Hey everyone!  This is where we are going to post information and news about the open source filament extruder that Russ and I have been working on.  There seems to still be a lot of interest in this project but we are going to need some help from you guys to keep it going.

The problem we are having now is with the calipers.  The ones from Harbor Freight reset the measurements to zero after 5-10 minutes.  Others will power off after 5 minutes.  You can see a brief video about it here.

https://youtu.be/i6OT4e8qpBQ

Does anyone know of a different sensor that is readily available?  I have seen some use an optical sensor but I can't find much information on the subject.

What kind of signal does the caliper itself output?  I'm guessing one pulse per tick mark of whatever encoder they use.  If that's the case, it should be fairly simple to just connect a little Arduino (or whatever micro-controller is your preference) to the encoder, then you can do anything you want with that data.

Quote from MakingStuff on June 22nd, 2017, 08:33 PM

Does anyone know of a different sensor that is readily available?  I have seen some use an optical sensor but I can't find much information on the subject.

Hi Bob,

There was a mention from a commenter of a readily available optical scanner sensor on one of Russ's videos that could be used to measure the diameter of the extruded filament. Unfortunately, I didn't have it bookmarked and that's all I remember.

Russ shrugged it off and favored the caliper.
:hide:

It linked to a Thing on Thingiverse.

Quote from haxar on June 22nd, 2017, 11:40 PM

Hi Bob,

There was a mention from a commenter of a readily available optical scanner sensor on one of Russ's videos that could be used to measure the diameter of the extruded filament. Unfortunately, I didn't have it bookmarked and that's all I remember.

Russ shrugged it off and favored the caliper.
:hide:

It linked to a Thing on Thingiverse.

The one I saw was on thingiverse and they had done a limited run on the boards.  Of course I was too late and the boards were all gone.  I'll have to look and see if I can find it again and put the link here.  Maybe someone else has made some.

Quote from Matt Watts on June 22nd, 2017, 11:09 PM

What kind of signal does the caliper itself output?  I'm guessing one pulse per tick mark of whatever encoder they use.  If that's the case, it should be fairly simple to just connect a little Arduino (or whatever micro-controller is your preference) to the encoder, then you can do anything you want with that data.

That's exactly what we are doing.  There is a stream of data that comes off the calipers and we are decoding it with an Arduino type board.  I made a video on it if you want to see exactly how it works.

It would have to be calibrated obviously, but something like this looks to me to be just the ticket:

https://www.youtube.com/watch?v=qWby7bKitsg

All you need is a cheap camera, a light source and an algorithm to compute the shadow area across one line of pixels.  At least it sounds simple in concept.

The optical version was made by Filip Mulier called the 'Filament Width Sensor Prototype' currently version 3, Thingyverse item # 454584...

It measures the filament by shining a light from behind, and measuring the shadow cast on a 1pixel wide x 128pixel long sensor. See https://youtu.be/5JmroyGb4qY?t=126 for a inline use with an extruder.

Quote from MakingStuff on June 22nd, 2017, 08:33 PM

Does anyone know of a different sensor that is readily available?  I have seen some use an optical sensor but I can't find much information on the subject.

http://objectswithintelligence.weebly.com/store.html

Currently $57, and claims to be 'available'

so a friend of  mine gave me 2 sensors similar to that one. i need to see if its the same or not.

he was having problems with the controller. this one has an external controller.

at this point i have a feeling that we can find something some where but i haven't looked.

57$ is prob about right by the rime you get all the parts and PCB...

should we try it? what i want to know is what is the tolerance?  The reason i went for the caliper is the fact that i had it working find on my other system . ( less than 10$ on sale from harbor freight)

but i wasn't reading the data from it. i think we some how are having some feed back through the outputs. so we can try to isolate it and see what happens...


1. put it in manual mode and disconnected the caliper Data lines but still use power...

and
2. a test with no connections as all just battery power.

~Russ

the weebly site, has the tolerance of the optical sensor as 'Accurate to +- .02 mm'... you have 128 pixels on the actual sensor part, and its , runnint perpendicular to the filament, so depending on the length of the part... its claiming 400dpi... which i work out to be about 2.5 thou resolution per pixel.  this is a shade over 0.06mm, so unless the board is doing some trickery in code??

Personally, I would look at running the sensor at an angle, to maximize the use of the pixels available, and increase the resolution...

Build Of Materials = https://cdn.thingiverse.com/assets/46/8f/95/73/9e/PC_Board_BOM.pdf

Good stuff. Thanks for the input.

~Russ

Hello Bob, Russ
This is just a follow up on regards of the comment I posted at https://www.youtube.com/watch?v=i6OT4e8qpBQ



After testing for one whole night, I can now say that the Neiko Caliper I'm using doesn't exhibit the problem mentioned here. When I leave my caliper running, the lcd screen goes off after about 2 minutes of not receiving any variations/changes in the readings, however, the caliper keeps sending data through through the data port. As soon as any variation is detected, the LCD screen comes on again.

