Musings and Experiments on the Art and Science of 3D Printing

Followers

A Strategy for Obtaining Great Prints

By SublimeLayers Wednesday, August 1, 2018
I've published this Strategy on several forums over the past 4 years. It has been greatly expanded (content not number of strategies!) for my upcoming book but I wanted to share this here on my blog for my followers. Note that some of the information in this list is a bit dated, the updated version in my book is completely up to date and greatly expanded.

A Strategy for Obtaining Great 3D Prints

Like all new endeavors, there IS a learning curve with 3D printing. This is still the pioneering era for desktop printing and we are very fortunate to have such a great community here as well as other resources on the web. But the challenge with all the information out there is finding it when YOU need it and deciphering the many different opinions and practices - some of which are good and some of which are, well, let's just say "poppycock".

Another part of the challenge is there are many different means to the same end, but I assert that those who have developed a workable (AND reproducible) technique most likely took a disciplined approach rather than the shotgun approach of trying one thing after another. So, I thought it would be helpful to describe a method that you can use to 1) develop a reproducible approach to successfully printing the things you want and 2) improving the quality of your prints to meet your (realistic) expectations. Don't hesitate to join in or ask questions. As required, I'll consolidate any interesting information from follow-on posts into this initial post to help make everything easy to find.

Ready to go? Before we do, here is a little suggestion.

TIP: When you are starting a new print session, give the printer a little warm up exercise! Much like an athlete needs to warm up before a game, so does your printer! Don't just turn the printer on and start to print, turn it on and let the hot end get up to equilibrium, let the heated bed get up to temperature. I even like to print a quick part to make sure everything is up to temp, in equilibrium and working properly. It's quick and easy to do and can help eliminate a lot of problems.

#1 Get Experience. Start with the printer. This is more difficult than it seems because without experience, it is hard to know if you have a mechanical or electrical issue, slicing issue or if something else is going on. So, to that end, keep things simple until you have some experience. By "simple" I mean, don't print the Eiffel Tower model for your first print, print a simple, reproducible and small item many, Many, MANY times until you nail it. For me, I used the calibration cube. In retrospect, I should have picked something much simpler (see strategy #2). 

#2 Start Simple. We have a tendency to want to jump ahead to more complicated prints, faster printing, and bigger prints as quickly as possible. But a few hours spent working on a simple object or two will pay dividends. There are many aspects to successful 3D printing, everything from the printer (which in itself has a mechanical system, electronics system, hot end, extruder, heated bed, firmware), to the slicer (and all of the parameters available to control the slicing), to the filament itself, to the actual item being printed. With so many variables (100s, maybe 1000s of them) it is really important to pin down as many of them as you can. One very easy place to do this is with the model itself. Develop your experience printing the same model over and over until you nail it. Even with a simple model, you can (and should) approach printing it with a methodical approach from the ground up. That's the next strategy.

#3 Practice in Measures. I play guitar and was basically self taught. When I found new music to learn, I did what many untrained folks do and practiced the piece over and over again from beginning to end. If I made a mistake, I started over. Then, I took lessons from a trained musician. My very first lesson was worth every penny! My instructor watched me learn a piece and then said "you should Practice in Measures". What he meant by this was to learn the first measure (music is divided into small blocks of notes called measures which are small and relatively simple). Practice it until it is perfect. Then, practice the second measure until it's perfect. Next, combine the first and second measures until that is perfect. Continue in this way until you've learned all the measures and combinations of them. In complex pieces, there will be a few measures or sequences of measures where you need to put in a lot more practice.

The advantage of this approach, my instructor said, is that you are not wasting lots of time playing measures you already know. The practice of playing from the start until you reach a difficult spot and make a mistake is that you play, say, 30 seconds (or more) of music you already know to hit a 1 second spot you need to practice. So in a 30 minute practice session you are really only practicing what you need to practice for 1 minute! This completely changed my approach to practicing everything from guitar to 3D printing to machining to learning CAD, to ...

