aaronalai
-
Posts
470 -
Joined
-
Last visited
Content Type
Forums
Events
3D Prints
Posts posted by aaronalai
-
-
Hi Guys,
I found that if you overtighten the 4 headscrews this also causes under extrusion. I loosened the 4 head screws and now I can achieve 12mm3/s even with a half empty spool.
Try loosening the screws as far as it can, keeping together the plastics.
(could be necessary to re-calibrate the bed)
Very interesting find Mr. Waldorf. Since it's so simple to implement, I am excited to see the results of other members in this thread after making the adjustment.
By how much would you say you loosened the four bolts? Is the black plastic holding the head together noticeably looser or is it still snugly fit together just not too tightly?
I wonder if because the knurled portion of the bolt is so large, that over tightening it is a common mistake when UM assembles the units.
-
Ahahah, lol; I'm in :mrgreen:. I love all the ladies, and snuggling with cats.
-
Hey take your time Markus, most of us in the forums are adults with full time jobs, children, bills to pay ... and all those fun adult responsibilities; and probably see the forum as a luxury. We all understand
. The only reason I respond so quickly sometimes is because I don't have a girlfriend or children, and I have already done all of my chores for the day; I'm just a single dude with a cat and a deep fascination with 3D printers at the moment :mrgreen:.
-
Alright now for those calculations you have all be dying to see :mrgreen:
First some assumptions and some facts:
Assumptions:
A 25 to 100 degree change in temperature: delta T = 75C (delta means change in something and is often represented with a little triangle from the greek alphabet).
0.005 ohms of trace resistance (the resistance of the copper on the printed circuit board)
0.393%/C change in trace resistance (how much the resistance of the copper on the printed circuit board changes with each 1 degC increase in temperature)
The precision resistor is at it's full 1% error (more resistive by 1%)
The precision resistor is at it's full 200ppm/C thermal coefficient (this describes how much the resistor value changes as a function of a change in temperature; this says that the resistor will not change resistance by more than 0.0002 ohms per deg C if it is operating within its stated temperature range measured from room temp, which is understood to be 25 C)
Facts:
The precision resistor is 0.05 ohms
The function that describes how the stepper controller chip decides when there is enough current going to the motor is (Vref)/(8*precision resistor value)
Vref is the reference voltage facilitated by the microcontroller chip; this is how the current is changed in the firmware.
Alright, let's do some calculations:
delta T of 75 * 0.393%/C trace resistance change = 29.475% change in trace resistance
0.29475 * 0.005 ohms of trace resistance = 0.001474 ohms of increased trace resistance
0.05 ohm precision resistor * 0.01 (the percent error) = 0.0005 ohms of increased resistor resistance
0.05 ohms + 0.0005 ohms = 0.0505 ohms of resistor resistance
0.0505 ohms * 0.0002 (thermal coefficient) * delta T of 75 = 0.000758 ohms of increased resistor resistance
0.0505 ohms (a resistor that is 1% higher in resistance than its nominal value) + 0.001474 ohms (the increased trace resistance) + 0.000758 ohms (the increased resistance of the precision resistor as a function of its thermal coefficient and our temperature change) = 0.0527 ohms of resistance. (the astute reader will recognize I made a mistake in my last post, geez where is my head :shock:)
I don't know for sure what the Vref value should be, if Cohen did not include the trace resistance in his Vref calculation and only used the value of the precision resistor. Vref should be:
1.25 amps (max current) = Vref / (8*0.05)
Some algebra and we find Vref to be 0.5 volts
For now if we take Vref to be 0.5 volts and plug in our new resistor value of 0.0527 ohms we get a new max current of 1.185 amps
(0.5 volts) / (8*0.0527 ohms) = 1.185 amps
or 64.7 mA less than our target goal of 1.25 amps
This is where things get a little tricky and why I wish Cohen would come back to the thread. See if he did include the trace resistance in his calculations my Vref assumption would be off. But if he didn't include the trace resistance in his calculation for Vref no one would get 1.25 amps going to their stepper motor and everyone would be getting some small value less than 1.25 amps, which probably isn't a big deal until you include the compounding resistance increases calculated above.
Let's include the increase in trace resistance AND the trace resistance in our max current calculation.
(0.5 volts) / (8*(0.0527 ohms + 0.005 ohms)) = 1.082 amps
or a 167.4 mA reduction in the amount of current going to the stepper motor.
