print head thermal stability (or not!)

[daj1u06 0]

Hi, just been monitoring the print head temperature for a dual material print run (PLA / PVA support). It is currently printing with PLA and Cura reports the AA head as being within 1 degree of target temperature and it looks very stable. Similarly the print bed is within 0.5C of target and stable. However, the BB head for the PVA (currently not printing) is repeatedly rocketting all over the place from 25C over to 10C under target (100C when not printing) over a matter of a few minutes. When it is actually printing it seems better but the support area is currently small so barely seems to stabilise before swithing back to PLA head. Is this fluctuation normal or have we got a suspect BB head? We have had some issues with PVA clogging and burning and not sticking well and we know that it isn't temperature or humidity related as we dried the filament and RH is constant around 30-35%. As we tend to do complex, multiday prints it would be good to get this sorted if it needs sorting.

 

Thanks in advance,

[DidierKlein 729]

Hi,

 

When you say monitoring its by looking at the printhead temp reported.by the printer?

 

And if i understand well this fluctuating temp occurs when the bb core is up and not printing? 

 

How stable is the temp on the bbcore when its printing? I have never checked on mine but it always worked nicely.

 

Fluctuating temperatures would mean, either a too weak heat cartridge or a malfunctioning heat sensor, but then it should show the same behaviour when the core is printing

 

 

[tomnagel 126]

if your UM3 is connected to a network, you can monitor the temperatures via a webpage hosted by the printer. Type in the IP-address in your browser and click "temperature graph".

In a dual material print, you will see that temperature setpoints are constantly adjusted. The nozzle that is not being used is lowered in temperature to prevent burning the polymers. And the beginning and end of each layer are printed a little cooler to minimize oozing.

[daj1u06 0]

Hi Didier,

Yes, I'm monitoring it by what it reports back to Cura's monitor function.

Yes, I noticed it when it was printing a large area in PLA so the PVA head was idle. At the stage I wrote the post, there was only a very small amount of support being printed and this was quick to print and it didn't seem to stabilise at the target temperature. Its now onto a different part of the model where more support will be needed so I will be able to monitor it better / for longer

 

It has just started a PVA layer, target is 215C but  oscillated 192-205C for first 90 seconds of print time, then stable on 217 for a while then down to 203 only after about 5 mins of PVA printing had it stabilised on target 215 +-2 C but still shooting around, never truely stable.

 

Extruder 1 (AA for PLA rock steady on 99-101C (target 100C) when not printing, heating up whilst PVA layer prints

[tomnagel 126]

I have seen these spikes in the heater signal many times, and put it on the issue list here internally. But it did not seem too important since the temperature control was accurate (like in your lower picture)

 I see now for the first time a graph (upper picture) where the actual temperature seems to suffer from these spikes. That is alarming, because it will influence the print quality. Is this screenshot from a normal printing situation? 

 

[daj1u06 0]

Hi Tomnagel,

thnks for your fast response, yes it is a "normal" print situation as far as I am aware in an air conditioned room pretty stable in temp and humidity (currently at 24C and 39%RH). Default Cura settings except brim in PVA and support everywhere is on, to try to make sure that any base level support sticks well.

 

The model itself is somewhat unusual and complex though, hence the need for good, reliable support (it is a concretion full of cannonballs raised from the shipwreck of a Dutch eastindiaman - one of a series of 3D models recording the excavation of the concretion to recover the individual components - lots of support needed and at several days per print in finest quality, I need to get it working reliably. The printer is only a couple of months old, has latest firmare and use latest Cura to generate the print file.

[daj1u06 0]

Hope this was a misread event from the aa core temp sensor!!!!!

[tomnagel 126]

The nozzle in one sample going to 700 °C and back to 200 is physically impossible, so it seems safe to say this is a misread event.

[daj1u06 0]

phew!! though it has reported it more than once today, and it is the "working" PLA head, not the suspect PVA one.

[tomnagel 126]

We're working on the problem with your other nozzle. The problem has been reproduced internally.

