaknesses are. For example, the cross-3d infill gives similar resistance in all directions.
Hey Wesley, thanks for the content. To help save memory onboard you can buy a nextion screen. It does the touch controls onboard its own μc. Has a super easy interface to create the GUI. You can use pictures and such. Also it only uses a TX and RX wire. As an ME that had to make HMIs for my projects it made it super simple. Using a teensy could make things easier they have an onboard CAN and they run like 40x faster than an arduino. Much more memory. Downside 3.3 V logic. Going to CAN could help you use a through bore slip ring since you only need a CAN high CAN low and power and ground. Just my 2 cents hopefully something helps.
Teensy seems to be popular- I actually ordered one last week but I have a bucket of Arduinos so it's what I know for prototyping. CAN is seeming like the way to go as well, and I agree. I tried some wireless protocols but didn't have a ton of luck with a solid connection. I'm going to have to look into the nextion screen, that's seeming to be a much better resource for the graphics side. 3.3v would take some work.
Absolutely super cool, keep it coming! The steering wheel part is really interesting since I’m working on the same type of system for my RSX project car. Printing TPU as we speak and yes it is a mess, but I think I’ve almost got it figured out.
Great video! Have you looked into a CAN-network as an alternative to the I2C-bus? It is also only 2 wires (differential pair) and should be able to run through the whole car. There are shields for arduinos. We use CAN in the Formula Student (like FSAE) team I'm in!
Somewhere between Premiere and UA-cam I lost some audio between 6:02 and 6:15. Sorry about that everyone! It's a short explanation and rant about how difficult TPU is to print.
Also another at 10:48 when you start to talk about the wheel's screen
I actually thought it was a joke. "Okay, We're going to talk about the spaghetti here..." then nothing about wiring at all. lol.
TPU is pretty material dependent. Ive used solutech tpu to much success on a stock ender 3, just cut all my speeds to about 40% of my pla profile
Tpu and abs and petg are all ambient temp dependant. Having an enclosure, even just a cardboard box, around the print bed is advised. Yes extra heat will kill your hotend and fan, and your power supply if it is anywhere near the print bed/enclosure, but it is the best way to prevent the edges from peeling off the bed (and thus warping) as well as keeping layer adhesion consistent. I learned this the hard way in a temp controlled room. Enclosures are key.
@@ryanhebron4287 same lol
You the man Wesley! We gotta get our formula cars together!
That is a Collab I would watch!
Thanks Casey! I agree, as soon as this is together I’m going to set up a track day, we need to!
You guys can touch exhaust tips.
@@WesleyKagan If I lived closer (I'm in Sweden) I'd ask to join :) I'm building my version of a Lotus 72E
holy smokes! what an invitation!
I’m in high school, one of the things I may go to college for is engineering. That is solely because of you and this build, thank you!
That's great to hear! Thanks for watching, and best of luck!
I went to school for engineering and I have to say I would have much rather just gotten a shop and started tinkering around and started learning with hands on experience instead. All you do in engineering school is book work and equations and very rarely anything hands on so when you finally go to do your project that inspired you to go to engineering school for, you would in actuality be a thousand times more skilled at that project if you had just been building projects like that for 4 years straight instead of reading books about how the modulus of elasticity and laminar flow. Or maybe you could meet half way, there are cool tech schools that do 2 year programs and have some good hands on stuff.
@@El_Chompo FSAE tho
Find a school with an FSAE team, if you like this kind of stuff you will not regret joining, I’m currently a freshman and joined my schools team and we are doing many of the things he did in the video. :)
Consider getting employed in a Machine shop along the way as you attend school. You'll likely start with menial task's, but you'll also be starting to practical app allot of stuff as well.
Always love to see practical use for 3D printing. Great video as always, and looking forward to the next two!
can´t wait for the next video
Stop watching youtube videos and get your next one out 🙃
@DevinderAthwal306 ayoo pause.
Welcome back, Wesley!
What a wonderful series. I'm so excited to see the Kart take shape. This thing is one to pass on to your kids; it's that great of a project.
Also, love the Synthwave background music.
Thank you! Same here, it's going to be a great project, I can't wait. Already is in a lot of ways!
