Uhh so to clarify, without your custom made ram guard, it was using the cooling fan to push prints off??? I've been using metho or IPA to clean the same garolite sheet for almost a year now and it's good as new using default settings, im not sure that custom coating is doing them any favours. Super appreciate the review! I still prefer to change by hand but a print farm application would be a different story.
Yup, they're not including any sort of ram or shield with the kit. When the bed releases like it's supposed to, the parts slide off very easily, so it does work under ideal conditions.
@@MadeWithLayers still not optimal! Doesn't deal with purge lines or skirts either, though you could disable them. Still reckon your bed swapping design is the safest approach but way more challenging to implement 😂
Purge lines are dealt with by... not using them. The instructions call for always printing a brim (or a zero-distance skirt) instead, which works for priming, but then you've always got a brim to peel off.
At my previous workplace, we used three MK3S printers to print enclosures for our custom made electronics for our clients. I would create two G codes for the printers: one which would run for 8 hours (or a bit less), and the other which ran for 16 hours (or a bit less again). This way the printer was running 24/7 during weekdays, and a colleague would start the 16 hour G code on weekends as well, since he was every day at the office, even on weekends, at least for a few hours. To equip all three printers with a Qinly would cost us 1650 dollars, or 1620 euros. I could have two more MK3S+ printers for that, or four clones, and still have enough money left over to buy a Raspberry Pi 3B+ to run a second instance of Repetier Server Pro (which we've bought) and some original Prusa sheets (if we'd go for the clones). It would be better, since there's no need to tweak profiles and parallel printing is faster and more flexible.
I tried this but the gcode fiiles got to be so large my text editor started to complain. Why the heck marlin can't add a qnty=X option I'll never know.
Thanks Tom, I used one of these to create parts for a local 3D printer part manufacture because they were in a pinch. It was a part they used in packaging. I used mine on an Ender 3. After I got it dialed in, it was set it and forget it. I lost about 3% of the parts due to first layer inconsistency, but I ran it for weeks, turning out over 400 pieces and I didn't have to do anything at all. I would do it again for sure.
FYI you can also do this with the Prusa Textured build plate. I run a 3D printing business, and we have about 8 machines using them. They work great. We find that 70 degrees results in zero issues with PLA sticking, and prints pop right off around the 25 degree C mark.
@@TheJustinist you can't search for anything other than phrases. Negative search is good, but without an ability to filter out categories, licenses and meta and not being able to even sort the results, it's just another useless feature.
Real talk, Tom: I love that your reviews are so honest and unflinching. I hate how objective information gets twisted for profit, so massive kudos for sticking to your integrity on these!
As a proof of concept it seems really promising. Get a G10 plate, hook up a gcode operated fan for rapid plate cooling and some kind of air cylinder with a bumper to push parts off the plate at a less aggressive angle and it just might work.
In the beginning of the pandemic, I was printing PPE face shield headbands in PETG. I've got an Anycubic Chiron with the Ultrabase bed. I had built a temporary enclosure to aid with energy usage reduction, and I set up a 24 hour job that would use g-code and an Octoprint plug-in to trigger a gpio relay to power a cooling fan mounted to the side of the enclosure between parts to assist with bed cool-down to release the part (took about 5 minutes), and had a small bumper attached to the toolhead to remove the part, and a paper guide sheet to direct the parts to a basket sitting on the floor under the printer. Worked reliably like this for weeks churning out hundreds of headbands.
You're totally right about the price / convenience aspect of deciding whether something is worth the money, trouble free solutions can be worth their weight in gold. Great sponsor, Mk3 owner for 4+ years which is still going strong, so very glad it's my first printer.
don't forget to DISABLE pruge lines, brims, skirts and basically everything that needs supports or a rafts (as rare as rafts are, they are sometimes useful)!
Great video, Thanks Tom. Wow, that is crazy expensive. Our Autopilot program is only $50 or free with MatterControl Pro. And the bed surface that auto releases is just Garolite with a scratch improvement coating. I did a ton of testing on this surface, it is not proprietary. You don't need to tilt your printer, just add a flexible plastic sheet to the front and that's it.
Mingda Rock3 Pro has an amazing print bed. It's textured glass and DOESNT have some odd sticky coating preapplied to it. The result is a bed that sticks when hot, releases easy when cold. Mine is perfectly clean after thousands of prints unlike a lot of other beds which always have traces of prior prints left behind.
The continuous print plug-in for octoprint is great. You can even manage multiple networked printers running non-stop from a single instance of octoprint.
I reviewed the Quinly system last year and had many of the same issues as you. While I was ready to accept the finicky settings, I'm sad to see that the software still is as unfinished as when I tried it (admittedly it was a beta back then, but doesn't look much different from what you showed). The release times also change dramatically depending on your ambient tem. I tried it during winter with ambient around 18C and it only needed around 15min. But in summer it took often an hour or more to finally cool down. They argue that most printfarms would be climate controlled anyways but who wants to pay to cool to 18C when 23C would be fine otherwise too?
I remember seeing a reddit post where someone built something similar but with an arm that came down to remove the parts from the bed. Seem better than have you x carrige crash into your prints everytime.
Doesn't hurt the printer much to be honest. The printer is pretty rigid and the motors are no where near powerful enough to do any real damage. If it's too much the motors just stall and skip steps.
Considering the price and the reduced build volume, it's probably worth pointing out that for the price of a mk3 plus this kit you can also buy 3 Prusa MINIs. If your goal is rapid mass production of sufficiently small parts, I would guess that that's the better option.
