How the #speedboatrace helped me slash time from quality prints | Rat Rig V-core 3 Part 5

in this final part of the rat rig v core 3 build series i'm going to take you step by step on how i managed to print an eight and a half minute benchy and why you might consider this challenge to learn how to slash your print times without any loss of quality [Music] that is a lot of benchies and they were required for my final installment in the rat rig vcore 3 build series if you want to get up to speed the full playlist is linked below this episode is about pushing the limits as far as i can go to print a thready benchy as fast as possible and even if you hate 3d benchies this should be enough here to help you learn and improve your own 3d printer but first let's answer some more of your questions a common question was can input shaping be used on my printer a common misconception seemed to be that this was only available for the right rig but it's actually a feature of the clipper firmware and that means it can be used on any printer that's running clipper to test this i've got everything i need to clipperize another machine and try input shaping with the target being the cr-10 max which has the biggest heaviest bed of any printer i own look out for a two-part series on this in the future how much did this build cost now that the printer is finished we can answer this question i've put everything in this spreadsheet and i'll link it below in the description however you should know that this is only a guide and the build is going to be different for everyone i've paid for everything here except the freight which was free from rat rig also free for me was the petg filament from my sponsor x3d other prices are gonna be approximate for instance my orbiter extruder which i ordered two off from aliexpress and that included tax and shipping in this spreadsheet i've simply halved the total price so if you order only one your price may vary you'll also notice that in australia we pay through the teeth for a lot of these things and if you're in north america ordering through amazon you'll probably get a lot of this stuff much cheaper with all of this in mind let's show the final price building the printer to the spec that you'll see in this final video cost me almost 2 200 australian dollars that's about 1650 us dollars and 1400 euros as i said this will differ for everyone so mileage definitely will vary on to printing a 3d benchy as fast as possible and we're going to start by covering what the challenge is the speedboat race was made official by annex engineering not only is it fun but it really helps you identify and then fix weaknesses in your 3d printer at the bottom of the nx engineering video is a link to a discord server and that's where we'll find the rules you can pause the screen if you want to see them in detail but all you need to know is that this is what i followed i first became aware of this challenge thanks to my patrons who linked me to this horrendously impressive attempt from miraj c creator of the heborg the only other video i'd seen heading into this was from stefan of cnc kitchen attempting this challenge for me what makes his video so great is that he had many challenges and failures and did an excellent job of breaking down the problems as well as testing solutions from the start my goal for a fast benchy was to do it in under 10 minutes i had issues and bottlenecks that i had to identify understand and then solve in this video i'm presenting my journey organized into chapters which you can access with the timestamps below i have to give a big thank you to michael schmidt here member of the rat rick team for this prophetic post on my last video as it turned out i ended up doing everything that was suggested starting by strengthening the shroud it looks similar but you can see there's a lot more material up at the top and all of the walls are actually thicker the first phase for me was to add stepper motor driver cooling and then optimize the performance of the drivers i needed part cooling for the stepper motor drivers and this is what i came up with it's to suit an skr pro version 1.2 and as you can see it clips around two of the mounts and holds a 5015 blower fan which outputs over the stepper motor driver heatsinks in particular x and y which work the hardest that leads us to the speed and acceleration settings that are set up by default for the rat rig vcore 3. as you can see we have a linked file called speed limits basic and when this is running it limits the feed rate to 200 millimeters per second and caps our acceleration at 1500 millimeters per second per second we can significantly up the ante by uncommenting the speed limits performance file making sure to disable the original reference if we examine this file we can see that all of our limits are raised but our stepper motor drivers also optimized with increased currents and an earlier transition to spread cycle mode when we save and restart we can see that our printer limits have the velocity up to one thousand millimeters per second and our two acceleration values to ten thousand so by adding active cooling to the stepper motor drivers we can work them a lot harder and push the performance limits which brings us on to pushing and pushing and pushing speed and acceleration let's have a look at a typical slicing profile we have a default printing speed or feed rate here set to 100 millimeters per second but we can see that our outline is actually only half of that and solid infill 80 percent of that and if we switch to another tab we can see that the first layer speed is only 30 percent of that