Custom heater block to fit E3D nozzle on Ultimaker 2 ("The Olsson Block")

[swordriff 119]

Hi twistx!

I think your sensor si fine.

Here is your problem: Due to the 3dSolex "Olsson Block"s higher mass than

the standard, the heating is a bit slower. When you lower the temp and the

heater comes on again at the target, you get a low temp "overshoot" and it takes

too much time for it to get up to your new target temp.

This is a safety feature kicking in, prenventing you from burning down the house

in case the heater spins out of control.

These rules are more relaxed in the newest firmware; http://software.ultimaker.com/current/Cura_15.02.1.exe

Download, install, connect your printer via USB and upoad the new firmware.

This should take care of it!

[twistx 0]

I tried that and there doesn't seem to be anything visually wrong with the heater or the sensor itself.

With the printer idle at 210C I issued these commands twice with one and four minute intervals to slowly ramp up the speed. It bails out after about a minute M106 S200 is issued.

>>> M107

SENDING:M107

>>> M106 S50

SENDING:M106 S50

>>> M106 S100

SENDING:M106 S100

>>> M106 S150

SENDING:M106 S150

>>> M106 S200

SENDING:M106 S200

Error:Printer stopped due to errors. Fix the error and use M999 to restart. (Temperature is reset. Set it after restarting)

 

I do have firmware 15.02.1 installed.

 

[swordriff 119]

Did you check heater connections on the motheboard?

Else.. I think Ill have to leave this one for gr5.

BTW; do you never sleep?

--haha

 

[twistx 0]

Nope. I'll have to check those connections.

What is this sleep you speak of?

 

[reibuehl 42]

I managed to get my heat sensor out using WD40 and heating the block up to 100°C. Now I just need to wait for everything to cool down, clean off the WD40 and then I can mount my shiny new "Olsson Block"! :cool:

 

[gr5 2,178]

>>> M106 S150

SENDING:M106 S150

>>> M106 S200

SENDING:M106 S200

Error:Printer stopped due to errors. Fix the error and use M999 to restart. (Temperature is reset. Set it after restarting)

 

Are you monitoring the temperature during this time? Is the temperature dipping much?

I might have to look up everything in Marlin that causes "heater error". This is what I know so far but I don't think it's the same error because this is when heating it up I think and the temperature must change by more than X degrees C in Y seconds.

14.09                - does not have the feature

14.12   oct 16, 2014 - feature introduced nozzle must change by at least 20C in 20 Seconds

14.12.1 dec 15, 2014 - from 20C to 10C (still in 20 seconds)

15.01   jan 14, 2015 - from 20 secs to 30 secs (now 10C in 30 seconds)

- George

 

[gr5 2,178]

Okay - I just looked at the code carefully. There's only one small bit of code that generates the "Heater Error" message. It generates this message if the PID controller is outputing 100% power to the heater for more than X seconds and the temperature hasn't increased by at least Y.

Once you are up to the goal temperature, the "Y" factor isn't important. The point is the PID controller put out 100% power to the nozzle because the temp was well below the goal temp and it tried to do this for more than X seconds (30 seconds in 15.01 as shown in table in my above post).

This means the fan is blowing on the nozzle or heater block too much. A few simple fixes come to mind:

1) Try to aim the fan differently.

2) Play with the M106 S amounts but leave them for at least 40 seconds before trying the next speed. It takes 30 seconds of temperature low (minimum - could take longer for I term to max out). In fact once you find the max fan speed that passes, repeat the test but leave it at a particular speed for at least 90 seconds. It would be nice to put an oscilloscope or volt meter on the heater power circuit to see when the heater is at 100% versus switching in PWM mode to know how close you are coming to being at full power for 30 seconds.

3) Put some insulation around the heater block. Maybe as simple as some polyimide tape (kapton tape). Or put some kind of air-barrier around the nozzle to keep the fan away.

I really don't think the larger heater block is the issue - unless it has significantly more surface area The fact that it is *clean* might be more important as having a dirty surface adds some insulation (as would some polyimide tape) or preferably a better insulator that can withstand 250-300C.