So happy to see you back at this project, Wesley. I wouldn't be terribly disappointed to see an entire video dedicated to the equal length exhaust so we can hear the V12 scream... just saying
Oh it's going to sound good- I'm debating on whether I'm going all mandrel bent on the stainless or if going to weld it together. Cost will probably be the question there.
@@WesleyKagan don't forget the time component. Every time I fire up the tig, it feels like the sun is about to come back up as I finish the job.
I liked the steering wheel bit too btw, it's cool to see the design in how communication will be handled and what data will be read based on what condition is active on the wheel.
ALso you should strongly consider using a Teensy 4.1 and Canbus for all comms between systems on the car. I also have my ECU directly interfaced to the Teensy4.1 datalogger as well through CANbus. cuts cables down hugely and getting different systems talking to each other stops being a new problem each time.
I was just gonna comment this. The CAN-bus is very robust, quite fast and it is easy to add stuff later on. Also as you say, with a Teensy or something similar it is very easy to implement!
what blows my mind is in like... 2015-16? i used Teensy 2.0's to build a few mechanical keyboards... really has never been a better time for DIY
I'd be interested to see a fleet update. You seem to have an interesting mix of vehicles. Similar to Superfast Matt, you should collab.
I probably should. It's been a while since I've done an update, honestly.
@superfastmatt you watching?
I'm excited to see this trilogy or however many episodes it will have. I enjoy the way you share your process of testing and building. The iterations on the steering wheel add lightness!
Exactly! I didn't mention it, but It's almost the same weight as a regular Wheel- ABS and carbon don't add a ton of weight.
@@WesleyKagan It's an absolute joy to see you do things that I have no way of doing. Be well!
I originally got suckered in by your mechanical logic gates. While I’m disappointed in the lack of further episodes of that this far, I’m more than impressed by the effort that you give to us on your channel and each new video you release is exciting anew! I’m a graduate mechanical engineer. Great to get inspired here!
I can't understand why this channel doesn't have more viewers/subscribers. Wesley is a very smart, innovative creator.
Post them videos!!! I’d watch them and I bet a lot of other people would too
omg, i was waiting sooo long for this!!
If you would like to use the i2c bus in higher noise environments, or environments with longer runs, I would look into a bus accelerator. Analog devices sells the LTC4311 which I have seen used to fortify i2c busses on cat 5 cable. Signals competently and reliably can be transmitted for distances over 30m. I know you've found a workaround at this point, but in case you feel like it, you could look into this. Also: I would point out that the arduino uno/micro uses an ancient 8 bit microcontroller with extremely low speed and memory resources. I highly recommend looking at the arduino zero or nano, which use a much more modern 32 bit arm microprocessor and have tons more space in them.
Alternatively move to CAN. There are some simple/cheap CAN transceivers like MCP2515, with corresponding Arduino library. Similarly to I2C, you can put a (large) number of devices/uCs on the network, which consists of only 2 wires, but is much more robust.
Bus accelerators do work, but they're an EMI nightmare - if you're running that I2C parallel to any other busses or analogue signals crosstalk can become a huge problem.
I second using a modern 32 bit micro. I spend a lot of time working with the pi pico and I'm a huge fan.
@@huzeff CAN would be the logical choice here, since it's pretty much designed for this purpose...
Also later in the video he refers to an MS3Pro by which I presume he means MegaSquirt, which supports CAN for clusters...
Oooohh! some CanBus folks!!! Hey ah, can someone tell me why we even use CAN? I dunno just doesn’t make sense with mini and micro PCs, GPUs, etc. One would think all that processing speed would be beneficial to engine performance. And then watching movies, ya see kids racing with friggin laptops in the passenger seat. Yea I’d rather ask the question than look it up and tumble down the research rabbit hole - kinda old school, but not old school enough that I wouldn’t mind a few 100 core processors (or whatever we’re up to now) on-board. Na don’t want a Tesla! Rather breathe car exhaust than mine the Earth for all she’s worth. (can’t wait for them to leave for Mars. I hope they can hold their breath!). Anyway, I ask because I own a Tech II. Boy! it’s painful to use, especially since GM only let’s you use a IBM Thinkpad from the ‘90’s and RS-232 serial connects (pre- USB). I got it new from China (can’t get a letter to my neighbor within a week, yet get pkgs from China next-day. I just don’t get it). Yea, don’t ding me for mining silicon - I got nothin’…😊
@@pistonwristpin1 CAN was designed from the outset with automotive (and similarly electrically nasty environments) in mind, it's got very high noise immunity without needing a bunch of shielding and such. It's also resilient enough that it'll keep working even if you break one of the two CAN lines (though at diminished performance). It's also *relatively* simple to implement in hardware.
