I had a nice big post written up yesterday and my web browser ate it because I let my login time out....
Completely my fault, but kinda ruined the mood.
So I've had the printer for two weeks now, and been printing for a week, almost nonstop. Overall, I've quite happy with it. The fit an finish it quite good, with one exception, which I'll get to in a minute. For a printer that has large structural and mechanical components composed of printed parts, the accuracy and repeatability is quite good. The mesh auto leveling does a great job of accounting for tilt in the bed. My bed has ~ 0.75mm of height offset from front to back.
While I could pull the bed heater and install shims to bring it closer to level, that just provides an opportunity to screw something else up, and the software is handling the bed variation just fine. On the subject of bed leveling... this is one place where the printer firmware falls a little short. The mesh auto leveling takes nine measurement points. However, the firmware only allows for five points of compensation and they are not independent. The five points are live Z, front, rear, left, and right. Each point effects all three points along the perpendicular axis. So a -10um offset to the front will offset front-left, front-center, and front-right by -10um. Further, offsets are linearly interpolated across the bed. So that same -10um front offset will offset the center-left, center, and center-right points by -5um. This makes bed offset adjustment very difficult. I am not the only one to notice this and at least one person has developed a firmware patch that lets you supply independent offsets for all nine points via g-codes. I haven't made it far enough into the thread to see if this is being incorporated into the production firmware or not.
On the topic of firmware features, while I was typing the preceding paragraph, the printer was busy printing bed calibration squares and I got to test another feature of the firmware -- power loss detection and recovery. The printer is currently sitting in my home office and plugged into my UPS -- a cheap APC Back-UPS 750. The UPS is quite sufficient for my office workstation, but the transfer time is slow enough that the MK3 power loss detection will kick in before the UPS can transfer to battery. Twice during the last print, the UPS switched to battery and the MK3 power loss detection triggered. In both cases it was able to recover and continue printing where it left off. There are minor surface imperfections due to some ooze while the extruder was re-heating, but fall all but the most complex or detailed prints, it would have saved the print.
Oozing. With the stock I3 MK3 Slic3r profile, I get pretty significant oozing on print start up. Once the extruder comes to temp, residual pressure will cause filament to ooze out of the nozzle. If just pre-heat and let it set, I can get a 6cm string before it stops. Where this really becomes an issue is during mesh leveling. If there is any ooze on the nozzle, it can leave blobs at one or more of the measurement points. You can fine several version of custom start and finish g-codes to help address the ooze. I've got my own version that seems to almost completely fix it.
I mentioned one fit and finish issue. Actually, I came across two. One i mentioned earlier in the thread -- the heat bed screws having varying sizes of heads. The other is the power supply. If you go searching, you will find a number of threads on the MK3 power supply and failures of the supply, sometimes rather energetically. One of the PRusa forum threads is over 55 pages now. Most of the power supply issues have been reported by people with 120V AC grids, like me. A common report is that the power supply clicks when the bed heater is enabled and it is quite noticeable when the heat PID algorithm is holding it at temp. The bed power can be pulsed several times a second and the power supply clicks each time it is pulsed. Several people have reported that heating the bed and the extruder can pull right at the supplies rated 240W. Running the motors at the same time would exceed its rated capacity. This can obvious be avoided at the expense of extended start up time by sequentially heating the bed and extruder and keeping the motors off during that time. However we don't have any control over the PWM phases and rates during print so it is entirely possible that power draw exceeds the supplies rating routinely, for short durations, during a print. At least my power supply is audibly noisy beyond the clicks. With the printer at idle, when power draw is minimal, I can hear the distinct whine/hiss of the PWM circuits in the supply. A number of folks have ordered replacement 350W power supplies as an "upgrade" and I count myself in that group. I have a MEAN WELL LRS-350-24 arriving tomorrow. There are brackets to adapt the slightly different size MEAN WELL supplies to the MK3 mounts up on Thingiverse and I will be running a print of one over night tonight. As noted, the problems, including the click, seem to be mainly related to running from 120V AC. Prusa is aware of the clicking and states that it isn't anything to worry about and is mainly noticeable due to how quite the rest of the printer is. For me, $30 seemed a reasonable expenditure to remove the possibility of an energetic power supply failure, especially since I'm not monitoring the printer at all times. While I think the possibility of anything more than a pop, flash, and bad smell is incredibly low and I'm not really worried about that, I don't want to come home and find a failed power supply and have to wait for a replacement supply to show up. Should Prusa have addressed this from the beginning? That's kind of up to individual opinions. The included 240W supply is fanless and finding a 300-350W fanless supply is much harder. Likewise, the LRS-350-24 I ordered is not an auto-ranging, active PFC unit like the one that comes with the printer. Such a supply, is more like $75 instead of $30. Prusa made the choice they made, and based on the amount of testing they do, I am pretty certain they believed the included supply to be sufficient. However, it is certainly the weak point of the printer.
I can't compare the I3 MK3 to another printer as I don't have one, though I do have prints from an Ultimaker 2+ from my local library. Time will tell whether I can get equivalent quality prints or not. The MK3 is certainly capable though. I haven't finished calibrating and adjusting yet and my printer still successfully printed one of the torture tests
you can find on Thingiverse, including 80deg unsupported overhangs. The back side of the overhang tests gets rough at 60 degress and pretty ugly by 80 degrees, but the print was successful.
Now, back to fine tuning my print bed.