How does this apply to 3D printing? Easily! Start with a simple object to print and practice nailing the first layer. Too often folks will print a poor first layer and allow the print to continue. Why print on a bad foundation? You might be able to salvage the part but more times than not, it will peel from the bed or warp badly. Instead, nail that first layer. Once you have that perfected, move on to print the rest of the object. Once you have the entire object printed successfully, change slicing parameters to print faster, or at higher resolution and start over (nail the first layer, ...). Practice in measures.

I can't say enough about getting that first layer right, the subject of the next strategy.

#4 Nail the First Layer. I don't believe folks spend enough time learning to print a perfect first layer reliably. If there are defects in the first layer, they will invariably come back later to bite you later - the part separating form the build plate, warping, or a defect in the part. Print a good (or great) first layer is probably one of the most frustrating experiences for most, it is also the most critical. Here's where strategy #3 comes to play, don't continue a print on an inferior first layer! Abort the print and restart that first layer again and again until you nail it. Why waste time on a part that will most likely fail or not be useful? Each time you print a first layer, measure it! If you configure your slicer to print a 0.20mm first layer, then it should be pretty darn close to 0.20mm. If it isn't, you've identified a variable that you can easily fix and nail down (Z height). 0.20mm is not a lot and unless you have highly calibrated eyes, you can't tell the difference between 0.20 and 0.15mm, but your printer sure can. At 0.15mm the first layer is going to squish onto the print surface. It may even seem like you are getting a great first layer and great sticking (which you are) but later, you'll discover the part is nearly impossible to remove or your extruder will start making that all too familiar TICK, TICK, TICK sound from missing steps. A perfect first layer will go down smooth and consistently time after time.

TIP: polish the tip of your nozzle! Chared filament and scratches on the very tip of the nozzle are dragged over the layer as it moves around. Best case, these leave a visible mark on the print; worse case, they rip the first (or higher) layer off the build plate. 

#5 Slow Down. Back to my guitar lesson example... The other thing my instructor taught me in that first lesson was to practice slowly (using a metronome) until I nailed the measure(s) at a slow tempo. Then, gradually and consistently, increase the speed. The same applies to 3D printing, print slowly at first. This gives you time to observe what's going on (strategy #6) and just simplifies everything. I like to start new folks at 20 to 25mm/s print speeds. What's the hurry? If you print 10 aborted prints at 50mm/s what have you gained (or lost)? Printing slow helps all parts of the printer, from the mechanics to the extruder to the plastic filament coming out the nozzle, stay in balance or equilibrium. Fast movements can highlight mechanical issues, extrusion issues, etc. But when you are first starting out, you don't know how to identify and isolate these issues. In fact, even with all of my experience, if something starts to go wrong, I slow down. That removes a lot of variables and gives me a chance to see what's happening. I've identified everything from loose pulleys, to a stretch belt, to a worn joint on a delta printer arm! And, I've helped a lot of folks identify other issues simply by slowing down.

#6 Watch What's Happening. Especially in the early stages of learning, watch all aspects of the printer. Combined with strategy #5 you'll start to develop an appreciation for how the slicer does its magic, how the printer does its magic, and it is just simply fun to watch! I highly recommend putting a flag of some type on your extruder motor shaft so you can actually watch retracts and advances and watch the steady push of the filament. A piece of masking tape stuck to the shaft is fine or print one of the pointer models. Watch that first layer print, that's how you'll see if there is a problem and maybe even figure out why. For example, I noticed that the first layer wasn't sticking in the same spot on my build plate. Turns out that I had some potato chip grease there (don't ask)! A little wipe with isopropyl alcohol and I was back in business. Watch what happens when the layer fan comes on. Is it coming on too early and causing the part to peal from the print surface? Pay attention to the details of what's going on and then...

#7 Keep Notes. I can't stress how important it is to keep notes. I have a word processor file I add notes to as I go. In particular, I keep a section on the filaments I use and the detailed printing parameters for them (strategy #9). Perhaps I'm becoming forgetful in my advanced age but I don't like solving the same problem over and over again. If I keep a note about a problem and my solution, I can usually find it again pretty quickly. Once comment on notes, don't be afraid to purge! After a few years of doing this, my file got quite big. Recently I archived all of my H1 and H1-1 notes. I don't refer to them any longer so why keep them in my working notes?

