Driving Printer Head from PC

As a second look, I would say that as long as you can find a print head that has a consistent delay time would be usable. The known delay can be figured into the programming.

Periodic maintenance (calibration) can be performed to ensure the delay time has not changed. When the software is produced, ensure the delay can be programmed through a simple database change or the like (reference a text file). This way, your maintenance worker can make the changes without having to spend more money to get the program changed every so often.

Another variable is that quality control must ensure that the nylon is meeting specs and elasticity is not affecting the measurement from the outset.

1. Yes - you are right about consistent delay being the key so that it can be programmed in.

2. As regards the nylon elasticity not affecting measurement the customer has told us that it won't as the threads undergo an anti-stretching process. We will have to check this out though.

Fishing nets come in mesh sizes. Length and width are measure stretched to that mesh size. If you know what your mesh size is. Couldn't you count then? Or use the camera to count them? I would also look at loom some where in the process of weaving one complete enclosed mesh you could place a counter.

What we are trying to measure here is the distance between consecutive knots.

I'm not grasping completely what you are trying to do here but from what I do understand I have a few concerns.

You're not taking into consideration the time it will take for the chemistry of your printed mark to cure, dry or otherwise stabilize.

The 0.1% accuracy seems like an absurdly tight measurement specification, particularly for a flexible item like a plastic net. You might find that thermal conditions and stretching of the net itself exceeds this tolerance.

1. After the line is printed on the net it travels quite slowly without touching anything else for about 800 mm. The ink should dry out during this time. Actually, even if it is not completely dry it does not matter too much as once it is imaged at the down stream position its purpose is served. Of course, we would like to leave a clean, thin line and not a dirty smudged one.

2. The inaccuracies I mentioned are of measurement which ideally should be around 10% of allowable inaccuracy in the part (net) dimension. In this case, though, the customer wants to get a much tighter accuracy on the net - +/-2. We have told them it is not possible to get this level of accuracy and since the accuracy they are getting currently is way over this limit they have agreed.

I am not sure why 0.1% accuracy is required? But if I have to do this, I would put my spooling unit on a weighing unit, manuplate it for Meters and tare it to zero.

The take up roller is very heavy and it is not practicable to weigh it and arrive at a measure of the length.

Silly question: Why do you need a PC printhead anyway? For me an ink pen and a relay like mechansm can do the job too. Or use a marker with a UV reflecting ink so the mark is also invisible, but can be detected under "black light".

In a similar project in the textile industy we used a springe & needle assembly and a micromotor with gearbox to move the plunger to form a small drop of the ink. The ink was sort of "sticky" and cured with the IR heat-up available in the machine anyway. The analysis can be done with a UV diode to activate the ink and a photo diode to sense the light emission. The measurement results were in the range as required for your project.

Two GA's from me, one for Rakesh and one for Uli.

I figured a rubber stamp with a solenoid could do it. And Rakesh's weight solution is even easier as the takeup spool could just be weighed at end of process and marked accordingly.

Seems like the OP is making a technical mountain out of a molehill.

You are weaving, so I suppose the machine criss-crosses the lines back and forth? if so, the the machine itself is the measure if ONE thread in the feed is a dark color! Has this been tried to see if it can be seen?

Certainly the same technology that enables my computer mouse to detect motion speed and direction as I slide it over a table, mouse pad, or my blue jeans, could optically detect the same of any regular mesh fabric.

Who knows, you might even be able to use a mouse in the application? Cheap enough.

Thanks, RufusVS. This has set me thinking. Let me check if this principle can be applied for a wider array (a camera?) kept at a distance of say 200 mm.

Passing any one of input fiber through a pulley that is coupled to a rotary incoder can do the job.you can process O/P of incoder to your requirements.

Measuring traverse of individual thread is not sufficient as there is not guarantee that all threads get pulled equally. A knot is being formed with adjacent threads and when the machine tightens the knot each thread may get pulled differently.

A weaving machine uses a repeating pattern. Change the thread color of one spool and do a test run and see what the resulting pattern is. You may find you need to only count the times the color returns to the origin to know everything you need to know.