Friday, March 23, 2012

Counting Mechanism Error

Yesterday I made reference to a terrible problem with the blue beam. Here's what happened...

Before I go into it, I want to talk about a concept called "Trusted Systems". It's one that I embraced years ago in order to free my mind of worry over lots of little details. Instead of perpetually worrying about every little aspect of everything that I do, I set up systems that I trust to handle the details for me. For show management, it's Daylite. For tasks, it's OmniFocus. For accounting, it's QuickBooks. Each of these have a method that they I use to let them do their work. I follow the method and everything just works.

In weaving, there are numerous systems that I use for the various steps in the process. In setting up a trusted system, I test and test, scrutinizing every little detail... and then I give the system my trust and let it do the work.

Well, my trust in one of the systems turns out to have been ill-bestowed. It is a tiny part of a big process: the counting mechanism that helps me wind consistent amounts of yarn onto the separate sections of a beam. This mechanism consists of a threaded rod with a bell that hangs from it. I mount a tape line to indicate how many revolutions I will be winding onto the beam, and hang the bell at that line. Then, as I wind, the bell moves toward the end of the rod until it falls off with a clang! to tell me that the winding is done.

This rod is attached to the beam by means of a wooden plug that inserts into the end of the beam's axle. I insert a screw through a hole in the axle to be sure that the plug doesn't "precess", rotating slightly in its place and getting out of sync with the axle.

I've used this system for over 600 yards of cloth. At first I scrutinized it, but as time wore on I just came to trust it. Each beam had been more accurately wound than that last, indicating that my skill was increasing and that the counting rod "just works".

Until now.

There's a new factor in the basement studio. Even though this is not the first beam I've wound here, this one factor has just arrived. It's humidity. As the year has become more and more wet, I have cranked up the humidifier higher and higher. The air is dry, but there are still some spots in the concrete that remain wet. Moisture breeds mold, one of the worst nightmares of a weaver.

Low humidity causes wood to shrink. Like the wooden plug that holds the counting mechanism in place.

At first, when I moved the winding tools from the humid garage to the dry studio, the mechanism worked as expected. Then, over the next couple of days it dried out and got smaller. This made it get looser and looser in its connection to the axle, letting it flop more and more as it rotated. Then, the flopping bell started to jump on the rod. I always insert the bell in the same direction, so it always jumped in the same direction, skipping over threads on the rod and reducing the actual number of rotations before it told me I was done winding.

But I had never seen it behave badly in the past so I trusted it. I noticed the sections getting smaller and smaller, but thought it was the fault of a thread that was slightly thinner than the others making the finished diameter a little less. Then I noticed a noise coming from the bell. Tink. Tink. Tink. I watched it as I wound and was horrified at what I saw. After reaching the half way point, each turn of the beam was being counted as two.

I taped the thing to the axle and went on winding, determined that I would fix the resulting bad beam later. No matter what, I wasn't going to remove hundreds of dollars in thread and days of my time from the beam.

I finished winding last night, and here's the final result:

When I measure the circumference, the widest sections are 48" long. The shortest are 44" long. This is a problem.

I've thought long and hard about the solution and I think I've found one. I'm going to thread the "long" sections through a reed and wind them off onto the beam of my "spare" loom, the 46" production dobby. This will be laborious, for sure, but it will allow me to save that thread and those gradients in a way that they can be used for another project. It's 20 sections that are too long, and they're too long by about 20 yards. This will eventually give me 20 yards of 40" wide cloth in a gorgeous array of blues. But more important to me, it will capture the work that I've already put into this and help me to eventually get paid for it.

And I'm still moving in a few days so I'd better get down there and get to it.


Andrew Kieran said...

I was thinking about this today, and it just occurred to me when I was on my way out the door with the dog (she's bouncing off the walls just now as i stop to type) that you could use a row counter from a knitting machine. You know what I mean aye? it's a wee counting device that sits on top of a domestic knitting machine and there's a wee arm on the back of the carriage that clips it every time it goes past, advancing the number on once. It also makes a wee "click" sound which would help you keep a mental tally

Anonymous said...

I use a wireless bicycle odometer/counter for my skein winder. They are fairly inexpensive and work really well.


Blossom Merz said...

Thanks for the suggestions. This is kind of a tricky engineering problem.

I've thought about lots of different options and settled on the one I use because it fits with how I work. When I'm winding, my attention is on the threads as they go toward the loom. I need to watch for cones running out and be sure that the threads are all feeding well. I'm usually running two loud cone winders at the same time to replace the cones as I empty them.

This all means that watching a counter or listening to little clicks as a way to track revolutions doesn't work so well for me. I need a clear and loud signal to tell me that I'm done. Like a two-ounce bell landing on the floor.

What I'll do to fix this mechanism is to replace the wooden plug with tight-fitting plastic and replace the floppy bell hanger with something that needs to actually be threaded onto the rod, preventing it from slippping. I'll work out these problems for the next beam, due to start winding in about three weeks.