Interesting experiment; your name and having written some dutch words makes me think you work in Veldhoven ;-). I visit there regularly.

Anyways, just to add some remarks:

• I'm not sure you can "feel" 833Hz with your hands. That's typically lower frequent. I would imagine the concrete drastically changed the vibration part right?
• changing the system "that much" with the frequency remaining about the same is quite strange. It makes me believe it's something else then the frame
• with the probkem being audible, I indeed think that acoustic measures (enclosure('s)) will now help you a lot!

You could also perhaps look at Tune Mass Damper principle. Make one as and test it at different places ;-).

But overall, really nice job man! Cool and it sure was worth reading to the end ;-).

Dumb question, but why neopixel if the duet3 supports Dotstar?

I relate to your ideas and findings. I have a nimble though, so bit different situation. A nimble has trouble to keep up wit PA due to its high gear ratio (as noted on duet website). So although direct drive it has some "bowden like" behavior in PA. My findings:

• using Marlin routine I need S0.54
• using the Python script approach I need S0.6
• in real printing I find better results with S0.4

So lower than in the tests with I think relate to your observation on for the under extrusion on very short lines.

I have no need anymore for coasting in cura, and overall I have better quality. I hope and think you do too right? It's about improving it even more.

So I think your idea on the minimum amount of time could work well. I'm wondering if it should be time or travel length though? That would take out the need to tune each time when you print with different speeds for different prints, or even when using different speeds while printing different features.

I think having both options would be best to tinker with.

Also I would think it would be even nicer if one can give two parameters:

• length/time No PA
• length/time Full PA.

And an interpolation in between to give a sliding PA value?

No idea how hard to implement, but I bet it would be a selling point when this works well (I'd be happy to test when indeed working well.

While think about it: anybody know how to modify a gcode file to search for short line length and putting an M572 in front (and after) to change the settings with some existing scripts or something to prototype test?

Cheers Q

Just to feed the discussion I've computed for my CoreXY printer the amount of force that could excite the printer due to the instant velocity change with a Gantry mass of about 2.2kG . Highest level is 40mm/s change and lowest only 2.5mm/s change.

This is quite theoretical as there is a mechanical system in-between and the stepper itself, so in reality it'll be a bit different yet. In all cases the forces results in lower torque requirement than a typical NEMA17 delivers though.

So still very curious if the pulse train (Step Dir) indeed changes "instantaneous" from one pulse to the other before I continue to dive deeper in this subject. THX

When searching for M566 etc I get over 50 pages on hits. And as my question is rather fundamental I thought to ask in a separate thread.

I'm trying to get an idea of the forces acting on the 3D Printer during operation. One important and relatively simple move is the instantaneous speed change setting (M566). So if we go really basic we could write:

F * deltaT = m * deltaV

where:
F is the force acting on belts from Printhead / Gantry
deltaT is the time the force is acting to make the speed change
m is the mass of the Printhead / Gantry
deltaV is the velocity change.

So m and deltaV are known from the printer and M566 setting. So what is F and deltaT?

I guess this is a question for @dc42 as it goes into the code but I'm assuming:

• There is a pulse train (step/dir) running to the stepper. As an example with a frequency of 10kHz at a certain moment with a constant speed of the printhead.

• Then all of a sudden the next line arrives and the new frequency (of step/dir) becomes 8kHz.

• Assume this change in frequency is permissible with the M566 setting as the actual speed change indeed relates to a step/dir frequency change that falls within 2kHz speed change in this case.

Then I wonder if indeed the change in frequency is from one pulse to the other or if there is still a (very short) interval where the frequency transitions from 10 to 8kHz.

If not: then deltaT is only ±0,1msec in this case (and speed dependant)
If yes; then deltaT is roughly the frequency transition time.

With DeltaT I'd have F what I'm after.

Personally I think it'll still make sense to have a transition time btw. even if ever so slow as it would limit the jerk severity a bit when moving faster and faster.

Hope my question can be answered, thanks a lot!

Q

@mrehorstdmd

Thanks for the nice additions a remarks! Wear and tear is not something I was think of to be honest ;-). I've also studied some of the Gates documentation, which is really nice till you want to do the math yourself: have you ever found the Specific Stiffness of GT2? I didn't unfortunately and once you start searching for the specifics I got lost in the fast amount of info ;-).

I thought driven pulleys require a lot of teeth, but was not away non-driven should be too. Makes sense when it comes to compression of the belt on the idlers though.

Most pulleys are 21 tooth on Ali, bit strange in view of your comments. These bearings are tiny indeed ;-0!!!

@gtj0 thanks for the answers! Great, I'll go twisted in my core to be too. So thanks for sharing, you helped me a lot!

Thinking about the idlers... Yes toothed could be better actually that way . I go toothed too, but saw some discussions bearings would be better. Nah, I like the toothed stuff!

Cheers!

@cdthomas9 looks nice, where will you buy this beauty? And what is the max temp it'll deliver?

You can get the same line thickness as the nozzle size, but it's typically better to add some 20-25% (you can choose but this is a good indication). It makes printing easier and can also be more esthetic.

So I use 0,3mm on a 0.2mm nozzle and 0,5mm with 0,4mm nozzle. Put the 0,5mm in Cura as line width and the flow should then be around 90% (you trest for yourself) to achieve this width and that's then good. Don't go too quick when printing the "vase mode) cube (wall).

Then you want to do the non linear extrusion speed test (m593 or something in that range from the top of my head) to get consistent extrusion amount with different extrusion speed. There are excel sheets here in the forum to help you with the math if you like.

Should cover the line width. Other things to calibrate is retraction and bridging. There are nice tests you can print found on thingiverse.

Holes are typically print too small, cura now had some horizontal expension parameter for getting holes print just right. Might be right for you but I typically use a drill if possible.

Cheers

PS: https://youtu.be/rp3r921DBGI

@mrehorstdmd Ok, so your feeling is that wear was at three places: left and right Idlers and in the middle the belt contact. THX.

@mrehorstdmd Interesting wear on the idler. Figures me a bit; would you agree that probably the debris there is cause by the belt rubbing over each other, with debris falling right down at the spot and another part at the idlers where the belts "Bend open"? Or do you think the wear is caused by the belt rubbing on the idlers itself?

@gtj0: did you notice any debris on your printer yet? You have the spacer in the middle that prevents the rubbing between the belts, so...

Also what is the minimum length you feel needed for the "twisting trick"?

I also do hard to figure out reading post if it is really better to have smooth idlers v.s toothed ones. Pffff. Feels to me that toothed idlers are not an issue because there is no load on them...

Ok thx, no dbc I suspect. Makes a lot of work, let’s see.

Dear @dc42,

I have a dynamic analysers that is used for vehicle development and would like to connect it to the Duet3 board's CAN-FD. It'll allow me to do some vibration measurements with acceleration sensors and map it to the CAN-FD information. To listen to a CANbus typically a .dbc file is used to decode the data stream. Do you have such file or can I find a list with address ID's somewhere on Github perhaps?

THX,

Q