This is the caliper I'm using: https://www.amazon.com/gp/product/B000GSLKIW/ref=oh_aui_search_detailpage?ie=UTF8&psc=1

Quote from xXxOlivierxXx on October 26th, 2017, 09:43 PM

Hello Bob, Russ
This is just a follow up on regards of the comment I posted at https://www.youtube.com/watch?v=i6OT4e8qpBQ

Quote from xXxOlivierxXx on October 26th, 2017, 09:43 PM

After testing for one whole night, I can now say that the Neiko Caliper I'm using doesn't exhibit the problem mentioned here. When I leave my caliper running, the lcd screen goes off after about 2 minutes of not receiving any variations/changes in the readings, however, the caliper keeps sending data through through the data port. As soon as any variation is detected, the LCD screen comes on again.

This is the caliper I'm using: https://www.amazon.com/gp/product/B000GSLKIW/ref=oh_aui_search_detailpage?ie=UTF8&psc=1

Fantastic! 

The ones we have seem to work. However. When Conected to our control  bord it resets randomly.

This is problem #2

I may get this pair and test it. Or Bob might.
One uf us needs to. ;)

Thanks for taking the time to respond!!!   

~Russ

My pleasure.
Just fyi, if you or bob get this caliper, you will need to find out the value of the cycleTime variable (sorry, I don't have an oscilloscope). The code I'm using at the moment and that works fine for me is: https://sites.google.com/site/marthalprojects/home/arduino/arduino-reads-digital-caliper

I realise this may be a dead topic and I am unsure if it was resolved but a reliable filament extruder is the main reason for joining this forum. A few stupid ideas and questions. Would there be anything wrong with using a cheap potentiometer as a voltage devider 0 to 5v on an arm running on the filament. This would give a proportional reading. A rotary encoder may be more accurate if you had a high PP rev but unless you used an absolute encoder would need to be calibrated every time it was switched off. A photo diode array could be used in a non contact way and this is what is being used in the video people have been referring to. I do worry whether translucent filament may be hard to measure this way. An analogue hall effect sensor could be used in a similar way to a pot or encoder. I hope this project has not been shelved indefinitely and I hope that we can push forward with this. Thankyou for this opportunity to further this goal.

Hopefully my resurrection of an old thread is not a concern.  Posted on YouTube as well as I wasn't sure how long forum access would take..

I didn't come across any software solutions.  I'm going to guess that the calipers are sampling at a rate of at least a few times per second.  This would seem to indicate that measurements should vary by less than  0.01mm from sample to sample. This means that the current sample and the prior sample are likely very close in size if not the same.  When the first zero reset occurs, the change is massive in comparison.  Can't you just adjust the software to detect when this first zero reset has occurred? You would track a global offset that you apply to the current measured value.  This starts out at 0.  When you detect this massive change ( e.g. 1.73 to 0.00 ) your global offset is set to the last known good value ( 1.73 in this case ).  Now when you read 0.00 + the offset ( 1.73 ) you have a current value ( 1.73 in this case ) .  Further when you read negative 0.01 you would know that it's actually a negative offset from the value (  1.72 in this case ) .   There does remain the issue of yet another zero reset when you're already near a zero value.  It now depends on how accurate the system is at keeping within certain range in a narrow time window.  If the resets are very infrequent and if the feedback system is able to maintain +/- 0.01 normally from sample to sample then a reset would likely only result in at most +/- 0.01 shift.  Assuming these resets are randomly distributed this should still keep you close depending on the frequency of the resets and the length of your production run.

So, I know this has been idle a long time... Have you considered it may be static? pulling the plastic through the calipers may be building up a static charge, then after a while it zots the electronics on the calipers... It may be as simple as adding a big ground wire to both sides of the calipers.

The other option to consider is to just always zero the calipers with a reference gauge piece in the middle at start up, Then you always try to keep it at zero -- if it resets, you're right back where you started.  (if you zero it open, does it read negative if you close it more?)

So from my memory,  you couls not "gage" it to zero. Andnif you did, when it reset. It still went out of zero. 

Good idea though. 

Thanks for the input.

~Russ

hey Russ, you have done amazing stuff with your 3d printers, , that geodesic dome you took too matter hackers meet was crazzy huh...  im sure some ppl could have stood on it and gotten away without, or if you put your helmit down...  -  wish i could remember what its called, -   cross between a rodin coil and a rubix cube,

i want too know do you think i can reuse a controller board from my anet printer, and just extend my drive belts and bed size... too make a bigger printer,,   update the slicer, - is that all there is too it ?, other than matter hacking ? or do i need a new mainboard

Yeah. Its called the QSN  Quasicrystal spin network.

Google it.

Yes you can you will need to update the firmware on your printer controller.  But that should be eazy now aday.

~Russ

thankyou.

was that one done in acid russ ? crazzy print bro

no, it was done on my build from scratch printer.

https://www.youtube.com/watch?v=AHH-6EqQCB4

~Russ