theoretical maximum speed

Hi All!

I am looking for the formula for the maximum theoretical travel speed.

Currently my settings are 160 both in X and in Y direction for the
"steps per mm" value.

I do NOT want to know what maximum speed my machine
can possibly deal with - THIS IS AN ENTIRE DIFFERENT TOPIC.

I want to know the maximum speed that Arduino or the PrintrBoard
or the Teensylu can handle with its usual Processor Frequency.

This is - so to speak - the THEORETICAL MAXIMUM SPEED
that MARLIN can handle with a 3 axes printer -
This is the theoretical speed I want to know.

I do not need to know exact values - I only want to know some
clues on how I can estimate the maximum theoretical speed
a microcontroller could possible provide.

Also I need an estimation on how big this maximum theoretical
travel speed could be with a faster ARM processor like in
smoothieboard.

many thanks in advance

Gerald

PS:
of course I am aware that the Microcontroller must send
many step commands to the output pins and that the higher
the microstepping the more will it be difficult for the
microcontroller to keep track with it,
Anyway I am not so deep in this matter and so I
am unable to find the estimation formula by myself.

Edited 5 time(s). Last edit at 01/04/2017 06:41PM by GeraldTrost.

... IIRC there are some rough numbers for the max. step frequenzies - below 80kHz for the smaller Arduinos and 150kHz up to 300kHz with the Due.

So you have to calculate the speeds with your settings for "steps per mm" ...

---

Viktor
--------
Aufruf zum Projekt "Müll-freie Meere" - [reprap.org] -- Deutsche Facebook-Gruppe - [www.facebook.com]

Call for the project "garbage-free seas" - [reprap.org]

I've measured around 125 mm/s max speed on Repetier, using 200 steps/mm. Note that this was a worst case scenario using all 3 axes at once. If only using two axes at once (you're unlikely to be slamming the Z value around), then I'd expect it to be higher. Ballpark figure would be something in the order of 225 mm/s when only using the X and Y axes, and your steps/mm value. This agrees roughly with figures given by DJDemonD on his CoreXY printer.

The limiting factor is how fast the step lines can be toggled using software. You need to divide this figure by the number of axes you're driving at once. I've seen this figure available somewhere... not sure where it is though.

As for 32 bit boards, the speed is effectively unlimited. This is because your printer is unlikely to ever be physically able to throw your head around faster than 200 mm/s. I don't know what the maximum would be, but I'd expect it to be stupidly, ridiculously high.

Edited 2 time(s). Last edit at 01/04/2017 10:53PM by nebbian.

take a look at [reprap.org]

... see the two linked videos - there I was testing the fastest possible speeds with a setting of 425 steps/mm ...

[vimeo.com]

[vimeo.com]

---

Viktor
--------
Aufruf zum Projekt "Müll-freie Meere" - [reprap.org] -- Deutsche Facebook-Gruppe - [www.facebook.com]

Call for the project "garbage-free seas" - [reprap.org]

I can move my delta effector on the large kossel with duetwifi, at 20000mm/min or 333mm/sec. The Z height is 465mm so it can accelerate to this speed in that distance with fairly severe acceleration settings such as 6000. I am using 200 steps/mm, 0.9deg motors 16t pulleys and 24v. I haven't tried to go any faster, that's scary enough for me.

As for whether this speed has any practical application for printing, even for travel moves I'd say not. I'd be concerned with travel speeds this high that it would pull parts from the bed if there was any friction from blobs or slightly over-extruded areas even with 0.5mm z lift on retract. You just don't give the nozzle time to heat its way through any filament which is cold and slightly above the nozzle's level.

Edited 1 time(s). Last edit at 01/06/2017 03:59AM by DjDemonD.

---

Simon Khoury

Co-founder of [www.precisionpiezo.co.uk] Accurate, repeatable, versatile Z-Probes
Published:Inventions

... my gut feelings says, that FDM/FFF-printing with 150mm/s should be enough for rough parts and even slower to get higher accuracy ... but for SLS-printing I'm trying to go as fast as I can - this could be up to 1m/s with a cartesian mechanics with belts, but then not with free XY-vector moving (this won't be accurate enough for mechanical oscillations/wobbling), but only with "hatching" along X with single-steps in Y per hatch line ...

---

Viktor
--------
Aufruf zum Projekt "Müll-freie Meere" - [reprap.org] -- Deutsche Facebook-Gruppe - [www.facebook.com]

Call for the project "garbage-free seas" - [reprap.org]

hi all!

thanks for the suggestions.

As I said - I dont want to print faster than the printer can move.

but in case I added 6 more extruders then I need to know
wheater I need a 32bit board or can I still use an Arduino - class board ?

will the microcontroller keep up with so many step
signals if all the motors used 32bit microstepping
or 128bit microspepping?

I will look at you answers in detail

many thanks
Gerald

The more axes moving (inc extruders) at the same time the more step pulses the processor has to generate. So cartesian machines put less strain on the processor as the z axis isn't moving much. Corexy the same. Delta is a trickier prospect as all three movement axes are constantly in motion. 32bit controllers can output more step pulses. So for example ramps/mega is probably limited to around 40,000/s - 40khz. If your x and y are 80 steps/mm and you want to move at 100mm/s thats 80x100=8000 x 2 = 16000, add extruder and z axis and you can see that its quite possible to be under 40khz. However if you begin using smaller pulleys i.e. 16t instead of 20t that adds 25% more steps pulses but improves resolution. 0.9deg motors double the step pulses again but improve resolution. Other things like graphical LCD's add even more processor overhead.

If you want to move fast and have lots of axes then 32bit will allow you to do this. For example the latest reprap firmware for duetwifi should allow 800khz step pulses, which means deltas with small pulleys and 0.9 deg motors and lots of extruders can still move fast.

If you start using 1/128th microstepping though you can still run into the controllers limits as your steps/mm will increase by a factor of 8.

On my large delta (0.9deg motors, 16t pulleys) 200 steps/mm I can still hit the processor limits on reprap firmware/duetwifi if I set 1/256th microstepping, which means my steps/mm are now 3200. As it happens this controller uses TMC 2660's so the stepper drivers can interpolate 1/256th but only actually need 1/16th step pulses from the processor, result ultra quiet motors but high speeds possible.

You could always build it running on an 8bit controller see how fast it will go, then swap the controller for a 32bit one later when you want to up the speed.

---

Simon Khoury

Co-founder of [www.precisionpiezo.co.uk] Accurate, repeatable, versatile Z-Probes
Published:Inventions