Well this is the post that most of us have been waiting a long time for! Time to make stuff! This topic is only about 3D printers but it belongs in my RC Project category because of course, what kind of a custom designed RC car can you make without a 3D printer? Not much, or at least not affordably for me. This post is going to be long and as thorough as I have time to make it because I haven’t seen anything else like it yet. Let’s get started.
Choosing a 3D Printer
Anyone new to 3D printing first learns that there are two main types of printers to choose from for hobby use and they are distinguished by the material form that they print from. The two common materials are resin (commonly called DLP/SLA) and filament (FDM/FFF). Resin based printers are typically more expensive, more accurate, fewer material vendors and smaller build volumes. Filament based printers are typically cheaper, less accurate, almost infinite number of material vendors and a larger range of build volumes. Filament printers range from about $150US-$6000US whereas resin based printers range from $100US (ONO3D) – probably $10,000US. The ONO is the definite exception with the rest of the prices mostly sitting between $2000-$5000.
Commercial printers range from house sized cement material based printers to nano scale ridiculously accurate printers in labs that can print details much smaller than you can see. The common commercial printers are SLA (resin), polyjet and SLS (selective laser sintering uses a laser to form powder into a shape). SLS can print both plastics and metals. There are also more specific FFF printers that can print with filaments with fibre in them (MarkForged). A special mention should also go to the Carbon 3D printers as their continuous laser system doesn’t have a parallel for speed and surface smoothness that I know of.
Basically unless your budget is over $2000 USD you are looking at the ONO 3D printer (small build area and relatively expensive resin) or a filament based printer. $200-$1000 USD approximately is the range I call Hackable. That means the printer you buy is able to be modified easily (open source components and readily available hardware) but the downside is that it almost for sure will require some modification/fixing poorly manufactured parts to get it working well. Over $800 my guess is that you will find a printer (either used or new) that should work reliably out of the box. Think Ultimaker or countless others. It should be noted that there are many experienced and qualified companies that have tried to produce an intuitive and reliable 3d printer consistently and have failed (E3D BigBox, Tiko 3D…). So basically unless you don’t actually want a 3D printer I would suggest not buying one off Kickstarter. You will regret it when all your friends have their printers already and you are still getting the occasional update about why your printer is delayed again.
Others have probably hashed this out in more detail but delta printers seemed finicky to me and harder to troubleshoot. I just chose to stay away and stick to Cartesian based (rectangular) 3d printers because I knew I would have to problem solve the printer at some point considering the price range I had. Cartesian printers normally have a larger bed than delta printers do which is the most significant print size factor. For a common/larger delta bed diameter of 180mm (~7.1in) that can only print a 5in square. Most models are only designed/end up practical to print one way up (without tripling the amount of filament you need for supports) and so if you want to print common models on a budget that leaves you with a cartesian filament 3d printer! I know there are tons of things you could print that would be smaller than 7.1in but sooner or later anyone looking to print functional items is going to want as big a bed as they can afford. It should also be noted that choosing a much bigger bed than I chose (200mm x 280mm) is somewhat pointless unless you want to be waiting multiple days for a print to finish. Multi day prints also have a lot of other problems, all of which will lead to your print(s) being garbage. The most basic one is simply not knowing exactly how much filament you have left. Changing filament mid print is still considered experimental in my opinion.
My Printer Priorities List
Most people will have different priorities when it comes to a printer but everyone will share some of the below. I am pretty sure that infinity has been broken and there are actually an infinite number of filament based 3d printers for your purchase so listing out your priorities is critical to making a decision at all! The priorities list should reflect what commonly goes wrong with 3d printers. In my opinion the order of most common issues are
- under or inaccurate extrusion,
- bed not level,
- temperature wrong/inconsistent and
- motion not constrained/smooth enough.
1 and 3 are solely related to the hotend and extruder. 2 and 4 are related to the frame and moving parts design and quality. So without further ado here is my priority list.
3D Printer Priority List
- Easy to upgrade/fix
- Anything less than $1000 roughly is going to make you want to tweak and or upgrade certain parts of your printer. We are all at one stage of learning about 3d printing, what it can do, and what we want to do with it but with something like the Tiko 3D printer you are screwed. There are no standard profiles, parts aren’t compatible with off the shelf components, and if something breaks you are at the mercy of the supplier if they even still exist. Since many if not most companies that have started manufacturing their own 3d printer designs have gone out of business or stopped producing the printers, this is a significant factor to consider.
