Hi 3D printer lovers! I'm a high school student working with a group at

Hi 3D printer lovers! I’m a high school student working with a group at the MIT Launch startup accelerator and I was hoping to do some market research regarding some of the current problems with desktop 3D printers. I was hoping to get some feedback from all your experience with 3D printing and the hours of troubleshooting you’ve likely encountered.

What are common problems that new printer users encounter and how hard are they to identify and fix?
What workarounds and aftermarket modifications are the most useful?
If you could change one thing about your printer what would it be?
How do you troubleshoot issues and how long does it take?
What would make you more likely to 3d print more often (ie never clogged, didn’t have to watch first layer, etc)?
In your opinion, what are the biggest issues the desktop 3D printing industry faces?

Thank you so much for your time and sharing your experience, we appreciate all feedback!

f0b4a3ca012bbdfad7ae2838e57e0c82.gif

Problems that new users include:
Clogging.
Overextrusion
Underextrusion
Bed adhesion
Skipped steps
Non-square frames

Thanks!

i’d say non-square frame is the first you will hit and one of the hardest to observe. belt tensioning goes hand in hand with it. then is getting the bed levelled - a sensor may help with that but it still may take a bit of work. finding the sweetspot of temperature for extruder and bed for a particular brand and type of filament would be the next and then getting it to extrude correctly (under extrusion, over extrusion, oozing).

Oh yeah, bed levelling is a bastard. Good call.

Most hobbyist printers are severely hindered by poorly selected components.
In just one printer (that cost $1,200) I had to replace the following :
Extruder, stock one was heavy, clogged a lot, and used poor internal gears.
Hotend, stock hotend jammed, and was slow to heat up.
Belts, stock belts didn’t meet specifications (dimensions) of that belt type. They also introduced ringing.
Belt clamps, stock belt clamps were loose, leading to backlash.
Motors, stock motors were 1.8 degree, and didn’t give enough resolution on the x/y axis causing gaps in infill.
Build plate, the stock build plate was not flat, and glass is a poor heat conductor.
Build plate interface, kapton doesn’t have as good of adhesion as PEI.
Firmware, stock firmware is grossly out of date, with bugs and missing features.

I have another printer that cost 2,500 that i did the following modifications:
Extruder, stock extruder was heavy and inaccurate.
Hotend, stock hotend is prone to clogs, and used 3mm filament which is far less common than 1.75mm.
Linear rails, stock y axis rails could not support the heavy bed without sagging or deviating. I still need to replace the z rails (they do not have the strength to withstand inaccuracies in the z leadscrews)
Replace the x axis ends (the mounting mechanism allows the belt tension to pull on the z rods, bending the z rods and making the belt loose).
Firmware, the stock firmware is buggy and missing features.

Bed adhesion is number 1
I have read countless threads if why hot ends fail, this is because of the industry of consumer 3d printers use a cheap thermistor, which is both inaccurate and breaks down.
That causes some to completely give up or spend a lot more money and time to replace the thermistor with the same device that will break again.
I have talked to Ultimaker, MakerBot, Diamond Age, which have all had a lot lot of customer dissatisfaction due to thermistors.
As such Ultimaker decided to use a Resistive Thermometer (RTD), Makerbot and Diamond Age decided to go with Thermocouples.
Thermocouples should be used as they have a higher resolution, and higher reliable temperature variance tolerance.

Otherwise things people spend a lot of time on is poorly set up Bowden extruders.

Lastly.
DUST buildup
Only a few consumer 3d printers have actually addressed this, and is generally the reason why nozzle clogs happen.
Periodic Cold extrusion, will help a lot.
Brass nozzles I find, tarnish too quickly, preventing heat to effectively pass from the cartridge to the filament. They are very cheap however, but results in downtime and frustration to replace or clean.
Optimal solution would perhaps use Tungsten. However tungsten nozzles are more expensive, but generally last a LOT longer

Thanks, you’ve all been very helpful

There is one last little thing, a lot of people are expecting a lot from a little 8bit micro, expecting it to behave like a 32bit one.
There is a definite movement in the industry to 32bit micros, however only a few OS firmwares support them, Smoothieware and Repetier. Marlin, which I follow on GitHub is currently in the transition to 32 bit.