#8 Be Consistent. A CEO friend I worked with many years ago was fond of saying "Consistency is the hobgoblin of small minds!". I understood what he was trying to say but it has to be taken into context. When you are first learning any new activity, it is critical to be consistent. If too many things are changing at once, you have no idea what contributed to a good or bad result. Don't change too many things at once. In fact, if you can isolate and change just ONE thing, you will have a much better chance of success and understanding. This isn't always possible so lock down as many things as you can. If after a run of successful printing you run into a problem, go back to a known good state (see #7 - you did keep notes on what this state was didn't you?) and start there. Many times we try to change too many things in our frustration and that almost always makes things worse. Step back and think about how to isolate the problem areas with as few changes as possible.

#9 Know Your Filament. This strategy is a little lower level than the previous eight but important and often overlooked. I see a lot of folks just assume that they should print filament X at temperature Z - for instance, print PLA at 200°C. This might get you in the ball park but if you really want to get consistent and GREAT results, profile your filament. It's easy and if you write it down (see #7) you'll never second guess how best to print that filament again. It's important to realize that higher temperatures are not always better, they can actually lead to issues - parts that are just a little too large, parts that stick to the bed too well and can't be removed, blobs on the print, stringing, and a host of other problems. In general, I like to print at the lowest temperature possible for PLA and ABS. Then, as I ramp up print speed, I also need to ramp up the hot end temp a little since the filament is not resident in the hot zone for as much time. I suspect little details like this cause people more problems than they might anticipate.

Here's how I profile a new filament:
  • Start with a reasonable target temperature - 200°C for PLA and 225°C for ABS (one quick note, it is ideal to have a calibrated hot end, so when I say 200°C I mean 200°C. One easy way to do this is to make a little table with the hot end set temperature (what you see on the temp display) and the measured temperature (with a thermocouple). Do this in 5°C increments from 160° to 240° C (or so). Keep this chart in your notes (#7) and you will always know what the actual temperature is.)
  • Now, use the manual controls of your host to extrude 50mm at 50mm/s and watch and listen.
  • If the filament extrudes nicely, reduce the temperature by 5°C and wait for the temperature to stabilize.
  • Test again by extruding 50mm at 50mm/s
  • Repeat until you reach a temperature where the filament does not extrude well. At 5°C to that temperature and note this as the "low extrusion temperature" for that filament. Use this low temperature whenever you are printing slowly (20-30mm/s). You might find some filament need to be bumped up a bit more than 5° so don't hesitate to experiment and find that lowest reliable extrusion temperature.
If you want to get really serious about profiling your filaments, do the melt-flow test at higher extrusion rates - 60 mm/s, then 70mm/s, etc.

Don't forget to measure the diameter of your filament too! Not all filaments are created equally. Measure in several locations to get a sense of variability. Most of the slicers let you enter filament diameter and they will calculate a reasonable flow for you.

Finally, once you've completed the filament profile, print the Simple Single Layer Test object in the Layer Tuning section at the end of this post. 

#10 Know Your Bedfellows. Probably one of the greatest mysteries in 3D printing is "the bed". Metaphorically, this is where the rubber (filament) meets the road (bed) and getting "it" right is absolutely critical to successful fused filament 3D printing. All sorts of folklore on bed materials, coatings, coverings, concoctions, and juju exists here and elsewhere on the internet. It is also one of the areas that there is no one right way to do it. If you have discovered a special incantation and bed preparation that works, by all means stick with it! But, for those of you struggling, here are some strategies you can use to make improvements. One comment before I begin...

I am VERY persnickety about the aesthetics of my 3D prints. My 3D printed fly fishing reel is seen from all sides and so it is important that the first layer is flawless and visually appealing. The photo below is the bottom surface (first layer) in both the outer teal ring and the inner white spool plate (you can see more of my work here). A perfect first layer finish is not required for all objects - consider the base of a Yoda or vase - but if you practice getting a great first layer on these non-critical pieces you'll be prepared when you need a visually perfect first layer on another project.

