Continue to Site

Welcome to our site!

Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

  • Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

Problem with a Loom!

Status
Not open for further replies.
If you mount 8 (or 16, 32, whatever is required) flat magnets with opposite polarity outward around the end of the bobbin then when it passes the coil multiple pulses will be produced. A microcontroller can easily identify a stopped bobbin by looking for the missing pulse stream. This can be done with a $1 pic chip. Trying to do it with discrete logic is adding unnecessary expense.

If you need help with the pic code then there are many people here willing to help and in your position you should learn this area of electronics as it will become invaluable.

Mike.
 
BTW, where about in Aus are you? If you're anywhere near Queensland then I would be interested in helping in such a fascinating project.

Mike.
 
About the Pic..

Hi Mike, the logic is simple here and I am bit old fashioned (I will get around those PICs one day :)).

It is not my project. If you go to the origial posting it is - Wizard's, not mine (see the 1st page). I am only trying to understand the process -then help, if possible.

I believe, the magnet solution is most cost efficient here and there won't be a train of pulses to speak of, just a few at best - and that's good enough for the presented solution because only one pulse per bobbin is needed in any 90 deg pass providing 4 sensors are used to define the mechanical starting point in time.

I belive, I have only enhanced slightly solution presented by elbc1388.

With a PIC and one single pick-up sensor, the difficulty sems to be not so much with the recognition of pulses but rather - with the anticipation of the time frame when the next pulse train should arrive and for that you would have to continually monitor(measure) the horizontal velocity of these bobbins.

As they are already mechanically set spaced 90 degs apart from each other - I proposed 4 sensors fixed at the exact separtion of 90 degs around the outer perimeter. This way I no longer have to anticipate the next sensing cycle (i.e. do not have to sense the horizontal speed).When the first pulse comes - so should the other three, i.e. - the time is now.

The first of 4 arriving pulses begins the cycle and the absence of 4 - does end it. By detecting the the first pulse I can set the max time needed to receive the remaining three without giving the horizontal speed much of the consideration. Following that time comes the "sampling". The absence of any of the 4 pulses within the begin to.. sample moment - results in a "stop". The moment to "sample the AND gate" has to fall somwhere in the middle between the begining and the end of each 90degs rotation cycle.

I am located in Nowra, NSW


Regards,
xanadunow
 
Last edited:
Addendum.. (we need more than one pulse!)

A passing magnet on a bobbin will produce a pulse even if it is not spinning. The answer here is to increase the number of pulses. To achieve this we need to increase the resolution i.e. we need more magnets..

Next, we count them (the pulses).. Two would be enough to set the monostable.

The end of the 90 deg cycle will also reset each counter.

xanadunow
 
Last edited:
A passing magnet on a bobbin will produce a pulse even if it is not spinning. The answer here is to increase the number of pulses. To achieve this we need to increase the resolution i.e. we need more magnets..

I believe I suggested that 4 posts ago and eblc suggested it about 2 pages ago!!

However, it is good to see that some posters are finally starting to see the simplicity of this approach.

Mike.
 
The common sense should always prevail..

Most of this thread was focussed on the "sensing" rather than a complete solution. Altough the sensing was the key, I have seen little other evidence of the Yes/No answer (solution) to achieve the result required.

xanadunow
 
Last edited:
the rest

As for the rest of this thread.. the problem is solved. Should this solution produce unexpected results.. the pattern recognition is the other way to go.. Other than that, I do feel redundant here.

Regards,
xanadunow
 
Last edited:
As for the rest of this thread.. the problem is solved. Should this solution produce unexpected results.. the pattern recognition is the other way to go.. Other than that, I do feel redundant here.

Regards,
xanadunow

I am sure the OP will be pleased that the problem is solved. However, I think this is his/her call. I got confused earlier as you posted as though you were the OP and you are doing that again. You should refrain from posting in a manner that makes you appear to be the original poster.

Mike.
 
Thanks

Thanks Mike,

It is solved - to my content; the originator has left for the Xmas I believe and he/(she?) will return to see the outcome. I do apologise for the image you have received, I have clarified it the instant it was mentioned. I am a newbie here and I am sure that with time I will learn the ropes.

It would help me a lot if you make it a habit to look at the first message in the thread - as I did.

Regards,
xanadunow
 
Thanks Mike,

It is solved - to my content; the originator has left for the Xmas I believe and he/(she?) will return to see the outcome. I do apologise for the image you have received, I have clarified it the instant it was mentioned. I am a newbie here and I am sure that with time I will learn the ropes.

It would help me a lot if you make it a habit to look at the first message in the thread - as I did.

Regards,
xanadunow

When people post in a style that suggest they are the OP then I assume that is who they are.

You have now posted in an authoritative way suggesting you know the OP and his/her whereabouts.

It would help me a lot if you stop posting in this way. Unless of course you do know the OP and speak on their behalf.

