Generate Shaper graph kills printer
When I try to generate shaper graphs, I get an internal error on command: "test_resonances".
57 Replies
please try to add this to your printer.cfg and try again
@dyocis
ohh i see, its a klipper bug
@miklschmidt found a bug in the klipper CHIPS name parsing
id doesnt allow to use other chip types than adxl, so atm idex doesnt support lis2dw chips
or am i wrong?
Ahh, ok. I thought I had run shaper graphs since installing IDEX...Hmm. guess not
yeah, it looks like current klipper code doesnt allow the use of the orbitool accelerometer chips
I know I've run it in hybrid
thats of course a problem
well, it works in single toolhead mode, but for idex we need to switch between the chips and klipper expects or adds the adxl345 prefix to the chip name
but you need the lis2dw prefix
Makes since. Well, just add it to the list of problems I've created for myself because I can't leave things stock....lol
i will try to file a bug for klipper, but i cant promise anything
Sounds good. I'll test it occasionally and let you know if it gets fixed
Should I close this post, or keep it open till it's fixed?
no, keep it alive please
@dyocis you can use the realtime analysis tool instead
I'm doing that as I type!
Ah nice! š
GENERATE_RESONANCES AXIS=X and GENERATE_RESONANCES AXIS=Y will do it, two separate macros š
Otherwise it'll be combined and you won't get the shaper results. You'll get a tension comparison instead.@dyocis i shared the finding on the klipper discord with theg guy who coded it, but as i said before i cant promise anything. i wouldnt have high hopes for it to be fixed
Yea, I saw the debacle around the z hop problem a few weeks ago. I don't have high hopes either.
Do the shaper results get saved somewhere or do I just need to look at the realtime graph? I looked in the input shaper folder, but nothing new is in there
@Helge Keck looks like this was already fixed https://github.com/Klipper3d/klipper/pull/6726
GitHub
resonance_tester: Fix chips selection, add accel_per_hz selection b...
-The current code does not have the option to "live" select a non-adxl345 chips when running TEST_RESONANCES, since it adds the adxl prefix by default.
-Added the ability to &...
You can look at all your macro recordings, just click on "view all macros"
Or you can click the link in the notification that pops up when it completes
Yea, I figured it out. It's the first time I'm using the macros in the tool, so I'm learning
The navigation needs a bit of love š
wut
We can update the pinned commit, but i have not had the opportunity to test the past two months of klipper changes
ahh, was jsut confused bc his logfile clearly showed this error. so its fixed but not part of RatOS atm, got it
exactly
Would the belt comparison macro be GENERATE_RESONANCES AXIS=XY ?
Or two separate sequences with X in one and Y in one?
GENERATE_RESONANCES AXIS=1,1
and another sequence in the same macro with
GENERATE_RESONANCES AXIS=-1,1
For tensioning the Y belts you can do:
and another sequence with
Remember to select the T0 accelerometer for the first sequence and the T1 accelerometer for the second.
Those are coordinates for a 400mm, change accordingly for a 300 or 500 š
Ie Y in both and X in the latter.I have a similar error after mounting an AXDL345 on my rat rig and I'm not as technically inclined as a lot of other people on here, my printer appears to keep crashing after doing a frequency test - this one occurred after I did the "MEASURE_COREXY_BELT_TENSION" command
Start your own post please
Any idea when this fix will be implemented in RatOS? I was hoping that the most recent update would contain it, but I tried it last night and I'm still getting the shutdown when I try to run IS
Any idea when this fix will be implemented in RatOS?Just pushed it this morning š Are you on 2.0 or 2.1?
2.1
That works.
Are you sure you updated?
Post a debug zip
And klipper is updated?
Aight, i'll need a debug zip to see what's going on
OK, I'll post one in a bit.
I clicked on the Generate Shaper Graphs macro button and it shutdown. This is the debug
Ah, you've selected the toolboard on t0 as your accelerometer in the hardware configuration, but you don't have a toolboard with an accelerometer
Not exactly sure how you got yourself into that situation though
The O2 board doesn't have an accelerometer?
it does but that's a lis2dw, not an adxl345
Aaaaah.
The plot thickens
i forgot the input shaper macros don't support that (they don't support anything but adxl345's right now)
let me check if it's an easy fix
We discussed that earlier this year in this chat. You said that klipper had a fix for it, but you hadn't had the time to get the commit added to RatOS
they changed it
should work now, thats klipper master
It's not an idex printer
I do have IDEX
Then i'm looking at the wrong debug zip
š
happend to me multiple times already
:kekw:
I was indeed. Okay so your RatOS.cfg is wrong for some reason... It has this
variable_adxl_chip: ["adxl345 toolboard_t0", "adxl345 toolboard_t1"] # toolheads adxl chip names.
I'm looking into why
I'm creating a test for this to track down the issue.Cool, let me know when you figure it out. I've got the macros in my realtime analytics, so it's not a show stopper, but it would be nice to be able to use the macros!
Curious why you'd rather want the static graphs?
(this is a bug, and it needs to be fixed, just curious as i'd hoped to eventually replace the static graphs with the realtime tool once it's more mature and polished)
@dyocis Fix pushed, ready to update in 5 minutes š
It's not a preference of one over the other, just that the macros don't work. The real-time graphs work well.
Aight just wanted to know if there was some use case where there were issues that i could do something about š
Nope, not that I can think of!
@miklschmidt It works now! I like that it does all the graphs at once. I didn't know the toolhead moves back and forth by like 20mm now. I thought something was broke...lol
That's the new resonance testing method that landed in klipper a few weeks ago š
Hi Helge, wanted to ask if you can specify a variable with only one head, i.e. not idex