A number of factors affect adherence of the first printed layer to the bed. These include:
  • surface material
  • surface texture
  • surface treatment/coating
  • bed temperature and uniformity of temperature
  • air temperature
  • chemical bonding or cohesion
  • print speed (see #5)
  • filament temperature (see #9)
  • first layer height (see #4)
cleanliness (of bed and filament)
This isn't an exhaustive list but it does include the big hitters and, as you can see, there are a few of them so it is very important to take a methodical (#2 and #8) and documented (#7) approach when solving bed-related problems. This is also a place where careful observation (#6) can play an important part.

I'm not going to go through all of these in detail now but did want to comment about the last one - cleanliness. Whatever you do, make sure everything near and on your printer is clean and grease free. Silicone greases and lubricants are especially problematic since they are invisible and very difficult to remove. Keep them away from your machine.

Your fingers are a prime source of contaminants. Every time you touch the filament or bed, you risk leaving a greasy print (see my observation in #6) and these can (and will) cause issues. I try not to handle filament with my bare fingers, I use cotton gloves. If you use a plastic or rubber glove, make sure it isn't coated or powdered - we're trying to eliminate sources of contamination, not introduce them. On the occasions that I do handle filament with my bare hands I wash and dry them thoroughly first. This is one area that I think affects a lot of user's and is completely overlooked. How many times have you loaded filament right after eating chips? It introduces a big variable that can be difficult to track down, so develop good habits and eliminate contamination as a variable.

Your fingers can also leave contaminants on the bed when you remove a part or brush off stray filament strands. Don't touch the bed surface if at all possible. If you do, clean/degrease it with an appropriate cleaner. For uncoated surfaces like borosilicate glass, PEI, the various 3d party surfaces (PrintInZ and BuildTak), and films (window tint, Kapton) you can use isopropyl alcohol. I like to use the little packages of wipes as they are convenient and safe. You can also do a quick wipe of your fingers before tossing it in the trash. It is more difficult to deal with coatings like PVA glue, glue stick, and hairspray since these can't be cleaned. If you suspect a contaminated coating, your only recourse is to remove and reapply it. 

Finally, don't overlook filament storage, keep it clean too. I store mine in large zip lock bags to keep off dust. You can put packets of desiccant to help remove moisture in the bag too.

#11 Learn to Diagnose. 

Patient: "Dr. it hurts when I move my arm like this."
Doctor: "Then don't move your arm like that!"

The first point of this joke is, many people do the same thing over and over again without making any changes or stopping to think about what to change (see #8: remember, change one thing at a time) - as if just repeating the same print with the same parameters will magically solve the problem. It won't (see my footnote below).

The second point of the joke is that the doctor didn't attempt to actually determine why the patient's arm hurt, he just had him avoid the problem. I see that a lot too. Usually it takes the form of "I tried printing it with my red PLA and it failed but everything was fine with my blue PLA". There are many other variations on this (changing slicers for example).

Learn how to diagnose problems. This requires careful observation (#6). Once you've identified where the problem occurs (let's say getting the first layer to stick) then PRACTICE that piece (see #3) until you sort it out. No need to run through the entire process over and over. Isolate the problem, formulate a hypothesis on what you think might be happening and design a test to prove or disprove your hypothesis. If you see a problem and can't formulate a hypothesis THEN seek help! Or, pre-test your hypothesis here to get some experienced feedback. But, whatever you do, try to work through the diagnostic process yourself first, that's how you learn.

Footnote: Many years ago (20) my company had an annual laboratory safety week (I worked in a corporate R&D lab with lots of nasty stuff). One of the annual favorites was a gentleman from OSHA who talked about electrical safety. He started his presentation with a black and white video from the 1940s (I think) of a speaker walking up to a microphone on stage. The presentation was being filmed. The speaker reached up and grabbed the mic and was immediately thrown back and fell to the stage unconscious. Members of the audience rushed up to help him. This was all on video. As 4 or 5 people worked to help the victim, you see a gentleman casually walk up to the mic, reach out his hand and touch the mic. He was immediately thrown back and collapsed on the stage next to victim #1. Literally 30 seconds later a THIRD audience member walked up to the mic (now there are 2 victims on the stage and a hoard of people working to revive them) and carefully reached out his finger (looked like the scene from ET) and very, very gently touched the mic with just the tip of his finger. He was immediately thrown to the stage as the third victim. All of this was caught on video. No one died (we were told). Neither of the second two victims stopped to think about the problem, consequences or solutions.

