print head thermal stability (or not!)

[Daid 306]

Quote

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

That's to be expected. We fixed the wrong temperature from effecting the control loop, but properly fixing the bad readings was more complex, so we decided to push out a simpler quick fix first while to better validate the full solution.

 

The 2nd graph looks all like bed leveling. During leveling the heaters are off, as they are negatively effecting the bed leveling sensing.

[cjs 34]

@tomnagel @Daid Correct, thats bed leveling. Thinking about it,the behaviour makes sense. I believe I have seen the same graph not related to bed leveling, but I wouldn't bed my UM3 on it ;D 

 

@Daid Any kind of timeframe for the full solution and the release of a new firmware? I ask because the update from 3.7 to 4.0 took quite a long time ( 6 month ?! ) 

 

How stable/safe is the test version? 

 

May I ask you to look at the temperature graph below? I marked the ER15 spot. It happens on 3 UM3 running the firmware version 4. The printer is configured to print PLA and Breakaway. The fourth printer, which runs firmware 3.7 doesn't have the problem. It happens on print slot 2. It's not the fault of the print core as swapping in a new one or swapping print core 1 and 2 doesn't solve the problem. I measured all my print cores and they are all about 28 Ohms, which to other information published is out of spec. 3+ bad print cores seems to be very unlikely as I think Ultimaker has good QC.

 

I would also like to mention that I don't like the programmed error behaviour of the UM3. 

Behaviour: Everything stops and the UM3 can only be recovered by a power cycle and the print is basically fused to the nozzle because the bed didn't go down a few millimeters.  

 

Do you want me to open a new topic for the error as it may not be related to the problem? 

   

Edited December 28, 2017 by cjs

[Daid 306]

Testing firmware should be quite stable, every printer in our office runs the testing version. Release shouldn't be far off. The 3.7 to 4.0 took long due to focus on other projects, and integration of Cura Connect.

 

ER15 is almost certain related to the print head cable. Due to the cables not properly clamping in the top of the print head, cause communication to fail with the printhead. A quick fix is opening the top of the print head and putting some tape around the cable where it is clamped. So it's properly clamped, and no longer pulls on the connector itself.

It gets worse over time due to the cable compressing on the only part where it is clamped.

As you can imagine, this is quite a big problem for us, and we only just finished our investigation into this problem. Which I took part in, which is why I know so much about the cause. I don't know how support is handling the problem right now, due to that being out of my usual work area.

 

We're also improving the cables, which where not exactly in specs, and the clamping, which could have been designed a bit better. To fully solve the ER15 issue. But that's all mechanical, in software we already did all we could.

[IRobertI 520]

1 hour ago, Daid said:

I don't know how support is handling the problem right now, due to that being out of my usual work area.

I can't speak for everyone but I've been recommending pushing down the cable a bit so that the individual conductors gets a bit of a curve to them so that the can move around a little bit more freely. I also recommend, like you said, to put some tape around the cable to give the strain relief a bit more to bite into. I've been doing that for a couple of months now and it has fixed the problem 99% of the time. I had a customer call me the other week who reported he had had no further issues since he made that "fix" a couple of weeks prior.

[elinovia 1]

Hi,

 

I just wanted to let you know that I also have this problem described earlier, that the temperature drops 10-20 degrees while printing using core 2. I've been in contact with the reseller that advised me to check this topic. I've installed the test firmware and as far as I can tell, the problem is gone, or at lest not affecting the print. I still have drops in temperature on the chart, but they must be faulty readings (they go from 200 to 30 in less then a second). And it was at those points the print started to look bad in the early firmware.

 

What I wanted to add to the discussion was that for me, every time there is a faulty drop in temperature, one of the fans looses speed. So I can hear when the drop occurs. This was also the fact in the previous firmware, but every time I heard a drop of RPM of the fan I knew the print would be affected. This is no longer the case. But what can be the cause of the drop in fan RPM?

 

 

Here is a screenshot of the print with the old firmware

[Marco_TvM 34]

On 12/29/2017 at 6:56 PM, elinovia said:

What I wanted to add to the discussion was that for me, every time there is a faulty drop in temperature, one of the fans looses speed. So I can hear when the drop occurs. This was also the fact in the previous firmware, but every time I heard a drop of RPM of the fan I knew the print would be affected. This is no longer the case. But what can be the cause of the drop in fan RPM?

 

That's an interesting thing to note. I'm not aware that this happens in the cases we've seen and examined so far.

If it would happen again (with the fixed testing firmware) I'd love to get some logfiles to analyze!

[elinovia 1]

7 minutes ago, Marco_TvM said:

 

That's an interesting thing to note. I'm not aware that this happens in the cases we've seen and examined so far.

If it would happen again (with the fixed testing firmware) I'd love to get some logfiles to analyze!

The RPM-drop still occurs at the same frequency in the test firmware, and all the times it happens, there is a faulty reading. I can make a testprint with nozzle 2 later today and send you logfiles and graphs if you like. I'm at work for 9 more hours, but after that I can prepare it.

[Marco_TvM 34]

That would be great... Just use something like wetransfer and post the link here.

[Daid 306]

Could be that the whole 5V power is dipping on a slight lose connection, which is causing the communication problems and the fan speed drop. Or could be that a communication error erroneously talks to the fan driver instead of the temperature sensor, momentarily adjusting the fan speed. In both cases, I would invest a tiny bit of time to seat the print head able better and putting some tape around it, as IRobertI also noted, this solves a whole bunch of issues.

 

We've just pushed the fix to stable. (Together with language selection)

[elinovia 1]

3 minutes ago, Daid said:

Could be that the whole 5V power is dipping on a slight lose connection, which is causing the communication problems and the fan speed drop. Or could be that a communication error erroneously talks to the fan driver instead of the temperature sensor, momentarily adjusting the fan speed. In both cases, I would invest a tiny bit of time to seat the print head able better and putting some tape around it, as IRobertI also noted, this solves a whole bunch of issues.

 

We've just pushed the fix to stable. (Together with language selection)

I've already fixed the cable with some tape, and changed cable, and changed the PCB connected to the cable. :-) But I can double check later this evening.

[elinovia 1]

I can confirm that the new firmware seems to work. The fan-RMP still drops and in the graph I notice that the Heater1 has some disturbance when this happens. But nothing that affects the print anymore. 

 

Are you still interested in logs and graphs? I will be able to supply this tomorrow evening.

[Marco_TvM 34]

Yes please. As @Daid mentioned, it would be interesting to figure out the source.