My 3D Printing thoughts and lessons learned

As I learn tips I will add to them.   I have used PLA, PETG, ASA, Wood filled PLA, Glow in the dark PLA and clear PVB so far.  

• Your time is valuable. 
  • Saving cost may end up costing you hundreds of hours tweaking, learning, and finding an answer.
    
• Use well supported printers.   That includes, slicers, firmware, help and user communities.   It is the fastest way to get up to speed. 
• Bed adhesion 
  
• Keep the sheet clean.  Dirt, oil, or any residue can mess up the adhesion.  Your fingers will leave oil.  A clean sheet, well cleaned, works well.  If you do not use a bed adhesive, you should thoroughly clean the bed sheet before each use.
  
• How well a print sticks is part design and Slicer print order dependent also.  You need to make adjustments.
  
• 0.4mm nozzle and 0.3 mm first layer setting.  Even if you print other layers at  0.2, this appears to help bed adhesion on the first layer
• Many suppliers are now recommending "drying" new filament before use, even though it is sealed. 
• Brims can help keep corners down and keep small parts attached to the bed. A Brim adds gripping power to the bed.  Warning, with PETG brims are much harder to remove.  A raft or glue of some kind may be best in these cases.  A good debur tool is very helpful when you need to cleanup a brim.
• Perimeter draft shields can help on larger parts to prevent edge curling.  The perimeter helps to prevent edge cooling which causes shrinkage and the edge to pull up. The need for draft shields is reduced in an enclosure.
• Just to be sure, I use Layerneer Bed Weld to help the filament adhere.  Clean the bed, then use Layerneer.   It may last many prints. Thin layers of adhesives are better than thick ones.
• Design to fit the printing   
• I had a thin wall , 2.4mm wide.  There was no infill, all profile passes.   One end kept making whiskers that stuck out the side.  As if it was just falling out.   I changed the wall thickness to be exactly divisible by the filament width.   0.4mm tip had a .45 filament width * 6 =2.70mm.   I made the wall 2.70mm wide and the whiskers disappeared.
• Know your printers positioning capability.  In my case, 400 increments per rev. on the servos.   8mm lead on the Z lead screw.  8/400=.02mm'.0008" positioning ability.  However, XY are belt driven.  2mm tooth distance on the belt, 16 teeth on the drive wheel.  2x16=32mm per revolution.  32/400=.08mm.  .003".   This is not small enough to handle making good holes under 3mm.  The "stop-sign" effect from the line increments make the arc close the hole.  That is why we often have to make small holes bigger, 2.4mm, to get 2.0mm actual.  A finer XY grid could possibly do better small details.
• Thru Holes with axis in the XY plane of the filament printer need to be 5mm or more from the end of the wall.  If there is not enough distance to lay down a good bead, it may look more like a blob between the hole and the end of the wall.
• Steep overhangs near the build sheet may not have room for supports.   Use a chamfer, not rounds, on edges next to the build sheet.
• Use high resolution .stl files.  Trying to get a good curve from a low resolution .stl file will not work.   Check your CAD system output settings for .stl files and make sure they make high resolution .stl files.
  
• Slicer settings
• Avoid Grid infill.  I would recommend cubic as a default.   Grid infill crosses over itself while extruding.  With PETG this often causes a scarping sound and buildup around the nozzle.  Grid can also add stress at the crossover causing warping when cooling.   Why is Grid often the default then?  It is fast.
•  Make sure you have the correct profile for the filament in use.  
• Organic supports save material and post processing cleanup time.   They do take longer to print, but post processing savings make up for it.
• Set the slicer for the best output resolution.  On Prusa slicer it is named G-code resolution.   In manufacturing it is called intol/outtol.  
• Filament
  