- Even the most simple problem can be a show stopper if nobody has the same component as you. I was terribly worried that I was shorting out my MKS Base control board after hooking up the wires and seeing little lights flashing over the board. Turns out the board has tiny led lights in different places and I had a loose wire which was making the lights flash so quick it looked like sparks. If I had a non-standard board then looking up whatever little thing is going wrong gets a lot harder. I had never hooked up a screw terminal before and wasn’t sure if I was putting the wire in the right place. I found an engineering drawing of the exact screw terminal on my board at sparkfun.com and convinced myself that indeed I was putting the wire on the wrong side of the little elevator piece (it clamps down not up). If I had some custom or proprietary component I would be at the mercy of whatever customer support was available.
- A cheap upgrade to my printer is one or two part cooling fans (as opposed to the hotend cooling fan). I got 5 blower fans from china for like $12 shipped roughly. The cords were too short which was a downer but nevertheless I can make my printer go from not being able to print cross sections smaller than about 4mm in diameter and no overhangs to being able to print 5cm bridges well for $15-20. That is really awesome instead of being stuck with only larger models.
- A good hotend and heated bed
- The temperature your electronics are capable of (hotend, heated bed, power supply) are capable of directly decide the which materials you can print. Look here for a quick list of print temperatures for various filaments. Note that once your printer is working well you may want to print quite a bit hotter so that you can print faster.
- I will say a lot more about this below but the hotend is the most critical part of the whole printer!!!! It is easy to get lost in the discussion about v-wheels vs linear bearings or about the frame design but the hotend should get your attention first and last. Another criteria of hotends that is critical is a short and efficient transition zone. This is based on the metal parts (heater block, heat break, cooling fins) shape and materials, the presence of a cooling fan on the cooling fins and typically insulating the filament via PTFE tubing up until the heat break. A long transition zone results in clogging/extruder skipping/under extrusion because the plastic melts too far from the nozzle and then there is too much friction to push the melted filament plug through the nozzle. Another criteria is low friction all the way through the filament path.
- A frame design that is rigid enough
- Having the most rigid frame ever is not enough. The key is having the most rigid frame you can make while still being relatively light. 3D printing is slow no matter which way you look at it and everyone is going to want to print as fast as they can without compromising quality. Every sub $1000 filament 3D printer has basically the same NEMA 17 stepper motors and drivers. Speed at good quality is based on light weight moving components so that you can have fast, accurate acceleration without vibrating the crap out of the frame.
- The extruder is really important but also easily upgradeable which is why it is number 4. Buying a 3D printer with a poorly manufactured extruder may make you want to jump off a cliff trying to initially troubleshoot it though.
- The tricky part about most of these priorities is that they mostly aren’t spec based. That means that you can’t really know them outside of user feedback. This puts less common printers at a huge disadvantage. Troubleshooting is also a lot easier when you can search answers online and ask questions.
- LCD Screen
- Being able to change the settings and control my printer from the lcd screen is nice. I like manually speeding up the print for easy sections and seeing how much time is left. Manually leveling the bed might actually be easier from Repetier Host now that I have it set up but it was nice being able to do that from the printer before I knew what software I needed.
The Tevo Tarantula!
You knew this was coming and may have skipped here anyways. So based on my priorities I chose a Tevo Tarantula (Large Bed option is $328 CAD shipped + sales tax on $119 USD item).
I like the v-wheels as opposed to smooth rods because aluminum extrusion is way stiffer than a couple smooth rods and is quieter ( or so I hear lol). It is of course a fantastic price with the promise of a large build area and a decent chance at a fully working printers worth of parts shipped to your door in two weeks along with 1kg of filament to practice with. There are no proprietary filaments to work around and the facebook group/multiple websites/forum is available 24/7. It comes with a heated aluminum bed and has a really nice surface that prints stick to (like pei/buildtak or similar). If you get lucky the manufacturer themselves even sometimes help users by sending replacement parts for duds that got shipped. The aluminum extrusion design is easily hackable. The parts are all common except for the hotend heat sink (which as far as I can tell is a good thing). For the many people that just want to know what a budget 3d printer is actually capable of… you have come to the right place. Happily satisfy your curiosity below with pictures of prints from a non-upgraded Tevo Tarantula with about 20-30 hours of assembly/calibrating/learning done. This would be about the minimum I would expect anyone would have to put into this printer to get it working well. Later I will talk about Assembly basics, Getting Started (with all the many and hard to find community links) and then Common Issues and Tips. The manufacturer updates the kit a few times a year it seems so you never know when they just might fix some of these issues.