once we slice the model and color it based on the feed rate we can see that red should be our top 100 millimeters per second but when we inspect the g code there's actually very little that's printed at this highest speed one way to speed up your prints is to up these values to get them closer or even match the base print speed there are definitely gains to be had with speed but the real focus should be acceleration if you've cranked up the speed in your slicer profile for most people your 3d printer is going to be like a land speed record car the top speed you've specified is high but not necessarily optimized for acceleration let's keep this simple our car is on a road with the designated standing start point and a designated finishing spot so here's the thing alfredi benchy is a pretty small print with a longer straight section around 30 mils back to our analogy it's like giving our land speed record car only 100 meters to get the job done with its modest acceleration the car begins to speed up but gets nowhere near its top speed before it has to hit the brakes and come to a standstill if we're going to print something as small as a benchy fast we need to massively increase our acceleration back to our short test track and the top fuel dragster with its stupendous acceleration can still hit top speed over the short distance analogy aside how can we work out the actual speed and acceleration we require for our 3d printing this excellent acceleration calculator is found on the prusablog website it lets us put in some parameters and then communicates what would actually happen let's take our typical acceleration value of 500 and a typical feed rate of 60 millimeters per second our benchy however is only around 30 millimeters long in a straight line so we can see from the graph the printer is sitting at top speed for a small percentage of the time let's say we bump up our print speed to 100 millimeters per second we can see now that we're only just hitting top speed if we go up to 120 we can see we pretty much spend the whole time accelerating and then decelerating just like our example graph with the land speed record car of course putting this number higher makes no difference if we push the speed all the way to 200 we can see from the graph that's irrelevant because it'll never get there to fix this what we need to do is up our acceleration let's double from 500 to 1000 there's no flat line so that's still not enough we'll double it again and we can see with a value of 2000 we're finally able to hit our top speed just not for very long the more and more aggressive we make the acceleration the greater percentage of time for our move that will be at top speed and obviously the higher average top speed the faster the print will complete calculated theory aside what do these settings actually mean for the print time of the benchies all of these benches were printed in pla with the exact same g-code setting the feed rate to 200 millimeters per second before starting each one however i use the printer limits tab to manually override the acceleration values and this gave me a really easy way to run the same g-code but compare acceleration back to back benchy number one had a sedate acceleration of 500 millimeters per second per second and we know from our calculator it will never hit the top speed and that's why it still took over an hour to complete benchy number two up the acceleration to 2000 and half the time to just under 34 minutes benching number three up the acceleration to 5000 and that slashed off another 10 minutes the next version was the first time i got under 20 minutes by upping the acceleration to 10k remember what you're seeing here is the exact same g code and feed rate the only thing that changes is the acceleration and with this we have a 60 reduction in print time this challenge is not about print quality but inspecting the benches will highlight our next bottleneck we can see as the print speed goes up the part cooling just isn't up to the task luckily for me pre-designed upgrades were available as you'll later see it's important to improve part cooling even for filaments besides pla the standard part cooling solution on the rat rig is actually quite good as evidenced by the test prints that i showcased in the last video however for these speed benches we need more the one that i went with was this dual 5015 part fan mod by super sniffles we add a second fan to the back and we have an additional outlet at the back of the nozzle and for anyone who really wants to push it michael schmidt has a 7530 part fan mod remember that the ever carriage is modular so all we need to do is print out the three new parts to replace the three old ones therefore installation is simply unbolt the old and bolt on the new with the hardest part being soldering on a second fan connector in parallel for the new fan one of the cable management parts was too short so i had to remix my own simple variation and that allowed me to get everything back together the way it was nice and tidy and here we can see the new duct is pointing exactly where it should around the nozzle so back to printing and i upped my feed rate from 200 to 220 and kept the same acceleration and that shaved off another minute and a half and the new part cooling system worked really well with a significant improvement in surface quality despite being a faster more challenging print and i probably could push things a little bit further with the 7035 mod but pla is inherently a running material and needs a lot of part cooling so instead i decided to try some different filaments i used pla for 90 of my 3d printing