Firmware 3.9.0-FINAL for Original Prusa i3 MK3S & MK3  

Page 2 / 3
  RSS
bobstro
(@bobstro)
Illustrious Member
Posted by: @graham25fox

[...] put the SD card in to the printer and it tells me i have got to put the 3.9 FW in!

could it be that the final release if different some how?

I will put the new 3.9 full release in,but have never seen this before, and i have updated at least 5 or 6 times.

There are new checks that you can enable in the firmware to verify your firmware, printer and nozzle size. It's a nice idea, but I find them more annoying than anything else. The specific lines you want to remove from your startup gcode are:

M862.3 P "[printer_model]" ; printer model check
M862.1 P[nozzle_diameter] ; nozzle diameter check
M115 U3.9.0 ; tell printer latest fw version (version changes)

The firmware check makes sense since newer firmware may implement features you won't have with older firmware. In any case, you can also disable or change any of these to warning only in the printer setup menu.

is this something new?

any ideas, at least i can carry on printing

 

My notes and disclaimers on 3D printing and miscellaneous other tech projects
He is intelligent, but not experienced. His pattern indicates two dimensional thinking. -- Spock in Star Trek: The Wrath of Khan...
Posted : 29/05/2020 3:47 pm
patrick.g16
(@patrick-g16)
Active Member

Modified my GCode to remove the M115 line and reverted back to 3.8.1. Printer is working again, but would like to find out why I have an issue upgrading the 3.9.0.

Posted : 29/05/2020 4:31 pm
Graham25Fox
(@graham25fox)
Trusted Member

Great thanks for that, now i know what is going on

will see how i get on with the standard code before i change anything.

G

Posted : 29/05/2020 5:56 pm
JIJM
 jijm
(@jijm)
New Member

@alessandro-m8

Hi, I have the same problem, I've just updated today and only can select English languaje. I don't test another functions and I haven't print any models at now.

Posted : 29/05/2020 5:58 pm
bicio2002-1
(@bicio2002-1)
New Member

Installed the new firmware without any issue, language selection worked fine (italian).

I printed 4 small items in a row without problem, so far so good!

Posted : 29/05/2020 11:23 pm
alessandro.m8
(@alessandro-m8)
Active Member

As suggested on another discussion I tried with linux, Xubuntu version and prusaslicer for linux, and it worked, it installed firmware 3.90 without errors. now I can change the language in Italian without errors . But using windows 10 I was unable to install it correctly.

Posted : 30/05/2020 4:20 pm
Karl Herbert
(@karl-herbert)
Famed Member

@alessandro-m8

Same problem - language changing was not pssible. Maybe a driver problem on W10. I installed linux ubuntu 18.04 with PS 2.2.0 on my old HP laptop. After the first flash with it, I could change the language without any problems. Currently I am still working with 3.8.1 and prefer to wait for 3.9.1.

Statt zu klagen, dass wir nicht alles haben, was wir wollen, sollten wir lieber dankbar sein, dass wir nicht alles bekommen, was wir verdienen....
Posted : 30/05/2020 6:19 pm
michel.c3
(@michel-c3)
Active Member

I have 2 printers one is MK3S-MMU2S which is still running FW-3.8.1 and an older MK3 with updated FW-3.9. I use OctoPrint and noticing a strange unresponsiveness to Control button clicks on the MK3 with FW-3.9. MK3S-MMU2S with FW-3.8.1 does not behave the same way. I am not sure if my button is giving me issues or the new FW3.9 is causing a delay in reading button clicks. It is most noticeable and annoying when trying to do Live-Z and then switching back to Main screen.  It does act as if there is some sort of random delay. I had no issues with this printer with FW-3.8.1, about a week ago, doing similar Live-Z tuning. Just to remind again, I am running OctoPrint, can somebody verify the Controller Button issues I am having with their machine and FW-3.9?  Thanks

Posted : 09/06/2020 4:55 pm
alessandro.m8
(@alessandro-m8)
Active Member

I confirm that in my case after updating the firmware via linux, the language problem has been solved, also now I can install the firmware also through windows 10 without errors.

Posted : 09/06/2020 5:29 pm
Karl Herbert
(@karl-herbert)
Famed Member

@alessandro-m8

Strange happenings, but it's good that it is working now. After I had the same problem as I said, I stayed with linux. No problems anymore since over a year.

Statt zu klagen, dass wir nicht alles haben, was wir wollen, sollten wir lieber dankbar sein, dass wir nicht alles bekommen, was wir verdienen....
Posted : 09/06/2020 7:30 pm
Kaiede
(@kaiede)
Active Member

I know Linear Advance 1.5 will change how the motors tend to behave, but I've noticed something a little unusual. 