Barrier to entry for slicing.
This has been discussed by @Richard_Horne ​ with suggesting that all filament manufacturers provide a set of parameters optimally suited for their filament.
I’m still hoping that the slicing side of things will be an invisible layer, while 32 bit micro could do the slicing on chip, this is something best left to a supplementary device, like a Raspberry PI/OctoPrint combination.

Warning: this post is too long, pack a lunch or move along.
Brook

I have worked for years to remove barriers. I’ll tell you what I’ve avoided:

Bowden. we’ve always done direct drive

Plastic, acrylic, printed, wood frames. we now only use metal frame construction.

Plastic parts.while I’ve dabbled in injection molding, and currently offer one model as a low cost kit w some printed parts (mostly non mechanical)… machined Delrin and aluminum are far and away a better way to mass produce printers. Quality! Everything is square and straight.

Cheap print beds. While price pays a part, the best bed is mic 6 (cast alluminum). A second is .25" aluminum. Flat beds cover a multitude of sins

Manual adjustment on print bed. This is a big one. We use metal beds and for years have used them in connection w induction sensors for “auto z leveling” (tramming) in software. Life saver.

Chinese made hotends. One man, Carl Ubis, has made and/or personally inspected every single hotend we have ever shipped. That’s about 60,000 hotends. He is a failure analysis expert by trade (30 years at HP) and continually improves even the slightest inefficiencies in manufacturing and assembly. They are hand assembled and tested thoroughly (no kits). The devil is in the details.

Chinese electronics. Our pcbs are from china, we manufacture in house. While we have made mistakes, we have made enough and have experience to spot a small number of common problems (usually a bad part placement or selective solder joint). So we are in house experts who load the firmware and test the boards. The burn in happens w the power supply we ship with… so QC is the message here.

We don’t buy anything we can make locally. Knowing the guy who made it and him/her being on the hook throughout the whole process brings an amazing amount of accountability.

What we DO:

Admit it and own it the second we are wrong. We fix it or replace it. Sometimes stuff has been out of warrantee but we do it anyways. I have a couple painful instances that cost me a lot of money, but all recoverable, b/c we protect our reputation and brand. You must have trust w customers. Period.

Transparency. Truth. We don’t over-hype. We don’t lie. We don’t hard sell. I’ve even refused to sell to some (big box stores, even potential customers). If I sense 3d printing isnt for you, I’ll tell you and discourage buying. Maybe point to a cheap Kit from another manufacturer to save their wallet. I’ve told more people that “3d printing is hard” than I can shake a stick at. It IS getting easier btw. (Simple Pro, anyone?)

Set expectations realistically. I show prints done by mere mortals. I never show post processed prints. If I do, I’ll show before and after.

The problems that plague us all are:

Set an expectation that light maintenance is needed. You gotta be ok to clean or lube rails/lead screws, tighten belts, troubleshoot mechanical parts, etc. on this point, for a while I was using lock tight on a bunch of parts… but people fiddle and can ruin a part that has lock tite. It’s better to leave the machine ready to be fiddled with.

My machines are workhorses that can run for years. They aren’t throw away toys. They can be fixed. We encourage fixing over replacing.

Set an expectation that tips are consumables. They wear out. Replace them every 3-6 months. Of measure the hole w precision gauges and adjust your slicing. (No one does this).

Repair them at the factory. We repair and tune them very quickly and efficiently. Not everyone wants to learn how. That’s fine. Ship it here and we will. We have seen every conceivable problem. And we have any part needed w/in reach if some mystery problem is found (intermittent) and we just replace stuff, working around even understanding it. Rare, but we are the repository of earthly knowledge on Printrbots. There are some superusers that rise to this level tho.