but it was time to move on and first up abs the main advantage being that typically you would print it without any part cooling except for bridges things started well with the initial layers bonding well to the bed but it wasn't long into the print that we ran into a limitation and that was the part warping even on a print this small the base of the benchy was warping enough to make it dislodge from the bed next up was asa similar in properties to abs but less prone to warping and better uv resistance and compared to abs it was a lot better although not perfect as you can see once it got to the slight overhang at the top of the cabin it would curl up and then the nozzle would smash it off when it got to doing the bridging and solid infill even so i let the print finish just so i could see what time i was down to and with 300 millimeters per second an acceleration of 15k we were down to 14 and a half minutes with pretty good surface quality but that was irrelevant because if the benchy doesn't finish then it doesn't really count so i needed to keep experimenting the next filament i tried turned out to be perfect volcano pla from form futura basically this is a more expensive pla that prints with higher temp has higher thermal stability and is designed to be annealed for greater strength for speed benches here's why it's so good firstly you can print at 100 of your normal feed rate for the first layer i've found that despite this intense start to the print the bench is still stuck extremely well to wham van pex sheet the other thing was that despite being printed hotter than regular pla the exterior surfaces of the benchy looked a lot better with less deformation and sagging don't get me wrong part cooling still needs to be on max any time you're printing this fast but this filament is less runny than pla during this phase i did encounter my next bottleneck in testing these other filaments i found i had the need to upgrade my heater power as well as my hot end flow that upgraded pipe cooling system will inevitably blow across the nozzle and make the job of the heater cartridge much harder when i first put together the rat rig i reused an old e3d v6 i had lying around and that included an e3d thermistor but most importantly an e3d 30 watt heater cartridge what i was finding with the high attempt filaments is that there simply wasn't enough heater power to maintain the required temperature and have the part cooling fan on at 100 at the same time in fact when i tried to pid auto tune for these conditions the hot end struggled to get up to temperature and then had an error because there wasn't enough power step one was to steal the 50 watt heater cartridge from one of my other printers in step two was to change to a more high flow hot end that being the slice engineering mosquito magnum to go with that i had an e3d nozzle x also meant to be suited to high flow printing the other thing i purchased because i couldn't find any from slice engineering was a silicon sock from triangle labs on aliexpress this was an easy swap because the rat rig runs the ever modular carriage so to convert all i needed to do was print the parts for the mosquito i assembled the hot end the torque wrench doing its thing with the satisfying clunk and i'm happy to report that the triangle lap silicon sock is a perfect fit for the mosquito magnum as usual the ever carriage parts were a really nice fit and the modular design of the evercarriage made switching between the two hot ends a pretty straightforward affair it's really nice to have these upgrade parts already designed and easy to implement after running a new pid auto tune and recalculating the z offset i was often printing again and thanks to the more efficient combo i actually found i could increase the speed but print 15 degrees cooler print after print i raised the feed rate and acceleration until i got it up to 400 millimeters per second with acceleration of 30k and this got the print time down to 10 and a half minutes given how fast this was printed i'd say that this is a pretty impressive benchy there's some issues in the corners for sure but that new part cooling system is really working well and overall you can tell it's a benchy which satisfies the challenge criteria i was getting so close to my target with pretty clean results and one thing i've been chipping away with the whole time was slicer settings when it comes to this part of the challenge i'm definitely still a noob but here are some tips and tricks that i know firstly you gotta pay attention to the rules having the maximum layer height possible and the maximum line width to cut down on the amount of extrusions to make up solid infill we also want to stick to those minimum walls top and bottom layers and infill you'll likely find that each slicer has some quirks such as simplify 3d adding these little bits of infill at the front of the boat and that means wasted print time so for this challenge i actually switched to prusa slicer as i can put in a really small threshold for solid infill to minimize the areas where it's printed unnecessarily we want to turn off all skirts we want to make sure our speed is not lowered in any case and we absolutely want to make sure there's nothing on like auto cooling which as it states will slow down the print time for each layer if it ends up being too short if your slicer has settings for quality there's a good chance you want them turned off and you want to set your scene position to random or nearest to keep the travel distance as short as possible finally you want your attraction to be as minimal and fast as possible a combination of