I did the same with octoprint and the "Continuous Print" plugin. The Creality glass build plate worked great with PLA, the textured Prusa build plate with PLA and PETG. No need to tilt the printer or anything. I just built a slide for the parts kicked off the bed.
seriously i've been doing this exactly like this for around 5 years. you don't need a fancy bed just use glass. I don't even have the printer at an angle, just use the hotend with a shield to push the print off when it's cooled down. you put the cool down gcode in the post script in octoprint. bit of python and you can grab orders and print it. I've not touched my printer for around a year, just walk in the room, pick up the prints off the table, clean them up and post them. job done
I agree that out of all my printers, my first choice for “it’ll just work” is my prusa. Also, it has the best bed out of all of them, and any solution that gets rid of that would be a non starter for me. If I were to emulate this system, I’d just use petg and the textured plate, octoprint (as you suggested), and your guard in addition to the printed bits they have used, and I think that would work consistently due to the petg always popping off the bed when it gets cool enough
I just got a kit in too, great to see your thoughts Tom. I can definitely see that ROI like you said, but yes, it is not cheap. Had the 3D Que team on a podcast and I definitely see the value in the product but yes, very specific use case. Love the ram, will have to do something like that for ours!
So in the end, you get about 20% increased productivity (using the Skull and plate of Benchies as the more realistic scenarios for a print farm), for a roughly 50% increase in price. A second printer gets you 100% more productivity for 100% more money. Additionally, the productivity increase of the Que system is basically negative when you are in a tight deadline (and it's okay for you to babysit the printer 24/7 until the deadline), because you have to wait for stuff to cool down. So things don't really line up here... Thanks for the review though!
I feel like... the half hour Cooldown is very unimaginative; in that... seriously just add a PC fan aimed at the bed to help it cool faster, and use gcode to turn it off and on (like you would LEDs), and that would allow you to improve that cool down time really well.
Everyone have his own use case for a 3D printer. I usually design something and print 1, 2 sometime 3 copies of it. The design process requires prototyping, which requires printing parts of the thing to see if the measurement and clearance is right. Overall, I don't have any use for something like that. However, I appreciate that some people help to improve the technology by trying different solutions for some of those issues. We are slowly getting closer to a "public ready" solution. Although, I'm not sure if I will see it in my life time...
They should sell the print bed separately. A print bed that reliably holds all filaments and releases them after cooling for about 30 € or so would certainly be a success.
They refuse to sell the bed alone, which is really what you’re going for. The controller is a gimmick to justify charging more- the “sweeping at bed temp x” can be accomplished with end gcode very easily. I bought a CR6 SE kit, and it was obvious they did zero practical testing with it. They wanted me to stick the $300 sticker on top of the bed screws- sealing them in forever. It pissed me off so bad I haven’t tried using it. It’s a gimmick cash grab. If it worked as well as they claim, everyone would have it. It’s been around for years.
In that case, their offer would totally fall apart though, as then they are basically selling you a couple of aluminum extrusions (and printed parts that you might just as well just do yourself) for like 420€.
@@adriansue8955 the vaapr is not just a g10 sheet. Not saying g10 won’t work for some filaments but they have a proprietary blend of things that is different to handle a wider array of materials.
I mean.. with PLA, PLA plus, or PLA-F my plain old glass beds stick like glue at 60-70 and slide off zero effort at room temp. My old school original ender 3 bed can actually print without heat as if its blue tape. By dialing in the bed level I can get some of my products to stick just enough with zero heat that it can be bumped off easily by the fan shroud into a bin (using the old sheet o' paper taped to the bed to keep things from getting jammed under there.) With no bed heat up and down saves a huge amount of time. as long as there are no major shifts in ambient room temp I can keep running for a full spool no problem. Cost $0. OH! and these systems are running non-stop thanks to YOUR awesome tutorial on how to run octoprint on an old android phone ;) so much appreciated.
8:54 I like that the spaghetti mostly just slides down the slanted bed instead of building up around your extruder and eventually gunking up the hot-end.
Tried using this system for a while for an ongoing production gig, the firmware was awful, and it wasn't much special. The price was pretty obscene for what it was, and bed adhesion was awful. At best, we managed to get about an 80% success rate. The lift kit is something you can replicate a few pieces of wood to prop the thing up. Any prints that need brims are out, as they have issues releasing. Considering the cost, it was downright dumb to not throw in an actual bed. The 'surface' is basically a FR-4 sheet with a sticky back. Cooldown and heat-up cycles are an issue, as mentioned. Can probably build a fan to cool the bed or something, but even in a cool room, it was an easy 20-minute gap between prints. Certain filaments just DO NOT stick, even some relatively common PLA. The support Discord based support basically leaned towards cranking the first layer temps higher and higher. Had some luck at 230C, but that was damaging the bed and discoloring the plastic. After messing with the thing on and off for about a year, features seem to be frozen in limbo. As stated, there is nothing to run this thing remotely in a safe manner, no master control type program to manage the farm, etc. Given the price per printer for what's about $80 in stuff, I could understand markup for R&D and good software dev. This just feels like a rehash of an old idea with a small bit of polish and little follow-up. Take the cost of the 3DQue stuff and the printer itself, you're better off using a belt. Still can be dang frustrating, but cheaper and easier to get running. Better yet just get a large bed printer and print either sequentially, or a long full build-plate to maximize production between shifts.
I did the "clearing the bed with the machine" a couple of times while being remote from the printer. Printer was just horizontal like normal. And a normal bed. I haven't done it enough to automate it yet. For sticking to the bed: small prints stay put after printing, while bigger prints tend to come loose after cooling. Current print: 18 hours, 324 small items. I need to go there to take them off.
use a conveyor belt and a nuc/old pc + pico for prints wich is cheaper and more modular as controller and the belt gives it a really really big efficiency increase wich has been done before
I have used an MQTT plugin with Octoprint, Node Red, and a smart plug with Tasmota firmware to automatically turn off my 3D printers a few minutes after a job completes. That way if I am there I can swap print beds and run another job right away. If not the printer shuts down so it is not powered up for hours while not in use. Saves wear on the fans and power supply. With this system, you could use a similar setup with a fan pointed at the print surface to reduce cool down time between jobs.
