Noisy temps, melted PEEKs, firmware 0.92 & heater decouples

Post Reply
kerog
Posts: 117
Joined: Mon Jun 09, 2014 1:54 am

Noisy temps, melted PEEKs, firmware 0.92 & heater decouples

Post by kerog »

It seems to me that there has been an awful lot of discussion on this forum about a number of inter-related issues. I'd like to see if I can summarize the situation.

As I understand it, a wiring/shielding issue with the thermistors is causing erratic temperature readings on some, but not all Felix machines. When using Firmware 0.91 or earlier, these erratic readings can cause a runaway heating that results in a melted PEEK part in the hot end, with the metal barrel falling out of the hot end. Firmware 0.92 includes a heater decouple test designed to catch this before matters get that extreme, but people have difficulty finding the right PID parameters to avoid false alarms prematurely terminating long prints. A number of people have reverted back to 0.91 to avoid these false alarm heater decouples.

Recent posts have indicated some preliminary hardware solutions. I've seen one poster who used a shielded USB cable for the fans and thermistors, and another who was able to get a cleaner signal by rerouting wires. In conversations with Felix tech support, I had heard speculation that the noise was coming from the cable wrap rubbing on the frame, but I was unable to consistently demonstrate a connection. A friend and I have been working on a Felix slated for a local school that has had 3 PEEK meltdown incidents, on both extruders. We have replaced both board and thermistors without having any affect. Recently, reading the post about cable rerouting, my friend tried just undoing the cable wrap, and this had the effect of totally cleaning up the temperature signal.

My conclusion is that the positioning of the wires inside the cable wrap is what determines the thermistor signal quality. We will try to isolate this to determine the exact culprit, but I suspect it is the proximity of the thermistor to the motor wires that matters, as the noise seems to get worse once the print starts. The community could help with this isolation by reporting for individual printers (and individual extruders) whether or not they are seeing noise in the temperature profiles and what the cable routing looks like inside the cable wrap. Specifically, it would be good to know which wires the thermistor wires are in close proximity to.

My question for the community is what is the best action to take in response? Can we come up with a simple solution - better than just undoing the cable wrap - that preserves some of the aesthetics?

Also, does my analysis hold water? Am I missing something?

-Keith

User avatar
DDME-Marc
Posts: 71
Joined: Sun Nov 24, 2013 11:04 pm
Location: New Zealand

Post by DDME-Marc »

Hi Keith,

I guess we all aspire to have simple plug and play machinery so I applaud your post.
kerog wrote:As I understand it, a wiring/shielding issue with the thermistors is causing erratic temperature readings on some, but not all Felix machines. When using Firmware 0.91 or earlier, these erratic readings can cause a runaway heating that results in a melted PEEK part in the hot end, with the metal barrel falling out of the hot end.
The fact that most printers have been self assembled would suggest no two wiring looms will be exactly the same. Further compounding differences will be what version of firmware are being used, along with the grade of PEEK (huge grade variance is available), thermistor details, board version, etc. Details such as these will be difficult to establish since components are supplied without documentation (yet all can have an impact).

Additionally when you consider all of these variance factors it is likely that not one set of parameter values, in particular PID values will suit everyone.

Personally I would consider improved insulation for thermistor and optical sensor cables a must if they remain with the main cable trunk.

For the record my thermistor cables are lying next to the fan and heater cartridge cables only (and the optical sensor cables run entirely separate).

Cheers,

Marc
Felix 3.0 Dual Head * E3D Titan V6 * Repetier-Host V2.1.3 * Repetier-Firmware 0.92.9 (01/08/17 - Modified) * KISSlicer Pro - 1.6.3 * Arduino 1.8.5 *

satoer
Posts: 436
Joined: Mon Oct 01, 2012 8:00 pm
Location: Netherlands

Post by satoer »

Interesting. I haven’t encountered electrical signal problems myself besides a bad USB cable. Maybe a simple solution is to put the noise sensitive wire’s a couple times through a ferrite core to eliminate the noise?
My Felix 1.5 printed designs on Thingiverse
E3D V6 Hot end upgrade for your Felix
Tweezer bracket for your Felix
Led bar upgrade for your Felix

frankjoke
Posts: 150
Joined: Sat Dec 28, 2013 6:07 pm

Post by frankjoke »

In my case, without any special cabling and all signals together (I did not look at this when had to re-wire because of a broken motor cable), I had same problem but after fastening the thermristors and the heaters when they were hot the problem was gone.
Using the last firmware from downloads section.

kerog
Posts: 117
Joined: Mon Jun 09, 2014 1:54 am

Post by kerog »

frankjoke wrote:In my case, without any special cabling and all signals together (I did not look at this when had to re-wire because of a broken motor cable), I had same problem but after fastening the thermristors and the heaters when they were hot the problem was gone.
Using the last firmware from downloads section.
What exactly do you mean by "fastening the thermistors and the heaters when they were hot"?

frankjoke
Posts: 150
Joined: Sat Dec 28, 2013 6:07 pm

Post by frankjoke »

On the hotend (Aluminium block part) there are the screws on the bootom where the Thermistor (temperature sensor) and the heater are fastened.

Make sure they are not loose, because after some 'hot' printing time they seem to get loose.

The only other reson I had was broken cables.
I was already thinking to change printer to energy chain having of 3 times broken within in a year.

Post Reply