Z Banding on N1 with V2 Hot End

[mitchfx]

Ok, ran the PID auto tune, max temp 230 at 5 levels. Running another test print now and can already see the diagonal lines are still there with no visible improvement. What's next?

[Jetguy]

Flat of course. The idea is that the Z nut and move and slide around in XY (front to back, left to right) because the mounting holes are larger than the capscrews used. The nut still lifts the Z stage via contact by pressing up against the screw shoulder. The idea is a flat washer gives more area and smooth contact for this sliding to happen on.

A lock washer would defeat the entire purpose of the mod.

[Jetguy]

I make no claim I can magically solve the diagonal variation. It's NOT due to PID and not due to the hardware Z axis.
The most likely candidate is the stepper driver current setting and the feeder motor, and that all coupled with firmware interaction.

It may not be something you can solve. It may require more in depth knowledge than I can convey in a forum post.
It could be a limitation in the electronics of the printer.

[mitchfx]

Well that is unfortunate. This kind of artifacting makes the N1 pretty much useless to be due to the type of prints I'm doing, so I might have a $2k paperweight on my hands unless Raise can correct/replace it. I'll be calling them Monday to see what my options are. Thanks for all your help!

[Jetguy]

This kind of artifacting makes the N1 pretty much useless to be due to the type of prints I'm doing

To put it bluntly to you, the grass is not always greener on the other side.

I own more printers than some folks dream of. I have one of almost every major brand on the market.
I've yet to find a printer sold that requires no tuning, no tinkering, no upgrades to achieve some level of perfection you seem to be seeking.

When I said limitation, the reason is, the next layers of tuning- Raise 3D tech support starts getting antsy with me when I post them.
Like I mentioned, stepper drive current does affect microstepping. This is because full steps are the magnetic hardware pole alignments in the motor. Microstepping is this idea of stair stepped lesser current settings that balance the rotor between 2 poles of the motor. Every single consumer 3D printer uses microstepping. The problem is, if the motor and stepper drive current are not accurately matched, just like driving a speaker too loud, you get harmonic distortion. The problem here lies the factory is going to say don't touch the vref stepper current.

The next thing is upgrades. Just like we know the stock extruder feeder setup is not exactly light- there is also the fact that 3rd party upgrades like the Bondtech exist and offer benefits. Examples are finer steps per mm meaning that microstepping error has way less impact. Then the added torque and grip, not to mention the lesser moving mass- these all affect these final artifacts in print quality in a positive way.

Again what I'm telling you, this isn't the ONLY printer on the planet to ever show that pattern. Troubleshooting such fine final defects in a print surface finish requires patience, understanding, background knowledge, willing and able to change electronics and mechanical variables in a very controlled manner.

[Jetguy]

Again, some searching in other forums I participate in came up with rather nasty artifact, but the point was the root cause is what I'm talking about.

The root cause was relatively badly chosen specification motors (high inductance and resistance) driven at an incorrect stepper driver current setting compared to the motor specs.

Step back to the Raise 3D printers- there have been changes over time in suggested settings based on what motors your printer specifically has. And Raise 3D used different motors depending on when your printer was made and may be a variation between models.

So again, the artifact could be related to stepper driver current and at the specific speed, rate and layer height, you hit this spot in tuning range of the printer that exhibits the pattern. The answer to that is lowering the Vref experimentally and running test prints to see if the artifact suddenly goes away. Another telltale sign would be the extruder motor temp. If it's burning to the touch- that's the upper limit of current and likely to cause such an artifact.

https://hackaday.com/2016/08/29/how-acc ... ng-really/

[Jetguy]

Also, point of note, Raise 3D does use A4988 drivers for the extruder and Z axis, which is the best (per that article and my own personal beliefs and experience concur) for accuracy. Luckily, Raise 3D also avoided the DRV8825s which are plentiful and cheap - but perform miserably and the 24V of the N series would make it even worse.

But with any stepper driver, the key is tuning and matching the current to the motor, and then the application of the motor and driver as a "system".