Take these values with a grain of salt. The trace resistance may be estimated too high, the Vref may be calculated with the trace resistance in mind meaning the overall change in current isn't as big as calculated; but inversely the resistor temp may be getting higher than 100 C and the trace resistance could be higher than estimated meaning the overall change in current is bigger than calculated. The take home message is that if there is a resistor soldered to a board that is within spec then there is a chance of having a roughly 13% drop in current going to the stepper motor under normal operating conditions from the expected 1.25 amps we all think is reaching the stepper motor.
Edited to make a clearer distinction of which resistances I am computing.
Edit for the more visually inclined:
-
Regarding the stepper accuracy; it's dictated by the stepper hardware itself. A 1.8 degree/step motor cannot get the same resolution as a 0.9 degree/step motor. The pole resolution is more coarse on the stator, this site has some pretty good pictures:
http://www.shinano.com/stepping-motor-operation-theory-pg2.php
The stepper controller chip will still do 1/2, 1/4, 1/8, or 1/16th a turn, but the turn is dictated by the hardware inside the motor.
It would be interesting to see how well you can print something with a lot of retractions; I'm not sure if reducing the retraction length in half and reducing the flow by about half can compensate for the courser stepper resolution. Not saying it's impossible, I really simply don't know. Perhaps you could try reprinting stuff you successfully printed with your old motor to detect a difference in the quality of your prints.
From your posted video the print seems to look okay, have you been able to control your over extrusion yet? or are you still trying to find a suitable setting?
Don't worry though, as all is not lost; if the more course motor does not serve the purposes you need it to, you can simply solder your old motor back to the motor lead wires. But you may end up in the same boat as everyone else who has under-extrusion issues, but there seem to be many solutions out there and one of them may remedy your problem; increased stepper current, different extruder design, smoother spool holder.
Please report your findings with your course motor though; I think since you already have it installed it wouldn't hurt to print out some of your previous prints to gauge the success of your new stepper motor.
Regarding this development about the stepper though, the circuitry overheating hypothesis may not pan out. This being said I'm still going to look more closely into the matter (I'm going to probe the hell out of my machine), and I hope gr5 still does his temperature manipulation of the resistor components with the mass pull test. I agree with him, and also believe it would yield valuable insight into the thermal sensitivity of the precision resistors.
It seems like increasing the current going to the extruder motor can help in some cases. If the current to the extruder motor needs to be increased, I would ask myself why this is the case. I still think the precision resistors drifting out of spec could explain the reason; but still nothing conclusive yet.
-
Hey illuminarti, you are correct. The stepper resolution of the motor Markus purchased is 1.8 deg/step or 200 steps/revolution.
The plot thickens...
-
@Aaron. Thanks for doing the calculations. 55.4mA does not seem anywhere near enough to cause issues, and certainly not in my case.
There is another "feature" of the A4988 no-one has mentioned so far: Thermal Shutdown, which happens at a junction temperature of around 165C. I don't think its caused any issues so far as it would cause a total stop of the extruder motor.
I think I made an error in what to include in my calculations:
No problem, I was headed out the door for work and a thought just occurred to me so I rechecked my calculations. I forgot to include the trace resistance in my calculations and only included the % change in trace resistance. If I include the trace resistance then I get that the stepper motor is only receiveing 1.090 Amps of current or a reduction of 159 mA; this could be quite significant. I'll post all of my calculations later today for scrutiny, I could take a picture of my notebook if anyone wants to try and wade through my handwriting though :shock:.
-
I seriously doubt this is a problem. For one thing, the other steppers all run at a higher current. The feeder is at the lowest setting. For another, resistors tend to be pretty tough. We are running them at .078 watts max typically but in reality if the feeder is moving then the current goes up and down depending on the current step and so on average should be at half that power. And it's rated to .125 watts.
It's possible but it just doesn't seem likely. Eventually I plan to spray cold spray and heat gun on those 2 resistors and repeat my "pull test".
I've been thinking about this as well. If this resistor phenomena is happening to the extruder, why isn't it happening to the other steppers with the exact same circuitry. I think that it could be happening to the other stepper motors it's just we can't see the results of the lower current flow as a function of how the steppers are setup inside the machine. Also, now that you are saying that these steppers use more current than the extruder motors, it may be even more difficult to perceive the problem.
The extruder stepper motor is constantly building up potential energy in the form of a flexing bowden tube and corresponding filament. This is why it has to retract so much, to dissipate the potential energy then the filament physically translates a bit. Throughout the entire print cycle the extruder stepper is the only motor with constant opposing force being applied to its axle. The other stepper motors are not burdened with this build up of potential energy/constant opposing force, although they are to some extent via the momentum of the print head, but this may be much lower and intermittent. So I don't know if a comparison between the two different operating procedures can be made.
