Home |
Search |
Today's Posts |
|
Woodworking (rec.woodworking) Discussion forum covering all aspects of working with wood. All levels of expertise are encouraged to particiapte. |
|
LinkBack | Thread Tools | Display Modes |
#1
Posted to rec.woodworking
|
|||
|
|||
Why Good Drawings Are Important - Long Boring Story
(WARNING - THIS IS A LONG AND BORING STORY ABOUT **** UPS IN THE
MANUFACTURING WORLD - YOU HAVE BEEN WARNED) A current thread on the use of teflon tape for air fittings brought up a nightmare from my past. Several years ago, when working as a senior project manager for a store fixture manufacturer, I was tasked by my boss with solving a problem that we were having with rolling fixtures. A rolling fixture is something that is used in the retail world that has casters on its feet, so that it can be easily moved from one location to another on the store floor. A caster typically has a wheel, a housing of bent gage metal that holds the axle for the wheel, and the attachment bolt, and the attachment bolt itself, which screws into the bottom of the tube steel of the fixture via what is called a 'weld nut'. The initial problem that we were having was that the welds were breaking on the weld nut. In the States a weld nut is usually a piece of gage metal bent into a U shape whose legs contact the interior sides of the square tubing, which are then welded to the steel tubing. A hole is previously punched in the bottom of the device and a nut, sized to fit the thread of the caster bolt, is welded to the interior face of the device, over the hole. This has good welding properties and is very strong. What the offshore guys who won the contract were doing was taking a hex nut that more or less fit into the tube steel end, and they were attempting to tack weld it at the points of "intersection" between the hex nut and the and the interior faces of the tube steel. When you think about it, there are a finite number of possibilities for points of contact between a hex nut of a size appropriate for a tap in fit to a square tube. These guys weren't even doing that good. The hex nut was a loose fit into the tube and, to this day, I am not entirely sure how they positioned the nut so that it could be tacked to the tube. I guess, if I had to do it, I'd use a magnet to pin it and do a sloppy weld on the loose side. Our PM, had gotten first article on all the pieces, assembled the unit and gave everything a pass. Of course, this only proved that the fixture could be assembled and could carry its own weight. When the fixtures were sent out into the field, we began to get failure calls. When the fixtures were loaded with merchandise (and these people were clothing retailers - not really loading the lbs on the units) and store personnel tried to move them - they were getting failures at the tack welds of the hex bolts. Go figure. So, the PM gets on the horn to China and the finger pointing begins. The Chinese guys say that they had no definition of the weld bolt and that there was no detail to describe to them how to manufacture it, so they invented it. Re-read that last sentence. The PM gets ****ed off and takes a picture of a typical weld bolt, showing the legs being bent and the good steel to steel contact potential for a good weld - and he fires it off to the Chinese guys. It was a great picture of the bottom face of the weld nut. Re-read that last sentence, too - it becomes important. The Chinese guys agree that the welded hex nut idea was poorly conceived and executed. They agree to compensate us for onshore remediation of the problem for the units shipped and they agree to use weld nuts that look like the picture that the PM sent to them on future shipments. There is not enough time to get first article on the modification because we are shipping stores and are already behind because of the work that needed to be done on the fubar'd first shipment. The PM unpacks the goods and assembles the units. He sees that the offshore guys have used the bent gage metal configuration described in the photo. He marks them for shipment and all is good. Then we started getting the failure calls. The good news is that we no longer have a weld failure. The bad news is that now the threads are stripping out on the caster bolt threads. The caster bolts are ripping out of the weld nuts and bottoming out - leaving the fixtures movable only with great difficulty when loaded - in the best case. Leaving the casters fall out of the weld nuts entirely - in the typical case. Did I mention that we were doing 8 million dollars worth of business a year with this customer? So, my boss (the owner of the company) pulls me off what I was doing and told me to get to the bottom of this. Let me say that the PM in question was not, fortunately, one of my PM's. I believe that my boss's exact words were, "Fix this ****ing thing, Tommy. And make sure we don't come out on the **** end." So, with these words of encouragement ringing in my ears, I proceeded to perform a destructive test on the offending joint. Sawing off the tube steel just above the weld nut showed that the Chinese