I want to add heated fan control G-Codes and controller functions to the firmware to allow me to create Heated Part Cooling Fans. It may seem strange to have heated part cooling but there is a method to this madness.

When printing ABS lack of part cooling creates poor print quality because a new layer is extruded on top of still hot and melted layers. Setting part cooling fans to a low setting such as 20% helps but for best results on some parts with overhangs 40% or 50% is needed.

However, at this setting layer adhesion is affected due to cooling to a too low temperature and warping for the same reason. I think temperature controlled blowers will allow rapid cooling to a temperature low enough to solidify the layers but above the glass temperature of ABS (~105C). The goal is to cool the extruded layer to 110-120C as fast as possible but not to risk cooling it below this temperature. Having temperature controlled blowers set for 100-110C at high speed should do the trick.

I think that heated blowers will also help keep the whole part above the glass temperature reducing or preventing warping and serve double duty as compartment heaters.

I tried using a temperature controlled plastic welding gun as a proof of concept parts blower and got spectacular results in ABS, as good or better than I was getting with PLA. Excellent layer adhesion, no deformation due to extrusion on top of melted layers on corners and overhangs.

I now want to add heaters to my part cooling fans but need to make these heaters work in conjunction with the fan setting to prevent over temperature and meltdown when the fan speed is changed.

Here is what I have in mind:

1. G-code to define that a heater is fan driven, effectively linking the heater and fan controls for new type of temperature controller.

More than one heater can be linked to a single fan. Two fans controlled by a single PWM will require two separate heaters. Each heater will have to be controlled separately to compensate for unbalanced heating elements and detect fan, heater or temp sensor failure.

2. A G-code setting to pause until the heated blower reaches the new temperature (or a new fan setting). This will allow delaying operation until proper operating point is reached.

3. Changing the heater temperature or fan setting will be handled by the controller in a single operation :

• Minimum fan setting before heater is allowed to turn on (probably 20%)

• Fan is turned off only after the heater is turned off, probably after a short delay to allow heater to cool down a bit.

• When the heater temp is set, fan setting changes cause heater PWM can to be adjusted for new fan settings right away and then let the PID (corresponding to the new fan setting) tune it the rest of the way.

4. Calibration procedure to determine heater parameters for fan settings from 10% or 20% to 100% in 10% increments.

I have setup the development environment but could use any pointers on what would be the best way to proceed with the implementation from someone comfortable with the RepRap firmware architecture.

Any suggestions on which G-codes should be used would also be appreciated.

It would be good to have an end-stop test to detect if an end-stop sensor locked in the "triggered" position.

I have had this happen twice on my TEVO little monster and it is not a pleasant experience. I have since changed the homing command after a print to a move to the top position but this is a mitigation of possible damage but not a solution.

The minimum nice to have would be a g-code to test all end-stops and if any of these are triggered then a sensor error is triggered, all axes are set to un-homed and no movement allowed until reset.

This would prevent the printer from homing if any of the sensors are triggered at start of operation but in normal operation this never happens and to recover the user would need to move the effector down a few mm manually. It is a small price to pay for additional safety.

I have experienced a sensor failure when a bad contact developed on the cable from the sensor to the DuetWifi. Both times this happened after a successful print and was nerve wracking because it was unexpected and threatened to damage the printer.

The homing of the printer resulted in one of the carriages to remain motionless (it's sensor was already triggered) while the other carriages kept moving up. The effector first pushed out to one side then rose above the carriage, finally everything locked and finished with the pulleys skipping over the timing belts.

Time between these failures was over a month of no issue printing so it appears to have developed in the normal printer operation, probably due to vibration.

My problem of unresolved references was solved by resetting .cproject and .settings/language.settings.xml in RepRapFirmware project to repository version and then changing the project properties as per instructions. Subsequent build was successful.

If it is of any help, I was getting the exact same error because I made the mistake of having the projects not under the workspace directory.

When I created the workspace in the parent directory of the project sources (CoreNG and RepRapFirmware directories) the error went away as expected. The ${workspace_loc:} is expecting them to be under the workspace directory.

However, I am getting undefined reference errors during linking for: lprintf, sqrtf, floorf, etc.

I am running under OS X with gcc-arm-none-eabi-6-2017-q2-update tools. Is there a library missing from the linker options that the IDE should include based on dependencies?

Just got my TEVO Little Monster printing with Duet Wifi with the LCD interface, and the WEB UI. Wifi seems to loose connection all the time. Sending [c]M562 S0[/c] to disable it then [c]M562 S1[/c] to enable it on the LCD console, reconnects every time. A bit of a pain but I will deal with it later.

The upgrade was not complicated except for a couple of loose pins in connectors that only showed up after a few days of working with the printer.

In the end the real cause of my problems was poor temperature control of the MKS-SBASE controller in the machine, which resulted in the PTFE in the hot-end melting and lining the nozzle (two actually because I switched to a new nozzle hoping to make it work) with thin ribbons of PTFE that would peel off at random and cover the 0.4 mm hole, causings extrusion to become very uneven, sluggish or blocked.