Also being an addressed, bus-based system it means you can stick all the stuff in the vehicle onto the same (or more often a couple of separate) CAN network(s) and have them talk to one another without having to build and wire dedicated point-to-point communications interfaces between each system and all the stuff that system needs to talk to.
Means you can just lego stuff together, for example you can stick something like a Haltech Nexus ECU, several Haltech PD16's, a pair of Wideband controllers, an electronic dash and whatever other gubbins into a system and just use the same two wires to connect the lot together rather than having to individually every circuit through the car separately.
Great to see you back! Awesome work as always!
Thank you! Glad to be back!
Marvelous project... congratulations, you are great!! parabens from Brazil
Good to see you posting again 😀
You're supremely underrated, keep on keeping on.
Very nice. Wow that is a lot of wires to wire through a steering wheel connector though.
I did try something similar once (ended up parking it for now) but used a 6-pin mini-din in the inside of the steering shaft. Lined up so that when the steering wheel was in the correct place it would just slot in. Only used 4 pins though, gnd, +12v, CAN L and CAN H. Used a custom circuit board to use CAN with an arduino. I'd link pictures but youtube won't let me.
Wesley, you are hands down the most fascinating automotive youtube channel. Well done, can't wait to see this come together!!
Wow, thanks! That means a lot!
I am purely impressed. Very nice project and it is well done. Full professional. Thanks for sharing and inspiring. All best to you.
Congrats on the write-up in Road and Track my man!!
Thanks! I'd be lying if that wasn't serious motivation to get this project more on the road haha
6:00 "We're gonna talk about this spaghetti here" and absolute silence follows lol
Surprised none of the other comments mention the 15s silence from 6:02 - 6:17, had to scroll a bunch to find this 😅
Anyone with a V45 Interceptor in their garage is cool. I had the 2nd one in the state of NC in 1983. Now I have someone else to follow and learn from, great channel.
I wish I were even a fraction as driven and as talented as you to be able to do something like this, but it makes me very happy that people like you can concoct things like this. Great work.
Duuuude, good to see you back. I was thinking the other day I had not seen a video of yours for awhile. It's interesting to see you are 6'2". Just a shade taller than me, so I can about how I would fit in that car. I like how you are designing it for your size.
Glad to be back! Thanks! Yeah, It bothered me for a long time because the roll hoop on the Lotus was 31" and the original kart was 36, so even at unsafe height it was still taller. I gave in and made it to fit the one who is driving it hah. With the adjustable pedals, ideally anyone should be able to drive.
This is one of my favorite projects on UA-cam.
Glad to hear it!
Roll hoop: check out the Crosslé 90F, it might be a happy medium between your two versions. They've been using them at a driving school for 5 or 6 years and nobody died yet.
Don't know a thing about mechanics and engineering, but I love these
Just nice listening to a competent man doing things
For running the length of the car, try RS485 instead of i2c, if you can.
I am up for this as well. Or at least in the case, you have 2 microcontrollers communicating with each other. The other option would be CAN, but most CAN modules for small microcontrollers use SPI, which will be busy controlling the display. The RS485 is basically uart, so it uses uart module/block inside the uCPU.
If you use RS485, it is handy to implement/use some messaging protocol to ensure message transmition and validity. This though is actually quite easy.
Glad to see you back!! Would love to see the lathe video you made and never shared.
I was just thinking about your projects a couple days ago. So glad to see you posted!
Thanks for the update Wesley. I really enjoy keeping up with the progress on your projects.
Thanks for watching!
