Posts made by RCarlyle

@rflulling We had Duet (@T3P3Tony and @Phaedrux), E3D, Lulzbot, Matterhackers, Sublime Publications (Michael Hackney and me) and SeeMeCNC all in the northeast corner of the main building. Was a good crowd, everybody was swamped, particularly E3D.

Hmm, maybe one of these days I should change my handle too, since I don't do any remote control stuff anymore, and what's an "arlyle" anyway?

Congrats on your retirement, I'm jealous of the time you'll have!

@arnix the reason no one is answering your specific size/power/speed questions is that nobody knows how “picks per minute” translates into speed/acceleration for your system. A pick is not a well-defined load, it’s a performance target. We all have no idea what you need because we don’t know enough about your machine, working loads, tolerance for slower picks when crossing the whole working volume, etc.

The required motor power comes from the max value of (torque * angular speed). You can probably get this from your simulation based on the worst case of picks in many different locations around the working area. Torque in Nm multiplied by speed in rad/sec gives power in watts. If you have the exact ideal gearbox for your application, this wattage is how much motor power you will need. But the farther your gearbox ratio is from ideal, the more you need to oversize the motor to make sure the desired speed/torque point lies within the performance curve for the motor.

@genghisnico13 Dang. That's not in my manuscript but it is in the final press file. Must have been introduced in one of the post-layout editing rounds. (Probably when we made all the "open source" vs "open-source" formatting consistent.) Thanks for pointing it out, I'll let my publisher know, maybe we can get it for the second print run.

Can't catch everything. If there's only one typo per chapter, I'll be pretty happy!

@nikker Once Volumes 2 and 3 are complete (they're written but need illustrating/editing) I'm planning to turn Vols 1-3 into a single textbook for a mechatronics course or similar.

@fma We will be working on the e-book after we get through the paper book launch. I agree that overseas shipping makes the price a challenge. Part of the reason the pre-order price is only $34 is to help offset shipping cost to EU and Canada. I am hoping we can show enough demand to get an overseas reseller to help cut distribution costs. (So far so good.)

@CaLviNx Your response tells me I didn't explain the book very well, which is useful feedback Most people in the US are telling me it's underpriced. It's a 370 page full color 8"x10" book with almost 200 illustrations and tables. There's a lot of stuff people around here probably know (like why Ultimakers can move fast) to stuff you probably don't know, like the original CoreXY patent from the 1950s, bearing capacity de-rating factors, rod and extrusion sizing rules, detailed lubricant recommendations, application-specific bearing life calculations...

Check out some sample pages and full table of contents here: http://www.sublimepublications.com/store/p1/3D_Printer_Engineering_Volume_1%3A_Motion_Platform_Design.html

Totally understand if we didn't hit someone's price point. The folks spending $180 on an Ender 3 and keeping it stock probably aren't going to be interested. But the book price is the equivalent of about two spools of filament... cheap filament if you're in the US or good filament if you're in the EU.

[One-time plug, thanks mods]

After years of work, my 3D Printing book is finally in press!
3D Printer Engineering Volume 1: Motion Platform Design
You can pre-order here for 15% off early-bird discount: http://www.sublimepublications.com/

This is a book for anyone who designs, modifies, or maintains 3D printers. Which is everyone here, I think!

The first round of printing is supposed to be done in 1-2 weeks, so it will ship soon. The early bird discount is planned to end at that point.

Table of contents for Volume 1:

• Foreword (it opens with a quote from Jetguy, because hell yeah it does!)

1. Introduction
2. Machine Architectures
3. Popular Machine Architectures
4. Frame Construction
5. Popular Frame Types
6. Motion Hardware
7. Popular Motion Systems

• Closing
• Further Reading

Some of y'all already know this, but yes, this is the first book in a whole series of 3-5 or more books. Volume 2 (Drivetrains) and Volume 3 (Stepper Motors) are fully written, but need illustrating and editing, and that takes a healthy while. So I think Volume 2 is maybe a year out. (I wouldn't wait on more volumes to come out before you buy; I wrote Vol 1 to be a good resource on its own.)

