Prusa Bed Leveling for Dummies

Take a look at the nyloc mod for the Prusa. It's dirt simple and will solve your problem. Here's some relevant info:

 
slodat said:
Take a look at the nyloc mod for the Prusa. It's dirt simple and will solve your problem. Here's some relevant info:

Click to expand...
Thanks. This was the first one that I've seen that was presented clearly. I found lots of references, but hadn't stumbled upon a clear set of pictures and steps. I'm not sure I want to use nylon washers, but may simply put the smooth side down of a washer. Nylon changes dimension a bit over temperature. A lot of these stamped M3 washers have terrible sharp edges. Even ones from McMaster.

Next order from McMaster will include a box of M3 nylock nuts...

So my takeaway is do this hot, and try the Prusa Leveling Guide Plugin. I had used the Bed Level Visualizer before, which is why I knew things where not all that flat. Roughly speaking, the first layer print quality matched the Bed Visualizer results. Where BV said it was low, was where the first layer had insufficient squish. You could see the bed side of the print was not flat there - not like other parts of the print.
 
If the bed isn't flat, everything else will suffer. This is worth the effort, as it is described, in my opinion. Including the nylon washers. I got my stuff from Amazon for $20.
 
It appears there's a few ways to perform this modification. A guy by the name of Dombi3D has a video using steel washers instead of nylon, while the "revisit" video by Chris uses silicone tube instead of the nyloc hardware.

Picking a bed temperature for the alignment process appears to be a really good idea.
 
Seems I'm back to the drawing board. Had to replace the thermistor a few days ago in the hot end, which cascaded into various calibrations. I did add the nyloc mod. I also added (and maybe this was a mistake?) peek washers between the nyloc nut and the heater board.

Here's the basic question: can I do the Octoprint bed visualization without the steel sheet but with preheat? I need a faster way to dial this in. Grabbing a hot sheet or waiting for things to heat and cool is pretty much a time suck, as you have to wait for things to come to thermal equilibrium. Waiting for the heat to soak in can takes 10's of minutes, actually one should wait many (10) thermal time constants. Was thinking of doing this at PETG print temperatures, 240C print head and 85C platen temperatures, which is the setting for the first layer. Does that make sense?

Second question, how repeatable is a measurement/visualization in your experience? Trying to figure out if the tool itself is good enough to get me close to the right place. I will check it with the 3x3 single layer pattern later to get things closer.
 
PEEK is an insulator, correct?

Having read through this thread again, I wonder if heat is getting to the sub frame and causing some of your distortion?

I read somewhere that over tightening on initial bed assembly could cause print leveling issues. Heat could certainly amplify the effects of that.

If PEEK isolates the bed from the frame it may resolve some of that distortion you are seeing. I would try to print a full bed layer again and compare your 3x3 grid and see if any improvements.

Through the entire thread, I haven’t read much about the adjustment of your PINDA. I’m curious if fidgeting with those adjustments have given you fits or not. If so, you might consider reinstalling the position. I read (somewhere else) that while you can calibrate with it grossly out of position, I wonder if there is such a thing as adjusting for too much “gain” on that one sensor to get desirable results? During my bed leveling calibration, I wasn’t happy with how much gain was required and moved my PINDA slightly so that it calibrated much more quickly than in its original position. If that doesn’t make sense, let me know.

Question: this is rabbit hole bait and could take us off topic so apologies ahead of time…functionally, your prints are good? Part corners aren’t lifting, warping, etc? Even with those small defects, parts functionally work. Sometimes with 3D printing good enough is simply just that.


Sent from my iPhone using Tapatalk
 
vtcnc said:
PEEK is an insulator, correct?

Having read through this thread again, I wonder if heat is getting to the sub frame and causing some of your distortion?

I read somewhere that over tightening on initial bed assembly could cause print leveling issues. Heat could certainly amplify the effects of that.

If PEEK isolates the bed from the frame it may resolve some of that distortion you are seeing. I would try to print a full bed layer again and compare your 3x3 grid and see if any improvements.

Through the entire thread, I haven’t read much about the adjustment of your PINDA. I’m curious if fidgeting with those adjustments have given you fits or not. If so, you might consider reinstalling the position. I read (somewhere else) that while you can calibrate with it grossly out of position, I wonder if there is such a thing as adjusting for too much “gain” on that one sensor to get desirable results? During my bed leveling calibration, I wasn’t happy with how much gain was required and moved my PINDA slightly so that it calibrated much more quickly than in its original position. If that doesn’t make sense, let me know.

Question: this is rabbit hole bait and could take us off topic so apologies ahead of time…functionally, your prints are good? Part corners aren’t lifting, warping, etc? Even with those small defects, parts functionally work. Sometimes with 3D printing good enough is simply just that.


Sent from my iPhone using Tapatalk
Click to expand...
Peek is an insulator, or at least an insulator compared to a stainless washer! Cheap M3 steel washers are stamped and have sharp edges, so I wanted to avoid any chance of the steel cutting through the coating on the PCB. As it turns out I used carbon filled PEEK which is claimed to be stronger and more moisture resistant. My thinking was I wanted a non sharp washer and a tiny bit of give. PEEK was the highest temperature plastic material that I could find easily in M3.

