Pulling force of UM Extruder

[Nicolinux 288]

Hi,

I have repeated the test again. This time with a proper digital scale. Another thing I changed was the knurled bolt. I noticed that it didn't align with the passing filament so I moved it a bit down the shaft, closer to the stepper). This made a BIG difference. The extruder can now pull 5.7kg! (<5% variance).

I have repeated the test at different extrusion speeds (in Pronterface/Printrun). When extruding too fast it would grind filament at some point. @gr5: it was a smart idea to extrude slowly at 50mm/min. I used the same filament and the same tools (except for the digital scale). Also didn't touch the extruder current.

 

[Nicolinux 288]

Just for reference. Here are the filament bits. The one on the left was ground down.

Oh and hat tip to Sigi. He had suggested to move the knurled bolt mantle down.

 

[gr5 2,172]

I did another test today to see how sensitive the current control is to the temperature. The quick answer: temperature doesn't seem to affect the current control for the extruder stepper.

First I tested with it set to 500ma mode (weaker than normal). I prefer to do it at weaker than normal because that way I am less likely to grind the filament. I used the same test setup as before although I zeroed the scale this time when the bungee's were mostly loose but the extruder was holding the weight of the vice grips. It was an arbitrary zeroing.

I also pulled the board off the bottom of the machine and twisted it so I had access to the back side. I measured temperatures with a infrared temp probe (doesn't touch the board).

I did multiple tests at each of 3 temperatures:

28C 536g 540g 544g

-5C 558g 550g

40C 560g

Notes about the temperature:

28C: initial temperature when I started measuring. Aiming at resistors I got 28C aiming at chip I got 30C.

-5C: used freeze spray to get it to probably -20C but by the time I finished two tests and measured temp it was warmed up to 5C.

40C: used heat gun - probably heated to a little warmer than 40C - maybe 45C but by the time the test was over it had cooled to 40C

Measurements were made in the order shown above. In other words the 40C test was done last.

These pulling strengths - around 1/2 kg aren't the full amount pulled. Unfortunately I walked away and lost where my zero point was but it was already pulling about 900g (very very rough) so the total pull strength was maybe 1.4kg which conflicts with my earlier graph I believe which pulled closer to 2kg but the 900g is not accurate.

 

[aaronalai 2]

Greetings gr5, and thank you very much for performing the test. I do have a question though? Why didn't you go higher than 40C on the heated part of the test?

If you check out my calculations here:http://umforum.ultimaker.com/index.php?/topic/4542-comedoone-for-ultimaker2/?p=39695

I don't think 40C would do much to the resistors; your delta T would't be that much. My calculations use a delta T of 75C, and users are experiencing much higher operating temps than 40C. The differences in our two temperature differences may explain why you didn't see much of a difference in your pull results.

Thanks again for sharing your personal time with us.

 

[gr5 2,172]

Why didn't you go higher than 40C on the heated part of the test?

I figured you'd ask this - you can do this test yourself when you get your own UM2.

Two reasons:

1) Testing from -5C to 40C is a 45C range of temp and I felt that it was unlikely that Nico's machine or anyone's machine was 40C hotter than mine. In my experience temperature coefficients don't change suddenly at say 75C. They tend to be constant - linear throughout temps below melting temp.

2) I didn't want to damage anything. My heat gun puts out about 200C even on low. I was spraying 200C air on the board while it was powered on for maybe 30 seconds. I could have easily melted insulation on a wire or damaged something else.

 

[aaronalai 2]

Roger that, through our PM I expressed that I even was hesitant of directly changing the board temp. I suppose you were as well. Again, I appreciate the attempt; but I'm still not convinced the resistors where getting hot enough.

Edit, both Mr. Waldorf and Markus measured board temps almost twice that of 40C. I do think the boards are getting twice as hot as yours.

Another Edit,

I don't think anything suddenly happens at 75C, I think that the temperature coefficients (TCs) were not allowed to multiply due to the small temperature changes. My calculations relied heavily on the TCs, remember for each increase in 1 degree C you have to add another ppm to the resistance. I'm saying I don't think your tests were high enough to express the full potential of the TCs.

Also another thing to note, is that due to the lower current demand on the motors you chose, the resistors wouldn't heat as much on their own from current flow as they would if you had gone the full 1.25 amps.

I'm not trying to discredit your experiment, and I really do appreciate all the time you put into it, but I simply think you changed too many variables to make any type of conclusion.

 

[geeks 1]

It is likely, indeed, that this change of variable is significant enough to cause huge differences. The Butterfly Effect!

 

The ideal, this would be of intercaller between the motor and the control circuit an ammeter. A good way to monitor what is happening and when.

 

That said, it takes a meter per phase. Both cables were twice. So two phases. So it takes 2 ammeter.

 

It could that be measured between each step and when the engine is maintennu forcefully. This test would be more reliable I think.

 

PS, use only ammeter with a needle display. Not only are they more accurate but varation is not subject to the display as slow on modern equipment.

