Missing Steps - Cant Print SpreadCycle StealthChop tuning help

droftarts

@sputnikoc3d We spent a large part of the day testing @carcamerarig Mini 5+ setup. I was able to reproduce the 'small bangs' he talked about, by setting the changeover from stealthchop to spreadcycle to around 32mm/s, and it was particularly noticeable on gyroid fill sections of the test print, just like his video. If set to one or the other mode in config.g, by using M569 D(2 or 3) parameter, it stayed in the set mode and did not bang (was not using stall homing at this point, so did not need to change M569 parameters after power on). I was unable to reproduce the large bang in either mode with the latest firmware (14/4), or in 3.2.2. I ran out of time to test 3.3b2, but have set up my machine to try it out when I get the chance.

Also tested movement with bed mesh on, and with high jerk (900mm/min or 15mm/s) but didn’t push acceleration, which was set at 1500mm/s^2.

I recommended and setup for him to use stealthchop all the time, as the printer was happy to run at up to 300mm/s in stealthchop. I spent some time setting up what I regard as a good stealthchop tuning series of commands, and tuning the stall homing, which was pretty finicky to get right, and I'm still not 100% confident with.

I moved the M915 stall setup to homeall.g, and the stealthchop tuning commands (see posted homeall.g). It follows the recommendations from the Trinamic datasheet; see page 35 and 36 here https://www.trinamic.com/fileadmin/assets/Products/ICs_Documents/TMC2209_Datasheet_V103.pdf#page35
If anyone has more questions about this, please do ask, though it may be a couple of days before I get back to you.

I have set up my printer now with 24V and motors that are very similar in spec to the LDO 0.9 motors. I ran a quick test in stealthchop, with the latest firmware (14/4) and found motors quiet and no banging. In our testing today with the latest firmware, we had no driver errors reported by M122.

Kind of annoying we couldn’t replicate the big bangs today, but maybe switching to an older 3.3b2 firmware will provoke it.

If anyone wants me to test a different setup eg with stealthchop/spreadcycle changeover, please let me know.

Any other questions, please ask and I’ll get back to you when I can; I’m a busy family man, as @carcamerarig experienced today!

Ian

oc_geek

This post is deleted!

droftarts

@oc_geek please post config.g. Looks like you have a low changeover speed to spreadcycle, does it mostly print in spreadcycle?

Ian

oc_geek

This post is deleted!

oc_geek

@droftarts something broken my config... deleting post...

Apology ... i realize the drivers went back to stealthchop.... damn

droftarts

@oc_geek nope, standstill is in stealthchop, changeover is at 0.5mm/s. From your M569 report:

mode stealthChop, ccr 0x00053, toff 3, tblank 0, tpwmthrs 2000 (0.5 mm/sec)

That’s still a changeover. What motors do you have? Please just post your config, hard to find it on a phone in a thread that is so long.

Ian

droftarts

@oc_geek However, a low changeover speed should still work, just be quiet when not moving. I tested that as well, yesterday, couldn’t provoke skipping.

Ian

droftarts

@oc_geek do you change the mode after config.g is run? @dc42 is not sure that M569 reports the current state of the driver, but the one before. So if you set spreadcycle in config with M569 D2, but switch to stealthchop for stall homing, then back to spreadcycle, M569 may still report the stealthchop mode. I didn’t get a chance yesterday to check this, instead sent command to set mode twice, just in case.

Ian

dc42

@droftarts said in Missing Steps - Cant Print SpreadCycle StealthChop tuning help:

@oc_geek do you change the mode after config.g is run? @dc42 is not sure that M569 reports the current state of the driver, but the one before.

I have verified that M569 does report the correct state, at least in the firmware at https://www.dropbox.com/sh/dupd9u1xppr1om5/AADYg4guuj2n8LJokVqtK-C7a?dl=0.

oc_geek

@droftarts said in Missing Steps - Cant Print SpreadCycle StealthChop tuning help:

@oc_geek do you change the mode after config.g is run?