I agree the average current going through the resistor would be less than 0.078 watts due to the nature of the stepping sequence and corresponding current application. But there must be 0.078 watts being facilitated by the resistor at discrete time intervals, and the Derating Curve may come into play at these specific instances in time, where the peak current is at its highest.
I appreciate your willingness to manipulate the temperature of the resistors, I think this will provide valuable insight as to how sensitive the resistors are to heat. Looking forward to your results, knowing full well you are committing your own free time.
Thank you
-
Greetings Markus,
I fear that bringing up the resistor idea has derailed your thread, and people may not look at the other modifications you made.
This was not my intention, I think the modifications you made were useful and I was trying to help explain why they worked. You also put fans on the motors which is something the thread hasn't talked about much, and my be important. From my experience though, I think steppers are designed to run hot so the fans on the stepper motors were lower on my list of explanations. That being said, the fans may help in conjunction with your other modifications.
I wanted to address three points you brought up in your last post:
Policy discussions are very important , I am sure. But I miss the simplicity of curious experimentation with possible alternatives.
I agree with you, I would like to see some more active testing and probing of the machine to see if there are consistent results from members in the forum. I also agree with Nicolinux though, that many people may be satisfied with how their printer is operating. This may be a temporary satisfaction though, as some people may find their print results degrading over time; so perhaps curiosity will grow over time. I also believe some people may not be as daring as you and do not perceive their UM2 like they did the hackable UM1.
Rest assured though, that when I get my UM2 I will be measuring and adjusting everything I am capable of and will report my findings in as detailed a fashion as I can muster. Even at work when we purchase new transducers and equipment that we pay 100,000+ dollars for we still need to test them. The tools are so important that we need assurance that they are functioning correctly, and I view the UM2 with such importance.
Some people in the forum may scoff at the idea; 100K+ dollars for something, you would think it would work without fault. I can assure you this is definitely not the case. I have found numerous very very important errors in very expensive equipment we have purchased. Without testing these pieces of equipment before a big test we would have not collected any data and would have lost several hundred thousands in time and specimen costs. I have learned that blind faith in something working as the manufacture intended is not always the best idea, and that rigorously testing the product is very important.
- I need to learn how to make simple and almost correct tests, has anyone have suggestions? What should I start? I am looking for a brief explanation and a test file.
Here is my suggestion, I know I brought it up earlier with you but I think this may yield some interesting findings.
1. Print out illuminarti's test cylinder here all you have to do is put it on your SD card and print.
2. Print out the cylinder with and without your active fan cooling turned on. I would print it with your added fans on first, then turn your added fans off to print it a second time.
3. Make regular notes of the system temperature provided by your snazzy monitoring setup.
4. I would then replace your more efficient extrusion motor with the original one provided with the machine, and repeat steps 1,2, and 3 again; providing a total of 4 cylinder tests.
5. For future printing I would purchase one of these http://www.robotshop.com/en/pololu-5a-acs715-current-sensor.html to constantly monitor current during a print. I can't be positive these will work for this application, but I've used them in the past for other stuff with good success. If you need help figuring out a way to interface it with your machine and read the data coming out of the device, I would be more than happy to help.
6. Check out Mr. Waldorf's thread here it's pretty interesting to see how others are either successfully or not successfully printing out the test cylinder.
- I can not fast enough absorb and understand information because I am occupationally very busy.
We very much understand, there is no need to feel rushed to respond or post something. Look at Nicolinux's mega post about under-extrusion; that thing has been going on for almost two solid months with 28 pages of responses. Take your time, we appreciate the work you have taken thus far, and I believe forum members will contribute to the best of their abilities.
:-P
-
Also, where did Cohen go? I hope he brings some news to the thread, even if he thinks it has nothing to do with the resistors; I would be very interested to hear what he has to say about this hypothesis either way.
-
Hey guys, I've calculated a worse case scenario if we make the following assumptions, WoofysPlace's trace assumptions (pretty reasonable, http://umforum.ultimaker.com/index.php?/topic/4542-comedoone-for-ultimaker2/?p=39341), a change in temp from 25 C to 100 C, and a resistor that is at it's full 1% error and it's full TC of 200 ppm/C as per specified by it's data sheet. The motor will receive 1.194 Amps of current, or a 55.4 mA reduction in current.
According to gr5's graph http://umforum.ultimaker.com/index.php?/topic/4222-pulling-force-of-um-extruder/?p=35887it should still be able to pull approximately 5 kg of mass, although his tests indirectly measured current as a function of the firmware.
