Stepper Smoother

I am going to try out some of these stepper smoothers added inline from the mainboard to the stepper motors.
TL-Smoother from Banggood

The idea behind them is that it adds a flyback/snubber circuit of diodes that smooth the voltage signal going to the stepper motor without adding any other characteristics to the signal. It is supposed to make steppers run quieter and cooler. The end result is less jittery running steppers and better surface finish because of the smoother running.

Here is a diagram purporting to show how the signal is cleaned up.


The R2 doesn't really need much smoothing as it is a pretty good and sturdy design, but since you cannot replace the stepper drivers (A4988) there is a bit of noise associated with the signal and the drivers make the steppers run louder and hotter than they should.

This is just an experiment and I don't expect miraculous results, at least nothing on par with the noise level achieved by TMC21xx drivers. But if it does reduce the noise some and/or reduces some heat then I will consider it a cheap win.

 

Cool, let us know what you see.
From what I see on the Pololu A series (A4988 for example) the RAMPS as an example is driving the steppers with a square wave. Not exactly clean or noise free.

 

I corrected my text, Robo also uses A4988's. I am not sure how well they work, you know you can't trust marketing people, but the are cheap, if you don't buy them from Amazon.


Not sure I believe all the hype. Since I wasn't in a rush for a new project and I didn't want to spend a lot I ordered them from Banggood. Amazon resellers are charging a healthy premium over current Chinese direct prices, for exactly the same product.

 

Theoretically a sine-wave would be a smoother motion and I can believe it would contribute to finer detail, but I have learned not to accept marketing over reality without proof Given the frequency of the wave-form I am unclear as to how much of an improvement they will generate, but I am interested to see.

 

Stepper smoothers are installed on the X and Y axes. Printing a pencil cup in Vase Mode. It was printed previously and I just restarted the exact same G-Code. Here is a section of the original

 

Just over 90 minutes later I have version two of the pencil cup. Printed with the exact same G-Code and the same filament spool. This is the same face as pictured above.


It is not just the faces, the edges appear sharper. So now the test is to see how much faster I can push this printer with the TL-Smoother's installed.

 

BangGood'ed

For that price it is a noticeable difference.

 

I am happy with the results. Now I am going to try to print stuff at 80 - 90 mm/s.

 

Friend of mine has played with them and warns over time you start skipping steps (changing the smoother to a new one solves it) might be something to watch out for if you notice them losing torque

 

I would think that would depend on the quality of the freewheel diodes used, some clones don't use adequately speced flyback diodes. Schottky diodes are preferred. Besides I bought back-ups, used two bought a 4 pack.

Attached is the spec sheet, these are Vishay S3G Rectifiers.

 

Did a bit more testing. I took my current S3D profile and just increased the print speed and increased extrusion temperature. No other changes. There are errors in the print with unknown cause, Primarily looking for ringing that is very evident in a stock R2 printer even at much slower speeds. This Kirby was printed at 85 mm/s print speeds, 205°C hotend & 45°C bed temperatures, 0.60 nozzle, 0.30 layer height. A few things show up at 85 mm/s that don't exist at 60 mm/s. Definitely need to increase the Coast value as the zits appear at the end of the lines. Need to turn off retraction hop except when crossing outside perimeters. Need to play a bit with extrusion multiplier, and infill settings.


The moiré pattern displayed is a result of the lens and resolution used it is not evident when looking at the back side of Kirby with the naked eye.


Aside from some slicing issues and I believe you can overwork OctoPrint with capture settings that are too high, there is little to no ringing. When I printed this before ringing was evident even at 60 mm/s print speed. It took drastic reductions in Jerk settings to help minimize the ringing.


The model was printed with 0% infill so the top edges are not as smooth as they can be. That plus the fact that I am using 0.30 mm layer heights means noticeable jagged edges. All the zits on the model are the result of insufficient Coast length.



I think with a few tweaks on the slicer profile 100 mm/s might be passable. I am going to focus on 85 mm/s and see how much higher the Jerk numbers can go before I see a bunch of ringing on the print.

 

Playing around with various filament types now it seems like somewhere around 85 mm/s is the fastest acceptable speed I can get from ABS, PLA & PET-G on the R2. Above that and the corners tend to round a bit and details are lost. I have returned my slicer profiles for that speed using those filament types. I cannot run at the bottom of the temperature range either, I have to raise the extruder temps about 3°C to keep up with the increase in speed. Something with a longer hot zone, like a Volcano, would not have issues with the increased speed.

I have not tried the Smoothers on E or Z, just X and Y at the moment. The slow movement of the Z and the other factors involved in extruding, I just can't justify putting them on those steppers.

I have also not tried Nylon or polycarbonate filaments yet but those are generally printed slower.

 

@Shrey No firmware changes needed.

 

Got mine in yesterday. I will see about installing them by the weekend at latest.

 

No layer shifts here. Been printing a lot of stuff lately and doing a lot of testing, so my failure rate is increasing but that is because I stop a lot of test prints before they are finished.