I've got mine on an old printer stand I've been using for my 2D printers, which I know isn't stable, and I do get some resonance/vibration. But generally with 3.8.1 it's been fine. That's the sound I've grown to expect over the last month with this printer.

But with 3.9.0, I get some really ugly resonances, especially on the Y axis that's noticeable when doing longer straight moves. Especially noticeable when printing out a set of spare parts from PETG using default settings. It's been making me think something is actually vibrating to the point that I re-lubricated both the X and Y axis to see if that was the problem. No change. Went back to 3.8.1 and the sound is now back to what I remember. 

Does anyone else notice changes like this, or is it just me?

Posted : 16/06/2020 7:47 pm
bobstro
(@bobstro)
Illustrious Member
Posted by: @kaiede

[...] Does anyone else notice changes like this, or is it just me?

3.9 provides an automatic translation of LA1.0 values to 1.5 equivalents. There were apparently some issues with early RC versions that have been cleaned up in the final. However, I had some really ugly results when initially switching. A couple of things I've learned:

  • The order of M900 Kxx statements matters. I forget the exact details, but I recall that using a LA1.5 value switches to 1.5 mode and will not switch back.
  • The automatic translation between 1.0 and 1.5 tends to produce slightly higher values than you might get with manual recalibration.
  • A power cycle or firmware reset may be required to clear old configuration data.

I've not noticed any audible differences but I've got my printer on a solid paver stone.

My notes and disclaimers on 3D printing and miscellaneous other tech projects
He is intelligent, but not experienced. His pattern indicates two dimensional thinking. -- Spock in Star Trek: The Wrath of Khan...
Posted : 16/06/2020 8:29 pm
Karl Herbert
(@karl-herbert)
Famed Member

If I change from LA1.0 to LA1.5, I would do the LA calibration again. Before doing so, perform a factory reset when flashing the 3.9.0.

Statt zu klagen, dass wir nicht alles haben, was wir wollen, sollten wir lieber dankbar sein, dass wir nicht alles bekommen, was wir verdienen....
Posted : 16/06/2020 8:38 pm
Kaiede
(@kaiede)
Active Member
Posted by: @bobstro
Posted by: @kaiede

[...] Does anyone else notice changes like this, or is it just me?

3.9 provides an automatic translation of LA1.0 values to 1.5 equivalents. There were apparently some issues with early RC versions that have been cleaned up in the final. However, I had some really ugly results when initially switching. A couple of things I've learned:

  • The order of M900 Kxx statements matters. I forget the exact details, but I recall that using a LA1.5 value switches to 1.5 mode and will not switch back.
  • The automatic translation between 1.0 and 1.5 tends to produce slightly higher values than you might get with manual recalibration.
  • A power cycle or firmware reset may be required to clear old configuration data.

I've not noticed any audible differences but I've got my printer on a solid paver stone.

It’s 1.5 first, 1.0 second. The models I’ve used with 3.9.0 have been sliced using 2.2.0 and I’ve confirmed they are spitting out both the 1.5 and 1.0 LA values, which are both using the default for the filament. So there shouldn’t be any translation happening here. 

It’s not really a print quality issue, but rather the noise difference for me is noticeable enough that it makes me think “something’s different/wrong, you should investigate”. It really does sound like something vibrating in a way it hasn’t done since I got it. I really wasn’t expecting it to go back to “normal” after reverting the firmware, yet it did, so it is a bit weird. Especially when it happens only on certain moves. The obvious one being a straight line that should only move the Y-axis.

I power off each night, but I haven’t tried a factory reset. Next time I update to 3.9.0 I’ll try that. 

Posted : 17/06/2020 2:29 am
bicio2002-1
(@bicio2002-1)
New Member

Had 2 issues happening that I would like to highlight in case they happen to anybody else for potential future corrections:

  1. after an electrical power interruption I chose to resume the print: when the print started Offsets were immediately  evident in both X and Y axes. The one on the X axle was in terms of centimeters to the point that the extruder went to the right mechanical stop while printing. Had to cancel the print and after that everything back fine.
  2. following to 2 subsequent crash detection WHILE ON 1ST LAYER (how that is possible?) the printing went offset on Y axle. After the print was cancelled everything back fine good, crash detection off since then.

I tested the belt and they spot on on the new numbers provided by the new firmware, I think a great improvement has been achieved with this new setting.