Here's an idea for a Voron 2.4 abe system. Take a pei sheet without the spring steel, and make a clamp on both sides so you can pull it in both directions. Take a magnetic sticker and adhere it to the pei sheet. Have a metal build surface ( that is magnetic) so the pei/ magnet combo can lay flat as if it was just a normal bed. Once the print is done, have the pei/mag sheet move (with a bend after the bed to unstick the print) and then build a sweeping mechanism to get rid of the print. Return all to their original state and begin another print
If you're printing thousands of the same part, it's probably time to investigate other production tools. Injection molding at the smaller scale has gotten a whole lot easier and cheaper in the last decade, between being able to print molds with high-temp UV resin, or casting your own molds in high-temp silicone. Sure, some things simply can't be made in a mold, but most things can and even more can with some tweaks. And then there's small-scale CNC, either to produce parts directly or to make aluminimum molds for injection molding. You can even recycle quite a few things from the waste stream to do that injection molding for free or nearly free.
Injection molding for a few hundred or even a few thousand isn't as cheap as you think. This greatly depends on the complexity of the model and resulting mold. Source: I've been professionally molding for years.
@@MastaSquidge True for commercial injection molding. But home hobbyist-level injection molding via non-machined molds has taken off in the last couple of years.
@@oasntet which falls back under the complexity of the part. The hobbyist can of course make multiple parting line molds and manually assembled cores and undercuts but at the end of the day it is still really only cost/labor effective with simple parts. And that's perfectly fine if that's what you're doing. Labor isn't free and 3d printing is mostly unattended time. If you go get a creality cr30 or sainsmart infi you don't need any fancy kit to do continuous 3d printing. It'll batch and remove prints from the bed right out of the box.
The idea looks great but in practice seems to need a lot of tweaks, and a RAM 3d printed to push off those parts. Think I'll stick with the plate changes for now. Thomas you did give us the honest truth and thanks so much !
I have the 3D Que Autofarm set up with 2 CR6SE printers and I love it! There have been several upgrades to the software since this review, which is pretty typical with reviews and new systems to the market. I do print different parts quite easily by uploading them into the queue. The main purpose of this system is to make it more efficient for Print farms. Which you do say in your review that it makes sense for people wanting to print on repeat. It has saved me a TON of time and money to use the system. They also do have a way to connect remotely so you can control things from the Autofarm dashboard. Still a great review, but I just wanted to point out that the developers have some a bunch of upgrades over the last 10 months. They are also very responsive if there are any questions or requests to adjust things, 😊😊
I did modified my Prusa mini's hotend fan duct, so now it has a lifter part (Thanks open-source!). The idea is roughly the same. The flex bed is taped to the heating panel just to be sure, and a sheet of paper between them in the front so the part won't fall into the mechanics. Only problem is, when i lift far from the supported side, then the whole x gantry twists and stays like that that makes the lifter's bottom lower, and catching into the print itself. However I tried it, and works as intented, if i set up correctly the temps and the cooling also. One note, that I used often. Turn down the bed temp after the first layer. It's already stuck, just need to maintain not to fall mid print, then when it's done you maybe need to wait third, or quarter times as originaly.
I was thinking of a print bed with release pins, like you get on injection moulds. Would be extra weight on the print bed, but you could use them only where needed.
Something else you didn't mention is that this printer setup only, is that it can only ram printed items that are above a specific height (i.e., the lowest rod holding the hotend).
Buy or build large 3D-Printer and print instances all at once or one by one in single print. Same scenario, but much more stable result with any fillament.
The bed looks promising, but the angled printer is overengeneered. I just printed a light slide I can attach to the front bed. Others used a sheet of paper. Works perfectly and is low cost.
Looks like a Pelletier cooler or blast of CO2 liquid would get the parts off more reliably as you need a difference in thermal expansion between plate and part to get it off. Also a thermal shock of active cooling would allow you to locally cool the part or bed quickly, and resume printing much faster while the thermal mass of the bed recovers operating temp.
Ya the bed type is a G10 or PEI bed, both do the same thing having the print self release after the bed is cold. PEI you usually want about a 10C hotter then suggested temp for any material FOR the first layer, you can then let it back off to 50c to finish the print.
Video idea for you, running klipper on non raspberry pi single board computers and/or old android phones. A lot of people are waiting on raspberry pi’s for klipper when rock/orange/banana/etc. pi’s may be good enough. I know you also did a video on running octoprint on android phones and it seems like it’s possible to do the same for klipper, though it’s a longer and more tedious process. Just some ideas! I think with the raspberry pi shortage and peak of klipper interest going on now there would be a lot of interest for low cost raspberry pi alternatives
You should be able to do this with marlin Gcode and a SD card. The code in marlin has a g code to load a program from a SD card and start printing it. With some simple edits to the end of your Gcode inside your slicer you should be able to handle this with just a base printer
For reducing the time between time because of the slow cool down of the plate, a common big fan that turns on after the print has finished would dramatically decrease this time.
I could be completely wrong because I’ve only had a printer for a couple of months, but I think the ender3_v2 has a similar print plate that grabs during the print and releases after the print is done. At least that’s kinda what I’ve noticed.
The Video was very interesting. I ordered an set for one of my prusa minis without a hub. Today I had the time to install all and i haven't any issues with the gcode transfer. Ok you can't change the quantity of the running que but you can extend the que with new quantity of the same gcode.
if they made it like $150, maybe $200 or $250 with the pi, I'd say this would be fine, but theres no way i could justify spending almost $600 on a raspberry pi and a few printed parts that a beginner could remake in fusion in a few hours,
heh, guess pushing sticky parts out is one case where my lead screw monster with closed loop steppers would get to flex it's muscles a bit! I've found out the worst way possible that if it ends up crashing on auto-leveling and then rams the head on the bed, it can just keep pushing the entire 310x310mm glass out of the 4 clips and crack the corner...
First generation set up. obvious place for improvement. Seems like a good thing. The biggest advantage to a system like this is someone who produces a product that uses a lot of smaller identical parts. It essentially can stop you from having to run to a printer and break your work flow. also. I only use one print material for the reasons you talk about. I don't care about other colors (i have this thing called "paint") and I'm perfectly happy with the mechanical properties of my print material.
In my eyes, I don't see how this idea can compete with something like the Creality CR-30, which does almost the same thing, using a conveyor belt, is more reliable in removal, and can even print parts much longer than the length of the bed, the only downside i see is the potential of having to relearn positioning for prints due to the 45 degree angle of the head. Otherwise, just put a bucket down, and the prints will peel themselves loose as the belt moves along.
