Ender 3 S1 Pro: Day 2

It gets better, right?

You would think but it was another frustrating day. We had company coming over for Boxing Day so we moved the printer into Q’s bedroom and decided to attempt a more complicated print. Before printing, I decided that it needed to leveled again, so I went through the process of manual levelling, probe (z-axis) calibration, and then auto-levelling.

And then we attempted a print. The thing we were printing was quite complex on the base layer: a Tie Fighter Kit Card Ornament: https://www.myminifactory.com/object/3d-print-snowflake-tie-fighter-kit-card-ornament-199485

And the first layer didn’t lay down nicely at all. It was lifting up in the rear-right corner, debonding from the bed, generally creating a mess. So, I went through the leveling procedure again. And tried printing again. No better. Level, print. Worse!

One thing the Sonic Pad shows you are the compensation values that indicate how “unlevel” the bed is, in a 5×5 grid. And I could see that each time I leveled the bed, these compensation values were getting more and more skewed. It was like I was compounding the problem each time I was trying harder and harder to get the bed leveled manually.

It was super frustrating.

Eventually (5th? 6th? attempt) I hit Google. It seemed like the auto-levelling just wasn’t working, and it seemed that the printer wasn’t using the auto-levelling values.

I came across this reddit thread which sounded exactly like the experience I was having. There doesn’t seem to be a great solution but there were a few things that stood out to me:

  • It’s not an uncommon problem
  • There seems to be some problem when the Sonic Pad saves or loads the automesh data. Someone posted some G-Code they say fixes the problem
  • Other people tried to delete the auto-level table from the printer, but had problems with deleting it
  • Some people suggested to avoid auto-levelling all together and just manually level the bed
  • Some people suggested printing a different mount for the z-axis sensor and using that mount

I wasn’t going to do something more time consuming or complex, so I decided I was going to delete the auto-levelling table and just manually level the bed so that we could get this print going. I went into the web interface and cleared the levelling table, went through the manual-levelling procedure again, and did the z-axis calibration.

Side note: Manually level first, then probe calibration? Or vice-versa? I can’t find a definitive answer on this. The manual calibration uses the probe, so shouldn’t that be calibrated first? Most places seem to indicate the probe gets calibrated after, however.

At any rate, the next print attempt started and I checked it out in the web interface. And the auto-levelling table was back! I killed the print and purged the table again. I don’t recall if I power-cycled the printer or not, but this second purging seemed to do the job.

Another round of manual levelling and another attempted print. This time, it was the front-left corner that was a bit wonky, so killed the print and manually-levelled again.

Honestly, I lost track of how many attempted prints I had. It was a good 4-hour process.

Eventually, using just the manual levelling (and probe calibration) using the thinnest paper I had on hand (a thermal paper receipt) ensuring that the contact force between the nozzle (make sure it’s cleaned of any filament residue!) and the bed was as uniform as I could make it in the four corners — it laid down a clean first layer. Well, mostly clean. There was a small bit of weirdness in one of the more detailed areas, but I just let it go.

And in the end, this is what I ended up with (sorry, didn’t take any photos during the print attempts):

Assembled TIE fighter Christmas ornament, printed in white filament

The print quality is generally good, but I noticed a few things: The small dish was somewhat warped and looks like it’s slouched a little bit. And another of the parts was a bit warped too.

Modelling warpages like this in thermoplastic manufacturing is literally my day job, so these warpages are both not unexpected and not insurmountable.

But, fixing this calibration problem and getting better prints is a problem for another day.

Some other notes:

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.