Idea for filament management implementation

Hello RepRappers!

I was wondering if there is any idea out there for a filament management system? If not, I have had a thought on one.

The idea goes like this.

There would be two rubber pinch rollers that are in place before the filament enters the extruder, which would be connected to a monitoring system calibrated for the amount of filament length for a given degree of rotation. You would feed your filament through these pinch rollers, and hen if the slicer sent a command for your filament to go 10mm, the monitoring system would know if that condition had been met or not. It could work for both extruding and retracting.

This would be helpful for many reasons.

1. You could enter your spool weight (I use mostly 300g spools, but most use 1kg). There are many calculators out there to convert filament length to weight. Then, there could be custom code in the slicer to ask what size spool you are using, and then subtract the filament amount intelligently based on how much filament had been put through the monitoring system. Then, after a build is done, you wouldn't need to guess if you have enough filament left on the spool. Then, if you changed spools, you could simply reset the counter on the system to whatever size spool you are using. This would eliminate having a bunch of half-used spools laying around like I do.

2. Filament grinding! This is a problem no matter what printer you use. In this setup, when the filament stopped moving the specified amount through the monitoring system, it would halt the print and turn off the extruder heater, and warn you that there may be an issue, and that the printer requires your attention.

3. Skipped steps on extruder. If the monitoring system detects that less filament is moving than is required, it would also allow you to investigate if your extruder was skipping steps.

4. The most obvious use is that it would stop the print if you were out of filament, which I think would be extremely helpful, because you could simply load a new spool, and then continue printing!

Anyways, let me know what you guys think of this, or if a system like this has already been implemented.

That's a great idea. I think it can be done but on some of the more powerful electronics, like a smoothieboard . Ramps is kind of overburdened as is.

I like it.

I think you will filament spool tracking in some forks of RepRapFirmware before long, along with additions to the web interface to report on it.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

Another thought,

There could be 2 sets of pinch wheels, which would allow for an automatic new spool command. When the filament is pulled out, there is a short time where one set is moving, but the other set isn't. After this condition is met, the firmware would notify you "filament reset condition has been met, did you load a new spool?" Yes/no you click yes, it asks you the filament weight. This eliminates the need of manually pressing a filament reset button.

If the extruder is reliable, you would need just one set of pinch rollers, at the input side of the extruder. Then the firmware could correlate the pinch roller movement with the extruder stepper movement.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

I like this. If anyone starts working on it and needs help let me know, I'd be happy to contribute.

Hello everybody.

I'm a newbie here. I just joined to add my idea to the existing ones. I've seen other, similar ideas as the one posted above, on other forums.

My idea to track filament use requires no additional wheels or complicated mechanics.
Just use the optical sensor as can be found in any optical PC mouse. They come with a resolution of 1000dpi (0.025mm) and cost less than $10 or €10. They could, possibly, be used directly on the filament itself, but concerns exist regarding transparent filament, and varying colors. So just point the sensor at the idler wheel (or pressure wheel, whatever you call it). The idler wheel has a consistent color and surface and should allow a reliable signal. Should issues exist with the reflectivity, just sand it to make it tad rougher.
When first calibrating/setting up the printer, all the user needs to do is (with an empty hot end) send a command via the host to extrude 5, 10 or 20mm of filament. The host then evaluates the optical sensor's returned signal, and voila!, the system is calibrated. I'm pretty confident the received signal would be accurate to within +/- 10%, easily. That should be enough to discover the onset of an extruder jam and other issues. If the accuracy is even better, a possible feedback loop could be used to regulate extruder temperature. (What I'm referring to now is: higher pressure in the extruder leads to higher slip at the drive wheel - higher temperature lowers the pressure, reducing slip ....)

The signal the optical sensor delivers would need to be processed by the host software, such as Repetier Host, Pronterface, Printrun, etc. The signal is transmitted to the host software, just as if it were still a mouse, either by cable, or wirelessly, through an available USB port. The host software knows which GCodes have been sent (and how much the extruder should have extruded) and compares the returned signal to the expected filament movement (there will be a time-delay involved, due to the buffered commands). But fairly accurate tracking of movement should be possible.

A down-side is that that system would not work when printing via SD card. It would need to be implemented in the printer's firmware (probably insufficient memory for that), or would require a dedicated micro-PC (Raspberry, etc.) to perform the expected-versus-actual comparison, with subsequent sending of a signal to the printer's motherboard/firmware.

I guess one potential for failure of the system could be when the filament runs out, and the idler wheel contacts the hobbed bolt (drive wheel) and continues to be driven, the system would think everything is fine. But that situation can be avoided easily, via mechanical methods, by eliminating the possibility of the idler wheel contacting the drive wheel.

What do you think? Should I contact some of the Host folks (Repetier, etc.) to have them start thinking about the ways of implementing this?

mjh11

That's an excellent idea. I think it would almost be easiest to gut an actual mouse, seeing as you can get them for almost (or actually) nothing in most cases. You need some sort of interface electronics because an arduino can't just hook right up to the little sensor itself (well probably)

I have an old mouse i can gut tonight to see if i can get a consistent reading.

Does ramps have two analog pins avaliable? If so, someone with better marlin knowledge than I should see if they can get OptiMouse arduino library simplified to a single axis and slipped into marlin.

The idea of using an optical mouse sensor has existed for a while but has never gained traction. Reprap even hosts a few datasheets ( [reprap.org] ).

One problem is availability. There is no point developing this feature if people need to buy and dissect a specific model of optical mouse to build one. There are also some questions regarding performance, especially for absolute positioning.

An encoder based design is already available here: [tunell.us]

I personally do not think these projects are interesting until they become precise and accurate enough to control actual extrusion volume. Using a sensor to compensate for hysteresis at the end of a bowden system could improve performance and allow printing with flexible materials. Optical flow sensors will never be suitable for this application because their sensitivity can vary with material and distance.