Plus it has zero downtime spent on cooldown/warmup. If it stays calibrated, the only time it isn't printing is the time you're spending installing a new filament (until you need to perform maintenance).
I suspected something like that. I want to salute the guys at Power3D which I support, Carlo has a Prusa like that with a very similar setup if not the same, and he told me something like real applications for this method are quite an exception even in his experience of professional farm/shop. Note that from 72 to 84 benchies full plate, and from 5 to 6 skulls or single mid-sized objects, the slight improvement is the same, around +15%. So, in case, auto plate swap is the way to go, more than continuous print like that. An alternative even better could (I am imagining) be the conveyor belt printer which has other alternative use cases (big linear objects ofc) and yeah it comes with its very own set of troubles and tricks to handle, compared to a normal printer. Interesting study tho, I liked it!
2 года назад
It would be easier to just use something like the CR30 with looped gcode for repeating parts, or similar queueing solution. That at least is made for automatically detaching parts and prints long in one axis, which could be used for repeated tasks instead
you just need to put a glass and octoprint to do this... I use the creality bed upside down so I print on pure glass, at 70 or 75 ºC instead of 60, when it cools, the GCODE push the print out. only thing necesary is to use a good bed-nozzle distance raspberry 60$, glass bed less than 30, so less than 100$
i say it on every of your video, but a glass bed with kapton will stick when hot et during the cooldown of the bed you hear a little pop from the piece.
Where i see the problem is that, while parts will easily come off if lifted from a corner, there may be a high resistance to pushing them from the side and parallel with the bed like it is attempting to do. I tested this on my anycubic i3 mega which has a similar bed by default: try to push the part parallel to the bed and it is hard to remove it. Grab it from a corner and the leverage increases and the part easily moves.
Sounds like a better system would be to have a printer with a lift kit that prints close to level and can then tilt up at the end of the cycle AND a separate arm to clear the print off the print head. Not many people are mass producing parts to the level that that would require.
This strikes me as more intended for cutting labour costs - for a very large print farm that's churning out hundreds of the same part, not having to mess with beds will greatly increase the number of machines one employee can manage. Outfitting that many machines would be eye wateringly expensive, though.
It would be so rad if someone made a robot arm to remove prints automatically! would be even cooler to see a solution like this working on some kind of belt and pulley system with multiple printers
I think is a bad desing: pushing object system put so much stress on the gantry, after some time it'll lose precision. Would be better replace the bed with a belt and when the printing ends, just advance to the front, a sharp edge could release the pieces. And you use all space abailable without "dead spaces" to print.
This 'waiting for the cooldown' issue seems like a job for a thermoelectric cooler assembly. Rather than some PTC heating element for the enclosure heating I have been thinking about using TEC so that the heating affects room temperature less in the house. With that and the ability to automatically reverse the hot and cold sides after each print, you've got a speedrun version of this. Also I really don't see the point of a very expensive fancy coated thing when a smooth glass bed does 70% of that if you have some sort of motor to shove the whole print off when the bed reaches ambient...? Well it does help if it has a smaller footprint than having a motor shove prints from behind the printer, but this whole tilt the printer thing seems awkward at best.
There is one thing missing with the calculation, the jump in time cost due to hiring help. I still think that having a farm makes more sense - just wanted to point out that not all minutes have the same value when talking about users with different level of experience.
Meh… a conveyer belt system would be much better. That would allow for printing of long parts and not have to cool down in between prints. I’m hoping that Prusa will offer this option in the future since conveyer 3d printing is the future of 3d printing desktop manufacturing.
Uhh so to clarify, without your custom made ram guard, it was using the cooling fan to push prints off??? I've been using metho or IPA to clean the same garolite sheet for almost a year now and it's good as new using default settings, im not sure that custom coating is doing them any favours. Super appreciate the review! I still prefer to change by hand but a print farm application would be a different story.
Yup, they're not including any sort of ram or shield with the kit. When the bed releases like it's supposed to, the parts slide off very easily, so it does work under ideal conditions.
@@MadeWithLayers still not optimal! Doesn't deal with purge lines or skirts either, though you could disable them. Still reckon your bed swapping design is the safest approach but way more challenging to implement 😂
@@MakersMuse Yeah, I've done G10 on my CR-6 SE.
@@MakersMuse Pallet Changing is awesome. Can't wait to see more of it come into the 3D printing world from the CNC machining world.
Purge lines are dealt with by... not using them. The instructions call for always printing a brim (or a zero-distance skirt) instead, which works for priming, but then you've always got a brim to peel off.
At my previous workplace, we used three MK3S printers to print enclosures for our custom made electronics for our clients. I would create two G codes for the printers: one which would run for 8 hours (or a bit less), and the other which ran for 16 hours (or a bit less again). This way the printer was running 24/7 during weekdays, and a colleague would start the 16 hour G code on weekends as well, since he was every day at the office, even on weekends, at least for a few hours.
To equip all three printers with a Qinly would cost us 1650 dollars, or 1620 euros. I could have two more MK3S+ printers for that, or four clones, and still have enough money left over to buy a Raspberry Pi 3B+ to run a second instance of Repetier Server Pro (which we've bought) and some original Prusa sheets (if we'd go for the clones). It would be better, since there's no need to tweak profiles and parallel printing is faster and more flexible.
Running 8 ender3s beats a prusa mk3 with this contraption anytime for the same cost.
I tried this but the gcode fiiles got to be so large my text editor started to complain. Why the heck marlin can't add a qnty=X option I'll never know.
@@electronicsandewastescrapp7384 Get a better text editor :P
@@fastundercoverkitgoogle7381 this is not the way
I like the robot voice edits. I really don't like it when there are *CORRECTIONS on screen. I'm often listening while looking at another window.
For the price, I'm surprised it doesn't at least come with a tool head shield like what you had to design
Or at the VERY LEAST an STL file.
