valve manufacturing problem 4

[MonicaLee]

We are manufacturing 2-way ball valves...For some reasons both thread ends are not concentric eventhough processes are verified. We tested concentricity by fitting one end to the lathe and the other end with a long pipe, let it spin and watch its rotation. It looks like the pipe is rotating away from the center line. We order 2" body from China w/threads. Please let us know why, we would like to thank you in advance.

 

[BobPE]

I have had a lot of problems with China and their concept of a circle there, let alone center lines. There is really no advice to give you other that buyer beware.

BobPE

 

[ApC2Kp]

To MonicaLee,
If the ends are not concentric, then the problem might be one of true position and geometric tolerancing. A verified process does not achieve the requirement of concentricity if there is not adequate control of the feature or true position of the threads or hole diameter. There probably needs to be a tolerance to limit the out of perpendicular of threads and hole diameters, as well as concentricity.
The use of lathe is not necesarily true indication of the concentricity. The long pipe will bend, deflecting from its own weight. The valve body as-cast surfaces if used for chucking in the lathe are probably not a good reference for the end concentricity, especially if there is not a tolerance that references the centerline to the body surfaces. A good reference would be machined surfaces that are toleranced and machined to the same centerline reference. A better method to check the valves concentricity might be to thread short 6" piece of pipe into each end of a valve, and then chuck one pipe end into the lathe, spin lathe, and observe the run-out of other pipe end.
If geometric tolerancing is new to you, then ASME Y14.5M will start the improvement of dimensional controls.

 

[MonicaLee]

Thank you so much Mr. ApC2Kp and I even gave you a STAR for your post. I have 20 out 100 good valve bodies and 80% are NOT concentric. My method of using the lathe (aaayyy) is not perfect but it works on the 1/5. So I think it might be a problem with dimensional tolerance at a factory in China. Could you please tell me how EXACTLY I should tell our CHINAMEN to do work correctly, can you or anyone give some recommendations to tell them to troubleshoot their own problem? Thank you.

 

[MikeHalloran]

The traditional way to tell a factory to do work correctly is to buy their product subject to a drawing, with tolerances. A sketch tells someone what you want. A drawing tells someone EXACTLY what you WILL NOT ACCEPT.

Now comes the hard part; DON'T ACCEPT JUNK. If a reasonable sample of the parts are not in accord with the drawing, do not accept ANY of them; send them ALL back, freight collect, for the factory to sort out and/or replace.

By sorting the parts yourself, you have gone to work for the Chinamen, and you're probably getting dirt wages for doing it. By the way, the 'good' parts are probably out of spec for the attributes that you haven't checked.

Yes, I know, you _need_ the parts to make production numbers. If you're willing to _buy_ junk, then you must be willing to _ship_ junk. That's a TOP MANAGEMENT decision, and if this problem doesn't become visible to them, it will never go away. The idiots who put you in the position of being reliant on a factory that can't deliver what you need, need to be suitably rewarded. Solve the problem for them, and they'll get bonuses for finding cheap sources.



Mike Halloran
NOT speaking for
DeAngelo Marine Exhaust Inc.
Ft. Lauderdale, FL, USA

 

[NGiLuzzu]

I agree with Mike Halloran about top management and... idiots ;-)

MonicaLee, please let us know (if you want) how you solve the case, at the end.

Bye to all, 'NGL

 

[MonicaLee]

Thank you for the Top Management philosophy, I appreciate it.

I respect to your Management Analysis but we also have an Economic Analysis that we would have to balance between Quality and Quantity. This is not the only problem that I am having with the Chinese factory. I am trying to put as many details on the drawings as possible. Maybe I should follow your instruction not to accept any parts if they have more than 40% defects.

Hope to receive more replies from all of you from any other angle of views about this type of problem.

 

[MikeHalloran]

You shouldn't have to give up Quality to get Quantity. In fact, with more quantity, you should be able to expect better quality, because your suppliers can afford more automation.