Is there anyone out there with something like this: http://www.robotshop.com/en/pololu-5a-acs715-current-sensor.html and can actually measure the current over long durations of time? I have one but don't have the printer yet, but will definitely be interfacing it with my printer almost as soon as it arrives just to monitor the current over time. I can make a pig tail connector so I don't have to cut any wires and make any permanent alterations to the printer.
This document will help provide some insight into how I am looking at things and why I brought this resistor thing up in the first place, it is a general overview of how resistors drift out of spec. http://www.vishaypg.com/docs/49865/selresis.pdf It's pretty interesting even if you don't have much of a technical background.
Edit:
Yeah, Mr. Waldorf has accurately summarized what I was trying to say earlier in post #41.
-
@IRobertI, I ordered one a while ago when you mentioned it earlier, and it arrived yesterday; a week before my printer is suppose to ship, thanks for the tip I can definitely see this being useful for the stuff I plan on printing.
-
Hey guys, before I go jabbering on again I just wanted to say I view this thread as a casual conversation with ya'll, and that I'm not trying to spam it with ideas or cause confusion. I hope you view it likewise
. We have meetings at work where we try to figure stuff out and all types of ideas are thrown around, I think it really helps to see things from many different points of view.Has anyone downloaded the datasheet for the precision resistors gr5 posted?
I thought the Derating Curve on page two of the resistor datasheet was interesting. If the ambient temp get's above 70C the rated power of the resistor plummets; perhaps heating up the resistor and then running 0.078 watts through it is slowly damaging the thing.
Up to 70C the resistors can facilitate 0.125 watts; with 1.25 Amps going through them at a nominal resistance of 0.05 ohms they are only facilitating 0.0781 watts. But if they are heated up too much and the power rating starts to drop off, perhaps trying to pump 0.078 watts through them is slowly killing them. It's hard to tell from the graph but it looks like at about 80C the resistors are only at 60% their power rating, which would be 0.075 watts.
Edit:
Regarding illuminarti's suggestion of increasing the current; I don't know if this would be the best idea if the resistors really starting to drift out of spec. Increasing the current may make the problem worse over time to the point where increasing the current won't work anymore and parts need to be replaced.
-
@Cor Kiekens
ifwhen your print is successful, you must share your settings :-P. It looks so intricate and I see almost no stringing. Are you using the spool holder that comes stock with the UM2? Great looking print so far!lololol @Xeno, nice print nonetheless. I have a cat with litter that doesn't work well with the scooper I have and was going to print out a better one, I'll use your print for inspiration :mrgreen:.
-
And another final thought before I go to bed, thank you for your post illuminarti. Woah, I didn't realize you have tried this on 10 machines without any problems. This is something to digest for sure. I also like your suggestion for the user intervention to tell the machine the motor is stalling, I asked myself the same question you did as to weather or not current monitoring could be used to detect stalls and the short answer is yes it's possible, but not as straight forward as just measuring the current here is the thread if you are interested: http://umforum.ultimaker.com/index.php?/topic/4466-monitoring-current-from-extruder-motor-for-feedback/.
-
Just one final thought before I go to bed, else it will be going through my mind and I won't be able to sleep. The ambient temp of the air around the chip may be cooler than 85C, the thermal transmission properties of the solid material that comprises the chips and pcb is going to be more efficient than those of the air around the chip.
-
Shoot, I knew it seemed to good to be true; the temp on the board increasing to the exact stated Absolute Max Ratings Ambient Temp, then failing. I wanted to believe, oh well; it is still curious though that the chip is clearly being exposed to temps larger than what can be tolerated according to the datasheet I still don't think this can be a good thing for that specific component. I wonder what Cohen's thoughts are on this?
Regardless, still interested to see the results of you next print; maybe it's simply a combination of the filament spool diameter and the chip as you are suggesting. It's been a long day at work, and I need to get some sleep and a clear head to think about this with a fresh perspective.
Edit:
At least we solved one thing this evening; what those damn kids were doing in your garden. I lived in a bad neighborhood once before and got stuff stolen from me, that really shot my nerves as well. Just a side note investing in an IP surveillance camera really helped put my mind at ease.
One more Edit:
I totally fell asleep earlier and woke up to respond, and didn't do those calculations I promised, I haven't forgotten about them and will get to them tomorrow.
-
If you or Markus are ever in Virginia, I'm going to treat you like kings; thanks again for your efforts.
-
Hot damn!! It could still be a coincidence; anyone else out there want to try this measurement? Mr. Waldorf could you post some images of where you took the measurement. I am very appreciative of your efforts and am excited to see how this progresses.