Haven't needed to change the position of the PINDA. I installed it according to the instructions (with a tie wrap for a spacer, if I recall) and haven't had an issue with it. Not really sure what you mean by gain. By moving slightly, do you mean closer to the base, or further?

On the rabbit hole question... Since replacing the thermistor (due to thermal anomaly error shutdowns) I have yet to get a successful real print. I did a small 1x1x1/2 print and it popped off the plate after getting to be 1/4" tall. Lots of spaghetti. Previously, prints were only ok on parts of the plate. Other parts had issues, ie they were too far away for correct adhesion. These problem areas matched (from memory) the areas that the bed visualizer said were too low. The first layer cal in the printer menu is not sufficient for me to dial it in, as it is only one tiny spot. (Before replacing the thermistor, I was able to get a somewhat ok first layer on about 50% of the plate. Large parts would lift on corners.). I'd like to be able to print larger parts, like needed for a printer enclosure.

I had set up the nylocks + peek months before replacing the thermistor. The plate was higher in the center and lower at the edges. I could get an ok print on the bottom center, if it wasn't too large. After the replacement of the thermistor, the plate appeared flatter but a lot more tilted, but I had not adjusted anything on the plate. But the printer had been moved and physically turned on it's sides to do the thermistor replacement.

If I am using a 0.4 mm nozzle and have a 0.2 mm "first layer thickness" setting, does that actually mean that that layer should measure 0.2 mm with a micrometer?

I don't like this eyeball stuff, the pictures that are provided online are not in focus, and I honestly cannot tell if a first layer is ok. I can tell if it is obviously bad, or obviously good, but not if good enough. It would be good to know some quantifiable values to be within a range, say pad thickness is between 0.199-0.238 mm. Have a Quantum micrometer, so physical measurement is preferred. I'd guess that many bracketed in focus photos ranging from bad to good, with gradations in between would work, but I have not found something like that.
 
If you set the first layer height for 0.2mm, and print a single layer, yes. The part will be the first layer height setting. Well, it should be anyway. That's how I usually set my z-offset.

You do need to be a bit forgiving and average the readings. Printing surface finish isn't perfect. I also get divots or blobs occasionally depending on how everything is tuned. There are always good clean areas to get measurements. Textured plates can also make this interesting.

And be careful with removing the part to measure. It's quite possible to distort it.
 
I think an enclosure might be the thing you need. It made an incredible difference for me.

You can test this by making some cardboard shields to build up around your printer.

I had some significant drafts in my shop from the mini split unit during winter months. Parts would fail via bed adhesion, warping.

Regarding the PINDA “gain” portion of the discussion. I mean gain as in like a pot adjustment on an electronic device or maybe better said as the fine tuning of a sensor. I remember as a kid adjusting gain on my speed controllers for top or bottom end acceleration for my RC pan car racing. It could make all the difference in how the controller reacted to throttle input.

I felt like when finding the sweet spot for the PiNDA, it’s like that gain adjustment. I read the instructions on PINDA fine tuning and getting to the nominal location was important. It sounds like you have it set properly so probably nothing else to do there.


Sent from my iPhone using Tapatalk
 
vtcnc said:
I think an enclosure might be the thing you need. It made an incredible difference for me.

You can test this by making some cardboard shields to build up around your printer.

I had some significant drafts in my shop from the mini split unit during winter months. Parts would fail via bed adhesion, warping.

Regarding the PINDA “gain” portion of the discussion. I mean gain as in like a pot adjustment on an electronic device or maybe better said as the fine tuning of a sensor. I remember as a kid adjusting gain on my speed controllers for top or bottom end acceleration for my RC pan car racing. It could make all the difference in how the controller reacted to throttle input.

I felt like when finding the sweet spot for the PiNDA, it’s like that gain adjustment. I read the instructions on PINDA fine tuning and getting to the nominal location was important. It sounds like you have it set properly so probably nothing else to do there.


Sent from my iPhone using Tapatalk
Click to expand...
I had bought the kit with the acrylic panels to do the enclosure - and got a little confused about what to print and not print. So I put it on the back burner. I will get back to it this year.

I did manage to use the nylock nuts + peek washers and things are a bit better - but not perfect. The top left corner near the DC input is too low. A printed calibration pad is see through and thick, because there was insufficient squish. There was not enough squish for the filaments to bond to each other. Adding corrections to the screw doesn't seem to fix it. Maybe that hole is stripped, not sure. I don't want to take everything apart again to check it yet. But I will get there pretty soon as you will see. Out of the 9 pads in the 3x3 calibration structure 1 is poor, a second is barely passable, and the others are good.

In a fit of madness, aka Black Friday, I bought the MK4 upgrade kit (it was discounted and there was free shipping). The box arrived last week. A couple of days ago, I realized I needed to print the parts. (In the box was two small reels of filament.). So that is what I have been doing. When the parts are printed (and that will take a while), I will be taking the printer apart. So I will check if that hole is stripped. If it is stripped, I will drill out the hole and install a steel threaded insert. Hopefully that will help.

Slow progress, but at least it is forward.
 
Last edited:
You must log in or register to reply here.
Back
Top