Thanks Tom, I used one of these to create parts for a local 3D printer part manufacture because they were in a pinch. It was a part they used in packaging. I used mine on an Ender 3. After I got it dialed in, it was set it and forget it. I lost about 3% of the parts due to first layer inconsistency, but I ran it for weeks, turning out over 400 pieces and I didn't have to do anything at all. I would do it again for sure.
Hey Chris, check also 3D Print Flow
FYI you can also do this with the Prusa Textured build plate.
I run a 3D printing business, and we have about 8 machines using them. They work great.
We find that 70 degrees results in zero issues with PLA sticking, and prints pop right off around the 25 degree C mark.
Have you tried active cooling the build plates to reduce cooldown time?
@@cinobro6393 I haven’t actually, sounds like it would be useful.
@@cinobro6393 I was thinking the same. A GPIO-controlled relay and a fan might cut that wait time in half.
PLA + textured = fail
@@jonhaze7537 Why are you having trouble with it? I've had ZERO issues.
Cool to see the prusa sponsorship! Hopefully more people move on to printables
Not until they fix the search
@@goury wdym? Search works fine for me. Can even search for exact phrases or negate search terms now.
@@TheJustinist you can't search for anything other than phrases.
Negative search is good, but without an ability to filter out categories, licenses and meta and not being able to even sort the results, it's just another useless feature.
@@goury that's called "filtering". Not search. So you're saying search works fine, but you want more filters...
@@ThantiK that's called "verbalism", not replying. I am saying that the search functon is useless.
Real talk, Tom: I love that your reviews are so honest and unflinching. I hate how objective information gets twisted for profit, so massive kudos for sticking to your integrity on these!
As a proof of concept it seems really promising. Get a G10 plate, hook up a gcode operated fan for rapid plate cooling and some kind of air cylinder with a bumper to push parts off the plate at a less aggressive angle and it just might work.
In the beginning of the pandemic, I was printing PPE face shield headbands in PETG. I've got an Anycubic Chiron with the Ultrabase bed.
I had built a temporary enclosure to aid with energy usage reduction, and I set up a 24 hour job that would use g-code and an Octoprint plug-in to trigger a gpio relay to power a cooling fan mounted to the side of the enclosure between parts to assist with bed cool-down to release the part (took about 5 minutes), and had a small bumper attached to the toolhead to remove the part, and a paper guide sheet to direct the parts to a basket sitting on the floor under the printer.
Worked reliably like this for weeks churning out hundreds of headbands.
You're totally right about the price / convenience aspect of deciding whether something is worth the money, trouble free solutions can be worth their weight in gold.
Great sponsor, Mk3 owner for 4+ years which is still going strong, so very glad it's my first printer.
I highly appreciate the length of the video :)
It is elite !! 1!!
don't forget to DISABLE pruge lines, brims, skirts and basically everything that needs supports or a rafts (as rare as rafts are, they are sometimes useful)!
Great video, Thanks Tom. Wow, that is crazy expensive. Our Autopilot program is only $50 or free with MatterControl Pro. And the bed surface that auto releases is just Garolite with a scratch improvement coating. I did a ton of testing on this surface, it is not proprietary. You don't need to tilt your printer, just add a flexible plastic sheet to the front and that's it.
Mingda Rock3 Pro has an amazing print bed. It's textured glass and DOESNT have some odd sticky coating preapplied to it. The result is a bed that sticks when hot, releases easy when cold. Mine is perfectly clean after thousands of prints unlike a lot of other beds which always have traces of prior prints left behind.
The continuous print plug-in for octoprint is great. You can even manage multiple networked printers running non-stop from a single instance of octoprint.
I reviewed the Quinly system last year and had many of the same issues as you. While I was ready to accept the finicky settings, I'm sad to see that the software still is as unfinished as when I tried it (admittedly it was a beta back then, but doesn't look much different from what you showed).
The release times also change dramatically depending on your ambient tem. I tried it during winter with ambient around 18C and it only needed around 15min. But in summer it took often an hour or more to finally cool down. They argue that most printfarms would be climate controlled anyways but who wants to pay to cool to 18C when 23C would be fine otherwise too?
I remember seeing a reddit post where someone built something similar but with an arm that came down to remove the parts from the bed. Seem better than have you x carrige crash into your prints everytime.
Yeah Saw that too looked way better
Doesn't hurt the printer much to be honest. The printer is pretty rigid and the motors are no where near powerful enough to do any real damage. If it's too much the motors just stall and skip steps.
Considering the price and the reduced build volume, it's probably worth pointing out that for the price of a mk3 plus this kit you can also buy 3 Prusa MINIs. If your goal is rapid mass production of sufficiently small parts, I would guess that that's the better option.
Oh that noise it made when it was trying to ram the part off made me shiver
the price is outrageous for what you get
I did the same with octoprint and the "Continuous Print" plugin. The Creality glass build plate worked great with PLA, the textured Prusa build plate with PLA and PETG. No need to tilt the printer or anything. I just built a slide for the parts kicked off the bed.
seriously i've been doing this exactly like this for around 5 years. you don't need a fancy bed just use glass. I don't even have the printer at an angle, just use the hotend with a shield to push the print off when it's cooled down. you put the cool down gcode in the post script in octoprint. bit of python and you can grab orders and print it. I've not touched my printer for around a year, just walk in the room, pick up the prints off the table, clean them up and post them. job done
I agree that out of all my printers, my first choice for “it’ll just work” is my prusa. Also, it has the best bed out of all of them, and any solution that gets rid of that would be a non starter for me. If I were to emulate this system, I’d just use petg and the textured plate, octoprint (as you suggested), and your guard in addition to the printed bits they have used, and I think that would work consistently due to the petg always popping off the bed when it gets cool enough
I just got a kit in too, great to see your thoughts Tom. I can definitely see that ROI like you said, but yes, it is not cheap. Had the 3D Que team on a podcast and I definitely see the value in the product but yes, very specific use case. Love the ram, will have to do something like that for ours!
HMU if you want the file for the ram!
@@MadeWithLayers Oh sweet, definitely will!
@@MadeWithLayers Should upload it to Printables!