guys had followed the photo to the letter. The photo that they had showed the weld nut from the bottom and they had perfectly bent the gage metal to conform to the inner faces of the tube. They had performed a more or less perfect weld. What they did not include was the hex nut that was to have been welded to the side that was hidden in the photo. The weld nut consisted of nothing more than a piece of 16 gage metal, formed into a U, that was threaded to receive the caster bolt. It provided about one and one half threads of engagement. This is when my boss fired the PM and put me on this problem full time, with the guiding words, "Tommy, don't **** this up and make us look bad." I pulled the drawings from the RFQ book and all that they showed for the bolted connection was no more than something like, "3/8" x 3" - 16 TPI bolt to appropriate weld nut." No detail on the connection. I figured we were hosed. Our guy had sent them a picture of the obverse, without including the reverse and we had not sufficiently described our intent on the drawing. The one and one-half of thread engagement was niggling in my mind and I thought that I might have an appeal to standards that might be in place for thread engagement. That is when I bought Machinery's Handbook. For those of you who are not familiar with this volume - it is not a coffee table book. What I was able to find there was a reference to international standards that described no less than three full threads of engagement for a minimal resistance to lateral force. It seemed that I had one card in my pocket. What I also found was a description of slop tolerances that are standard for bolted connections. On a straight up bolt with what are called parallel threads ( the usual bolt that you and I deal with) there are three degrees of fit tolerance. These are 1A/1B, 2A/2B and 3A/3B. What these arcane terms mean is that there is a certain amount of slop allowed in a male thread to female thread connection and that these designations describe a decreasing tolerance for slop from 1A/!B to 3A/3B. I figured that I was on to something. It turned out to be the case that there are internationally accepted standards that require a manufacturer to assume tolerance 2A/2B in the absence of further direction from the drawings. Our supplier had submitted a lower tolerance. I turned my findings over to my boss and he used the two talking points in his discussions with the supplier. We were looking at involving this guy in another 14 million dollars worth of metal business. The guys caved and agreed to compensate us for onshore modification and they agreed to go forward with the next phase of production - with the caveat that they got detailed drawings on the bolted connection, and that better details were provided in the future. They proposed to only hold ten thousand dollars as a compensation for their efforts on our behalf to resolve this matter. Seemed like a cheap price, considering the gross amount in play. That is when the light went on. About a year previous to this we had switched our CAD work from AutoCAD to another AutoDesk environment called "Inventor". Inventor is a 3D program where you draw every little individual element and then assemble all of them into the final drawing. Every screw, nut, bolt, etc. must be drawn in order for the program to create an assembly drawing - and the assembly drawings are used to create what are called the "plate drawings", which show each part and which are included in the RFQ book. One of the benefits of using this program is that you draw once and use the element many times, copy and paste it into other drawings. We insisted that our suppliers used a current version of the program, so that we could provide plate drawings that were essentially design drawings, and so that the supplier could open up the file and see the details that might not have been included in the plates. When I opened up the design file on the computer I found a perfectly executed description of the bolted connection, including all of the measurements for the weld nut and its use in the connection. It showed the 16 gage U bend and the hex nut welded behind. We got our ten grand back. What should have happened? We should have had better details on our plate drawings. Who was wrong? Everyone. The supplier should have used the ability of the program to find the details and our company should have had a man in place to know when the details were not being followed. We should also have remembered the capabilities of the very expensive drawing program that we had bought, far earlier. What was the result? As far as I know, my former company still does business with the supplier and he, in fact, the last that I knew, provided about seventy percent of the metal for a company that used seventy five percent metal in a fifty million dollar a year business. Anyway - this is how I came to own a copy of Machinery's Handbook and know what I know about threaded tolerances. Sorry if I bored you. Regards, Tom Watson http://home.comcast.net/~tjwatson1/ |
Thread Tools | Search this Thread |
Display Modes | |
|
|