I have recently been involved with the steering group in our company for the implamentation of TS 16949. This is an ideal opportunity for us as a company to improve our current working practises.
Currently our shop floor workers have a tick sheet on-line which they must tick once per hour for each dimension on the control plan.
Has anyone evolved from this routine?
New to this site but am very impressed with some of the feedback from other discussion threads, please share your experience, thanks, Chris

Welcome to the Cove Chris.

The key to the question is what is required by your control plan? Sometimes, a checksheet as you describe documents the action of meeting a specific requirement of the job. From what I read, the requirement is there to do a measurement for each dimension. If the control is listed, for example, as "Hourly checksheet", I would say you have to continue to document the measurements as specified. If however, there is no specific requirement for the documentation, you could consider a single sign-off by the operator/QC showing it was done hourly throughout the shift.

I have found the use of checkmarks as documentation to be unreliable. I would much prefer the actual measurement to be recorded. It offers much more information about the process. If the operators use checkmarks, they tend to just go down the line with checks whether they measured or not.

Dave

Welcome Chris!

[If the operators use checkmarks, they tend to just go down the line with checks whether they measured or not.]

I tend to agree with you Dave, however I think anyone who would falsify a checklist would also do the same with a sheet where they have to record the measurements. It's real easy to write something down quickly and go on about their business. I think a good relationship with the employees is the best way to ensure checks are being done. They have to be involved!

Russ

Agreed Russ, but I didn't mean they would purposely falsify the checklist. I have found a lot of times one of the measurements may be skipped for some reason or other but get checked off as a matter of routine. Your point is valid either way though.

Dave

I agree periodic control measurements should be actual measurement, not just go/no go attribute.

That is a different matter from dealing with the possibility/probability of "fudging" the data.

I find that fudging the data becomes less of an issue if the Control Plan is specific and actually provides time for the operator to perform this task instead of forcing him to choose between activities in a rushed atmosphere of hurry-hurry.

Depending on budget, quantities involved, and Cpk targets, many organizations find investing in capability for direct wire readings of instruments into computer programs eliminates minor "read" errors caused by eyesight and dyslexia as well as the inclination to "fudge" a reading.

Production workers and check sheets = a dangerous situation without QA Auditing.

The intent is great, inprocess inspection at the machine…outstanding !

But you need to ensure that the operators are not going through the motions. So you have QA or some type of auditor do random checks through out the day and the operators know you are serious.

Actuals are always better then checks. But I have used checks with SPC Precontrol and it worked great.

Agreed - if you can get to the stage of using pre-control, a simple check should work well. Further to that, IMO, pre-control should always be based on a stable, controlled process with a high Cpk.

Dave

Good Points Dave

Agreed - if you can get to the stage of using pre-control, a simple check should work well. Further to that, IMO, pre-control should always be based on a stable, controlled process with a high Cpk.

Dave

If you have a stable, controlled process with a high Cpk you shouldn't need to do hourly measurements. What would the point of doing the statistical analysis have been? We need to put the emphasis on controlling the process, which is done by proving (statistically) that the chosen process controls will result in conforming output.

I can think of many valid reasons for hourly checks, regardless of other SPC or Cpk considerations.

Many times the frequency of these checks (hourly, per shift, per 24 hours, etc) is also dependent on the quantity of pieces produced during that time period.

I remember a time, before we switched focus to become a contract manufacturer, when we might be working on a tool for molding in which the same workpiece might be in the machining center for nearly a week. Even then, we sometimes stopped the process to measure progress, taking actual measurements versus making check marks.

Thank you all for your comments, my plan is to use the TS Satndard to move away from these tick sheets which have turned into a pointless paperwork exercise. Instead I like the idea of having a signed log that relates the operator to the cell and is a record that he/she has carried out the checks as per control plan.
This will however be a two way thing, with the paperwork being removed it will be used as a bargaining tool to record actual values on certain characteristics, I think it would be an ideal opportunity to put the emphasis back on the control plans which in turn need to be reviewed, wish me luck, I know it will be well recieved on the shop floor, it's just the people who originally instigated the tick sheets I need to convince!
Further comments more than welcomed, Chris

If you have a stable, controlled process with a high Cpk you shouldn't need to do hourly measurements. What would the point of doing the statistical analysis have been? We need to put the emphasis on controlling the process, which is done by proving (statistically) that the chosen process controls will result in conforming output.

It all comes down to the industry.

Cold heading for an example, where you can produce thousands of pieces per hour. To maintain the desired CPK the operator needs to monitor the process a a minimum of once an hour.

It all comes down to the industry.

Cold heading for an example, where you can produce thousands of pieces per hour. To maintain the desired CPK the operator needs to monitor the process a a minimum of once an hour.
You seem to have misunderstood my point. I think that operators should be constantly monitoring the process. But you do this by monitoring the process, not by measuring the parts.

