Beware of the Fan Control in Skeinforge Cool plugin

I noticed while doing a print a whine coming from my Mendel90. I traced the source to the operation of the advanced fan control within the Skeinforge Cool plugin.

The fan control logic attempts to control the fan speed using a technique called Pulse Width Modulation (PWM) where a variable width pulse of +12V is applied to the fan to modify its speed. Unfortunately the X-carriage fan does not like this mode of operation and it just emits a noise and does not turn.

The solution:

1) Set full fan speed: 255
2) Set Maximum Layer Time for Fan use (seconds): 60 <-- See Note 1
3) Set Full Fan Speed Bellow Layer Time (seconds): 60 <-- See Note 2

Note 1: Set this value to a sensible time for which cooling will not be required.
Note 2: Set this value to the same value set in step (2). This effectively disables the use of PWM and the fan will run at full speed.

This might account for some users seeing heat-related effects in their prints when they think that cooling was used. In fact the fan might not be running at all when cooling is required.

Regards,
Neil Darlow

---

I try to write with consideration for all nationalities. Please let me know if something is unclear.
Printing with Mendel90 from fedora 25 using Cura, FreeCAD, MeshLab, OpenSCAD, Skeinforge and Slic3r tools.

I can confirm this. Learning how to deal with ABS, I've had issues with layers pulling apart when the fan is used on full. So I've explored lowering the fan speed. With low fan speed values, the fan just whines.

I haven't fully explored the useable range of fan speeds on my Mendel90, but it seems to work fine with speeds of more than 63. It definitely does not work at 31 or below.

Hi guys,
If the 12v circuit to the work cooling fan is a pulsed supply, (Nophead may clarify this in relation to the Melzi board supplied with the kit), then I think accurate speed control of the cooling fan on the Melzi may be obtained by the simple addition of a diode and capacitor (basic soldering knowledge assumed).

I encountered the same scenario you describe above when using Slic3r cooling and also direct fan control under Repetier Host on my Mendel90, the fan cutting out below about 80% power, and basically no speed control. I'm using a different controller board but the principle is the same (provided the fan pin on the Melzi supports PWM.) There's a full post explaining how I got cooling fan speed control working perfectly on my Mendel90 here: [numbersixreprap.blogspot.ie]

Having variable speed on the cooling fan gives much better cooling control and better print quality as a result. It is excellent for that whole control to be automated via the slicing software, so you don't have to be keeping an eye on the print.

Hope the link helps.
Regards,
NumberSix

---

[numbersixreprap.blogspot.com]

I do not know how variable cooling fits in with Nophead's recent change to the cooling duct to direct air from the fan onto the, now insulated, hot-end. He may be assuming the airflow is constant.

Regards,
Neil Darlow

---

I try to write with consideration for all nationalities. Please let me know if something is unclear.
Printing with Mendel90 from fedora 25 using Cura, FreeCAD, MeshLab, OpenSCAD, Skeinforge and Slic3r tools.

Yes PWM doesn't work well on fans because they are electronically commutated motors. It just confuses the electronics. I always run them at 100% after the first layer for PLA and not at all for ABS.

An RC filter is probably better than a diode and a cap.

I think the silicone insulation is enough to stop the extruder jamming so it should be OK to reduce the fan speed.

Edited 1 time(s). Last edit at 10/26/2013 07:33PM by nophead.

---

[www.hydraraptor.blogspot.com]

Hi Neil,
Can you clarify something for me so I can understand this issue a bit better. Is there "advanced fan control" within Skeinforge or was it just that Skeinforge Cool was trying to turn the cooling fan on but the fan speed was set much lower than Max 255, perhaps in firmware?
I thought my Skeinforge was up-to-date but I don't see any settings that might vary fan speed within "Cool".

