So, I've gone nuts, and installed a 0.9 degree stepper on my extruder. I'm running Marlin 220.127.116.11 on the Predator, and had previously calibrated the extruder to 410 steps per mm. So I upped that to 820 to compensate for the new motor. I had to swap two wires, and the results are well worth the effort, even without fine tuning or calibration. No ghosting, no salmon skin, no blobs, no stringing.
Tomorrow, I'm going whole hog and replacing the axis steppers with 0.9 degree units, and replacing the drive pulleys with genuine Gates 16 tooth units, the belts with genuine Gates 2GT3, and the idlers with genuine Gates 20 tooth units (to replace the stock smooth idlers). I've also got an E3D V6 hotend, which will take a bit of machining to install. But steppers, belts, and pulleys first. Just placing the new Gates belt in the new Gates pulleys feels distinctly more "precise" by hand, so I'm looking forward to the results.
I'm also going to install Sorbothane 40 OO pads under the feet to help in killing mechanical crosstalk between the x,y, and z motions. With these mods, the machine is down to 5 micron resolution.
I obtained all of the parts (motors, pulleys, belts, hot end) from E3D-Online, and they all arrived 3 days after I placed the order. At least for the extruder motor, I didn't have to do anything other than re-set the steps per millimeter, and shorten three screws on the extruder because I'm direct drive (no bowden).
I'll follow up after the axis motor, pulley, and belt transplant.
These are excellent improvements. I've copied most of the improvements you've made, in order to get a more accurate machine. M48 reports 0.02, with the stock probe. With a bit of luck, I'll be able to piece together the marvelous piezo probe you've developed. Using an E3D hotend and extruder has helped a lot, in addition to the 0.9* steppers and gates belts. And Marlin 2.1.1 works very well with MPC. The only thing holding my machine back now, is the lack of a cooling fan duct and the stock fans. I want to thank you for all of the effort you put into modding this machine, and for the highly detailed and useful information you provide. Some people have a mix of good and bad mods and ideas. But every single idea you've had has a purpose, and notably improves printer performance
What model of the E3D stepper motors you install? It seems like their Compact but Powerful version should have comparable torque to the stock motors, and should be adequate but I'm tempted to get the High Torque version for the axis motors, or the LDO equivalent. I'll probably go with the LDOs TBH, just looking at the spec sheets they appear to be the same motors, and they're a few bucks cheaper where I'm at.
I'm going to do an overhaul on my printer this summer, Hayden's magball arms, 0.9 steppers all around, linear rails, new pullies and gates belts. Now that everything is starting to wear out I feel it's time. I've already installed a BMG extruder, Phaetus Dragon hi flow hotend, magnetic bed, a knockoff Duet 2, and a ton of other modifications.
As I write this I finally realized what I think is the best, and relatively simple way to install linear rails. Drill and tap holes into some flat bar steel for the rail, drill holes into the extrusion that matches the hole spacing on the rails, slip the flat bar into the tower extrusions and use the flatbars to clamp the rails to the extrusions. The other methods I have seen like tapping into the 2mm thick extrusions or relying 3d printed parts to attach the rails just seemed janky. It should be a heck of a lot cheaper, stronger and hopefully ear than using aluminum u channel like what hamrx8 did. 15 dollars worth of steel, a transfer punch, a decent m3 tap, a few hours of my time and I should be G2G.
I guess the rails will add some resistance now that I think of it, I'll probably just get the higher torque motors lol.
I used the E3D Slimline (1701HSM140AE) on the extruder, and the High Torque (1704HSM168RE) on the axes. All are 0.9 degree steppers, and of very high quality. Don't forget to adjust the drive current. I set mine to 75% of max rated for all four motors, which seems to work well.
I'm currently working on a completely different extruder design with a Sanyo Denkin NEMA 17 stepper that is only 10 mm thick. The entire extruder and hot end should come in at under 100 grams, although my current setup is direct drive and has no issues even at 300 mm/sec.
Should be interesting to see what you come up with, that is friggin' light.
I managed to shave about 50-60 grams off of the stock effector/hotend with the one I designed for my dragon hotend, probably could have made it lighter but I tend to overbuild things to make up for my lack of engineering knowledge. Not completely happy with the design, so I haven't bothered publishing it.
I've been digging around a bit and I have at least narrowed down what I'm going to use for my axis motors.
I'm probably overthinking this, just about any 48mm 0.9 degree motor should be more fine for what I'm doing. I'm going to go get some work done outside and let a random number generator decide for me if I haven't made up my mind when I'm done.
The 42HM48-2404 is a generic stepper manufactured by several reputable manufacturers. However, it is rated at 2.4 amps, so cannot be driven to it's full specs by the A4988 drivers in the Predator. I'm not familiar with the LDO unit.
For what it's worth, I'd go with the E3D units. They are available and known to work (at least by me). The prices are very reasonable. The price difference, even with shipping, is less than the cost of a spool of filament. E3D also has industrial quality pulleys, belts, and other critical parts so you can place one order with one shipping charge.
Even with stock Gates pulleys, I true them in a lathe before use. If you go for cheap on Ebay, I can pretty much guarantee that you will get what you pay for, and you will regret it. Accuracy, repeatability, and reliability are not cheap.
I have a Duet 2 the drivers go up to 3 amps, so no worries there. As far as the pullies idlers and belts I'm going with genuine Gates parts of off Filastruder, they also carry Hayden's Magball arms and as a bonus they're slightly cheaper than off of the magballarms.com site. LDO is from what I understand a very reputable supplier of motors, extrusions, rails ect. Regardless the random number generator has spoken, E3D motors it is lol. I also don't have access to a lathe, I wish I did, maybe some day.
EDIT: I poked around a bit more, got what should be the real specs off the LDO site for the LDO-42STH47-1684, they're identical to the E3D equivalent, so there's a good chance that it's literally the same motor. It looks like it's LDOs all around for me off of Printed Solid, saved like 20 bucks vs E3D. I also went with the LDO-42STH25-1404MAC for the extruder, same size as the E3D slimline but it's a bit more powerful.
Have you already upgraded the stepper drivers? If not, that will make a much more significant improvement compared to the stock A4988 drivers that come with the Predator.
By "upgrade the stepper drivers" I'm going to assume that you mean switching to one of the Trinamic/Maxim devices? I'm not sure where folks got the idea that they are somehow "better". Sure, they have an ultrasonic PWM chopper frequency, so they are quiet (unless you're a dog), and they have all kinds of nifty bells and whistles (on the fly coil current adjustment, bus communications, stall detection) none of which are of any interest to me. Nor, I would venture to guess, of they of much use to the vast majority of printer users.
As an electrical engineer, the Trinamic step functions are identical to those of the A4988 (as they should be, as they are determined by the physics of the motor). The decay functions for the internal H-bridges are also identical, with the same potential for "glitches" as with the A4988. It's just that the glitches happen faster, so you can't hear them.
A quick perusal of the manufacturer's data sheets shows that Trinamic drivers have the same slow decay issue as the A4988. You can see it on their switching function graphs. At least on the A4988, the decay function can be forced to avoid that glitch by reducing the oscillator pin impedance to zero. Not so on the Trinamic drivers. You're stuck with the waveform they give you, with no chance of modifying it.
Sometimes the latest and greatest isn't necessarily the best, especially in electronics. Functions get added "because they can". My iPhone is a pretty decent camera and GPS navigator. It certainly is not the worlds best telephone. Just food for thought.