This post was modified 1 month ago 2 times by bicio2002-1
Posted : 20/06/2020 3:42 pm
bobstro
(@bobstro)
Illustrious Member

I read somewhere that the presence of a M115 U3.9.0 command in startup gcode affects translation. I haven't dug through the source code to determine whether this is true or not, and I didn't spot anything reading through various github threads.

 

 

My notes and disclaimers on 3D printing and miscellaneous other tech projects
He is intelligent, but not experienced. His pattern indicates two dimensional thinking. -- Spock in Star Trek: The Wrath of Khan...
Posted : 20/06/2020 4:33 pm
Kaiede
(@kaiede)
Active Member
Posted by: @bobstro
Posted by: @kaiede

[...] Does anyone else notice changes like this, or is it just me?

3.9 provides an automatic translation of LA1.0 values to 1.5 equivalents. There were apparently some issues with early RC versions that have been cleaned up in the final. However, I had some really ugly results when initially switching. A couple of things I've learned:

  • The order of M900 Kxx statements matters. I forget the exact details, but I recall that using a LA1.5 value switches to 1.5 mode and will not switch back.
  • The automatic translation between 1.0 and 1.5 tends to produce slightly higher values than you might get with manual recalibration.
  • A power cycle or firmware reset may be required to clear old configuration data.

I've not noticed any audible differences but I've got my printer on a solid paver stone.

So, I just rebuilt my extruder (because I wanted to... mostly), and since I had to factory reset anyways, I went to 3.9.0 and did a factory reset there with full re-calibration. And again the noise is back.

That said, I'm noticing it's definitely the Y-axis bearings that are the source of the noise. The ball bearings themselves seem to be rattling or producing more noise during specific moves. Most obvious when printing the 20x20x20 mm "hollow" cube for checking extrusion width (using your recommendation of two perimeters), or when laying down the first layer of some board game organizer boxes I've been building in CAD. Both of these being slower Y-axis movements. In the second case, once it's on the second layer moving more quickly, it goes away, and starts to come back when the top layers (of the box's bottom) are getting printed. Larger box perimeters also don't trigger it on the outside perimeter like the 20mm one does. Interesting.

So now I'm wondering if it's just that the changes to Linear Advance are hitting some sort of sweet spot to vibrate/rattle these bearings, while 3.8.1 didn't. That said, this sort of bearing noise has me a little concerned about the lifespan of the bearings under these conditions.

Posted : 30/06/2020 3:49 pm
bdlau
(@bdlau)
New Member

Linear Advance e-jerk value

The conversion function between LA 1.5 and 1.0 doesn't refresh e-jerk value. As a result, the print might be up to 30 % slower

We had upgraded to 3.9.0 and had to revert back 3.81 because our normally 7.5 hour prints were taking 11+ hours and believe it is probably due to something with the e-jerk value changes we saw in the relase notes: https://github.com/prusa3d/Prusa-Firmware/releases  . We run a business that uses 3 Mk3S printers with two more printers on the way after the Prusa backlog. This upgrade was not ideal for our situation but hope to figure something out so we can utilize some of the new features. We couldn't find anyone in a similar situation in some searching, so we have a few questions.

Does anyone know if there a way to stay at the 3.9.0 without re-slicing all the models into new e-code files or if there is another way around this?If not we understand and will stay with 3.8.1, as we just need to know if this is just part of the improvements in the firmware and no other way around it without re-slicing, thanks. 

 

 

Posted : 01/07/2020 6:47 pm
johnbeck7799
(@johnbeck7799)
New Member
Posted by: @graham25fox

@frio

Ok thanks, but i just have put in the full and not the beta, it has stopped it

did not change any codes and used the same sd card and file.

The only thing i can think of is there must be a line of code in the FW that tells it which version it is here.

G

may be you programmers out there can tell me if this is the case, for me that is way over my head

G

yeah u r right like (travis scott merch)..

This post was modified 4 weeks ago 3 times by johnbeck7799
Posted : 06/07/2020 2:35 pm
Philippe D.
 phd
(@phd)
Trusted Member

Hi all,

My newly built i3 Mk3S came with the 3.9.0 firmware installed, but I have a question.

I tested the belt tension using the firmware, and it came out as 231-239 on the X-axis, and 271 on the Y-axis. So, according to information on this page, my X belt tension is a bit high and I should loosen it some.

Is this correct? Are the values still good with this firmware version? (I've been told that the belt test gives somewhat incoherent values in 3.9.0, hence the question)

Posted : 14/07/2020 6:42 pm
Page 2 / 3
Share:

Please Login or Register