In some industries the only wal to monitor the process is to measure the parts

In some industries the only wal to monitor the process is to measure the parts
Examples?

Cold form heading using older types of headers.

Operators charting critical dimensions can control the process. (Punch and die wear) Tool limits can be established perfoming a trend analysis.

In some industries the only wal to monitor the process is to measure the parts
Yep. What he said. In point of fact, the machining industry often goes further than merely measuring the piece parts. Often, the parts are examined under a microscope to look for chatter or other symptoms the tooling is erratic or wearing. We also dig down into the chip pit to examine the chips and ensure they are "chips" and not "strings" which slow the turning, create heat, and otherwise disrupt the process. We also examine the coolant and filters in the sump.

Regardless of how "automated" the machinery, there are myriad things to check at regular intervals. Piece part dimensions are just one of those factors and the only one the customer gives a rap about.

Often, the parts are examined under a microscope to look for chatter or other symptoms the tooling is erratic or wearing.

Can ususally be controlled by machine settings, and knowing when the tool will start to produce "bad" parts. Not always, but usually. In general, machining practice is to wait until the tool is producing bad parts.

We also dig down into the chip pit to examine the chips and ensure they are "chips" and not "strings" which slow the turning, create heat, and otherwise disrupt the process. We also examine the coolant and filters in the sump..

Both of these are process controls. The phenomena which produce strings can be controlled. (Again, usually.

Regardless of how "automated" the machinery, there are myriad things to check at regular intervals. Piece part dimensions are just one of those factors and the only one the customer gives a rap about..

I don't totally disagree. Sometimes measurement of parts is necessary. I'm just saying that it shouldn't be the prime method of control with the obvious reason being that once you've found a bad part, it means something failed in the process that probably could have been controlled. The whole point of SPC should be to provide substantial evidence that process controls will result in predictable output. If you're just going to measure everything anyway, SPC is, in general, a waste of time.

Cold form heading using older types of headers.

Operators charting critical dimensions can control the process. (Punch and die wear) Tool limits can be established perfoming a trend analysis.

Yes. I agree that the only way to qualify the process is to set process parameters and then measure parts. The next logical step is a statistical study to verify process stability and and predict performance. If a reliable Cpk value has been established, the amount of nonconforming product produced over time using the same process parameters will be predictable. Tool life can also be ascertained. The idea then is to run the process, and make changes as necessary to insure continuing conformance, and the ongoing measuring of parts should end. Not all measuring of parts, mind you, but it should not be the prime method of insuring against nonconforming output.

Can ususally be controlled by machine settings, and knowing when the tool will start to produce "bad" parts. Not always, but usually. In general, machining practice is to wait until the tool is producing bad parts.
Smart machine shop operators do NOT wait until bad parts appear. Who ever told you this is about 50 years behind current good practice.


Both of these are process controls. The phenomena which produce strings can be controlled. (Again, usually.

This is the entire concept of hourly checks: to assure processes ARE under control.

I don't totally disagree. Sometimes measurement of parts is necessary. I'm just saying that it shouldn't be the prime method of control with the obvious reason being that once you've found a bad part, it means something failed in the process that probably could have been controlled. The whole point of SPC should be to provide substantial evidence that process controls will result in predictable output. If you're just going to measure everything anyway, SPC is, in general, a waste of time.
No one has ever suggested in this thread that measuring parts is the PRIME control. It is merely the one which is most visible to customers. The point is not to find a bad part, but to identify TRENDS before the parts go bad. In-process measurements by taking a series of consecutive parts for SPC charting is performed precisely as the PREVENTATIVE measure so no one ever sees a nonconforming part. We don't measure EVERYTHING. We are sampling - if not every hour on some process machinery, then perhaps every shift.

The whole point of modern manufacture is to aim toward zero non-conforming parts because the process is under control.

I think that "smart machine shop operators" in this context might be tautological; what makes them smart is that they don't wait for bad parts. That doesn't mean that all machine shop operators are smart, or are allowed to be. You might be surprised at how many job shops of various kinds are firmly mired in the 1950's. That's not good, obviously, but it's reality. For many shops part measurement is the primary control, even for ones that think they're doing some form of SPC.

The one aspect of all this that hasn't been addressed here is the time factor. Good process control requires having enough time to firmly establish a stable process, and that luxury is often lacking, with the result that new jobs are set up and running before they should be, and part measurement is the only way to be sure of the output. American industry, with few exceptions (Bose is a notable example) don't bother with research.

I guess what some have trouble grasping about the Cove is its position as a repository for the BEST practices, the BEST answers, the BEST methods offered by experienced practitioners who have seen the best and the worst and the gamut in between and want to help the inexperienced or the questioning individuals AVOID getting mired back half a century behind the best competitors they will face in today's marketplace.

