Smooth walls

[IRobertI 520]

The play in the nut is "normal" and I've also heard that crunchy sound from the z-stage. I have it on the first UM2 I have as well which is ancient at this point.

As for the irregularities on the surface, I wish I knew. It's something I've seen on most printers I've used from time to time and there never seems to be a clear answer as to why. I know I had terrible banding like that with some Colorfabb I bought years ago so I don't know if the filament itself is to blame or not.

How's the temperature, is it solid or does it keep fluctuating?

[Nicolinux 288]

@Dim3nsioneer: I have the +Kit with the 35W heater and the new coupler type. That thing is only a few weeks old and I din't print much. But I had this "banding" problem with the old UM2 head too and there I already used a 35W heater (but an old-school coupler).

@neotko: The story is a bit different. I didn't swap the x/y rods, only the thin rods that came with the +Kit.

In my quest to find a solution for the "layers not touching" problem I swapped a lot of parts and tried many things. Then I had a very big print job with XT-CF20 which ate some couplers and a bit of my nerves

@IRobertI: The temperature is solid, even better than my regular print head with the 35W heater.

The weird thing is, when I got the printer for the first time and although it had that "layers not touching" problem from the get go, it could still print very smooth walls with zero "banding" (see the comparison pics in the first post here in this thread).

So something must have degraded/happened that produced this problem. At least I can say that replacing the linear bearings and smooth shafts with more expensive ones (and tighter tolerances) doesn't help.

I hate debugging through exclusion

[geert_2 557]

Couldn't this be caused by underextrusion or irregular extrusion too?

Possible causes of underextrusion that I experienced:

- White teflon insulator on the nozzle is worn out and deformed internally, causing an irregular flow. (If this is the case, replace it.)

- Dirt in the nozzle (if so, do an atomic pull).

- Irregular filament diameter.

- Filament is near the end of the spool, and/or wound too tight, so it acts like a tough spring resisting unwinding. I often manually unwind and straighten a few meters of filament, which greatly improves results.

- Irregular friction between the filament spool and the spool holder (black cylinder at the back), causing a jerky movement of filament, and thus irregular flow.

I am not saying one of these things is the cause, but they might be worth examining?

[neotko 1,417]

My new crazy theory.

Let's assume that new um2 uses a ferrite block (like I suggested to nicolinux long ago).

So, there's a chance of electrical noise. Let's (remember this it's a crazy theory) think that there can be other parts (due to the user electric ambient/powerplug/house) that can get electrical noise. So, what could happen if the pt100 readings get electric noise? Won't that affect the extrusion rate since when you print small details you need everything to be as constant as possible.

So my crazy idea, pt100 electrical noise. Very easy to test with a ferrite block. Ofc the chances of pt100 getting noise it's very very low, but it isn't hard to test.

[Dim3nsioneer 557]

Noise usually consists of high frequencies (except e.g. 50Hz pickup or similar). As far as I know there is a software low pass filter in Marlin for the temperature reading. But it can be tested of course...

Edited May 8, 2016 by Guest

[Nicolinux 288]

DAMN it is the heated bed!!!

Check this out:

White is old and has perfect smooth walls. Red and Teal are new and have shitty walls.

On the left printed with heated bed on (bad quality). On the right printed with bed OFF and has perfect smooth walls.

I still don't know how to solve it yet, but at least I know what was causing it... I upgraded the UM2 to an UM2+ with the +Kit. Firmware is the latest stable: 15.04.5.

Recapitulating - you don't need fancy (and expensive) new linear bearings and smooth shafts from Misumi. The stock ones in the UM2 are fine. You don't need a new z-nut either. My new one has a bit of play (up/down) but this does not influence quality at all (with the heated bed off...).

And one more thing. I disassembled the z-stage again because I thought the issue was mechanical. While doing it, I also loosened all screws that hold together the frame. I then pushed (more or less gently) on the frame from all directions and now the frame is square. The printer does not wobble anymore.

However, without the z-screw which is attached to the z-motor, the bed slides up and down very smoothly. But as soon as I attach the z-motor the bed doesn't move smooth at all and won't move down past a certain point. I have to exert quite some force to move it to the bottom. I can lift the entire printer just by grabbing the bed when it is at the bottom.

@neotko suggested to loosen the screws on the z-motor and wiggle the bed up and down while tightening the screws back one by one - but it didn't help. The z-motor is moving the bed up/down fine, but I still suspect there's something up. But whatever it is, it does not have any impact on the print quality.

If anyone wants to give the stretchlet a go - here is the .gcode file. Takes around 25 minutes @ 50mm/s, 230°, 0.2 layer height and uses spiralize, 0.4mm nozzle.

https://www.dropbox.com/s/9vlk55g7kyo781p/stretchlet_banding_test.gcode.zip?dl=0

Edited May 10, 2016 by Guest
updated print settings for the stretchlet

[neotko 1,417]

Indeed the bed can cause that, the link where anders said something about that then I started to dig on the old thread where @daid said that there was a change on firmware long ago that changed the PID-Bed to bang mode. Ofc this can't be fixed by standard firmware since the European commission doesn't allow that amount of EMF. So... Custom firmware! Does @Tinkergnome firmware allows to change the bed bang to pid?

[neotko 1,417]

Oh also about the z screw, if it can print like that I would not bother at all. But after seeing your z video, I might give it a go. The worse that can happen it's that you might need to recalibrate the whole bed bearings x8 screws. But I did that (after cleaning with wd40 the z screws and apply new liquid ptfe grease for bikes) and they are working just superb.

[kris 29]

Maybe a closed build chamber helps to stabilize the build plate temperature?