Mike.
 
Hmm..

Settle down Mike.. I don't even know who the OP (?) is ? .. I am guessing that you do refer to the member of the forum that started the thread.

My writing style should never be a predisposition to your assumptions. Just click on the first message in the thread and know the answer. The implications suggested by you are ridiculous. Instead of just simply acknowledging your mistake - you seem to be accusing me of doing something delibrate for the reasons I can not depict.

Is it not the truth that you are the one who got confused, mistaken and presumptuously wrong? Even a newbie like me knows to check the first message in a thread.

If you have forgotten about it then I say it again - check it.

It is not that "ego" thing again? We are here to give and get help - not to exploit our egos, I have mentioned it erlier in this thread and you have been in the midst of it then too.

Preach/teach.. do whatewer you desire but stay professional about it. You were the confused person - not me. You have made all assumptions - not me. I have checked the facts - not you.

I sincerely hope we both learn a good lesson here Mike.

Merry Xmas,

Regards,
xanadunow
 
Hello all guys,

Thanks a lot for your participations.

I am busy with my examinations and that’s why I did not get any time to continue this thread. I need to reread this thread again to survey the whole inputs and or ideas again. I will let you guys to know if the solutions are able to help me out to solve this problem or not.

Before starting this thread I was thinking about the solutions for 3 days, I was thinking about most of solutions that you guys suggested afterwards (Magnet, inductive, IR, IRFD, ultrasonic and even generating electricity by the shuttle itself). Most of those solutions had and now have incapacities. I must get time to reread your ideas again, especially since the page 4. I will do so tomorrow. So I will write down my idea tomorrow too when I evaluate the whole said ideas again.

Special thanks to you 'xanadunow'. Maybe you caused confusion to Pommie, but you helped me and this thread to get a logical conclusion. Thanks a lot for that.
Pommie is right though, but I like your method. It helps a thread to go ahead reasonably. That was my duty as the poster but I could not do due to my busies.

Till tomorrow.

P.s. 'xanadunow', I am too busy for now to know if Christmas is arrived or not. lol!
Oh, merry Xmas
 
Most of this thread was focussed on the "sensing" rather than a complete solution. Altough the sensing was the key, I have seen little other evidence of the Yes/No answer (solution) to achieve the result required.

xanadunow
I proposed a complete, reliable solution and got scoffed at!

Dan
 
I did not try to "scoff" anyone and if I did miss something - I do apologise.

I have asked some questions twice to understand the process. It was vital for me and following that, I came to draw my own conclusions. I have just only joined this site and this thread sometime soon after, it was already going for a while. I am a newbie.

All possible "sensors" presented and I can think of, do fall into one of the three categories (thinking of horizontally rotating and spinning bobbins):

A) no power supply needed,
B) power supply needed,
C) power supply needed and other means of the communication needed,

“A” is the obvious choice and the number of sensors required is presentable in the my (“A” sensor) solution utilizing 4 sensors spaced 90 degrees apart and very much inspired by other members of this thread (magnetic). (THE SOLUTION IS NOT MINE, THE 4 SENSORS AND WHAT FOLLOWS - IS.. )..It means - I am listening to what you do have to say and I had backed-off from other solutions I have proposed earlier. This is my learning curve I do need to respect and obey, and expect your reaction to it - in kind.

When I did say ALL, I did really mean – ALL but one.

The only other valid solution I saw, involves the optical “pattern sampling” of the "sock", and that one - is mine alone within this thread.

The OCR software is capable to distinguish between “A-to-Z” and “0-to 9”. I am very sure, it can distinguish a single “character” of the pattern detected on the production line against the expected (wanted) sample. For this, perhaps even a PIC could be usefull. This "novelty" has not been presented earlier, except by me.

I do fill quite bad for that, becuse I am taking away the job from a working fellow, but the macro camera would do the same job for no greater cost than required for the initial setup + some conehead type of - programming. The additional benefit is hidden in the the fact.. that - the camera will not fall asleep so much as may the "working fellow" and in the end, - I may get my "sock" cheaper :D.

The reason for me to feel redundant here is - that I do feel; the answer - has already been found. I have stated that already earlier.

Regards,
xanadunow
 
Last edited:
I did not try to "scoff" anyone and if I did miss something - I do apologise.
not a problem, I do have a tendancy to use harsh terms
All possible "sensors" presented and I can think of, do fall into one of the three categories (thinking of horizontally rotating and spinning bobbins):

A) no power supply needed,
B) power supply needed,
C) power supply needed and other means of the communication needed,
Actually they all revolve around one thing:getting the power in and the signal out
The only other valid solution I saw, involves the optical “pattern sampling” of the "sock", and that one - is mine alone within this thread.
While this sounds good on the surface, it was referred to as a "bag" last I knew and I was guessing that it was a 1M diameter burlap tube that was actually being manufactured.

