Temperature measurments of print bed and hot end

[plexus]

Here is a chart of my measurements of the hot end. I used an Apollo dual channel digital thermometer and J type thermocouples. The thermocouple was tightly mechanically coupled to the hot end nozzle and all temps were stable before readings were logged.

In order to get a true 230C I updated the firmware to allow a 275C max extruder temp (the default was 250C only allowing a 220C true temp).

[plexus]

Here is a temperature chart for the print bed. mine is aluminum and the thermistor was mechancially coupled to the middle of the heat bed with not other obstructions, shown in blue. also, in green, are measurements taken from the middle of the bed with a typical hardware store IR thermometer gun from about the same height (20cm) and pointing straight at the bed.

[plexus]

Here is a chart without the average and the sample points. the blue line is the thermocouple measured values. the red line is ideal. the green dashed line is the measured values, mapped onto the ideal line using a simple linear transformation. what this means is that if I take the value as computed by Marlin and displayed in the host software, called T, I can get a pretty accurate reading by: ( T + 14.94 ) / 1.14 so that the displayed temp more closely matches the true temp as measured by the thermocouple.

Any idea where the best place in Marlin is to do this calculation on the value that would normally be displayed?

[lwalkera]

I believe there is a python script in the Marlin source to regenerate the temp table it uses to linearize the thermistor readings.

[Mooselake]

Since you have the actual temps you could adjust the values in thermistortables.h rather than trying to approximate new ones with a script.

Impressive work!

Kirk

[plexus]

Since you have the actual temps you could adjust the values in thermistortables.h rather than trying to approximate new ones with a script.

Impressive work!

I hear you. however I feel the better (easier?) course of action would be to simply take the value being sent out to the host software and do the math adjustment on it. that way the firmware is only modified with one line that would be specific to my printer. i just don't know where in Marlin to do this.

[plexus]

I just did a PID autotune on the hot end. so Marlin on our PBrds does use PID. its all over the firmware source code. there are even comments on how to disable PID and apparently its not disabled for the PBrd. Doing the autotune really makes the temp much more stable. before i would get a wavy line in Repetier-host temp graph, variations up and down about +-5-10C depending. after running autotune for 10 minutes and using the new values, now its a rock solid flat line. seriously. wow.

How to:
Power everything up and connect. make sure the extruder is stable at room temperature.

Issue M303 Sx, x is your typical target temp. i used 250C because as per the posts in this thread 250C is really 230C. i use ABS mostly so I want my hot end at a nominal 230C. i let it run for 10 mins and then disconnect. make note of the P I and D numbers under "Clasic PID" (is this a typo? there are lots through the comments in Marlin).

You can either just put M301 Pp Ii Dd, where p i and d are the numbers from above, in your start gcode and it will use those. or, what I did, was issue M301 Pp Ii Dd and then M500 to store them in the EEPROM on the PBrd. I also, for good measure, put them in my Slice3r start g-code. I did M501 to make sure they are set right.

That's it!

[chris]

plexus - so you are still using a custom fw build? Wouldn't it be a good permanent change to allow an override of key constants if a certain file is present, or maybe boot parameters?

I'm a little confused, in the end did you decide 250 was high enough (which the regular fw allows)?

[plexus]

plexus - so you are still using a custom fw build? Wouldn't it be a good permanent change to allow an override of key constants if a certain file is present, or maybe boot parameters?

I'm a little confused, in the end did you decide 250 was high enough (which the regular fw allows)?

I think its a bit of a misnomer to call it a "custom firmware build" considering that its open source and its intended to be customized. also, the only "customizations" have been to variables that control prime aspects of the printer; variables that are intended for the end user to change. in fact, PBHQ themselves are basically customizing Marlin for use on the Printrbot. so its really customized from the get-go.

i agree it would be great to have perhaps all the variables in config.h and some in config_adv.h, the two files that have the what i will call "end user variables" available. but alas, that is not something supported by Marlin at the moment.

[chris]

I'm actually just saying wouldn't it be better to make it so that frequently changed things don't force a recompile. I guess noone else sees the value in that?

