Timer too close and Lost connection to MCU errors
Dear Community,
I am running a RatRig Vcore 500mm with USB Toolboard.
The specs:
Raspberry Pi4 B+ 2GB (reads 1,6GB RAM only) - 64GB Kigston MicroSDXC card - Active cooled (Mostly 45C all the time)
EBB42 v1.2 without MAX - 3DO 2m USB cable + 2x0,75mm wire for power - 2x1mm Piano wire for umbilical + zipped so the connector does not move
Octopus V1.1 Board
Phaetus Rapido HF hotend
24V Supply
BTT TouchScreen 5” with FCC cable + KlipperScreen
FW versions:
klipper
v0.11.0-303
KlipperScreen
v0.3.6-52
mainsail
v2.8.0
moonraker
v0.8.0-186
RatOS
v2.x: v2.0.2-15
ratos-configurator
v0.0.0-67-gc5b78c02-inferred
RatOSTheme
v2.0.2-0-g515851db
timelapse
v0.0.1-142-g57bec0de
Host Klipper version says dirty at the end: v0.11.0-303-g67499853-dirty
Also sometimes its failed to start up then I have to power cycle it 1-2 times. Also sometimes don’t get the wifi signal and starts as RatOS but the configuration is still good. After a restart it catches up with the Wifi. Signal strength should not be an issue as I have a Voron which is next to it and it starts all good. BTW they both share the same electronics and it never happened on the Voron.
My problem is sometimes I have this error message:
Klipper reports: SHUTDOWN
MCU 'toolboard' shutdown: Timer too close
And sometimes this:
Klipper reports: SHUTDOWN
MCU 'toolboard' shutdown: Lost communication with MCU
What I did already:
Before the umbilical upgrade to 3DO cable and piano wire a lot more error happened.
I lowered the micro stepping to 16
Only Input shaping is enabled ADXL is disabled (its only needed for measurement) - it improved a bit.
Restarted the printer before every print - No change
Speeds are: 350 travel - 150-250 print
Max access is 9000mm/s2
Can someone point me in the right direction?
Thank you guys, I will send virtual beers to all 🙂
Have a nice weekend!
Mate
48 Replies
Lost communication - sounds like you have a lost connection to the toolboard. How do you have the USB attached on your hotend?
I am using an EBB42 board using its USBC port with 3DO cable
Using a 2x0,75mm2 cable to connect the power
It printed without errors for almost 10 days without turning it off
I checked sysload from log and it was around 1.6 without printing a thing close before it lost communication with the toolboard
sysload is 1.3 maximum with my voron when printing not in idle mode
BTW thank you for helping me in advance!
I will check for mixed power maybe. "There have been reports of high USB noise when both the printer's power supply and the host's 5V power supply are mixed. (If you find that the micro-controller powers on when either the printer's power supply is on or the USB cable is plugged in, then it indicates the 5V power supplies are being mixed.) It may help to configure the micro-controller to use power from only one source. (Alternatively, if the micro-controller board can not configure its power source, one may modify a USB cable so that it does not carry 5V power between the host and micro-controller.)"
No its not the problem the Jumper is not on the toolboard on the VBUS.
Maybe the 0,75mm wire is the problem its current rating is 6 amp. The Rapido is 115W (4,8A) + the stepper is around 1,3A + the fans. Its around its edge but it does not covers it when its idle. It sometimes do it when idling.
What are these HEIC files? I can't open them
These are images
I will convert
Looks like the USB connected to the ebb42 is cocked at an angle
A little bit
Can this cause a problem?
Yes
Probably why it is losing connection sometimes
I will turn it straight and just use a ziptie on the upper part
the flexible part
But why did it worked for almost 14 days nonstop and it started to fail in the past 2 days?
Also if you dont mind. Why is it happening that when I power cycle the printer it always fail to boot on the first or the second attempt. And when it boots sometimes it start the RatOS SSID and not connecting to my wifi but the config is intact. When I restart it connects to the wifi. Can this issue be related? And also can this be the fail of the Raspberry?
Why did it just start failing - probably the connection finally started breaking from all the stress from the cable being at that angle and the big accelerations the machine sees
Why is it failing to boot - not sure, you would need to look at the logs and investigate more
Why is it creating Rat-OS wifi - it does this when it can't connect to the configured wifi
Is the wifi related - no
Is the pi failing - probably not
Okay. I will investigate it further and return to you
Thank you!
@kasztimate_additiq If you have no success try disabling crowsnest the next time your boot your pi and see if that helps on the timer too close / lost communication errors.
There's been a few users reporting that it fixed their issues, i'm still not sure why or if it's crowsnest's fault at all (I doubt it, but too many reports to ignore at this point).
Also - this can easily be a power delivery issue (5V to the Pi).
Also, make absolutely sure you have removed the V_USB jumper on your toolboard and your controlboard.