Switch to using CAN rather than I2C. There is an arduino library available and there are plenty of shields. You can keep using the CAT5/6 cables (in fact doing so will be good if you use a pair that's twisted for CAN. It'll make the electronics side more difficult but all your data transfers will just use the serial port and it's still 2 wires. If you're building a custom board for the steering wheel you can include the CAN chips directly there anyway
Because I'm already expanding the circuit board, I agree. I originally was expirimenting with wireless protocols (bluetooth, exc) and couldn't get a reliable connection. I2C worked well for this, but as I expand, it really needs more headroom. I want to use a board for all the connections on the wheel instead of all those soldered wire connections anyway.
I've used I2C in an industrial enviroment before and it was a nightmare! (Even using pullup resistors that are almost shorts and a baud speed of 1k)
@@WesleyKagan Yeah, you do NOT want to do wireless in this kind of environment and I2C is great for small amounts of data, but it's very slow, which causes problems when you want more than a few frames per second.
Thanks for sharing, happy that you are back.
Good to have you back! Hope you give us an update on all the projects
Good to see you back in the garage with some new content, Wes! Suggestion on the steering wheel rotary control knobs: (I know this, because I am also building my own steering wheel for my street car, and I am trying to print circuit boards to retrofit steering wheel controls from a ~2018 Hyundai into a 1998 Trans Am) the way that the factory avoids the spaghetti, is this--Instead of running each resistor to a separate wire out (a good idea at the start), each resistor runs into the next adjacent resistor, similar to how a rotary phone works. Select item 1, and the signal goes through one resistor, and then out. Select item 2, it goes through that resistor, followed by the resistor from item 1 for a total if 2 resistors, and a different resistance value. As expected, item 9 runs through all 9 resistors before exiting the steering wheel. In this way, only 1 wire need go in, and one back out again. Hope that helps.
Instead of TPU, have you considered printing molds and pouring up some silicone? I've done some of this, including making little silicone one-way valves like they have on the full-face snorkels. It's worked pretty well.
Hey man! Here are some tips for printing TPU:
n.1 (most effective) - use a direct drive setup. It makes printing tpu an absolute dream as it pulls the you more than pushing it through Bowden.
N.2 - print slowly, usually printing at about 40-45mm/s works best, possibly slower too.
N.3 - not often mentioned by high acceleration and constant flow is vital for getting good tpu prints, as the less pressure change in the nozzle with flexible filaments the better the end result. High acceleration means that there's less time spent changing extrusion amount and there is less flow / pressure change.
N.4 - use high flow nozzles, they allow for more consistent extrusion of tpu as the bigger melt zone can prevent funky tpu moments.
N.5 - bigger nozzles print better. A combination of a 0.8mm nozzle and a high flow hotend (phaetus rapido for example) means that you can get more plastic down quicker at the same amount of pressure, and smaller pressure changes don't effect the end extrusion amount too much.
These things have helped me print flexible filaments the most, though I would definitely say that the extruder matters the most. Look up printable extruders capable of printing 70-85a TPU and they will be more than enough to print any material reliably and at speed
I don't know what shots anymore, the fact that this channel only has 165k subscribership, or the fact that you do NOT look like a Gearhead and can do things that are pretty much considered an art! Subscribed!
Awesome to see @theMountainGoats being appreciated!
Wow there is some amazing detail in this! I love all the 3D printing effort to detail too
It’s awesome. Keep the updates coming. I don’t care if the video production isn’t “up to snuff” I just wanna see content. Love it.
@2:47 The normal way to do this is to use the same value resistor linking adjacent legs so that your voltage divider has even steps, then put 0v on one end of the chain and 5v on the other end.
For example:
0v to pin 1
1k resistor between pins 1 and 2
1k resistor between pins 2 and 3
1k resistor between pins 3 and 4
1k resistor between pins 4 and 5
1k resistor between pins 5 and 6
5v to pin 6
Signal from sweeper pin.
This should give 0v at position 1
1v at position 2
2v at position 3
3v at position 4
4v at position 5
5v at position 6
If you want some error banding then you can also put a resistor inline on the 0v and 5v lines to condense the steps further. This will allow you to program an error function if either 0v or 5v is read on that input.