[gr5 2,172]

Actually these spikes circled in red concern me the most - it implies the temperature sensor read the wrong value (as there is no way the core heat up that fast and then cooled down that fast):

 

[gr5 2,172]

Regarding the first major dip - realize that the UM3 has power management.  This means it can't have the bed heater at full power at the same time as the nozzle heaters.  I don't know how this works exactly but hopefully the print cores have higher priority and the bed has the lowest priority.  But it seems possible that the power management software gives the bed first priority and if the bed is on full power and the second core is also asking for full power it might be that the second core doesn't get the power it wants.  That would explain why it shows dark green as requesting max power yet at the same time it is cooling.  But this is just a guess.

[daj1u06 0]

For info, this is bed temp and bed heater record on the ongoing print (1 day in, 8 hours to go), it is very impressively steady for a flat plat with large SA to vol ratio. One might ask, once the first few layers are in position, why does the bed need to stay hot?

 

Unfortunately (for this scenario) Cura seems to log, rather than scroll the data so 1 day into the print it is now quite difficult to see the fluctuations in the print core temp record. Maybe something for the wish list, an option to log or scroll and save data to file.

[daj1u06 0]

Ah, refresh the browser page and it opens the log out somewhat. THese are the 3 heater records. Despite this, the print does seem to be going OK, albeit with the PVA on the LHS tidier than on the RHS.

[daj1u06 0]

Am also seeing this on the log file, can't tell if it correlates with the spikes in the AA head record:

 

Dec 21 17:35:37 ultimakersystem-ccbdd30009f7 python3[438]: 2017-12-21 17:35:37,190 WARNING materialManager No maximum temperature for material in hotend 0

 

 

[daj1u06 0]

There does seem, by eye to be somewhat of a correlation between some of the peaks in bed heater and dips in head 2 temperature (yellow lines on attached)

[Msuurmond 28]

How does your print look? Is the effect big?

[daj1u06 0]

Here's the final print before removing the supports, a fair amount of PVA "spitting" and some broken off chunks (these occured early in the print process) but overall it looks like it has worked out OK despite the temperature fluctuations. About to start a slow support removal as some of the overhangs are very delicate (bar shot hanging out of the side of the lump).

[kmanstudios 1,120]

That 'spitting' may be due to humidity in the filament boiling out tiny bubbles. Do you hear crackling and popping while printing? Also may need a good nozzle cleaning.

[daj1u06 0]

THanks but humidity issues seem unlikely, we have dried the filament at controlled 60C for several days and the printer's room climate is controlled 22C and 35-40% RH (we actively monitor it. I did a thorough nozzle clean not long ago as a precaution and I always unload and reload both heads before a print run to check the filament flow. The first few cm of PVA flow are always a bit bubbly but there after the extrusion is clear and consistent, I usually run out at least 30-50cm on both printheads and then immediately start the print run.

[Marco_TvM 34]

@Daid has found the culprit and most likely fixed it. It can be found in the TESTING release as we speak: 4.2.4.20121222.

 

 

 

[Daid 306]

I've been investigating most of yesterday and today. And I'm confident that I have a fix for the temperature fluctuations.

 

The root cause is a single bad temperature sample is throwing the temperature PID controller in a bad state. Cause the heater to go full on/off for a few seconds. (Generally off)

The bug is a combination of a bad temperature sample, the "functional range" logic and filtering on the "D state" of the PID controller.

 

The reason we are suddenly seeing this now, is because of a different bug introduced. Which causes, occasionally, the temperature of the other hotend being read for a hotend. Which greatly increased the frequency of this problem. We where already testing a fix for this wrong-hotend-temperature problem, so we where aware of that problem, but didn't notice the large impact it had on the temperature stability.

[daj1u06 0]

Many thanks to everyone in the community who took an interest in this and has contributed to resolving it. Still some support material removal to go, but the print turned out great, even with the fluctuations. I'll give the testing release a go for the next one.

[cjs 34]

Thank you UM team and especially @Daid for responding this quickly to this problem and working on a fix   

I installed the new test firmware version and observed a few things:

The first problem seems to be fixed, but there still seems to be a mixup between print core 1 and 2: 

 

Do you mind also looking into the start sequence temperature control, as to my observations the heat bed seems to be prioritized leading to huge temperature swings of the print cores at the start and eventually to an ER15.

[tomnagel 126]

Are you sure this graph is from during printing? Or is this graph from during Active levelling? Behaviour during Active leveling is different on purpose.