-
Hey Mr. Waldorf I would try to measure as close to behind the chip and resistors as possible. Try probing around and noting your findings, I think if you measure the temp with the cover off then the value may be much lower than it would be with the cover on, so if there is a way to adhere the thermal probe to the board then stick everything back together that would be most ideal. You are probably very aware of this, but I would feel bad if I was in anyway responsible for you shorting something out, but if your probe is conductive make sure you insulate it before sticking it to the board.
Edit:
Hey Mr. Waldorf, look at page 12 of the datasheet I linked to above, it has a cross sectional picture of the Thermal Vias which may provide useful when deciding where to place your probe.
-
Alright guys, not to keep flooding the thread every time an idea pops into my head, but has anyone downloaded the datasheet for these A4988 chips? On page two of the data sheet it clearly states under the Absolute Maximum Ratings table that the Operating Ambient Temperature rating is -20 to 85 C. If Markus is measuring 77C in the center of the back of the board I don't think it is too far of a stretch to think that that chip itself can reach temps of 85C. It's confined in a small space and the thermal radiation from all of the components is going to rise upwards into that small space. Any thoughts? The resistors and traces may not play as much of a role as the ambient temperature the chip itself is subjected to.
Link to data sheet:
http://www.pololu.co..._translator.pdf
Edit:
Let's say the ambient temp inside the confined area that houses the board and also the chip is 80C, then according the thermal characteristics of the chip it could easily start to creep into the 85C temp conditions when operating.
-
@WoofyPlace thanks for the suggestion, gr5 sent me a PM earlier suggesting the TCs of the traces could play a role as well. I'm going to redo some calculations with this in mind and post some worse case scenarios.
There are still a lot of other things to consider, especially as to why some people experience under-extrusion more so than others and how much local ambient conditions play a roll. I wonder what the actual percentage of users who are experiencing this issue is. I think this would really help pin down the cause, is it an error on the component manufactures part? It wouldn't be the first time something like this happened, but I don't think there is nearly enough information to start pointing fingers. Maybe UM should print out the test cylinder by illumnarti as the test piece before shipping :mrgreen: lol, probably a silly thing to do but it would help identify potential issues before shipping the units out.
Or this isn't the problem at all and we are still chasing our shadows; who knows, at least we are whittling down the potential suspects.
Edit:
Also, thank you very much Mr. Waldorf; I think this information will provide very useful :grin:. I would have checked myself, but still don't have my printer yet; although it is slated to ship the week of March 3rd, so for now the forum will have to be my eyes and ears for material I cannot obtain without a physical printer on hand.
-
Thanks Coen and gr5,
I calculated that if the temperature changes from 20 to 100C (just a worse case scenario, could it get worse though?), the resistor was at it's full 1% error, with a TC of 200 ppm/C, and a 4983 stepper driver chip, there would be 1.218 Amps going to the stepper motor when the device thought there was 1.25 Amps going to the stepper motor.
I'll recheck my calculations when I get home, but I think they are correct. I agree this small reduction in amperage shouldn't be enough to stall the motor. I don't feel completely satisfied with this though, I feel like something is still nagging at me and I can't quite put my finger on it. Is there anyway someone could confirm that the 4983 stepper chips are being used in the device?
I guess I'll know how much time to invest in this idea after a couple people report putting a fan under their suspicious units and print out the cylinder test.
Some thinking out loud, I wonder if there were a few bad apples in the resistor bunch sold to Ultimaker? and if there were any long term exposures to heat during manufacturing that could have permanently changed the value of the resistor? I'm also wondering what the actual temperature of the individual resistors are, the board may be 77C but the resistors themselves may be much higher. I'll make some plots when I get home, but it would be nice to get confirmation on the actual stepper controller chip.
Edit: gr5, didn't you say that you increased the current to your stepper through the firmware and it ground down the filament? Was the increase a small amount 10mA or so, if the motors are this sensitive to the amount of current applied to them, then I wonder how sensitive they would be with about a 32mA reduction in current.
-
Hey @ritzenthaler can you note a couple of things?
Which item did you print first?
The amount to time that elapsed between prints?
The approximate diameter change of your filament spool, for each print? Was the spool getting low on filament for any of the prints?
The room temperature?
I know you might not be able to recall all the requested information, but any information you provide can help support or discredit some hypotheses that have been generated in the forum.
Thanks
. 
. We have meetings at work where we try to figure stuff out and all types of ideas are thrown around, I think it really helps to see things from many different points of view.
Pulling force of UM Extruder
in Improve your 3D prints
Posted
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.