It's now available on Printables! www.printables.com/model/240057-mk3-bed-clearing-ram-guard
So in the end, you get about 20% increased productivity (using the Skull and plate of Benchies as the more realistic scenarios for a print farm), for a roughly 50% increase in price.
A second printer gets you 100% more productivity for 100% more money. Additionally, the productivity increase of the Que system is basically negative when you are in a tight deadline (and it's okay for you to babysit the printer 24/7 until the deadline), because you have to wait for stuff to cool down. So things don't really line up here...
Thanks for the review though!
I was always fond of the conveyer belt 3D printer design. I think this need to be revisited.
Have You ever Looked into the Ender CR-30 ? Its a 3d Printer with a small conveyor belt as a bed. You Can print innfinte z axis
I feel like... the half hour Cooldown is very unimaginative; in that... seriously just add a PC fan aimed at the bed to help it cool faster, and use gcode to turn it off and on (like you would LEDs), and that would allow you to improve that cool down time really well.
For the cost of that, you may as well just buy Prusa Mini. Puts the manual swapping numbers back along side the Quimby
Great video, I really appreciate a non-biased viewpoint when it comes to videos like these.
Everyone have his own use case for a 3D printer. I usually design something and print 1, 2 sometime 3 copies of it. The design process requires prototyping, which requires printing parts of the thing to see if the measurement and clearance is right. Overall, I don't have any use for something like that. However, I appreciate that some people help to improve the technology by trying different solutions for some of those issues. We are slowly getting closer to a "public ready" solution. Although, I'm not sure if I will see it in my life time...
They should sell the print bed separately. A print bed that reliably holds all filaments and releases them after cooling for about 30 € or so would certainly be a success.
you can just DIY that yourself. buy a FR4/G10 sheet, that's what they're using. Maker's Muse has a video where he does this.
They refuse to sell the bed alone, which is really what you’re going for. The controller is a gimmick to justify charging more- the “sweeping at bed temp x” can be accomplished with end gcode very easily. I bought a CR6 SE kit, and it was obvious they did zero practical testing with it. They wanted me to stick the $300 sticker on top of the bed screws- sealing them in forever.
It pissed me off so bad I haven’t tried using it.
It’s a gimmick cash grab. If it worked as well as they claim, everyone would have it. It’s been around for years.
In that case, their offer would totally fall apart though, as then they are basically selling you a couple of aluminum extrusions (and printed parts that you might just as well just do yourself) for like 420€.
@@adriansue8955 the vaapr is not just a g10 sheet. Not saying g10 won’t work for some filaments but they have a proprietary blend of things that is different to handle a wider array of materials.
@@carpenecopinum1665 Implying its not a total ripoff and just propping up your printer on some books doesn't achieve the same?
I mean.. with PLA, PLA plus, or PLA-F my plain old glass beds stick like glue at 60-70 and slide off zero effort at room temp. My old school original ender 3 bed can actually print without heat as if its blue tape. By dialing in the bed level I can get some of my products to stick just enough with zero heat that it can be bumped off easily by the fan shroud into a bin (using the old sheet o' paper taped to the bed to keep things from getting jammed under there.) With no bed heat up and down saves a huge amount of time. as long as there are no major shifts in ambient room temp I can keep running for a full spool no problem. Cost $0. OH! and these systems are running non-stop thanks to YOUR awesome tutorial on how to run octoprint on an old android phone ;) so much appreciated.
8:54 I like that the spaghetti mostly just slides down the slanted bed instead of building up around your extruder and eventually gunking up the hot-end.
That price point is nuts, it’s bordering on the point where you might as well buy another printer.
Great video though!
Tried using this system for a while for an ongoing production gig, the firmware was awful, and it wasn't much special. The price was pretty obscene for what it was, and bed adhesion was awful. At best, we managed to get about an 80% success rate. The lift kit is something you can replicate a few pieces of wood to prop the thing up. Any prints that need brims are out, as they have issues releasing. Considering the cost, it was downright dumb to not throw in an actual bed. The 'surface' is basically a FR-4 sheet with a sticky back. Cooldown and heat-up cycles are an issue, as mentioned. Can probably build a fan to cool the bed or something, but even in a cool room, it was an easy 20-minute gap between prints. Certain filaments just DO NOT stick, even some relatively common PLA. The support Discord based support basically leaned towards cranking the first layer temps higher and higher. Had some luck at 230C, but that was damaging the bed and discoloring the plastic.
After messing with the thing on and off for about a year, features seem to be frozen in limbo. As stated, there is nothing to run this thing remotely in a safe manner, no master control type program to manage the farm, etc. Given the price per printer for what's about $80 in stuff, I could understand markup for R&D and good software dev. This just feels like a rehash of an old idea with a small bit of polish and little follow-up.
Take the cost of the 3DQue stuff and the printer itself, you're better off using a belt. Still can be dang frustrating, but cheaper and easier to get running. Better yet just get a large bed printer and print either sequentially, or a long full build-plate to maximize production between shifts.
I did the "clearing the bed with the machine" a couple of times while being remote from the printer. Printer was just horizontal like normal. And a normal bed. I haven't done it enough to automate it yet.
For sticking to the bed: small prints stay put after printing, while bigger prints tend to come loose after cooling. Current print: 18 hours, 324 small items. I need to go there to take them off.
use a conveyor belt and a nuc/old pc + pico for prints wich is cheaper and more modular as controller and the belt gives it a really really big efficiency increase
wich has been done before
You could use a current sensor that turns on a cooling fan for the bed when current drops to cut down on cooling time.
I have used an MQTT plugin with Octoprint, Node Red, and a smart plug with Tasmota firmware to automatically turn off my 3D printers a few minutes after a job completes. That way if I am there I can swap print beds and run another job right away. If not the printer shuts down so it is not powered up for hours while not in use. Saves wear on the fans and power supply.
With this system, you could use a similar setup with a fan pointed at the print surface to reduce cool down time between jobs.