0v to pin 1 with 1k inline resistor
1k resistor between pins 1 and 2
1k resistor between pins 2 and 3
1k resistor between pins 3 and 4
1k resistor between pins 4 and 5
1k resistor between pins 5 and 6
5v to pin 6 with 1k inline resistor
Signal from sweeper pin.
This should give 0.71v at position 1
1.42v at position 2
2.14v at position 3
2.86v at position 4
3.57v at position 5
4.28v at position 6
If I could like this again I would. I have a 928 and looking to build a car with boxster parts
i'm so stoked you're back! thanks for sharing!
I appreciate your editing, stoked to see how everything goes
I appreciate it!
What really helped me print TPU was getting a filament dryer, TPU wicks up moisture really quickly from the air and creates artifacts and other print issues. I just got a cheap food dehydrator off Amazon and cut the racks until I could fit a roll of filament in there. I normally put the filament in for a couple hours before I print
Great to have you back on UA-cam. Looking forward to follow the ongoing work
This is so impressive. Rather jealous of your skills to build this thing!
Always awesome to see these vids. Can't wait to see this thing in it's full form!!
Was my guy listening to The Mountain Goats while filming? The Sunset Tree is an 11/10 album.
I think it's my favorite Mountain Goats album, and I absolutely was.
Absolutely awesome steering wheel build. Glad someone showed that a modern looking F1 style wheel can be made at home. I think another good route that’s less time intensive, but also less cool, would be to mount an AiM dash to a steering wheel
Pumped to see this revived! Steering wheel looks good though it's so close to the rest of the dash would be easy to accidentally hit switches and stuff while turning, I know all that's gonna change anyways but yeah, something to plan, a wheel looks good in it's initial position but you have to think of ergonomics mid-turn too and not hitting anything else with your hands.
Yeah, I didn't have room on the build now for gauges or dash switches or anything, so that's why it all moved to the wheel. Makes things easier for sure!
i'm an embedded software engineer, and the display is pretty dope. i've been designing a similar display for an F1 style go kart i have (made in the 70's). if you are still having memory problems, there are some tricks you can do. you can store some (or all) of the static image data like the lines in the program space using the keyword PROGMEM. if you haven't used it before, read the arduino page on it. you also want to live in the stack as much as possible, and use as few variable declarations as possible. in some cases, you can use the video memory to do calculations. if you have a chunk of video memory you know is all 0's or all 1's, you can write some math stuff in there, and then just put the 0's or 1's back when you are done (John Carmack perfected this method with Commander Keen years ago). Basically, try to avoid any alloc calls if you can, and reuse variable as much as possible. anything static, stick it in progmem.
you can also usually compress your graphics a TON. i noticed you have a lot of lines , and instead of storing a bunch of 1's in memory, you can store where the line goes and how long it is. typically storing compressed data in progmem is the way i save the most space when it comes to fancy screens.
you also might write your version of Wire.h that does the bit banging for I2C. you can often squeeze out some more space by doing this because the libraries have to be general purpose and often include stuff you don't need for this particular project. this goes double for screen libraries, which often have a TON of complicated and bulky code for screens that can rotate like a phone. The adafruit screen library has a huge amount of code that is useless if you don't plan on rotating your screen at all.
WOW! The amount of envy and admiration I have is unreal!
Took a peek at your code. IIRC, you don't need to redraw the whole screen every loop, you can update just the parts you want to change. In other words, draw it all at the start, or when you change modes, then just update the relevant bits. Hope this helps.
also, the pi pico is mostly compatible with Arduino libraries and is quite a bit faster, so you could look into that on top of loading images from the sd card to improve display speed.
I can’t wait for this car build. I wanted to create my own open wheel/f1 style car for the longest time, but make it the other way: instead of going back to 60s, I wanted to use as much banned tech as possible and make it modern. As I learned a few videos back, we’re the same age, so looking at your progress, I should get on it 😅
This is insane and I love every second of this video. I can do the mechanical work but I’d definitely need to learn the electrical side of this but it’s absolutely stunning. Best of luck I love this idea!
Keep it up, you rock! looking forward to see the journey of finishing this amazing idea!