Or simply having the nozzle stick down a little more into the air flow might be part of the problem

 

[gr5 2,178]

There's a great explanation for PID controllers on wikipedia:

http://en.wikipedia.org/wiki/PID_controller

Quick explanation: "error" is goal_temp - current_temp. P is multiplied by error to get power input to nozzle heater. That is P controller. So if you are right at the correct temp you put out 0 watts. If you are too hot you put out negative watts. The I term is the Integral of the error term. Multiply *I* by the integration of the error - the cumulation of the error (just add all the error up from all history). This is what finds the equilibrium point - this is what is causing the failure:

If the temp is 219C and goal temp is 220C and this lasts for a long time the integral_error*i_term gets larger and larger until equilibrium is found but if it stays too long at 219 this value never shrinks until the nozzle PWM is set to 100%.

The D term is multiplied by the differential of the error and is used to reduce overshoot and oscillations. It is looking at the "velocity" of temperature change and warns you that you are going to overshoot if you don't "pull back". These 3 values P,I,D are added together to get total power sent to nozzle.

 

[gr5 2,178]

Why did it take me so long to notice????? Versions of Cura that came out in 2013 start with 13.X

2014 14.X and

2015 15.X (and so on)!!

Wow - the month also. Version 14.12 came out december 2014.

 

[twistx 0]

Thanks gr5, I'll mess around with this some more. The temp drops from 210C sometimes by 5-10C almost immediately. I'll try some new PID values, or just modify marlin to extend that timer.

 

[pm_dude 27]

Hello,

 

I finially got some time to install the new heater block and its running like a charm. Did my first print with it and it's looking good!

 

 

Now this is not for self promotion but since my info are already available in the code I figured I it would not change much to leave it there

 

[gr5 2,178]

Print that upside-down on the glass, then flip glass over. Now it will be in every photo of things sitting on the print bed!

 

[pm_dude 27]

Haha nice idea.

I also did a test with FlexPLA with the 0.6mm nozzle

Result:

Ultimaker bracelet printed in a little more than 3 min. 0.2mm layers, 1.2mm walls, 90mm/s. So is that 21.6mm^3/s ?

Please ignore the horrible sound. My hand was placed on the mic and it made some weird distortion.

 

[swordriff 119]

Bonjour, pm_dude!

So cool that you are having fun with "The Olsson Block"!

I am not sure about the calculation, I leave it to you and gr5 (who is one of the 3 authorities in here on that matter),

But looks really interesting!

 

[ultiarjan 1,223]

So is that 21.6mm^3/s ?

I would say it is ... but obviously it's MUCH easier for the feeder to do this with a 0.6 nozzle than with a 0.4 nozzle. So basically you can only compare achievements between same nozzle size :-P

 

[pm_dude 27]

Yes of course but the heater block seems to perform really well and maintain the temp A1.

[IRobertI 520]

Wait... you sliced it with 1.2mm walls (with spiralize, so all in one go) @90mm/s and then increased speed by 300% and flow by 200%? That's a hell of a lot more than 21mm3/s... But then I'm not sure what your minimum layer time was set to so it probably didn't reach that speed.

 

[pm_dude 27]

No no. It was sliced at 30mm/s and 0.6 mm walls

[gixxer 1]

Why the 200% flow? Where you getting underextrusion?

Assuming the increase in flow had no slip in the feeder this would indeed be 21.6mm3/s.

[ultiarjan 1,223]

If you increase flow while printing spiralize you make the walls thicker..

[gixxer 1]

Ah.. I've not tried that before!

[pm_dude 27]

The only thing I dont understand is the hesitation of the head around letters. You can clearly see it in the video. Its causing the echo effect BEFORE the letter gets printed. You can also see it when doing the backside sometimes.

Is it just to much information to manage for the firmware?

The slower versions suffer less of that behavior.

[gr5 2,178]

Not sure - you'd have to post a picture. Probably some overextrusion as it is slowing down just before the letters. It has to decelerate a bit as it approaches those letters.

 

[swordriff 119]

The web shop is "barely" up and running!

 

Thank you for your patience.

If you have ordered parts AND paid, you do not have to do anything.

If you have ordered ( you are waiting ) and have NOT paid, I hope you

will place it through the store.

I am quite new at making web pages, so I hope it works!

http://www.3dsolex.com

 

[chrisr 18]

http://www.3dsolex.com

I like the new website! :-)