[Jetguy]

Again, look what a $2500 brand spanking new MakerBot does https://www.flickr.com/photos/90025904@ ... 2732987534
And then take another brand large size printer and how bad it's leadscrew wobble was stock that I had to create a complete mod for.

[mitchfx]

I totally get what you are saying. I've owned 6 FMD printers before the N1 ranging in prince from $300-$2500. Each has had their issues and most required some level of modification or upgrades, however I've never experienced anything as extreme as the level of banding I'm getting on the N1. I'm perfectly fine to pursue solutions to remedy these artifacts, I'm just hoping that this is possible because otherwise I won't be able to use this printer at all. All of my parts need to meet a certain level of surface quality which I've been able to produce on every other printer I've owned. The current results I'm getting from the N1 make my prints pretty much useless and destined for the garbage bin, so I either need to find a way to get it up to the same quality as my other printers or shed it at a financial loss. I plan to contact Raise3D on Monday and see what my options are. The ironic thing here is that his printer was NOT the unit I was supposed to have received in the first place. I purchased a demo unit which was to be delivered to my home and the tech was going to swap out my v1 hotends on the N2 Plus for the V2 and try to fix the extruder assembly wiggle issue on that same printer as well. Not only did he show up with the wrong unit (he grabbed a NEW unit instead of the one that was set aside for me) he mistook my V2 hotends I'd been holding on to for 6 months as V1s and took them back with him AND said he had no idea how to fix the extruder wiggle issue. So I got handed a lemon and the two issues on the N2 Plus that were supposed to be addressed got shuttled. Turns out I did indeed have V2 hot ends but he didn't know about the earlier design, thus took them back with him and had the updated ones shipped out. These are still sitting on the shelf since I've not had time to swap them out, and my N2 assembly still wiggles like a worm. All I can say is thank goodness for my Form2 since it works like a champ and produces flawless prints over and over and over.

Sorry for the rant...just needed to vent. Hopefully Raise3D will offer some kind of solution when I contact them. I've recommended them to several friends who ended up buying N2s so fingers crossed they will make good

[mitchfx]

All this being said, I'm definitely down to follow any advice you might have as to what hardware and electronics I should consider upgrading to. I've narrowed down my herd of FDM printers to only the N2 Plus and and N1 at this point and had expected the N1 to be my best quality FDM unit while the N2 Plus would be reserved for large parts. I sold off everything else and foster parented my RapideLite 200XL to a friend who did a much needed complete overhaul of the electronics and wiring. If it takes a few hundred bucks and several hours of my time to get the N1 in tip-top shape then so-be-it, but as it is now I have no use for it other than to run slop parts. Please don't take any of my previous complaints about this printer as in any way directed at you...I fully appreciate all your help and 100% understand your input and reality checks. At this point I will see what Raise has to say on Monday and then go from there. Thanks!

[mitchfx]

Here's a photo showing the results at 0.2mm sliced in Simplify. Right is vase mode and left regular layer mode. I called Rasie3D yesterday who said they are looking into the issue and will let me know.

[Jetguy]

I'm right in the middle of a 20 hour print on my N1. Can you do me a favor and share the exact gcode files you printed for those 2 test objects?
I want to see if I can reproduce and then tell you what to tune.
Also, details on your printer:
I assume stock feeder since you just asked about changing out to V2 hotend the other day.
My question is, what extruder motor is installed? Is it the all black body Bohong or the previous silver ended, black laminations in the middle Minbea?

If a newer Bohong motor or even the Minbea, we may have to slightly lower current setting on the stepper driver to see if that is the cause.
Yes, this thread is about one on a Bondtech, but the rules and general principals apply. On the N1, to get to the controller board, you lay the machine on it's side, remove the 4 M4 screws holding the lower plastic cover on, and then the motion controller box also has a lid under that with 6 M3 screws. Again, N1, you have to lay the printer on the side to access. Can't take a picture- mine is mid print. viewtopic.php?f=4&t=2705&p=17777&hilit=bohong#p17776

[Jetguy]

Just for reference only- not saying go to this stage, here was a thread on alternate plug in drivers and some machine details viewtopic.php?f=4&t=634&p=20754&hilit=bohong#p6712

[Jetguy]

Also, I feel like we now solved the Z banding, this is now a different issue. The Z axis is not causing the patter you are seeing now, this is an extrusion volume compared to XY movement from what I know, but until I run your file, see if my machine does the same thing or if I can tune things to make it replicate it, then I know what the problem is.