As an extreme example, automobile manufacturers today expect their suppliers to consistently achieve a defect rate in the single digits of Parts Per MILLION.

People who do this Quality stuff for a living have lookup tables and formal programs for systematic sampling. I don't do it myself, and I don't know your quantities anyway, but just off the top of my head, if you inspect a sample of say 5 pct of a lot or 5 parts, whichever is greater, and you find _any_ defect, the whole lot should be rejected. If your supplier can't achieve that level, and you can't seem to find a supplier who can, then your supplier selection team should be evaluating their career choices.

Maybe your outfit chooses to do business differently. But if you have to buy and inspect 100 parts to find 40 good ones, then the cost of the wasted parts and the cost of inspecting all of them needs to be factored into the cost of using that particular source. You should be able to find other sources, even domestic ones, who can compete on a fairly adjusted basis like that.

Quality is free. Inspection is not.



Mike Halloran
NOT speaking for
DeAngelo Marine Exhaust Inc.
Ft. Lauderdale, FL, USA

 

[ApC2Kp]

To MonicaLee,
The problem of dimensional control has to be discussed with the manufacturer, such that both they understand what the requirements are, and that you understand their process of machining and assembling the valve. The best manufacturing process has the production people trained to understand the requirements, equipped with tooling to make quality parts consistently, and empowered to inspect their own work so that the quality is ensured. Quality is not inspected into a product, as MikeHalloran might say.
The improvement of dimensional control for concentricity would begin with the selection of datums for use both for dimensioning and for machining fixtures. If the features are dimensioned from rough casting surfaces, then it will be difficult to get repeatable results. An investment casting surface would be better as a datum, but best datums would be machined flat or cylindrical surfaces. The machining fixtures should then be setup to refer to the same datums used for the dimensioning and tolerancing. Clamps and positioning jigs need to position the parts for machining and assembly in a consistent manner. Also, the dimensions of a component would have more consistency with less movement between fixtures (multiple fixtures result in each step adding variability). That is one result of modern machining centers where a casting is chucked one time and several machining steps are performed, then the finished part is removed. Some castings are designed with flats or shoulders just for locating the part in fixtures for machining or assembly.
One approach to dimensioning and tolerancing would be to establish a datum centerline based on machined bore through the valve. The end diameters and threads would then have centerlines with requirements to be within a tolerance of 0.0XX inch / 0.XX mm position of the datum centerline measured at the valve end and also extended from the ends at a reference length of say 12 inch / 300 mm from the ends to establish angle or measurement for perpendicularity.
The Machining Engineering forum probably would be better group for discussion of these tolerance issues.
The valve design also affects the final assembly results. A 3-piece valve body has the variation from the end pieces bolted in different non-concentric and rotated positions. A 2-piece valve has the one body joint to cause variations, and a 1-piece valve body could have minimal amount of non-concentricity with good machining fixture design.

 

[MonicaLee]

Dear gentlemen,

You have very profound knowledge in our product. I just cannot believe it there are too many people already went to what I am going through now. I do not flatter but sincerely thank you for all of your help and support. I also have the next problem listed in other post, but no one seems to explain. I just to calculate the strength of the opening where found the thinnest shell wall. I have the minum tensile strength and yield strength of the material. I want to apply the maximum allowable pressure to prove that this product will work. Please read more details below:

I am applying for Canadian Registration Number. CSA (Canadian Standard Association) asked me to provide strength calculations for all Fittings. This is what I have so far:
1) I have the Material Min. Tensile: 80,000 PSI
Minimum Yield: 30,000 PSI.
2) I am going to use the weakest shell thickness (where reinforcement), t = .25"
3) I am going to use UG 27, 34, 37 for strength calculation.

This is where I need help, how do I apply Min. Tensile of 80,000 psi (material) to the maximum allowable working pressure to the thinest thickness. Please advise.