Lost communication with MCU and MCU 'mcu' shutdown: Timer too close

Basic Information:

Printer Model: custom corexy 48v
MCU / Printerboard: Manta M8P v1.1 & CB1 v2.2

Fill out above information and in all cases attach your klippy.log file (use zip to compress it, if too big). Pasting your printer.cfg is not needed
Be sure to check our “Knowledge Base” Category first. Most relevant items, e.g. error messages, are covered there

Describe your issue:

For a few days I have been experiencing 2 problems connected to the MCU which appear completely randomly even with the printer stopped without work.

The most frequent one is “Lost communications with MCU” while the other error is “MCU ‘mcu’ shutdown: Timer too close”. Initially I thought it was a power supply problem so I replaced the power supply, subsequently I replaced the motherboard and CB1 with other twins and the problem remained and therefore it is not a hardware problem. Both errors are present in the attached log but I don’t have the skills to understand why they occur. I thought it could be a problem caused by the latest changes I made to the printer by inserting other extruders for multi-material and a second temperature sensor for the print bed, but how could this be the cause of the problem?
klippy.log (843.6 KB)

Hello @MikiFX !

You may have a look on these:

1 Like

Hi @EddyMI3D , thanks so much for the reply.

I tried replacing the hardware components right after I had already consulted the knowledge base. However I noticed 2 tests that I didn’t try, the elimination of the webcam and the advanced graphing.

In the log I posted, the additional multi material extruders are not present precisely to “reduce” the load on the MCU which however remains between 2 and 8% of use so it is not really overloaded.
Electromagnetic interference should not be present because the power supply, high voltage wiring (220V) and solid state relays are enclosed in an ABS box with IEC 61340-5-1 ESD classification and is placed far from the board. I’ve tried them all and I don’t know where to look anymore :frowning:

I did more extensive testing and finally found the problem. my print bed worked with 2 ntc 3950 100k sensors, one integrated into the thermal pad which works at 220v and an independent one embedded in the sheet metal of the print bed. Thanks to the Klipper diagnostic graphs I think I found something anomalous, some peaks where there shouldn’t have been just before the MCU crash, so as a first test I decided to delete the sensor integrated into the thermal pad from the configuration. I reduced the presence of the error by around 60% and then by physically removing the sensor I completely solved the problem. Now I ask a perhaps obvious question, can an NTC sensor really cause these problems? What could be the real causes? lack of adequate insulation? interference with the high voltage of the pad?

What do you mean exactly by saying that by “physically removing the sensor” you fixed the problem?

Did you pull the thermistor out of the pad or did you simply disconnect the wires from the sensor to your main controller board?

In any case, it sounds like there was some kind of voltage (induced?) that affected your main controller.

I mean physically disconnected from the board. and yes, I think the problem was the high voltage somehow affecting the sensor. I’m not an electronics expert so I can’t explain better than this. The error occurred only when the printing bed was turned on (it made no difference during the heating phase or while maintaining the temperature, since the printing bed was turned on the error occurred approximately 6 times in 3 hours. I believe that the voltage affected the sensor because after deactivating it from the printer.cfg (but keeping it connected) the error occurred only 2 times in 3 hours. After physically disconnecting it the printer completed 17 hours of printing without ever presenting either of the 2 errors. I’ll keep it monitored in the next few days to confirm what I just said, but I’m pretty sure I’ve solved the problem.

1 Like

Just to recap:
A connected NTC 100k was causing a “Timer too close” error on your side?

This is very strange indeed.

Just saw that you are using unoffical modification to the code. You should always debug such issues with a pristine Klipper WITHOUT any modifications. Klipper is complex and very sensitive to timing, so a slight oversight in any of these modifications can lead to unwanted and impossible to diagnose issues.

Yes, it’s damn strange, but I can believe it since the sensor isn’t of excellent quality… it’s the first time I’ve had a similar problem in 2 years of using klipper where I could solve every problem simply by reading the posts on github etc. I didn’t know that my version of klipper wasn’t official. using a manta board with Bigtreetech’s CB1 module instead of my traditional Raspberry klipper is probably dirty for that…now I delete everything and install the official version. Now that you tell me this, I am convinced that it can improve further.