Ian

No i don't.
I had force spread-cycle in config.g (or to say i did but it got messed up...) via M569 D2
I now restated it in config.g...
will continue trying that mode and report if anything happens (hope not)

I have a Voron 2.4, it doesn't run any stall homing thing...

thanks

Argo

@droftarts

So it does look like the main issue was solved with the new beta FW.
Interesting that he needs V2 in his config to put the drivers into StealthChop. With my Mini5+ StealthChop mode is by default on and I do not have to set any “V” parameter.
Still I’m curious why this hasn’t been in issue with my Mini5+ as we have quite similar Prusa Styles printer’s.
Maybe revision of the MCU? I remember having a skipping extruder with 3.3b2 which after flashing and compiling firmwares back and forth wasn’t reproducible anymore. During that time I also had timeouts but the fact that I could not reproduce it was annoying from a debugging standpoint.
I’m also glad you confirmed the capabilities of TMC2209s StealthChop (300 mm/s). Here and there you can hear voices that assume it’s only usable for low speeds but for belt driven 3D printer’s it’s well beyond what we need and what will also produce decent prints in terms of ringing.

Nevertheless I’m glad we have a working firmware version now. I also will continue testing the LDO 0.9 motors with the new FW version but that’s maybe worth a new thread then?

oc_geek

@argo working with SpreadCycle... which we are testing

Stealthchop caused layers shifts to me tonight

dc42

@oc_geek said in Missing Steps - Cant Print SpreadCycle StealthChop tuning help:

@argo working with SpreadCycle... which we are testing

Stealthchop caused layers shifts to me tonight

Stealthchop can't react to sudden changes in load, such as when the print head hits a blob or overhang when travelling at speed. For this reason, I advise against using stealthChop at high speeds.

oc_geek

@droftarts said in Missing Steps - Cant Print SpreadCycle StealthChop tuning help:

@oc_geek please post config.g. Looks like you have a low changeover speed to spreadcycle, does it mostly print in spreadcycle?

Ian

Hi Ian,
which speed do you suggest to set for the mode swicth over ?

Currently i'm forcing D2 but i may try differently

thanks

Argo

@oc_geek

Though it’s not RRF but my printer with Klipper I use 200 mm/s travel speeds for draft prints and 160 mm/s infill speed for infill. External perimeter speed is slower (75 mm/s) because it’s just a bed slinger printer and would ring too much.
Bord is a cheap SKR 1.4 + TMC2209 in StealthChop mode.
Cornering velocity (comparable to jerk) is at 5mm/s as Klipper advises against going higher with input shapers and acceleration at 3.000 mm/s (going higher would be possible but then fine details get mushy because of too much input shaping).
Luckily so far I haven’t had any issues with shifted layers. During travel speed I use a Z hop height of 0.2mm so it does not hit anything.

The last time I had trouble with StealthChop having layer shifts was with the old 8bit Prusa EINSY board and TMC2130 drivers.

I do not remember but are you having layer shifts with 1.8 or 0.9 motors?

oc_geek

@argo ...sorry to say but all that it doesn't seem to be much relevant in here...let's please try to keep the replies on issues found on RRF / Duet
the tread is already long enough for the folks to work it with..

besides i have 0.9 OMC on A/B/Z

Argo

@oc_geek

0.9 I see. I also had issues with 0.9 + TMC2209 on different boards and firmwares (Marlin, RRF).
At what speeds does the layer shift happen or is it somehow possible to easily reproduce it? I’m happy to add data to the matter with two sets of 0.9 motors from LDO and StepperOnline.

oc_geek

@argo said in Missing Steps - Cant Print SpreadCycle StealthChop tuning help:

@oc_geek

0.9 I see. I also had issues with 0.9 + TMC2209 on different boards and firmwares (Marlin, RRF).
At what speeds does the layer shift happen or is it somehow possible to easily reproduce it? I’m happy to add data to the matter with two sets of 0.9 motors from LDO and StepperOnline.

If you find issues on your end please report here.

droftarts

@oc_geek said in Missing Steps - Cant Print SpreadCycle StealthChop tuning help:

which speed do you suggest to set for the mode swicth over ?

Currently i'm forcing D2 but i may try differently

Yesterday I tested at D3 V40 (I think), which was around 3mm/s from memory. This was below the jerk value, and seemed to work well, but I was running short air prints and fast test moves, not an actual print. Again, i set this in config and didn’t change it after reset.

Ian

oc_geek

@droftarts @dc42

Configuration is :

Rel 3.3 beta 2+2
Forced SpreadCycle (D2)
Mesh active

Still getting layer shifts
(see below at back after couple of layers)
I stopped the print - it produced a sound when that happened

Dumps: