Thermocouple temperature fluctuations and bad soldering (solved)

[dc42]

@verminsup said in Thermocouple temperature fluctuations and bad soldering:

@verminsup No it is reading in the correct direction. When I increase the nozzle temperature the reading also increases. I just have no way of knowing when it's hot if the temp is correct or even stable

1. Does the reading depend on whether the heater is on or off, and on whether VIN power is applied or not? If so, that suggests there is electrical leakage between the heater and the thermocouple.
2. Does the reading depend on whether any of the stepper motors is activated or not? If so, that suggests electrical interference between the stepper motors wires an the thermocouple wires.

[verminsup]

@dc42 Nothing changes when motors are on or off when It does go down it goes down slowly as the chamber heats up. When this happens onlt the heaters for the chamber are on.

[dc42]

@verminsup can you confirm that you are using the proper thermocouple wire all the way from the thermocouple to the daughter board?

[verminsup]

@dc42 Just double checked the wiring and all seems to be in order. I truly wish I could convince my boss to just use some glass bead thermisters on the hotends because those would fit just fine.

[dc42]

@verminsup can you post photos of the wiring?

[verminsup]

@dc42 This uis the only part that is vissible without taking the machine further apart. The wires go through the cooling duct to the printhead (this is the original desing of the stratasys) The red light is from the config because I've turned that one off.

[rjenkinsgb]

@verminsup
Those wires do not all look to be thermocouple cable?
Examples of what it should be:
https://uk.rs-online.com/web/c/automation-control-gear/sensors/thermocouple-extension-wire/?applied-dimensions=4294876816

Also, even adding ferrules adds extra metal-to-metal junctions and can introduce errors.

There should be nothing other than purpose made, type-matched (eg. Type J, if that is what you are using) thermocouple cable and connectors until you reach that board.

A thermocouple is only a connection between two dissimilar metals, so add in any other dissimilar metal connections along the wiring and it messes things up in a big way!

(I've had a lot of experience with them in industrial machines over many years).

[verminsup]

@rjenkinsgb
That must be it, the whole wire is rated thermocouple wire until the last 50cm. My boss wired this machine and had no idea this was a thing with thermocouples, thanks for the reply this has finally convinced him to switch over to regular thermistors.

[dc42]

@verminsup if you use regular wire for the last section at the daughter board end, this will have two consequences:

1. The cold junction will be where the thermocouple wire joins the regular wire, instead of at the daughter board. This means that the temperature reading will be out by approximately the difference in temperature between that cold junction and the temperature of the daughter board.

2. If the regular wire is not tightly twisted pair (as thermocouple wire always is), then it will pick up inductive interference, for example from stepper motor cables.

In view of the wiring disadvantages of thermocouples, we recommend them only for use with temperatures that are too high for PT100 or PT1000 sensors.

HTH David.

[verminsup]

@dc42 The chamber and the cooling fans are 5 very large blower fans, could the temperature change when opening the door be explained by a lower pressure in the chamber causing the fans to spin faster and creating more interference?

[dc42]

@verminsup I guess that's possible if the fan PWM changed.

Is there any reason why opening the chamber door would change the temperature difference between the TC daughter board and the cold junction?

[ozgunkilicafsar]

@dc42 is the T"J" the correct command to define a J type thermocouple? In https://docs.duet3d.com/User_manual/Connecting_hardware/Temperature_connecting_thermocouples it says K"J" and I've seen both in various posts on the forum.

[ozgunkilicafsar]

@dc42 nvm found your reply on another thread from 2022 "...there was an error in the documentation. The thermistor type letter parameter is K not T. I have corrected the documentation."