Edited May 10, 2016 by Guest

[Nicolinux 288]

That's overkill... I am not printing with ABS so I won't use a build chamber only for that. I guess I'll need to switch to the @TinkerGnome firmware.

[kris 29]

That's overkill... I am not printing with ABS so I won't use a build chamber only for that. I guess I'll need to switch to the @TinkerGnome firmware.

Hmm...ok. Well, first thing I do if I get a new UM2; close it. Helps to stabilize the process, no matter if PLA, ABS etc. I have thus far not encountered the problem you describe in any of those machines. I print 90% PLA.

Edited May 10, 2016 by Guest

[shurik 87]

If anyone wants to give the stretchlet a go - here is the .gcode file.

 

Hi Nicolinux,

Printed the file in order to see whether to use the ferrite you sent me once.

The stretchlet came out smooth, like your old one.

There are some bad walls from time to time on my machine, however. So, I will follow the topic hoping to find a resolve.

[Dim3nsioneer 557]

Hmm...ok. Well, first thing I do if I get a new UM2; close it. Helps to stabilize the process, no matter if PLA, ABS etc. I have thus far not encountered the problem you describe in any of those machines. I print 90% PLA.

 

Sounds as if you get a new UM2 quite often...

Do you just close the front (with bubble plastic?) or do you have a more advanced setup?

[Torgeir 242]

Hi Folks,

Hmm. I must say that Neotko is up to something here…

I’ve read in one of the topics about the experiments of adding a more powerful fans for cooling the printing object and the upper part of the extruder (the 5 volt fan). They soon discovered that the extruder temp failed erratically. People assumed that the added power consumes from this more powerful fans lead to voltage drop and made this failure…  

However, here this is not the case! Those fans are brushless and have electronic speed regulator built in. Such controller/regulators often spread lots of harmonic around the feeder lines (the two wires feeding power to the fan).

The voltage from the PT100 (platinum thermistor) is amplified by an OP amplifier.    So which level (voltage) do you think the PT100 deliver to the OP amp? Well, this voltage calculated here is for the temperature span from (0-210) deg. Celsius. The answer is only 85.4 mV, yes millivolt! This is only 0.407 mV/1 deg. Celsius, or if you like; 407 µV/°C.

Armed with those numbers it is easy to see that high level radiation will/can jam out such a low level signal when wires are bundled together.

Another thing here is; the original temp sensor harness has a copper tinned braided jacket and also a transparent insulator overlay, covering the three conductors all the way from the stainless steel cap to the connector. Just where those three wires escape from the jacket, there is a heat shrink insulating the braided jacket. This braided jacket is meant to be a shield for this lo signal feeder, so it would be smart to ground this wire at the controller PCB. This wire is to be shielded all the way to the board connector…

Thanks.

Torgeir.

Just made a correction of the calculation.

Edited May 11, 2016 by Guest

[kris 29]

 

Hmm...ok. Well, first thing I do if I get a new UM2; close it. Helps to stabilize the process, no matter if PLA, ABS etc. I have thus far not encountered the problem you describe in any of those machines. I print 90% PLA.

 

Sounds as if you get a new UM2 quite often...

Do you just close the front (with bubble plastic?) or do you have a more advanced setup?

 

Well advanced is the wrong word I guess. I build a door and a lid with acrylic plastic. Hinges are printed. Gaps are sealed with tape. So it's completely 'air tight'. I add a cooling fan for the feeder and another to extract fumes from the build volume. It then gets hocked up to a ventilation system.

[Torgeir 242]

Hi Folks,

This is interesting; there are so many parameters that are to be checked when things go wrong.

Do you remember the zebra stripe topics? This one is yet to be solved, if I’m not wrong?

Well, this one is about the heat bed:

As I’ve been printing different test object in order to understand the influence the various parameters has to my 3Dprint object, I found this problem when tried to print the “3Dbenchy boat”. I found some strange printing in the forward part (at the bow) of the boat. It just look like the PLA is overheated at that first part of the bow, were the overhang is closest to the heat bed. To be honest I’ve always been suspicion to this heat bed, as it really radiates lot of additional heat to our printing object, especially to the part of object that’s close to the heat bed. Also, when the extruder is away from the heat bed, say an inch (25mm) or so, there is no sign of this.

Have a look at this picture.

This picture shows 6 ea "3Dbenchy boat”.

From left to right: First 3Dboat, head temp 208, bed temp 60 and layer h 0.2. Number two only change head temp down to 203 and position turned 45 deg. Clockwise. Number three; only turned 45 deg. more clockwise. Number 4 and 5 some extruder feeding change 103% and 98% ough.. Number 6 as number 2, whit Z-hop. Enough said here..

I’ve also happened to notice that the EMC test did not pass the test of the “UM2” (?) due to the high radiation from the “PWM modulation” of the heat bed. (This can be fixed by a hardware correction).

I’ve also monitored the bed temperature to be approx. 60 deg. C., same as the setting at the controller.

Have anyone seen this on their “3Dbenchy boat”?

Thanks

Torgeir

[Nicolinux 288]

I knew about the influence of the heating algorithm (pid vs. bangbang) but I didn't think it would influence the quality that much. Since the printer is able to produce superb quality just by switching off the heated bed (or changing the heating algorithm in firmware), at least we can say that the issue is non mechanical but software/firmware related.

Edited May 11, 2016 by Guest

[Torgeir 242]

Thanks.

Do not have any Capton or blue tape. Can I use a couple of hairspray layers on a cold glass plate?

This cause I’m only having the Verbatim PLA, that jump of the heat bed when it’s cold.

Torgeir