Nobody is served by reading obsolete or erroneous answers. Some of us spend a good portion of our day on the job pointing (leading?) our organizations to the best and most competitive position in their industries. We try to share that knowledge with others.

From time to time, we indulge in humor, but we don't attack individuals, we attack ideas, especially obsolete and outmoded ones. We see no purpose in forcing anyone to reinvent the wheel time after time.

We don't spoonfeed the answers, we try to show the way to a method of looking at and solving problems that can be applied throughout a career.

[QUOTE=JSW05]I think that "smart machine shop operators" in this context might be tautological; what makes them smart is that they don't wait for bad parts. That doesn't mean that all machine shop operators are smart, or are allowed to be. You might be surprised at how many job shops of various kinds are firmly mired in the 1950's. That's not good, obviously, but it's reality. For many shops part measurement is the primary control, even for ones that think they're doing some form of SPC.

As previously discussed ……

Part measurment can be an effective primary control for SPC.!!!!

That seems to be a mighty big "we" you're swinging around. With due respect, can you tell me how your latest post relates to anything I've said in this thread? Am I not allowed to disagree?

That seems to be a mighty big "we" you're swinging around. With due respect, can you tell me how your latest post relates to anything I've said in this thread? Am I not allowed to disagree?
Disagreement is fine and encouraged in the Coffee Break Forums. Here in the Quality Topic Forums, the thrust is toward factual information which helps show the questioning person the BEST answer, not an "alternate, old fashioned answer based on emotion rather than actual best practice."

WE are big, in the "group mind mode" and we strive to reach consensus on the best answer, not dissension and confusion and a multitude of off-hand comments.

I'm at work right now. I have at hand a PPAP submission from a QS-9000 registered supplier. In 4-page control plan, the Process Characteristics column is completely blank. Completely. It's an inspection plan, not a process control plan. The document is completely devoid of any evidence that there is any kind of process control other than inspection going on. I see more of this sort of thing than you would believe. My advice to the OP--that it's better to control the process than to measure the parts--is nothing more than common sense. That's fact, and is based on personal, successful experience, not emotion or what someone else thinks is "best practice." I respect your opinions and experience, but if we all agreed on everything, most of us would be unnecessary (to paraphrase G. B. Shaw).

I'm at work right now. I have at hand a PPAP submission from a QS-9000 registered supplier. In 4-page control plan, the Process Characteristics column is completely blank. Completely. It's an inspection plan, not a process control plan. The document is completely devoid of any evidence that there is any kind of process control other than inspection going on. I see more of this sort of thing than you would believe.

What was required for the PPAP and if a process cap study is required then why have you accepted it.

"My advice to the OP--that it's better to control the process than to measure the parts--is nothing more than common sense."

For some industries the only way to show process capability for a PPAP is to measure the parts.

What type of industry are you in?

QUOTE=JSW05]I'm at work right now. I have at hand a PPAP submission from a QS-9000 registered supplier. In 4-page control plan, the Process Characteristics column is completely blank. Completely. It's an inspection plan, not a process control plan. The document is completely devoid of any evidence that there is any kind of process control other than inspection going on. I see more of this sort of thing than you would believe.

What was required for the PPAP and if a process cap study is required then why have you accepted it.

I don't understand the question. I was talking about the control plan, not a capability study. Why have I accepted what?

For some industries the only way to show process capability for a PPAP is to measure the parts.

I agree.

What type of industry are you in?

Motor vehicle manufacturing.

Your Statment

"In 4-page control plan, the Process Characteristics column is completely blank. Completely. It's an inspection plan, not a process control plan. The document is completely devoid of any evidence that there is any kind of process control other than inspection going on. I see more of this sort of thing than you would believe.


Did you rec. a process cap with the PPAP ?

Your Statment

"In 4-page control plan, the Process Characteristics column is completely blank. Completely. It's an inspection plan, not a process control plan. The document is completely devoid of any evidence that there is any kind of process control other than inspection going on. I see more of this sort of thing than you would believe.


Did you rec. a process cap with the PPAP ?

Yes, but let me clarify my point. The capability study was based on measurement of product characteristics, and there were plenty of those referenced on the control plan. The AIAG form has a bifurcated column headed "Characteristics" with the subheads (and columns) "Product" and "Process." A control characteristic belongs to one of the two categories. Features of the part (size, finish, etc.) are product characteristics. Features of the process (feeds and speeds, e.g.) are process characteristics. Process characteristics, for the most part, control product characteristics. The capability study called for parts to be measured, not process characteristics. It is expected that process characteristics and their associated controls will be addressed in control plans.