This printer sounds like it may not have a power supply sized appropriately for printing ABS with the large bed. Many people (myself included) have received sub par parts and sub par customer service. The wiring doesn’t have key safety features and you probably have to learn how to solder to assemble the printer. There is no ground on the power cord. The green power in and heated bed screw terminal connectors on the board aren’t rated for high nozzle + bed temperatures. The power supply has fairly open terminals for you to short out your printer by bumping one of those many tools on your 3D printer table the wrong way when you are mesmerized by your printer. The cable management is a bit of an art. It is often a huge pain assembling the ptfe tube into the hotend and extruder properly. This was by far the biggest downside of this printer for me and almost stopped me entirely from getting it working. Acrylic is used everywhere to hold the components in place and cracks easily. The bed supports are scarily flexible and not symmetric due to the hardware supplied.
Assembly Basics and Initial Setup
Here is a picture of what the 3d printer looks like once assembled.
At this point you don’t really need the youtube assembly videos anymore since the manual that comes with the printer comes with everything you need to put the parts in place. What the manual doesn’t really show is how to wire the printer. The y axis holds the bed. The x axis holds the hot end. The z axis holds the x axis. A ratcheting screw driver with hex bits or an RC allen key set will come in handy with all the hex bolts to put in everywhere.
- Don’t worry about which side the y axis motor goes on as it is just changing one line in the firmware. I think the right side is what the default firmware was setup for even though the manual showed it on the left if I remember correctly. I have it on the left in the picture.
- You definitely want to get extra washers so that every bolt holding acrylic in place has a washer. I left the paper on all the acrylic except for the control board case in order to give myself a bit of grace if I cracked the acrylic a bit too much. Many of the pieces I heard some cracking sounds on while tightening so I am glad I left the paper on because only the z axis end stop cracked all the way through. Luckily I had it kind of stuck in place so that I could print a new z axis end stop. I don’t remember any of the pieces that had washers cracking. Hopefully all the other cracking sounds I heard were just the sound of each bolts connection friction with the acrylic breaking from static to dynamic as I tightened them.
- At the end of the assembly you will need to move your z axis along the y axis precisely in order to maximize the length of the bed that you are able to use. I found I could only get to about 275mm of it.
- You will need to install FTDI drivers I can’t remember which one I chose
- You need to install Arduino and download Marlin RC8 The stock firmware doesn’t allow you to change any calibration settings (and have them remain after reboot) and so is basically useless. Ruiraptor user on Youtube has great videos that walked me through getting the software installed correctly. You need to match the version of Marlin with the version of Arduino you install on your computer.
- Calibrate your extruder using Repetier Host by extruding 100mm through your extruder into the air (not through the hotend). Measure the length. Look up your current esteps value. Calculate new Esteps= old * 100/measured and set it. Reset the filament position, extrude another 100mm into the air and measure. Repeat as needed. Set the values in your firmware if using Marlin and reupload. Personally I don’t like this way of doing it because it assumes a correct and constant filament diameter. I think you should adjust your flowrate to compensate for differences between filaments instead of your extruder esteps.
PID CALIBRATION of the hotend thermistor
This is easy so you should do it just because but it may also be the difference between you being able to successfully print ANY model successfully if you received a setup like mine where the thermistor was deviating +/-10C throughout the print! Just open Repetier Host connect to your printer and under the manual tab run the following gcode command.
M303 E0 S200 C8
I got the gcode commands for starting the PID tuning process from here. Then put the resulting PID values into your controllers firmware. I tried setting the EEPROM values through Repetier directly and the PID values didn’t stick but the extruder values did for some reason. What did work was manually putting them into the Marlin firmware via Arduino and re uploading. I am curious about switching to the repetier firmware and slicer to see how it does. Since extrusion flow accuracy/consistency is the biggest challenge in getting high quality prints (in my opinion) having a slicer that moves the nozzle smoothly between motions/layers is really important. I haven’t tried it yet to test that promise though.
ASSEMBLING THE HOTEND
This is the most important step!