Great to see this project back on the front burner. I truly hope you reconsider the Top Fuel dragster roll cage and go with something a little more period-correct, just taller and braced better.
Great to see you back and on this build.
You’re open-sourcing it! That’s awesome! So awesome!
Always, It's boring if only I can use it.
Right on Wesley! Happy for you mate keep it up 💪🏽🏁
Thank you!
The really crazy part about all this is in the end you likely could have just built a second car and kept the Boxster racer. Evolved instead so survival of the fittest I suppose! ;) Awesome video as always lots of insights.
Great work looking forward to the missing videos and the next installments
Love it and your content, cant wait to see how you get on with it :)
Thanks so much!
That is an awesome steering wheel!
For good results with tpu, the best thing I’ve heard to do is get the straightest and shortest length direct drive extruder as possible. Things like ender 3’s are especially bad at tpu because of their Bowden tube setup where the filament has room to squish. Hope this reaches you and is of help 👍🏻
Hey man, just wanted to say you're one of my favourite home build engineering creators. Your videos are so, so great. Up there with the best, and I can't wait till your channel blows up.
Noooice. Good to see this project progressing!
I actually started a project like this myself, so this is perfect timing
Please post those stored videos! Would love to see them all. And great update for this mega project.
you can use momentary switch on the shifter for the gearbox to send signal to the arduino. my bike (honda supra x 125) use some sort of rotary switch on the shifter, which is pretty much the same
Very excite you got Bobby on board for this
I remember seeing a Honda V-12 F1 car at Watkins Glen in 1968 or 69. What I recall is that it was, by far, the best sounding car on the track. It had those insane blending harmonics at 9 or 10 thousand RPM. Yeah, I loves me a V-12 sound.. LOL
You can get good results with tpu. I have a Bowden setup and adding a PTFE guide inside the extruder stepper to constrain the filament made the biggest difference. I also dropped the retract setting down to absolute minimum.
this has got to be one of my favorite builds ever
There's some great ideas in this reboot, the trick is to finish it rather than constantly iterating individual parts. Thanks for sharing Wesley, I'm looking forward to seeing the build progress!
Why aren't you using deutsch connectors? Also, EEPROM for storing your graphics off-chip. Cheap and easy (SPI/I2C/Serial IIRC). Also also, why not use a CAN bus? That tinsy you got supports it IIRC.
Awesome project, welcome back !
Nice to see this project going again. Don't think I've ever seen an open wheeler with a dragster style roll cage. Aesthetically it looks a bit odd, but I suppose safety was a higher priority than looks here.
Keep up the good work!
Another potentially useful habit to have with I2C communication, or most embedded programming, is to stay away from transmitting text whenever possible. It looks like you are sending a raw concatenated string of values that you then unpack into your various numerical values. If instead of the ASCII text you are sending as a string, you just send the value as a byte you will see something like a 2-4x decrease in the data sent. Just convert it to a string when you need to print it. For example the example string you showed on screen is 14 bytes. But if you sent the values it would be 6 bytes. (each value besides tach is a 8 bit integer value, tach could fit into a 16bit integer).
Your code seems mostly fine, a lot of optimizations are not worth the hit to readability. But trading ASCII chars for integer bytes usually involves little change to the code, and often removes bottlenecks like you mentioned running into around 32 characters, so it's usually a good trade to consider.
That's a really good point- I didn't explore that, as I really am not sending _that_ much data- with everything I can't see much more than 50 characters. Still substantial, and I'll do some testing with switching them over. This was easier to read during testing.
@@WesleyKagan yeah that is one advantage of sending ASCII, is it can be easy to read in some situations. I fall back on that for a lot of web requests and such, especially since in those settings I am not limited by compute capacity usually.
Usually you can do a print command that will print a integer/float/ect value as text directly. For example printf(“d”, value) in C which Arduino essentially is. It adds an extra step, but is simple enough that it can become habit.
I have found that I often work best by writing the base fragments of my program first, such as a function to handle packing and unpacking whatever data I am sending, then a function that takes the packed data and sends/receives it, then whatever functions actually use that in the logic of whatever I am trying to do.