these things and probably many more i haven't discovered will help shave a few minutes off your print time and when we're printing this fast the estimated print time is completely bogus because the acceleration values predicted by the slicer are way too conservative i was knocking on the door of some minute benches and what got me over the line was remembering one setting that i hadn't actually touched yet the magic setting for me was square corner velocity which is similar to junction deviation or jerk in other firmwares let's go back to our acceleration example of getting from a to b except this time make it more realistic by making our dragster travel two segments with a corner in the middle with our previous understanding of acceleration we would decelerate to a complete stop for the corner before accelerating up to top speed again clipper has a setting called square corner velocity setting it to zero will completely stop the nozzle at the corner and setting it above zero will allow us to maintain velocity around corners if we go back to our analogy and this time apply square corner velocity we can see that we maintain more speed around the corner the more aggressive we are with this setting the faster that will corner and that means we're able to maintain top speed for longer either side of the corner too all this time i was still running the default five millimeters per second so the obvious thing to do was to edit my start g code and change this value to something higher i started by raising the value from 5 to 10 millimeters per second and this smashed through the 10 minute barrier coming in at 9 minutes 20 quality inevitably suffers but there's still more to come i upped the square corner velocity this time to 30 millimeters per second and this took off another 58 seconds rather than push harder i decided instead to tweak the settings and slow things down just a little bit with this high attempt reprint coming in at 8 minutes 31 comparing the two makes for an interesting diagnosis on the left hand side we have gaps in the hull where there was under extrusion in the fastest movements the hot end being 15 degrees hotter on the right fixes this but then we lose some clarity in the corners because the part cooling can't keep up when you're trying to print as fast as possible you can never have too much part cooling a quick edit here to respond to some of the comments after i posted up my entry video firstly this is no doubt strenuous on the printer but so far it's been very reliable and the rattling you can hear is not from the machine but rather everything moving on the bench underneath secondly some people didn't understand the spirit of the challenge and take it as a competition instead i mean there is a leaderboard and my entry is respectable at best compared to the sub 5 minute bungees we're seeing but it's not really about outdoing other people but rather learning about your printer and then sharing and collaborating with others my machine is close to stock and i could modify it a lot further but i'm going to try and be sensible and stop here even if you don't undertake this challenge you should know that those at the forefront are pushing things along and their ideas and solutions might trickle down to other printers in the future success finally after quite a lot of benches and you might be wondering why go to all this trouble so what exactly is the point of this challenge so we push our printer to go crazy fast and see what its limits are and all we get in return are some ugly benches it might seem like that but it's actually a lot more valuable after the upgrades to the machine and applying everything i'd learnt this is my new standard of benchy its base feed rate is 200 millimeters per second with acceleration of 5000 millimeters per second per second you'd have to say it compares favorably to the first bench you ever printed on this machine without much tuning when we reveal the print times it all makes sense i'm printing in 59 of the time with no loss of quality this new default print setting gives me fast beautiful prints no ringing no stringing no bad artifacts and to put it into context the whistle on the right was downloaded remixed in tinkercad to add the lanyon holder sliced and printed in half an hour recently i surveyed 3d printer users on what they actually wanted and the clear trend was that people wanted to print faster without any loss in quality in part due to this fantastic machine and in part due to undertaking this challenge i feel we're taking massive steps towards this goal let me know in the comments if you agree thank you so much for watching and until next time happy fast 3d printing g'day it's michael again if you like the video then please click like if you want to see more content like this in future click subscribe and make sure you click on the bell to receive every notification if you really want to support the channel and see exclusive content become a patron visit my patreon page see you next time

Channel: Teaching Tech

Views: 96,834

Keywords: 3d printing, 3d printer, 3d printed, 3d print, rat rig, #speedboatrace, speed benchy, quality, speed, fast, step by step, how to, guide, tutorial, lesson, klipper, v core 3, acceleration, part cooling, feedrate, jerk, square corner velocity, junction deviation, retraction, raspberry pi, cost, budget, price, noise, pla, abs, asa, volcano pla, filament, slice engineering, prusa slicer, tips, tricks, settings, parameters, mosquito magnum, silicone sock, triangle labs, ringing, input shaping, marlin, rrf

Id: 6kRjdprTjFc

Length: 21min 23sec (1283 seconds)

Published: Fri Jul 02 2021