#12 Be a Fanboy. I am probably going to lose some fans for this post about cooling fans!

Don't think of a part cooling fan as an object, instead, think about "air flow". If you need cooling on a PLA (or other material) part, then you need to understand air flow. Not all cooling fans are created equally. Consider this, some folks use a 40mm, some a 25mm, some (like me) a 25mm squirrel cage fan. Some are mounted to blow the full fan width stream at the nozzle area, some have a duct or some (like mine) have a very focused soda straw duct). So comments like "run your fan at 1/2 speed" are not specific enough to be useful information. Instead, you need to understand how your particular fan, it's arrangement, your material, etc, all relate to the air flow.

Using the previous strategies, try to minimize or eliminate the need for any sort of air cooling. Slowing a print down (#5) is one great way to do this. It also gives you a chance to see (#6) where any problem areas on a print might be. You can use this information to focus the right amount of air flow on the problematic areas. The tendency for many is to use as much air as possible. It is much better, more consistent, and more reliable to use as little air flow as necessary. This puts less thermal stress on the printed part.

When you do determine you have a problem that only a fan can solve, start conservatively. I also recommend using a duct of some sort to focus the air flow where you need it. Ideally, the fan would have the ability to follow the print nozzle and direct a small stream of air to the filament right after it is laid down. That is a difficult problem to solve, so most of us direct the air to area around and under the nozzle. But, by directing the air (duct) you can reduce the air flow significantly since it is now focused where you need it.

I suggest doing your own experiments and observations but start conservatively. I don't use a fan during the entire part. If you find you need to turn the fan on at full blast from no air flow, do it in stages so the hot end can equilibrate properly. You can do this manually, some slicers can support it, or it is easy enough to learn the simple "fan mcodes" to manually insert them where you need them in the gcode file (this is what I do for tricky parts). 

M107 is fan off
M106 S50 turns the fan on at 50% - the S parameter is the speed from 0 to 100

Using a focused air flow, lower air flow and the step up technique I just described, you won't see a significant drop in hot end temperature. PLA has an interesting property that if you change the extrusion temp at the hot end, it has a visible effect on surface sheen of the part from matte to gloss as you raise the temperature. RichRap has written an excellent post about how he uses this phenomenon when printing decorative vases. Although he was varying the hotend temperature, a similar effect can occur with improper air cooling.

I'm also an advocate of using off-platform cooling. By this I mean strategically placed (ducted) fans that direct air to problematic areas of a print. These can be mounted to your vertical columns or simply sat on the bed if it is not too hot. With ducting, you can reduce the air flow considerably and keep the cooling right on a "hot spot". This technique does require manual adjustment, repositioning, etc. But, it you are trying to print a really tricky part, it might be the only way to do it. Frankly, the part cooling capabilities of desktop 3D printers is extremely primitive at this point. It's fine for the majority of objects you might print but as we push the envelope on what's possible, part cooling is one area that needs some more work to automate it.

Consider this, the way I maintain very tight tolerances on the rotating spindle and hub assemblies on my fly fishing reels is to use a low beam of air cooling on the spindle as it's printed. This "locks" the filament in place in a very predictable way. Once I printed a few parts and measured them to make sure there was little variation, I incorporated that into the design to get exactly the tolerance these parts required.

2 comments to ''A Strategy for Obtaining Great Prints"

ADD COMMENT
  1. "extrude 50mm at 50mm/s" - That seems a bit fast. Did you mean 50mm at 5mm/sec?

    ReplyDelete
    Replies
    1. I was just searching back this profiling info, and thought the same thing exactly. Extruding 50mm second at least takes a BMG ;-) and is probably not a real world value.

      Delete