Maximum speed

Hi,

Is there any way to measure max speed of components ? I assume I could use a g-code after removing hotend from extruder. But is there limits in the Marlin firmware ?
How about the printhead ? is it limited by firmware ? If so, can I up the limit, move it say 20cm and time the middle 10cm (say using video, and count frames, then I have it down to 1/60th or 1/120th of a sec.

I assume these values could then be used as new max for firmware

If I know the max of my extruder with no load, then I assume this is only reachable if the heating element can melt the plastic fast enough. So I might need to adjust temp up to get the higher speed ? Say I run at 100mm/s = 0.1m/s, PLA, then it is 0.3g/s I need to heat 180 degrees (20 -> 180). 0.3g*180C = 60 calories/s = 250 J/s. That is 250 Watts. This is more than many smaller printers are running at all inclusive, and a 360W PSU is not unusual. All in all I think 100mm/s would be pushing the heater for most printers ? Of course we preheat, and we have many periods of recovery - movement, retraction etc.But the heater will try not to accumulate extra heat in these periods. Since many printers are <200W in normal use, this would lead me to think the nozzle melting capacity is one of the limiting fanctors ? By turning up the nozzle temp, we can actually accumulate some of the energy when the head is not printing.

Assume I make a long line at max printing speed (for practical reasons, lots of lines running full length of X), would it become thinner and thinner if the heater can't keep up ? Or at least at some point the extruder will start to skip ?

How do I find the maximum speed of my specific printer ?

I think many would agree that limiting factors, at this time, are hotends and extruders. Here's a slow motion video of same hotend depositing plastic at different speeds. Looking at the video I see three possibilities:

o extruder unable to keep-up (unlikely in this video but possible in other circumstances)
o firmware not extruding at the proper time - not taking to account delay between extruder and hot-end - this applies to bowden setups mostly
o hot-end not keeping because nozzle has low heat capacity and not enough heat passes from the heatblock to the nozzle.


BTW, I calculate that 9.5 joules are needed to heat 10cm of 0.4mm extrusion of PLA from 20 to 200degC (assuming 1g/cm^3 density).

Edited 3 time(s). Last edit at 05/08/2017 09:25AM by newbob.

Nice video. I think some of what happens at 50 mm/s is, that it is priming the the nozzle (after retract), and depositing a bit too much. And it is clear at 50mm/s that the extruder can't keep up. Maybe because there nozzle is not warm enough (I have fine results to 220+C) to melt it at this speed. But this should result in noise from a skipping extruder, as the extruder is supposed to push enough filement thru. In the video I do not think there is a delay. It primes and moves.

PLA density is around 1.25g/cm^3.
And all I have read about the extruder is that the extruder speed is measured in mm of filament passed thru the extruder per second. So calculating on the output is wrong, if the speed is measured in mm of 1.75mm Ø filament. By measuring the extruder without the nozzle, you can calibrate it by feeding say 10cm, and measuring how much was actually sent thru.

My calculations are based on ~330m of 1.75mm filament per kg, hvis seems to be pretty common. So 10cm = 0.3g or about 3g/m. And 0.3*180 degrees heating ~ 60 cal, then multiply by 4.184 to get joules, and we are around 250 joules. And if this needs to take place in 1 second, it is 250 J/s = 250W. So at 50mm/s, we need 125W continously deposited directly into the melting PLA, and since there is loss, the heater needs to deliver more. A bigger heavier nozzle might help a bit.

Apart from the extruder, I think other limitations often are the acceleration and jerk parameters. If we can add a bit of xyz speed, we gain a bit in the bottom as well, but not much. Going from 25 -> 50mm/s does not give a huge speed increment. But increasing layer from 0.1 to 0.2 will almost cut time in half. Yet, it will need to deposit twice as much filament per movement distance.

Basicly you should be able to speed up at low layer height if the nozzle is the bottleneck, yet most people slows down their printer. But since we are measuring input, 50mm/s should should be twice as long a line for 0.1mm layer heigh.

I am no expert, but trying to understand and tune things. I want to be able to print as fast as possible (like most) while maintaining the quality.

Your sums are way off. PLA isn't water and it's specific heat capacity is very different (about 1800 J/Kg.K, vs 4200 J/Kg.K). I suspect that figure varies quite a bit for different colours and makes of PLA too.

Calculating energy requirements on the output is no more wrong than the input, because they produce the same result. But it's probably easier to calculate the output.

No extruder can extrude 100mm/s of filament, but it's not necessary to either. 100mm of input on a .4mm nozzle is about 1900mm of extruded output- what printer can move at that rate in XY?

I`m not that expert of calculation energy stuff, but when you calculate 250W for the hotend i think there probably is some mistake in the calculaton.