Thank you very much. Todays to do list:
- Disable Crowsnest
- Redesign extruder mount to optimize USBC Connector
- Checking V_USB Jumpers
- Checking 5V and 24V supplies on PI, MCU and Toolboard MCU when drawing full power.
What is crowsnest btw?
Webcam script.
Cool
I have designed this mount:
The USB C is supported from the back and constrained from the top + supported with zipties
See if it help
Okay so made the extruder modification as attached on the picture, also found a jumper on the Octopus on VUSB and removed it. Started a print and got Timer too close error after 4% of the print. Attaching you the log also and a picture about warnings with crownest path
Attaching the warning:
And also found a strange issue that when I plug the printer it wont boot. Just black screen on the touch panel. Then I unplug and wait replug and then a colored screen but still no boot. After the third time the printer boots as normal. I attach a picture about the screen it shows on the second attempt
Also attaching the RPI dmesg output
You have Crowsnest 3 installed, you need to update it to Crowsnest 4 (I am assuming you are running RATOS 2+). Go to the Crowsnest site to get info how to uninstall the 3 version and install the 4 version, it will not update automatically from 3 to 4.
This could cause this issue?
Can we fully uninstall this service? I never use webcams with my ratrigs and vorons
When I plugged in a screen to my RPi I started getting MCU timer too close issues mid print.
(It boots into the default vanilla klipperscreen).
Without the screen I only get MCU timer too close issues when I start a print but before it starts printing (rarely and only after I added an ebb42).
Okay I completelly removed crownest service. Checked the voltages and for the PI its 5,18V from a 3A Meanwell PSU and for the other its 24,5V from a Meanwell PSU. I start the same print again. See what happens.
Its 14% and still working 🙂
I think I will delete the timelapse function also as I dont need it
The strange thing BTW is that my Voron shares almost the same hardware (same RPI, same PSU, same toolhead (less power hungry dragon sf) and never had these problems. Even runs the old crownest v3...
How can be such differencies
Okay so its gone again :/ at 29%.
What now :/ should I rebuild a new sd card and change the raspberry or the usb cable or the ebb42 the last two is almost new :/
I am out of ideas 😦
Okay I ordered a new Raspi 4 8gb + Sandisck Extreme and Endurance cards + Active cooling for the PI also trying the new Pi5
In the meanwhile I will swap the card and the raspi from my voron and see if it helps.
I rebuild a new RatOS image on a 16GB Sandisk Ultra and put into the Raspberry and have this error: mmc0 error -110 whilst initialising sd card
Did you try running a print with the screen unplugged?
Not yet. Why would it be a problem after all when I printed without a problem for almost 14 days non stop and did not change anything?
Okay so changed the RPI to a New 8GB Rpi with a new 32GB Sandisk Extreme card with a clean install only using the printer.cfg from the old version. Pi is activelly cooled. Changed the 24V wire to the toolbard to 1,5mm2. Used the Official 27W (5,1V 5A) powersupply for the PI. Sliced a new file:
Same issue with timer too close. Now I will change the EBB42 and then if not I will change the USB-C Cable to a new one.
Changed the ebb to a new one. No luck again timer too close issue. Now I will try to fully uninstall crownest and disconect the screen.
I see you're running the 64-bit version. Try 32-bit.
Also please monitor your system resources when that timer too close issue happens
It's an issue on the pi (can't keep up). Should be evident from the error message.
How do you see it? I made the card using a 32image.
Strange
wait
Ubuntu/linaro
What system is that thing from
😂
?
What do you mean? I downloaded the image from Github. And used balena etcher
Well it's 64-bit, it should look like this:
download the 32-bit version and flash that
I bet that fixes your problem
I should really delete that 64-bit version there seems to be some major issues with it
(it should say
arm instead of aarch64)Can you check that the latest version where it says rpi32 is really the 32bit version as I have double checked it and this was the one that downloaded and flashed
really.. that's the one i'm running too
Not sure how that's possible
Hmm some devil is under the hood. I will check it then
Linux ratos 6.1.21-v8+ #1642 SMP PREEMPT Mon Apr 3 17:24:16 BST 2023 aarch64 GNU/Linux
Made a new card downloaded this:
Still shows this for uname-r:
Linux ratos 6.1.21-v8+ #1642 SMP PREEMPT Mon Apr 3 17:24:16 BST 2023 aarch64 GNU/Linux
For getconf LONG_BIT I get: 32
The superfunny thing that my Voron writes the same but for getconf LONG_BIT it says 64 and there is no error on it
I will change the 3DO cable to a new one as my last resort by I really dont think that this could be my problem.
Hi All 🙂 I solved the issue:
- Changed the cable to a new 2m 3DO cable (The old one was also 3DO)
- Removed ARCs from sliced file
- Removed the top hat while printing PETG
Dont know which one has solved it but its OK now. I will design a top hat raise for it because the umbilical has rubbed against the top panel so I will lift the whole top panel with 150mm. I will share the results there 🙂