Electronics are planned for Volume 5, sorry that's going to be a good while out

I'll post one more time when shipping starts, and then stop spamming y'all. Thanks to everyone here for years of interesting posts and discussions -- you were part of developing this book!

Closed loop servos also have their own engineering issues, like feedback tuning, jitter, hunting, inertia matching, polling speeds, encoder resolution matching...

The way this typically goes is that someone builds a closed-loop feedback system for 3D printers, everybody goes “ooh, ahh” and then it gets minimal uptake due to added cost and complexity. Mechaduino is a good example; it has a really slick way to read stepper shaft position and had a successful round of sales, but it had kinks to work out and more complexity to deal with and after a year or so everybody kind of lost interest.

If you want to play with servo motion systems, the best ways to start are to either buy a prepackaged motor+encoder+drive that accepts the step/dir signals your Duet already outputs and handles the feedback loop on its own, or get into MachineKit with a Beaglebone setup where servo loops are baked into the controller code. Reinventing the wheel to do it internal to the Duet with some makeshift encoder rig doesn’t make a lot of sense to me.

@zbeeble this discussion goes back at least a decade to the early RepRap days — are open-loop steppers acceptable or do we need some kind of servo feedback?

The short version is, the physics of the stepper motor allow it to execute billions of steps with no drift whatsoever, as long as you configure the system properly (eg motor current and speed limits) and don’t run into anything. You can confirm for yourself by homing at the start and end of the print and seeing if the offsets differ. (They will match within the precision of your endstop switch.) The mechanical drivetrain in a typical Cartesian printer is a lot more reliable and accurate than the noodle of molten plastic squirting out by the extruder at high pressure.

Many, many people have worked on various servo solutions to get closed-loop performance with position feedback. The dream is either an encoder tape on the linear axis, or a non-contact rangefinder to absolutely position the hot end to the build plate. Most people give up on that true linear motion measurement and instead use a rotary encoder of some sort on the motor shaft. That’s a bit less accurate in terms of absolute position (eg doesn’t see belt stretch or backlash) but it plenty good enough to detect skipped stepper steps due to collision etc.

For the rotating sensor options, there’s everything from high-end precision quadrature encoders to simple rotating flags tripping an optical endstop.

My personal opinion is that simple open-loop steppers are more than adequate for small and medium printers. Occasional failed prints are cheaper than closed-loop positioning. Very large printers should probably incorporate some sort of servo position feedback because the loss caused by a failed week-long print is so large.

@dc42 oh, that’s awesome, thanks for the note

@cvmichael I am 100% positive you were never getting 50,000mm/min with Smoothieware and that hardware setup. Some firmware limit was keeping you way below that speed. Might have been a cap on step pulse frequency, or a queue depth limit (since Smoothie never exceeds a speed it can decelerate to 0 from during an E-stop), or some other setting.

0.9 degree steppers wired in series on 12v absolutely cannot run fast. Each motor is only getting effectively 6v when you do that. There is not enough voltage being applied to each coil to overcome inductance and back-emf at high speeds.

On the microstepping note, you never LOSE torque by using finer microstepping. Microstepping can only help until you go so fast you hit a firmware step pulse generation frequency limit. This is a commonly misunderstood issue. It's true that each individual microstep provides less torque with finer microstepping, but at the same time you take proportionately more microsteps, and the torque is additive, so finer microstepping ends up being the same running performance for the motor.

A quick rule of thumb:

• Full-stepping:: never do this, you will get mechanical resonance and lose lots of torque due to torque ripple
• Half-stepping: barely acceptable if you're using a very high gear ratio and you can't generate enough step pulses any other way... still somewhat prone to resonance
• Quarter-stepping: Basically the minimum you should try to use; this is where mechanical resonance instability largely drops off. Fine for geared extruders
• 1/8-stepping: Noticeably better than 1/4 stepping in every way: quieter, smoother, finer motion resolution
• 1/16-stepping: A little better than 1/8 stepping in every way. This is the point where motor angle errors are likely around the same magnitude as your microstep size, and so is a pretty reasonable stopping point
• 1/32-stepping: Quieter than 1/16th but diminishing returns kick in for smoothness and resolution. This is typically the finest microstep size where the human eye will see an improvement in print quality in a very well-tuned Delta.
• 1/64, 1/128, 1/256: Continuing reduction in audible noise, no other meaningful impact.