• Stick with the proven filament as much as possible.  Cheap filament may be expensive to use.  Filament can be bad.  Colors can change how filaments work.   I opened and dried a roll that was 3 years old.   For the first 30 meters of filament it would not stick and turned into a ball.  I checked and the color had been discontinued, non-sticking is probably one reason.    
• Filaments are sold with diameter measurements, but not density and moisture content guarantees. 
• Humidity matters, but don't go crazy about it.  How much it matters varies greatly depending on the material.    
• Old exposed filament is bad filament.  Don't stock pile too much.   Use up or protect what you have.
• When using a buffer for multi material printing, put them in a dry box also.   Many people keep their filament in dry boxes, but then leave them exposed in a buffer.   Sadly most dry boxes do not have room for a buffer or a re-winder. There are exceptions, Tool Chest dry box.
• Be aware of the health hazards a few materials may pose.
• When using tubes and buffers to guide the filament, eliminate drag.   Resistance to feeding the filament  can cause problems.  Buffers may need fine tuning or smoothing to eliminate sharp edges that catch the filament.   Cheap PTFE tubing is cheap for a reason, it often has more drag.
• Each part printed has quirks.  I was printing bridge pieces with "rivets" shown.  The printer made the rivets, but it also caused build up on the nozzle since small details take a small squirt of filament with hundreds or retractions in a few mm of material used.  Resistance in the filament supply caused the extruder gear to wear a grove in the filament and fail.  There are limits to each type of process.
  
• Learn the difference between I did this vs. Engineered parts.  If you are building tools, upgrading the printer or complex designs, look for engineered designs.   Look for comments that state if supports are needed etc. Even Prusa has been improving part designs to aid the ability to print well.
  
• Reward and praise good designs, ignore the rest.   Constructive criticism is an art that does not involve anger, complaints or insults. 
• Nothing is free.  Many have graciously made the models available to everyone.  Some designs have taken hundreds of hours.  At least praise someone for the models you use.
  
• Diamond nozzles once you have experience with 3D printing.  10/2023 You may need to adjust temps.   Expensive, 2-5x, compared to others. It can be used for a wide variety of printing.  The Polycrystaline Diamond doesn't shatter as ruby does, it transfers heat very well and has extreme life compare to brass or steel.    Even a hard Z bed crash usually does not harm a Diamondback tip.  You may be able to drop tip temperatures by 5-10C, since the diamond conducts heat better.  One exception may be small plain PETG parts with corners or overhangs, the diamond may induce some curling.  In most case I have found the standard temperatures work with a diamond nozzle. 
  

What is needed to advance 3D printing to be a reliable process?

1. Designing for past manufacturing methods was very mature.   Designers must learn how to design for 3D printing to make it useful for production work.  3D printing wont replace everything, it has it's niche and use.  Many very good uses.  To take advantage of it Designers will have to learn new design rules and methods.   Just changing the orientation of a 3D part while making the part could have a major effect on the design.  Most importantly, design leaders must learn to use and validate 3D designs and parts. 
   
2. Organic design tools are needed.  CAD software that can also check and implement design rules for 3D modeling. 
   
3. The "post processing" of the data to make manufacturing code must continue to develop.   To be able to design a part 10mm square with a 5mm hole and have it print out to 10mm square and with a 5mm hole within a design tolerance.  Adjusting the model from nominal dimensions to fit tolerance needs to go away.
   

The industry must stop re-inventing the wheel. 

1. The very first 3D machines in 1986 did not use G&M codes and did not use tool paths.  They printed the completed table as a grid and deflected the laser.  They were very slow and expensive.  Customers had to teach the 3D companies the basics of NC programming at the time.   Those designing "additive" machines need to  investigate machine tool design, movement compensation, welding and other overlapping technologies.
2. Trade secrets and patents too easily obtained and widely applied are limiting advancement.
3. Data, data and more data is needed.   The rocket industry embraced 3D printing to help make very expensive and costly rocket engines.  Traditional processes took hundreds to thousands of steps and 9 months to years to build all the complex parts and assemblies.   With 3D printing the process is simplified, overall faster and potentially 10% of the overall cost.   It has taken the rocket industry 30 years to be successful with 3D printed parts. 

Return to 3D printing