1 Take the nozzle block and heat break out of the cooling fins. Insert the ptfe tube as shown below. If you can’t get the tube through the connector you may have to drill out the connector by sketchily holding it with one set of pliers while using another set to compress the fitting top while drilling it out carefully. Believe it or not this did work for me. There seemed to be an internal plastic shoulder that no matter what I did I couldn’t get past until I drill it out.
2 Make sure the tube can get through to the end of the heat break (Even sticking out a bit).
3 Back the nozzle out a quarter turn and thread the heat break into the nozzle block. Make sure the tube is pushed up against the heat break. Then use pliers or something to compress the tube fitting while inserting the nozzle/heat break assembly back into the cooling fin block. Heat up your hotend to temp and then tighten the nozzle back up and you are done! This method guarantees that there are no restrictions for your tube to get caught on and is fairly painless. It also makes it easier to drill out your fin block if you need to.
INSTALLING A PART COOLING FAN
I didn’t even include this in the list of upgrades below because it is more of just finishing the assembly as opposed to upgrading (even though a part cooling fan isn’t included in the kit) Buy this or similar 50mm x 15mm radial blower Part Cooling Fan for use with Fang Duct or Wide Nozzle Dual Blower Make sure whatever fan you get has about 1m of cord so that you can actually run the cables properly. I bought a 5 piece set which I think is the one the Tevo Tarantula official store is selling that only has .3m of cable. This won’t allow the hotend to reach the full travel much less have nice cable routing!
In order of usefulness…
Facebook Group https://www.facebook.com/groups/TEVO.3dprinter.owners/
New Forum http://tevo-3d-printers.com/
ruiraptor Youtube software videos (just Google it)
Helpful Website http://tevonewbieguide.weebly.com/
Another Website http://3dprinting.atomicego.com/build-notes-avoid-went-wrong
Manufacturer’s Website (for support tickets re actually sending working parts)
Common Issues and Tips
- Z hop on retract (cura slicer setting) will cause some stringing! Normally the nozzle gets wiped as it moves between each surface on the active layer even as the filament gets retracted to prevent oozing after the nozzle gets wiped. However, during a z hop the filament in the nozzle will be pulled out by the part of the print where the nozzle last touched. This will cause stringing. The only way to avoid this would be to mechanically wipe the nozzle (not practical although that would be a super awesome feature if someone could design it) or to calibrate your coasting settings to break the connection between the part and the filament in the nozzle just before it needed to z hop. The best results for small cross sections require a part cooling fan as far as I am aware.
- Don’t use the supplied SD card
- Don’t try to print from the computer via usb (although accessing the sd card from repetier via usb is probably fine). This is because your computer might not prioritize the usb communication once and it could ruin your print.
- Note that thicker layer heights are easier to print well than thinner layer heights because the layer height is closer to the nozzle width and for a given total print time you can go slower which means less acceleration vibrations.
To Infinity and Beyond!
So now you are bored, don’t have kids or some other problem causing you to want to push your printer to new levels of awesomeness…
https://github.com/KevinOConnor/klipper with Raspberry Pi / OctoPrint
Frame stiffening brackets
Single motor Dual Z rods upgrade
Something to stiffen the bed (carbon fiber, aluminum, or your own design). I have an idea for wood+3d printed support under the wood which I think will work well. I haven’t designed it yet though so you are on your own for now.
Stock Tevo Tarantula Example Prints
These were obviously at various levels of my understanding of the software and printer. All of the good looking prints were after I assembled the hotend correctly, turned off z hop on retract, set the first layers to 30mm/s, .3mm first layer height, .2mm layer height, 40mm/s print speed, 800mm/s^2 acceleration, 9mm jerk, 2 rows of skirt. I am starting to see that I can print faster for above layers and am leaning towards more like 100mm/s^2 acceleration for first layer because the corners and little profiles rarely print well on the first layer still. I am sure that my top print speed will be around 80mm/s or more after the first layers are done. Infill is my biggest problem currently. Even at same speed as walls and double the nozzle thickness I am still not getting prefect infill (but way better than infill printed faster than walls and 1 nozzle width).
I will update this post after I get more filament, install my part cooling fan and test some of what I have learned about slicing software. Hopefully I can make it clear how various factors have improved my print quality.
My next post will probably be on getting started with 3D modeling software…