I’m printing some change gears to check the fitment of the gear train on a banjo. I have been fighting with this particular brand of PETG trying to get a clean and accurate print since last week.
Notice on the 2 20T gears the poor bores and under extruded 5 layer top surface. They were printed with a hardened .40mm nozzle on a Bambu A1 Mini. They do NOT fit the spindles on the banjo at all.
Now look at the larger 80T gear. Clean teeth and bore with proper flow rate on the top surface. (Yes the infill is showing a bit due to the mere 3 top and bottom layers I printed it with to save on filament). This gear fits as it should right off the plate.
Same exact filament, same printer, same exact slicer setting. Just a hardened .60mm nozzle swapped for the hardened .40mm nozzle. And dried to 14%Rh for days and I printed directly for the dryer.
Soooo much time and filament wasted calibrating and recalibrating for that .40mm nozzle and it still failed. All I needed was the .60mm nozzle. But, still a new and interesting phenomenon to learn from for me.
Maybe something to do with your max volumetric flow rate? Could cap that lower and see if anything changes with the 0.4mm
I second this. I’ve had far more issues printing small parts than large due to printing too fast. I like to use minimum layer time, but max flow rate works too.
Been there done that. Didn’t work. It did print the calibration print just fine. It just sucks when I try and print these gears.
Just a question, but why is the large gear solid like that? It’s probably printed at some low infill, so it shouldn’t matter much. But I think you can massively reduce printing time and material needed, while at the same time increasing the strength of the gear, if you were to switch to a spoked design.
You know… I just had to test it.
Just threw together a model with some holes, it’s actually faster to print without the holes, even though it uses slightly less material, because there’s fewer retractions and nozzle hopping around.
I’m sure you can optimize it more, but for a quick test, just throwing a solid disc out there will most likely be faster to print.
edit: accidentally had the gear teeth bigger on the non-hole version - they’re the same now.
Thanks for trying it out. I would model something up, but don’t have time right now. It would also depend on the slicer used, the printer and the settings etc.
It probably doesn’t matter that much and not be worth the effort unless it’s really what you want. But for a quick prototype just having it solid is fine and not worth optimizing.
I’ve tried adding holes in some things and sometimes it uses more filament since the walls often use a lot more filament than sparse infill, even with the holes.
Part lazy design choice and part running these gears for a few minutes at speed to observe real world function. Spoked FDM plastic gears probably would not run long enough at 500rpms and there is no soft start on a lathe. And since most of the strength with FDM printed items lies in the skin of the part, removing half the skin would reduce the rigidity.
Do you think the dryer or the nozzle helped more?
Dryer didn’t do squat to make a difference. The filament was at a steady 14Rh for several days for both nozzles. The only change that made a difference that made a difference was the nozzle swap.
Filament drying is YouTube delusion.
Cue “hydroscopic, meaning plastic absorbs water”.
There are filaments that do need to be dried before use or even dried as you use them. But they are higher end engineering filaments and they cost real money. You tend to need a real reason to buy them. Despite the fact that many consumer level printers have the ability to print them these days. $50 to $100+ a kilogram or less, is a real barrier to the wide use.
Still, years of experience has taught me that I need dry PETG as it tells me it needs to be dried, and TPU is always dried before use. But PLA is a waste of watts.
No kidding! Thanks. I’m going to get a 6mm.
I seldom use a .40mm nozzle much anymore since I mostly do functional prints like these gears. I have found the .60mm nozzle is a good compromise for speed in printing, better strength, and good accuracy.
Interestingly enough, a .40mm will print nearly or as fast as a .60mm nozzle on Bambu printer in general. This because Bambu optimizes their slicer setting for the .40mm nozzle. What it does give up a bit is between layer strength. This is due to to the smaller contact area difference between layers. I can get the same strength with 2 perimeters using a .60mm that would take 3 perimeters with a .40mm nozzle.
Very interesting, thanks.
Have you considered just using gear molds? Or getting some machined if they’re super custom and you’ll be making several of the same gear and replacing them over time.
I need to know if the gear train will fit on the banjo of the lathe first. I might run the plastic just to see how long PETG might last. If I want metal change gears, I will either spring for the 100+ dollars needed to buy the proper gears. Or I will spend $60 for the proper cutter and cut my own gears. But I need to be sure they will fit first.
If it’s just for fitment then all you need at the critical dimensions. You could have just printed the ID of the gear centers to see if they fit on the shift or added a ton of open space on the big gear like how a spoked wheel looks to save lots of filament and print time.
I did say it was a lazy design choice. And change gears a mostly made as solid gears. You very seldom see spoke designs. It has to do with casting/machine time costs. So, solid gears are what I think about with change gears. The gears are still cheap to print and they only took a good afternoon to print all 4 gears and the job is done. And the post is about the weird crappy finish I got with GeeeTech PETG. And the simple nozzle swap solving the problem. And at this point, I just want to burn up this spool to just get rid of it. I won’t buy this brand again.
Turned out the real design sticker was that I needed a 16mm keyed center hole to fit the banjo shafts. And while 16mm hole just barely fits on a mod 1 20T gear, the 4mm key needed does not. I ended up modeling the 20T and 80T gears as one piece and the single 20T gear needed to be designed with a solid 30mm round attached to it to have some place for the drive key to go.
The upside to doing it how I did things meant that yesterday I used those 3D printed gears as made to do the lathe job I needed to do, burnishing a section of shaft in my lathe. Sometimes you just gotta go for it and hope it works out. I might even reprint them with a better PETG brand like DeepLee or better yet, MatterHackers and make them with 4 or 5 perimeters and 6 top and bottom layers with 80% gyroid infill. I have both on hand. For the few times I might need a .0025" per rev feed, they might well last for years. If I find a #2 and #6 pair of used gear cutters for cheap, I might buy them and machine a set from steel. But even decent gear cutters are stupid expensive. And if I don’t have a customer to bill the expense out to, it’s probably not worth thinking about.