The stuff that has no easy answer is:

You can’t help people that buy, even though we say “experts only”… they buy anyway and end up frustrated.

You can’t make people read instructions. This one is hilarious, or maybe it just makes you beat your head against a wall. This might be #1.

You can’t make people buy ONLY quality filament. Well, some lick their machines down to only their filament. After 6 years, I actually understand why. I don’t understand gouging customers.

You can’t move power users to a system that is more reliable because it is more controlled (removing power features). Our Simple Pro is for newbies or people who find no pleasure in learning about materials, temperatures, slicing, host software. Most power users feel castrated w/o all the knobs and buttons.

Maybe the biggest need is design software that meets a user at his/her ability level to design or customize things they need. There IS simple design software out there, but some aren’t wired for any of it. Some can’t. And some who can, are paralyzed by the emotional fear of something seemingly complicated or w a learning curve. Our present society of instant gratification has doomed a large portion of our youth.

Last one: it’s difficult to train a newbie to look at s model and sense “that is impossible to print”… or to deduce HOW it must be printed or aligned, etc. newbies have this crazy knack to choose impossible models to print. Yes, education helps, and experience teaches it. I toyed w the idea that I could force customers to step through training levels- like a video game does. Their first print would be a calibration cube. Achievement unlocked! Now they can choose one of two models, demonstrating overhang limits. Achievement unlocked! Now they can move through the achievements until they are allowed to use PETG, then ninjaflex, abs… then infill, support, brim, spiral case mode… eventually, they could print their own design, do a print for a friend, fix something around the house. You get the picture.

At one point, newbies were buying my $2000 Printrbot Go Large: 3 extruders, 1’x2’x1’ build plate that folds up into its own suitcase. Disaster. They would try to print the hardest, biggest thing in abs with no heatbed, or in two colors of ninjaflex with HIPS for support. I stopped selling them.

Protecting customers from themselves is an art.

Good luck! It’s a very complicated goal. It crosses the streams of design, hardware, software and materials. It’s easy to do this business badly. Like a pilot, nothing replaces time in the seat.

If I locked the Simple Pro filament and settings down to only work w the factory approved stuff, and I approved ever model before printing (a database could do this), then I’d have the (almost) perfect printer. I’d have to add a .5 inch thick cast aluminum bed, put it in a temperature controlled case, give it redundant everything, upgrade everything and only let robots use it after reading the instructions. And operate it in space w/o gravity and underwater for cooling… that’s the perfect printer.

Brook

@Brook_Drumm I really appreciate the time you put into this response, it’s a lot of helpful information

other problems:
patterns that appear on the side of prints - hard to diagnose the cause and fix
skipped steps - usually an overheated motor or not enough current or physical resistance
bad overhangs - can be avoided with slicer settings for adding supports if the overhang angle is too much or by manually turning on support material
lack of bridging - use support material or tune speed and temperature
Lack of quality when there is little to print in a layer - multiple choice of :increase minimum layer time, add more objects to the bed to increase printing area per layer, use a cooling fan if applicable which it might not be for ABS although opinions differ there.
to slow of print time - getting the print time down is often difficult but printing more things on a bed, keeping the parts closer together, rotating parts, etc may help improve the overall print time or at least some bother
stringing - this is something that some of us will never fix. Some can fix it with speed, retraction and rapids speed adjustments

There are a few webpagea with titles like ‘20 3D printing problems and how to fix them’. They have pictures and have suggestions on fixing them. You should also look at those as part of your research.

Wh is pictured in the animated GIF seems to be a bed adhesion problem due to the nozzle not being set close enough to the bed during calibration. Some people have beds that are not level or have warps in the bed. When people can’t/won’t fix whichever of those two issues, the sometimes set a thicker initital layer or they increase the initial layer extrusion. A 0.3mm initial layer when you print 0.2mm layer prints can be the result. Oh yeah…imperfect delta calibration or delta numbers can cause a need for such a work around too. Delta issues are also a pain to figure out.