Well with a full PC and a macro lens, it is conceivable that you could use an edge detect algorithm to detect the holes in the fabric left by a single missing strand. I do not see any hobby caliber micro being able to keep up with the machine, even with a camera meant for this type of application.

Dan
 
This message is only about the pattern recognition..

Hi Dan,

The sock/bag/tube.. we do know what these terms do refer to..

Your estimation of around 1m in diameter does sound pretty good, I wish we had some more details refering to the speed of the production line. An estimation can be made looking at the only video attachment provided by Wizard early in this thread.

I would like to pick the brains of an expert like you Dan. You did state that no PiC will process that amount of information and I do agree.

The input from the camera does come in at least 24 frames per second and this is 24 frames of pixel'ised information (you choose the resolution).

Let's stop for a moment here.

The time has stopped.

All you have in your hands is one single frame with "x" times "y" pixels.

This is the frame I would like you to process. How much time do you need to process one single frame. Forget the other 23 frames in that second..

Would you feel confident to make the decision to stop or continue the production once in every second?

If the answer is "yes" then you have 23/24 of a second to make the processing and the decission before you pick up and sample the next frame and this is a lot of time.

I have choosen 1 second as an example. The true question is - how much time do you need to process the information from one single frame and make the decission?

What I am getting at here is to pick up just ONE single frame and use the remainder of that secound (1 sec) to process the information and make the decission.. The incoming information is in abundance but you will process one frame only and discard the rest of them until you are ready to accept one single frame again.

If you can process and make the decission in just under a second, then you will "mimic" what the human operator does now, keep him out of the job and make the production line owner happy..

Regards,
xanadunow

PS. You have stated earlier that a missing string will produce a "hole". I do not believe this is the case. Somewhere earlier in the thread I suggested that an effect of a missing string results in two (instead of one) horizontal strings being captured together horizontally by the vertical strings and no one has corrected me on that as being wrong so I am guessing that I am right. This is important for me to know because I try to recognise the valid pattern..
 
Last edited:
Hi Dan,

The sock/bag/tube.. we do know what these terms do refer to..

Your estimation of around 1m in diameter does sound pretty good, I wish we had some more details refering to the speed of the production line. An estimation can be made looking at the only video attachment provided by Wizard early in this thread.
Well making a few assumptions we can get there. If the frame rate of the video is 24Hz then the shuttles are passing 23 or 25 times per second or some multiple of that. If the thread is 1mm in diameter you are making 3-4cm of tube a second.
The input from the camera does come in at least 24 frames per second and this is 24 frames of pixel'ised information (you choose the resolution).

Let's stop for a moment here.

The time has stopped.

All you have in your hands is one single frame with "x" times "y" pixels.

This is the frame I would like you to process. How much time do you need to process one single frame. Forget the other 23 frames in that second..

Would you feel confident to make the decision to stop or continue the production once in every second?
That is not our decision, it is the manufacturer's. How fast it the machine going and you much waste will he tolerate?
What I am getting at here is to pick up just ONE single frame and use the remainder of that secound (1 sec) to process the information and make the decission.. The incoming information is in abundance but you will process one frame only and discard the rest of them until you are ready to accept one single frame again.
If you pull, or omit in this case, one thread from a weave you get a stripe of vertical bars, but the contrast in a capture will be horrendously low.
PS. You have stated earlier that a missing string will produce a "hole". I do not believe this is the case. Somewhere earlier in the thread I suggested that an effect of a missing string results in two (instead of one) horizontal strings being captured together horizontally by the vertical strings and no one has corrected me on that as being wrong so I am guessing that I am right. This is important for me to know because I try to recognise the valid pattern..
Correct, technically a broken string in a fabric would be a hole. A missing string would be more like a run in a nylon.

Dan
 
That is not our decision, it is the manufacturer's. How fast it the machine going and you much waste will he tolerate?
Dan

He (the production line owner), will tolerate anything that gets a fellow worker off the payroll. The assumed by me 1 sec is pretty good approximation.

Assuming, you do have this 1 sec - what can you do to make the executive decision?

You have "x" by "y" frame and 1 sec.. where do we go next?

We're sampling the "sock" (not the bubbin) with (say) 1 square inch worth of video sample..

I had originally proposed to sample the average luminance of the "frame" and to make the yes/no decision basing on that (dynamicaly, using the sample and hold to "pick up" the change in the pattern)..

If you do insist to call it a "hole" - that's ok.. I am after the change of the pattern - I do have 1 sec to make the decision.

There are two things we do know: the "correct" pattern as seen by the line operator and, the "incorrect" one as it is being picked up by him..

There is 1 sec of time.

There is "x" times "y" pixels in the frame..

Can you make the decision using these PiCs in 1 sec?

Regards,
xanadunow
 
Last edited:
Status
Not open for further replies.

Latest threads

New Articles From Microcontroller Tips

Back
Top