Nothing worked consistently, not the Titan extruder, not the e3d hot end. Only when I cleared out all plastic out of the nozzles by heating them with a hot air gun and cleaning them out with a 1.94 mm drill bit, slowly by hand turning it. Then soaking them in MEK to pull all the crud out. Only then did I realize what the issue was when I saw thin strands of plastic floating around that MEK did not affect.

Now the printer is running and better than before. Finally, after a four week struggle trying to figure out what the issue was.

@dc42:

It's usually possible to convert 5V optical endstops to 3.3V. Identify the resistor in series with the IR emitter in the slotted opto switch and reduce its value by about 50%. If it's a surface mount resistor and you don't have SMD desoldering equipment then it's easiest to solder a second resistor of equal value on top of the existing one.

Does the Duet Wifi not have the ability to use the Stall detection of the stepper drivers for end stop detection?

I know this is possible according to Trinamic information and also by the Prusa MK3 implementing this exact feature.

This is definitely something I would like to use and if not available to work on the firmware enhancement to make it work.

I hate relying on the optical end stops. They are not as reliable as the trinamic stall detection because the latter will work if an obstruction stops the carriage from triggering the optical sensor.

@Emulsifide:

I've now concluded that I seem to have been provided with a defective MKS SBASE controller board.

The MKS-SBASE controller is unreliable garbage. Mine worked for about 6 weeks.

Thank you for publishing your success. I was debating whether I should change to the Duet Wifi from the stock MKS-SBASE disaster of my TEVO. Seeing your success and the fact that Duet comes with Trinamic drivers tipped the scales.

I made a board with TMC2130 step sticks for my TEVO and these drivers made a huge improvement in printing quality and printing is nearly silent. I could not go back to the whiny sound of standard stepper drivers.

I agree with you on all the shortcomings of the software and controller combination. There are also absolutely no fail checks or safety limits for anything on this printer.

I bought my TEVO Little Monster thinking that for the price I was getting a 3D printer instead of a good mechanical frame, stepper motors, power supply, some cheap cloned parts for the extruder and hot-end. Oh, lets not forget a small chance to have all these come together as a working 3D printer.

I had small issues:

1. The BL-Touch failing and causing the printer to drive the hot nozzle into the bed, melting the liner. I wound up checking its status every print, just in case it failed.

2. One of the end stop sensors getting blocked by a loose filament, causing homing operation to think this axis is already homed and running the other two at 400 mm/s until they locked because the other carriage was low. The printer wasn't the wiser and kept grinding the belts until I turned off power.

3. No supports for the umbilical and filament coming to the effector meant that the filament got caught by one carriage during homing and could not reach home position. The printer kept going, grinding the belt.

I got it set up and running decently. Added printed mods to support umbilical and filament, added shrouds to protect the end stop sensors. However, I kept having intermittent failures with the printer just stopping once in a while but I was learning to live with it because it was very rare.

Then two and a half weeks ago (six weeks after I got the printer) I got a "filament jam". Since then I haven't been able to get it to extrude reliably and all attempts to fix the problem only seemed to make the situation worse. I haven't had the printer working since then and spend over $200 trying.

I wound up replacing all the cloned parts with genuine e3d parts, hoping it will resolve my issues and wanting to eliminate variables. As I kept diving deeper into the workings of the extruder and hot-end I realized that it was a miracle I managed to print at all with such poor quality components. I can vouch for the much higher quality of e3d components but my real issue turned out to be a slowly failing controller board reading higher temp than actual.

If it all went at once I would clue in faster but it was a slow degradation process. The hot-end thermistor channel must have gone first and caused the original "filament jam" by lowering the actual temperature at the nozzle. Finally, the bed channel also failed. Not realizing the two issues were related, I re-configured the bed and extruder to use the spare channels TH3 & TH4. It seemed to work but the actual temperature was 30 C below what the controller was showing.

This morning, just when I thought I could print again, the two remaining channels started reporting 93 C at room temperature. This is when I finally clued in that my problem from the beginning was an erroneous temperature reading by the controller and the rest of the failures were symptoms.

I am glad I now have all genuine e3d parts. I did not want to blindly replace them without doing some testing to assure myself that the change to genuine parts makes a difference. My trust in 3D parts and components has gone to zero unless I verify them myself. I used a Fluke multimeter thermocouple probe which I can trust and lowered it into the filament path. Stripping the outer teflon insulation, which keeps the red/yellow wires together, reduces the diameter so the probe fits where 1.75mm filament does. By controlling the depth of insertion I could measure temperature all along the filament path, right down into the nozzle.

I ran combinations of TEVO stock components (with the TEVO fan and fan mount) against the e3D parts (e3d fan shroud, quiet 30mm fan I bought at DigiKey) and found that with TEVO stock parts and configuration the heat-break temperature just inside the heat sink is 48 C with hot-end at 190 C. With e3D components and the quiet replacement fan it is 39 C for the same nozzle temperature.

I will place an order for a Duet Wifi and hope to have the printer working again over the weekend. I have a backlog of a few projects that are waiting on printed parts. I never expected my printer would be out of commission for almost a month after less than two months of use.

---

Vladimir.