I test each as I go, and once I am convinced they work I treat them loosely as a “black box” that just works.
That being said, that is just what I find works well for me. YMMV
love the project,
did you think about using CAN for the dash ?
ms3 supports it, and you wont be limited by bandwidth or cable length (plus its only 2 wires from ecu to dash)
Run it please with on board camera. Youre the man. Thank you.
You can use a cam lever for the rear pushrod.
Open top steering wheel AND H-pattern gearbox?? You’ll go back to the round wheel or find a paddle-sequential kit for it in no time
So cool to see you do a project similar to what I've been working on, as far as the arduino and screen go, Been working on an oil temp and pressure on a similar Nextion 320x480 screen. Not sure if their GUI works with the Adafruit, but the Nextion Editor program is super helpful on creating those graphical backgrounds. Looking forward to the next couple videos of this!
I'll have to check that out! Honestly my method of lines and bars did okay, but it's definitely a bit heavy- code wise. Thanks!
@@WesleyKagan Oh this is lovely on code, it gets loaded on that SD slot and there's not much code to send the data to each box. You can even change designs and not touch the code as long as you keep the box ID's the same.
@@WesleyKagan I was going to recommend the nextion as well. Besides the editor the other benefit is there is an existing library of dashboards that the sim racing community has created. It should just be a matter of tweaking the identifiers in the editor to those the Arduino is using... at least in theory.
Great to see this build come back! Can't wait to see what you accomplish.
Thanks for watching! It's going to be fun!
Great to see a new video. Fantastic project as always.
Many thanks!
Quick recommendation, to get around the printing challenges with that material, you could print yourself molds and then cast your forms with material from Smooth-On. There's also expanding foams on the market now with many density options. You could always wrap the handles with something nice like Alcantara. 👍 cool project 😎
Dude. Great job. Your work is awesome. Keep it up👍👍
For printing TPU, my experience has been that you want a direct drive extruder, print relatively slowly and very importantly calibrate your extruder steps. I need to up the flow by 30% compared to PLA because the TPU gets squished on the extruder gear giving it a smaller apparent diameter. The bed doesn't need to be heated, cooling is dependent on what you're trying to print. I've only ever used NinjaTek NinjaFlex on a modified direct drive Ender 3 Pro for reference.
Also, if you want a part you can compress check the infill types, they have different properties in different directions. In cura there is a plugin that gives a lot of explanation about all settings including infill, including what their strengths and weaknesses are. For example, the cross-3d infill gives similar resistance in all directions.
Hey Wesley, thanks for the content.
To help save memory onboard you can buy a nextion screen. It does the touch controls onboard its own μc. Has a super easy interface to create the GUI. You can use pictures and such. Also it only uses a TX and RX wire. As an ME that had to make HMIs for my projects it made it super simple.
Using a teensy could make things easier they have an onboard CAN and they run like 40x faster than an arduino. Much more memory. Downside 3.3 V logic.
Going to CAN could help you use a through bore slip ring since you only need a CAN high CAN low and power and ground.
Just my 2 cents hopefully something helps.
Teensy seems to be popular- I actually ordered one last week but I have a bucket of Arduinos so it's what I know for prototyping. CAN is seeming like the way to go as well, and I agree. I tried some wireless protocols but didn't have a ton of luck with a solid connection. I'm going to have to look into the nextion screen, that's seeming to be a much better resource for the graphics side. 3.3v would take some work.
Absolutely super cool, keep it coming! The steering wheel part is really interesting since I’m working on the same type of system for my RSX project car. Printing TPU as we speak and yes it is a mess, but I think I’ve almost got it figured out.
That 83 Interceptor.... it and my same-but-in-blue one were the first real superbikes and still the prettiest to my eye.
Always a real treat when I see a new video of yours pop up! Great work as always, looking forward to seeing the progress continue.
What a fantastic project , can't wait to see it finished.
Great video! Have you looked into a CAN-network as an alternative to the I2C-bus? It is also only 2 wires (differential pair) and should be able to run through the whole car. There are shields for arduinos. We use CAN in the Formula Student (like FSAE) team I'm in!
So happy to see a new video!