So again, sharing those gcode files for those last test prints would be a huge help.

[mitchfx]

Hi Jetguy,

I have the Bohong. Michael from Raise3D is here right now doing some diagnostics. Yea!

Here is the Gcode from Simplify3D, one standard layer and one vase mode.

Thanks,
Mitch

[mitchfx]

Still no resolution from Raise3D on this issue. I've been waiting for 2 months to hear back on their engineer's diagnosis for the cause. I'm hoping that I don't have a $2k paperweight on my hands

[ABH]

The "wavelength" of the interference in a single layer seems to be around 6 mm. This happens to roughly correspond to one full-step of the extruder stepper motor which is 5.7 mm according to my calculation, based upon the gcode file you posted and assuming that you have the stock extruder (not Bondtech).
I would check
1. the stepper driver,
2. the wiring from the stepper driver to the stepper motor and
3. the stepper motor itself.

1. For a single extruder printer there is usually a spare stepper driver in the motion controller board that can be used. Simply switch the two drivers marked E0 and E1 and see if it helps.
2. Unplug the extruder stepper motor and remove the stepper motor driver module (E0). Then use an ohmmeter to test the wires all the way from the driver to the motor plug. Both check for connection and for shorts between wires and shorts to ground.
3. With the motor unplugged, use an ohmmeter to measure the resistance of each motor winding. There are two windings, each with a center tap. Windings should have identical resistance (approx. 4 Ohm) and the center tap should have identical resistance to each of the winding ends (approx. 2 Ohm).

You could also try to adjust the stepper motor phase current up or down, just to see if it changes the interference pattern.
I think this must be an extrusion problem, as the interference pattern in a single layer is continuous across 90 degree corners and also seems to be the same in the cylinders you first posted.

[JohnSays]

I've got to ask this question and hopefully JetGuy will see it. I have been pondering the use of the four rods that are in the corners of the build plate. I came to the conclusion that their job is to hold the plate from wobbling, but apparently the plate wobbles anyway and thus the topic of this thread. So, what are the rods for if not to contain wobble? In fact, I thought that is why the ball roller bearings can be so loose -- because they only move the plate up and down and the four rods locate it. I didn't see a mention of that here.

Is what is being discussed here, in part, such a badly bent ball screw that it exceeds the freedoms of the loose ball nuts and thus puts pressure on the four rods to deflect them?

[Jetguy]

Remember we are talking different printers, but there are some common themes.
The N1 only has 2 vertical rods at the back and is a cantilever shelf style axis. This means that because the rods and the bed front edge are some distance, any movement, even a fraction of a mm is amplified to the front edge of the bed.

The N2 and N2 plus DO have rods at the 4 corners of the bed. However, the total span of those rods (top to bottom attachment point to the frame) is longer, so even though you might think they are massive 12mm rods and could never flex, that is absolutely false.

In both systems, the Z axis ballscrew(s) is(are) constrained at BOTH ends by ball bearings mounted to the frame. If there is the slightest runout from damage, shipping, manufacturing or just stress, then as the rod rotates, it's like a crankshaft and the nut follows this orbit. This is why for a defect free linear stage you must decouple the ballscrew in some way. Either you can make a complex nut system where the nozzle can wobble in XY and not in Z or simply removing a top bearing and allowing the Z shaft itself to deflect from center while the nut maintains a straight path. This minor deflection is so small, there is no concern. The reason for NOT just removing the bearing is the way the frame is constructed, you must take the ENTIRE printer apart to get to that bearing. Hence why alternate methods of dealing with the situation are recommended.

[JohnSays]

Thanks JetGuy.