[PxT]

I'm actually just saying wouldn't it be better to make it so that frequently changed things don't force a recompile. I guess noone else sees the value in that?

A lot of things can be set via M-codes and either written to EEPROM or changed at runtime, so no recompile needed. Speeds, acceleration, temp settings, etc.

[PxT]

I believe there is a python script in the Marlin source to regenerate the temp table it uses to linearize the thermistor readings.

Source:
https://github.com/PxT/Marlin/blob/mast ... eLookup.py

[AlexC]

I understand the charts, basically the temps usually lower than the setting likely due to iffy contact of the thermistor to the hot end,. and everyone's contact is a little different than anyone else's,. etc,.

But is there an easy, "follow the directions", way to get your hot end temps calibrated more accurately,. i am not really a programmer, don't really know what marlin is, guessing a compiler of one sort or another, but altering programming/firmware is not really one of my strengths,... I am more a mechanical engineer,. really good at following directions though,.. so if this is outlined somewhere even if its over my head i can probably just go step by step it,.

I connected my hot bed thermistor to my hot bed using non conductive thermal epoxy, which increased the accuracy by about 8 deg, still a hair lower than reality but almost 50% more accurate. thinking of doing the same for the hot end if i cant find a follow the directions method of calibrating it,. though that's a semi permanent move, but at least ill be a little more accurate,. maybe,.?

[Tdeagan]

Since you guys are chatting about Autotune, I'm kinda stuck on an aspect of that.

The temp on my extrudrboard extruder is oscillating nastily (set to 235, oscillates between 207 and 249) I want to Autotune and know the syntax for extruder specific tuning:
M303 E1 S235 C5 (five cycles tuning to 235 on second extruder)

What I'm stuck on is that neither via M301 nor configuration.h (or the LCD EEPROM access for that matter,) is there any way to put in different Ki, Kp or Kd values for each extruder.

Is this like the extruder feed rate, i.e. it's one value shared amongst all (except the bed)?

if so, any tips on how to pick a good set of values to solve a problem that only existed on the second extruder without horking up the first one?

[marlus]

I have a problem with my bed thermistor, I wonder if anyone can help me.

I've already bought and tested a new thermistor
I've also reflashed the firmware of my Printrbot Plus v1 (Rev D board)
with Printrbot Firmware Updater (OSX)

But the problem persists: bed temperature is 80 celsius. If I heat it, don't pass 90 celcius

Is there a way to calibrate it in these conditions?
Could be a problem with the board?

Thanks in advance!

[RetireeJay]

There are many varieties of thermistors. The standard for Printrbot printers calls for 100k ohms at 25C, negative temperature coefficient. But even with those criteria, there are still several different thermistor tables to translate temperature into resistance. If you are capable of looking at the Thermistor tables in the source code, you should try to find your particular brand and model number of thermistor, and then specify the corresponding table in Configuration.h

That said, it's still odd that you can't get your temperature reading to go above 90C. Do you have a sense that it's anywhere near that hot in reality (near boiling temperature)?

[marlus]

Hi @RetireeJay

I've bought the new thermistor from Printrbot store, and they are both (old and new) showing the same behavior...
That's why I'm afraid it's something with the board.

I reached 90C by putting the thermistor touching the hot end tip at 150C.

[calex_fr]

Hi alls,

I am french and I have a Simple Metal Rev D with new hot bed upgrade, and I have a problem.

In Repetier I confugure hot bed to 60°C but in reality I have 50° mesured with my infrared thermometer, does it normal ?

[plexus]

Hi alls,

I am french and I have a Simple Metal Rev D with new hot bed upgrade, and I have a problem.

In Repetier I confugure hot bed to 60°C but in reality I have 50° mesured with my infrared thermometer, does it normal ?

IR thermometers are not very accurate unless you have a really expensive one and even then... you should use a conventional or digital thermometer that you know is correct.

[calex_fr]

Hi Plexus,
I controlled my IR thermometer, he has 1% errors, therefore not 50 for 60°C.