The one big caveat here is that if you run high speeds with a high steps/mm level, most 3D printer firmwares will start firing multiple step pulses at a time to keep up with the step pulse generation frequency requirements. This effectively coarsens your microstep size dynamically at higher speeds. For example, 8-bit Marlin on RAMPS can only fire the stepper interrupt at 10 kHz, so if you need to fire 10,002 step pulses at 1/16th microstepping per second, it will actually fire 5,001 double-pulses which effectively outputs 1/8th microsteps. Likewise if you hit 20,004 steps per second it will fire 5,001 quad-pulses.

RepRapFirmware will go up to octostepping at very high speeds, meaning if you're trying to do 1/8th steps, you could actually get full-steps. That's generally fine because inductance and back-emf at high speeds screw up the current waveform so much that microstepping doesn't really work anyway. It's all basically full-stepping once you get going fast enough, but that typically happens beyond the resonant frequencies for the motor so it doesn't matter much.

Some more clearance around the hot block sock would be good for the vertical mount pieces -- they're under load and not directly cooled by internal airflow so they're going to be most susceptible to excessive temps.

For example, the one support on the open side that looks like it's touching the sock -- you could support the screw mount off the hot end fan duct with an overhang rather than taking it all the way to the bottom plate past the sock.

@leblond if it’s flat when cold and high in the middle when hot, the problem is probably a physical warping of your build plate from differential thermal expansion. Can you check that with a straight edge with a flashlight behind it?

What’s the physical construction of your bed stack?

Does running auto level with the bed hot change anything?

@dc42 I think the one with zeners I’m remembering was just a homebrew board someone was trying to sell. I agree that flyback diodes are redundant with the driver H-bridge body diodes, although I have seen CLAIMS that additional external flyback diodes provide better suppression. (Lower voltage drop or conducting resistance or something, I don’t know.)

I move motors with the printer powered down constantly, and have never had any issues. Other people report damage. I have a few theories for this:

1. All my printers have screens (Vicki, LCD, PanelDue, whatever). If the motors generate a meaningful amount of power, the LCD backlight will come on and then there’s a load on the system to keep the board capacitors from overcharging too much. If your motor is generating a lot of power, it will act as a very strong brake, so the problem becomes self-limiting.
2. Some board designs have diodes for PSU reverse polarity protection. With such a diode, I think power generated can’t escape the board. Without these diodes, the motor EMF can charge up the PSU capacitors, so you get much more safety cushion before anything overloads.
3. Maybe people are using motors with bad specs (too many coil windings) and that generates more voltage for a given motor RPM. Frankly, it is HARD to hit a motor RPM that will generate 28v of back-emf to overload something.

Just a few theories. I don’t know what the truth of the matter is.

By the way, there ARE some add-on PCBs for Pololu-socket drivers that include zener diodes and flyback diodes to try to prevent motor EMF board damage.

The Smoothie guys had issues with S3D gcode generating very small segments a while back. They made a segment reduction tool, you may be able to find that online somewhere.

An S3D update 1-2 years ago also included a minimum segment size too, I think? Are you running a recent version?

@dc42 what’s the thinking on how you improve elasticity compensation with S-curve motion? (Not disagreeing; just don’t know what the approach is)

Any thoughts on this being a regular product? I know people (like myself) have been interested in the past.

@dc42 other firmwares use extruder jerk with no issues. High E jerk is commonly used to improve retraction performance with geared 3mm filament extruders. Sailfish even allows small instant position jumps with E axis pressure advance. It works great!

Remember, there’s a “torsion spring” torque/error relationship between the stepper driver’s coil energization angle and the physical rotor angle. A position jump of a few microsteps just rapidly changes the instantaneous torque on the rotor. There’s absolutely nothing wrong with doing that on an extruder which is highly damped. In fact, I would say it’s highly desirable since it lets you unload the built-up pressure/compression a lot faster.

Imagine an extruder pushing right up near stall (one full step of load angle) and you want to retract. Why would you gradually accelerate through the ”dead travel” of >1 full step of coil energization angle between peak forward torque and applying reverse torque?