Speed in terms of only movements is normally not the issue, for example i tested the extruder speed at 400m/s printspeed (0,5 nozzle 0.2 layerheight 1,75mm ABS Acc 9000, jerk 50),
the Extrusion rate at E is something between 21 to 22mm/s, so far away from 100mm/s, and most printer even don`t print over 150mm/s.
this tests have been done with an oldschool 5W resister which takes about 25watts, 265°C, a geared extruder with a 1,8 Ncm stepper, and the whole printer only having a 90W PSU.
Most Printers have 40W Heatercartriges which are also good for printing very fast.

Limiting factors of the extrusion speed are primary:
* Style of the Extruder / Geard/nongeard , stepper strengh (normally about 44Ncm)
* grip of the Bolt / Tooth design
* design of the hotzone
* lengh of the transitionzone between hot and cold zone
* design of the Nozzle inside
* materialchoice of the hotend

so before starting such highend calculations, it would be easier to show us your printersetup and what speed do you reach and what the specific failure will appear when printing too fast.

Chri

Edited 1 time(s). Last edit at 05/09/2017 05:47AM by Chri.

---

[chrisu02.wordpress.com] Quadmax Intel Delid Tools

Afaik you need latent heat of melting (∆Hm) for PLA. This won't be a fixed value as I believe it depends on the degree of crystallisation for the specific type of PLA. Either way it's already well known different PLA melt at different temperatures. Agree those figures seem well off, 30W or 40W heater is commonplace and 50mm/s wouldn't be categorised as especially fast at most usable layer heights for a typical 0.4 nozzle. If you really wanted an estimate of how many Joules of energy required for melting a common PLA you could probably measure the energy consumption of your hotend to melt a specific mass of PLA over baseline for the appropriate hotend temperature i.e. how much extra energy was consumed. Probably won't be very accurate but would surely give better figures than what you have. Knowing how much filament your hotend can melt per time period is probably more relevant e.g. crudely put calculate the volume for the maximum mm/s of filament you can reliably extrude for the highest usable print temperature. Extruding in free air is different than regular printing as there will be more back pressure when there is extruded filament cooling around the nozzle tip. So knowing the maximum rate at which you can melt a volume of filament still won't give the maximum build rate i.e. how much volume you can realistically deposit per time period.

Also remember that the ratio of volumes relating to layer height is non linear as it is a quadratic term. This means that you may not necessarily achieve the fastest usable build rate at the maximum head speed of any given printer / configuration. You'd need to test to find out what is achievable. For example with a single wall print at low layer heights you can potentially achieve very high feed rates but it doesn't necessarily mean that it would be the fastest print time. You might be able to print more quickly at a higher layer height and lower feed rates. I think the term print speed is ambiguous and possibly meaningless without more criteria. For real world applications on desktop FFF machines I think it makes more sense to time how quickly you can print a given model under certain basic criteria e.g. layer height, no of perimeters, scale etc.

---

If common sense is so common then why does no one actually seem to have it. Perhaps it should be renamed uncommon sense.

blog [plasticwrap3d.wordpress.com]

If you want to overcome the issues with melting filament fast and increase the volumetric rate of filament through the hotend then in general you want three main improvements.
1) increase heater power, 25-40w cartridges can only heat so much filament in a given time, how about 60w? How about mains powered?
2) Longer melting zone, this is where volcano heat blocks manage more like 35mm³/s than the 15mm³/s that a v6 block is supposed to deliver at maximum.
3) larger heater block to buffer the heat and ensure you aren't getting cooling when the volume is high. 3D printing is very stop-start, its rare to be laying filament down continuously unless your talking about the first few layers or top infill of a large flat object.

Get a volcano (or clone) and screw in the 1.0mm nozzle. Then print. "Great" you say it can lay down big wide filament, but the nozzle cross section has an area of 0.79 mm² compared to a 0.4mm nozzle with a cross section of 0.13mm² so its outputting 6 times more plastic at any given extruder feedrate and can't go any faster than 30-40mm/s or you start to get stringy (under extruded) areas, it just can't melt 6 times the plastic with the same heater cartridge even with a 3 times longer melt zone.

But heavier heat blocks add to inertia on your super rapid printer. This is why I'd favour a 60w heater cartridge on a volcano block as a starting point for a fast extrusion printer.

---

Simon Khoury

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

I have always thought a pre heater for the extruder would probably help, raise it from room temp to something higher, less work for the hotend. As long as you dot go to far so it just gums up the extruder.

Edited 1 time(s). Last edit at 05/10/2017 12:49AM by Dust.

An enclosure achieves this, at 50 deg C you are preheating your filament and hot end.

---

Simon Khoury

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

I'm even happier now about my print speeds after seeing that video doing 50mm/s
perhaps they had a bad printer? or mine is lying to me.(which I think it has sometimes)
many of the thoughts I visualized in my head are shown in the vid, some real nice lines at the slower end.