Here's an idea for a Voron 2.4 abe system. Take a pei sheet without the spring steel, and make a clamp on both sides so you can pull it in both directions. Take a magnetic sticker and adhere it to the pei sheet. Have a metal build surface ( that is magnetic) so the pei/ magnet combo can lay flat as if it was just a normal bed. Once the print is done, have the pei/mag sheet move (with a bend after the bed to unstick the print) and then build a sweeping mechanism to get rid of the print. Return all to their original state and begin another print
I also do same model Rack L connector to my customer, now I do plastic mold to produce it. Total high I put is 80mm,
If you're printing thousands of the same part, it's probably time to investigate other production tools. Injection molding at the smaller scale has gotten a whole lot easier and cheaper in the last decade, between being able to print molds with high-temp UV resin, or casting your own molds in high-temp silicone. Sure, some things simply can't be made in a mold, but most things can and even more can with some tweaks. And then there's small-scale CNC, either to produce parts directly or to make aluminimum molds for injection molding. You can even recycle quite a few things from the waste stream to do that injection molding for free or nearly free.
Injection molding for a few hundred or even a few thousand isn't as cheap as you think. This greatly depends on the complexity of the model and resulting mold.
Source: I've been professionally molding for years.
@@MastaSquidge True for commercial injection molding. But home hobbyist-level injection molding via non-machined molds has taken off in the last couple of years.
@@oasntet which falls back under the complexity of the part. The hobbyist can of course make multiple parting line molds and manually assembled cores and undercuts but at the end of the day it is still really only cost/labor effective with simple parts.
And that's perfectly fine if that's what you're doing.
Labor isn't free and 3d printing is mostly unattended time. If you go get a creality cr30 or sainsmart infi you don't need any fancy kit to do continuous 3d printing. It'll batch and remove prints from the bed right out of the box.
Prusa print bed has been the best option for me. Nothing else compared.
I use 3Dlac for adhesion now and can't fault it, stuff sticks. That said, a single fine spritz/mist of spray putty worked pretty great too!
The idea looks great but in practice seems to need a lot of tweaks, and a RAM 3d printed to push off those parts. Think I'll stick with the plate changes for now. Thomas you did give us the honest truth and thanks so much !
Man you make good reviews! I love learning about whatever you're testing on this channel even though I will never buy 99% of it lol
I have the 3D Que Autofarm set up with 2 CR6SE printers and I love it! There have been several upgrades to the software since this review, which is pretty typical with reviews and new systems to the market. I do print different parts quite easily by uploading them into the queue. The main purpose of this system is to make it more efficient for Print farms. Which you do say in your review that it makes sense for people wanting to print on repeat. It has saved me a TON of time and money to use the system. They also do have a way to connect remotely so you can control things from the Autofarm dashboard.
Still a great review, but I just wanted to point out that the developers have some a bunch of upgrades over the last 10 months. They are also very responsive if there are any questions or requests to adjust things, 😊😊
I did modified my Prusa mini's hotend fan duct, so now it has a lifter part (Thanks open-source!). The idea is roughly the same. The flex bed is taped to the heating panel just to be sure, and a sheet of paper between them in the front so the part won't fall into the mechanics. Only problem is, when i lift far from the supported side, then the whole x gantry twists and stays like that that makes the lifter's bottom lower, and catching into the print itself. However I tried it, and works as intented, if i set up correctly the temps and the cooling also.
One note, that I used often. Turn down the bed temp after the first layer. It's already stuck, just need to maintain not to fall mid print, then when it's done you maybe need to wait third, or quarter times as originaly.
I was thinking of a print bed with release pins, like you get on injection moulds. Would be extra weight on the print bed, but you could use them only where needed.
Something else you didn't mention is that this printer setup only, is that it can only ram printed items that are above a specific height (i.e., the lowest rod holding the hotend).
Buy or build large 3D-Printer and print instances all at once or one by one in single print. Same scenario, but much more stable result with any fillament.
Any system that uses the delicate print head as a clearing ram is an instant no no.
The bed looks promising, but the angled printer is overengeneered. I just printed a light slide I can attach to the front bed. Others used a sheet of paper. Works perfectly and is low cost.
"upfront price doesnt matter" said no one ever
if heated beds used peltier elements, they could swap polarity to cool the bed down?
Looks like a Pelletier cooler or blast of CO2 liquid would get the parts off more reliably as you need a difference in thermal expansion between plate and part to get it off. Also a thermal shock of active cooling would allow you to locally cool the part or bed quickly, and resume printing much faster while the thermal mass of the bed recovers operating temp.
Ya the bed type is a G10 or PEI bed, both do the same thing having the print self release after the bed is cold. PEI you usually want about a 10C hotter then suggested temp for any material FOR the first layer, you can then let it back off to 50c to finish the print.
Video idea for you, running klipper on non raspberry pi single board computers and/or old android phones. A lot of people are waiting on raspberry pi’s for klipper when rock/orange/banana/etc. pi’s may be good enough. I know you also did a video on running octoprint on android phones and it seems like it’s possible to do the same for klipper, though it’s a longer and more tedious process. Just some ideas! I think with the raspberry pi shortage and peak of klipper interest going on now there would be a lot of interest for low cost raspberry pi alternatives
"most of the time, it works 100% of the time"
This is the description of my life.
You should be able to do this with marlin Gcode and a SD card. The code in marlin has a g code to load a program from a SD card and start printing it. With some simple edits to the end of your Gcode inside your slicer you should be able to handle this with just a base printer
For reducing the time between time because of the slow cool down of the plate, a common big fan that turns on after the print has finished would dramatically decrease this time.
I could be completely wrong because I’ve only had a printer for a couple of months, but I think the ender3_v2 has a similar print plate that grabs during the print and releases after the print is done. At least that’s kinda what I’ve noticed.
Creality CR-30 is already a fiddley high-volume printer plus dat Z axis?
The Video was very interesting. I ordered an set for one of my prusa minis without a hub. Today I had the time to install all and i haven't any issues with the gcode transfer. Ok you can't change the quantity of the running que but you can extend the que with new quantity of the same gcode.
Interesting concept. A DIY mod like this could be a nice second live for my mk3 after I got my XL or something similar.