-------------------------------------------------------
> I noticed while doing a print a whine coming from
> my Mendel90. I traced the source to the operation
> of the advanced fan control within the Skeinforge
> Cool plugin.
>
> The fan control logic attempts to control the fan
> speed using a technique called Pulse Width
> Modulation (PWM) where a variable width pulse of
> +12V is applied to the fan to modify its speed.
> Unfortunately the X-carriage fan does not like
> this mode of operation and it just emits a noise
> and does not turn.

Where are you changing 1, 2 and 3 below? In Skeinforge Cool?

> The solution:
>
> 1) Set full fan speed: 255
> 2) Set Maximum Layer Time for Fan use (seconds):
> 60 <-- See Note 1
> 3) Set Full Fan Speed Bellow Layer Time (seconds):
> 60 <-- See Note 2

Tks,
NumberSix

---

[numbersixreprap.blogspot.com]

NumberSix Wrote:
-------------------------------------------------------
> Hi Neil,
> Can you clarify something for me so I can
> understand this issue a bit better. Is there
> "advanced fan control" within Skeinforge or was it
> just that Skeinforge Cool was trying to turn the
> cooling fan on but the fan speed was set much
> lower than Max 255, perhaps in firmware?
> I thought my Skeinforge was up-to-date but I don't
> see any settings that might vary fan speed within
> "Cool".

Vanilla Skeinforge does not have this functionality. Nophead's fork on GitHub has it.

> Where are you changing 1, 2 and 3 below? In
> Skeinforge Cool?
>

Yes. The enhanced Cool plugin features Fan settings where these are modified.

If you have downloaded Skeinforge yourself then you can get the cool.py (and fixed comb.py) files from Nophead's GitHub repository. Or take them from the SD card if your kit was recently purchased.

Regards,
Neil Darlow

---

I try to write with consideration for all nationalities. Please let me know if something is unclear.
Printing with Mendel90 from fedora 25 using Cura, FreeCAD, MeshLab, OpenSCAD, Skeinforge and Slic3r tools.

Hi Neil,
Thanks for the additional info. I now have the cool.py with fan related settings. I've played with it a bit, as I'm interested in cooling with varying fan speeds. It looks like it's behaviour is such that if "Turn Fan On at Beginning" is selected then an M106 S255 is inserted near the start, and no further M106 commands are used (presuming you have Max set to 255 as you suggest). If this box is unchecked then varying M106 S??? values are inserted down through the gcode file, giving fan speed variation (if your fan supports it), or causing the fan to stall as you experienced. Of course, your opening recommendations will have the same desired effect of not varying the fan speed, when the fan is activated.


(Values shown above are experimental)
Thanks again for the info.
Regards,
NumberSix

I think that this thread might benefit from a link back to where the cool plugin is discussed.
DeuxVis writes here: [forums.reprap.org]

The following is an extract ...

"Trying to explain priorities :

1) if "Turn Fan On at Beginning", well then fan is always on.
2) if "Force Fan on Bridge Layers" and layer is detected as a bridge, fan is activated.
3) after that only, "No Fan use before Layer" is checked. Yup it means even second layer will get some air if it's a bridge.
4) then the max and full speed duration of layer settings apply to calculate fan speed (and the time used here is the duration of the layer before cool slowdown is applied)."

Alan

Thanks Alan! Too many threads... to little time! . Bedtime reading!

With the settings I use the fan is just on 100% after the first layer. That is done by setting Full Fan Speed to 255 and having Max Layer Time For Fan Use the same as Full Fan Speed Below Layer Time. If those two values are different it ramps the fan speed from 0 to the max between them. It would be easy to add a min fan speed to stop it stalling.

---

[www.hydraraptor.blogspot.com]

The fan seems to reliably start at 25% PWM so setting that as a minimum should stop it stalling. Above that it does whine but PWM does work. It will also go less than that if it is already spinning but I don't think the extra complication of a higher starting speed is worth the effort because I can't see the point of running it so slowly. In fact I am not sure why full power after the first layer is not the best.

---

[www.hydraraptor.blogspot.com]