 

[aaronalai 2]

gr5 have you thought of trying to make a direct current measurement of the feeder stepper motor? I've got two of these hall effect-based linear current sensors just itching to be interfaced with the motors. Like Geeks said, measuring the the current of both motor coils is probably an accurate way to gauge the current. Here are a couple of datasheets for the units I own:

http://www.robotshop.com/media/files/pdf/allegro-acs714-current-sensor-datasheet-1185.pdf

http://www.robotshop.com/media/files/pdf/datasheet-1185.pdf

They have a very high bandwidth so hopefully will be able to capture the micro-stepping functions very accurately, and the output is 185mV/A so I think they would have enough resolution to see important differences in current. If you have an o-scope, which I imagine you do since you told me you are an electrical engineer, you can just hook up the output directly to that; this is at least my intention. I understand your hesitance to make any permanent modifications to the unit, but I think a simple pigtail or micro-alligator setup would work very well without having to make any modifications to the wire.

Just a thought, it would be nice to see someone else make this measurement so I have someone to compare my results with.

 

[gr5 2,172]

Sorry - I lost interest in current-related issues. Besides, my machine works great so everything I test works fine.

I think my 5kg pull test is more relevant to underextrusion as it is testing many things at the same time (motor current, motor strength, server controller, knurled sleeve, feeder mechanics).

 

[aaronalai 2]

Sorry - I lost interest in current-related issues. Besides, my machine works great so everything I test works fine.

 

Roger that, I'll start a new thread when I have some data. I also have a hot air reflow station for surface mount stuff; it has very finite increments of heat output so I can measure the effects of current as a function of heat without melting anything. I think this would be the most quantifiable way to test the precision resistor hypothesis; and put the idea to bed if it really isn't true.

 

[bertho 6]

I have UM1 so I cannot comment on its performance. I did many measurements on the original extruder after my upgrade.

See:

http://vinland.com/UM_Extruder_Bearing.html

http://vinland.com/Extruder_Drive_Force.html

Instead of using a platform scale, try using a luggage scale. They work great since they are designed for "pulling".

They are very low-cost. I have bought several brands both in China and in the USA and when checking the calibration I have always found them to be accurate.

Bertho

 

[aaronalai 2]

Cool analysis Bertho, thanks for sharing I don't think I would have come across your results on my own. Although your setup is for the UM1, the results are still pretty interesting.

Thanks again

Edit:

In fact there are several neat things on your site :mrgreen:.

 

[mr.-waldorf 0]

I did another test today to see how sensitive the current control is to the temperature.

 

Yes you did and you proved that decreasing the current consumption it will decrease the board temperature

I am sorry George but you "cheated" the test when you decreased the motor stepper current, this will generate less temperature.

"It is a matter of common experience that a conductor, when carrying current, becomes hot

after some time. As explained earlier, an electric current is just a directed flow or drift of electrons

through a substance. The moving electrons as they pass through molecules of atoms of that sub-

stance, collide with other electrons. This electronic collision results in the production of heat. This

explains why passage of current is always accompanied by generation of heat."

If we increase electric consumption this will generate more heat

I measured with a calibrated probe sensor (not IR) during the extrusions tests and I got this values:

rom temp:22ºC

Idle: 22ºC

end of heating up: 37,7ºC

1 layer: 47ºCºC

3mm3/s: 74,0ºC

4mm3/s: 74.1ºC

5mm3/s: 80.8ºC

6mm3/s: 85.1ºC

bla, bla,bla

12mm3/s: 87.1ºC

However I have to agree with you it seems that underextrusion (Stepper motor skipping) is not because of an overheated board, but I certainly will go to ventilate my board since its achieving high temps and this can deacrease the life time of some components

 

Sorry - I lost interest in current-related issues. Besides, my machine works great so everything I test works fine.

 

:cry: maybe a bad sleep night?

My printer works fine too, but I want to help others solving their problems

 

I think my 5kg pull test is more relevant to underextrusion as it is testing many things at the same time (motor current, motor strength, server controller, knurled sleeve, feeder mechanics).

 

Didn't understand how you tested the motor current?

I am sorry maybe my english is not so good so I don´t know what is the server controller, maybe you meant the stepper controller? but if it was that I cant understand how you tested?

In my opinion to test well the knurled sleeve and the feeder material it should be on normal working conditions with the filament on the stock spool holder

Don't bother to repeat the tests because will not get you anywhere...

 

[ArnoldLane 1]

Hi gr5.

I was impressed with the simplicity of your experiment and followed it for a simple test of a new extrusion system that I intend to attach to a UM and just wanted to thank you for providing the information.

I also thought you may be interested in the test results, we tried 5kg which was not a problem, so we tried 15kg's. The 1.75 PLA filament snapped at just over 10kg's before the extruder had any chance of slipping.

Link to video demonstrating test procedure.

Regards

Arnold