If you add a strong fan adjust the airflow to the bed it can make faster the slow cooling cycle 😅
if they made it like $150, maybe $200 or $250 with the pi, I'd say this would be fine, but theres no way i could justify spending almost $600 on a raspberry pi and a few printed parts that a beginner could remake in fusion in a few hours,
heh, guess pushing sticky parts out is one case where my lead screw monster with closed loop steppers would get to flex it's muscles a bit! I've found out the worst way possible that if it ends up crashing on auto-leveling and then rams the head on the bed, it can just keep pushing the entire 310x310mm glass out of the 4 clips and crack the corner...
VERY nice, I don't like the spool mounted on the machine, especially up so high.
1337
ok
Kek!
First generation set up. obvious place for improvement. Seems like a good thing. The biggest advantage to a system like this is someone who produces a product that uses a lot of smaller identical parts. It essentially can stop you from having to run to a printer and break your work flow.
also. I only use one print material for the reasons you talk about. I don't care about other colors (i have this thing called "paint") and I'm perfectly happy with the mechanical properties of my print material.
In my eyes, I don't see how this idea can compete with something like the Creality CR-30, which does almost the same thing, using a conveyor belt, is more reliable in removal, and can even print parts much longer than the length of the bed, the only downside i see is the potential of having to relearn positioning for prints due to the 45 degree angle of the head. Otherwise, just put a bucket down, and the prints will peel themselves loose as the belt moves along.
Plus it has zero downtime spent on cooldown/warmup. If it stays calibrated, the only time it isn't printing is the time you're spending installing a new filament (until you need to perform maintenance).
i wouldn’t trust a system like this running 24/7 tbh
The Linus style sponsor segues was a very nice touch 👌🏻
I suspected something like that. I want to salute the guys at Power3D which I support, Carlo has a Prusa like that with a very similar setup if not the same, and he told me something like real applications for this method are quite an exception even in his experience of professional farm/shop. Note that from 72 to 84 benchies full plate, and from 5 to 6 skulls or single mid-sized objects, the slight improvement is the same, around +15%.
So, in case, auto plate swap is the way to go, more than continuous print like that.
An alternative even better could (I am imagining) be the conveyor belt printer which has other alternative use cases (big linear objects ofc) and yeah it comes with its very own set of troubles and tricks to handle, compared to a normal printer.
Interesting study tho, I liked it!
It would be easier to just use something like the CR30 with looped gcode for repeating parts, or similar queueing solution.
That at least is made for automatically detaching parts and prints long in one axis, which could be used for repeated tasks instead
you just need to put a glass and octoprint to do this... I use the creality bed upside down so I print on pure glass, at 70 or 75 ºC instead of 60, when it cools, the GCODE push the print out.
only thing necesary is to use a good bed-nozzle distance
raspberry 60$, glass bed less than 30, so less than 100$
i say it on every of your video, but a glass bed with kapton will stick when hot et during the cooldown of the bed you hear a little pop from the piece.
Charts are Back!!! I love charts!!!
Where i see the problem is that, while parts will easily come off if lifted from a corner, there may be a high resistance to pushing them from the side and parallel with the bed like it is attempting to do.
I tested this on my anycubic i3 mega which has a similar bed by default: try to push the part parallel to the bed and it is hard to remove it. Grab it from a corner and the leverage increases and the part easily moves.
Sounds like a better system would be to have a printer with a lift kit that prints close to level and can then tilt up at the end of the cycle AND a separate arm to clear the print off the print head.
Not many people are mass producing parts to the level that that would require.
Anything for a Ankermake m5?
For that price I’m surprised they didn’t have a bed cooling solution to save that half an hour
This seems clever but of limited use to me. For mass replication I'd think that one of the conveyor belt printers would be a better approach.
man you went all out in testing for this video, i learned a lot watching this thank you!!!
(even how to replace the 5v fan loololol)
This strikes me as more intended for cutting labour costs - for a very large print farm that's churning out hundreds of the same part, not having to mess with beds will greatly increase the number of machines one employee can manage. Outfitting that many machines would be eye wateringly expensive, though.
Also, a bed cooling fan that switches on when the print is done would save a lot of time
It would be so rad if someone made a robot arm to remove prints automatically! would be even cooler to see a solution like this working on some kind of belt and pulley system with multiple printers
You can actually do this with octoprint and a camera,move the tool head when the print is over,you just need internet
I think is a bad desing: pushing object system put so much stress on the gantry, after some time it'll lose precision. Would be better replace the bed with a belt and when the printing ends, just advance to the front, a sharp edge could release the pieces. And you use all space abailable without "dead spaces" to print.
This 'waiting for the cooldown' issue seems like a job for a thermoelectric cooler assembly.
Rather than some PTC heating element for the enclosure heating I have been thinking about using TEC so that the heating affects room temperature less in the house.
With that and the ability to automatically reverse the hot and cold sides after each print, you've got a speedrun version of this.
Also I really don't see the point of a very expensive fancy coated thing when a smooth glass bed does 70% of that if you have some sort of motor to shove the whole print off when the bed reaches ambient...? Well it does help if it has a smaller footprint than having a motor shove prints from behind the printer, but this whole tilt the printer thing seems awkward at best.
Surely for high-volume production, belt-printers are the way forward? Or do they have their own set of issues?
There is one thing missing with the calculation, the jump in time cost due to hiring help.
I still think that having a farm makes more sense - just wanted to point out that not all minutes have the same value when talking about users with different level of experience.
They literally just make a conveyor belt printer that can just.... do this.
What do you do about removing skirts from the build plate? Is the idea that you don't use a skirt (maybe use a brim as it will release with the part)?
yes you don't use one
Skirts are set at 3 loops and a .24mm distance, they slide off with the parts, I do mine 3 layers high to ensure removal with the part
Meh… a conveyer belt system would be much better. That would allow for printing of long parts and not have to cool down in between prints. I’m hoping that Prusa will offer this option in the future since conveyer 3d printing is the future of 3d printing desktop manufacturing.
What if you make the bed into a conveyor that rotates whenever a print finishes?
I'm getting my prusa mk3s+ this week, lol
Sound liks they put a PEI coating on top of a FR4 sheet.
That synology NAS 😍😍😍😍