Limits that change (shift) over time

[#1]

Hi all,

User of JMP 18.

To trend our processes, we may use up to 4 different sets of limits:

• upper/lower alert limits UAL/LAL
• upper/lower action limits UACL/LACL
• upper/lower control limits UCL/LCL
• upper/lower specification limits USL/LSL

In theory they could all exist and be applied for a given parameter, and they could shift over time due to some change to the process - so suddenly from batch to another.

I'm wondering how I can easily visualize this using Control Chart Builder and Graph Builder. Do I need JSL? Can I rely on Column Properties alone?
Any tips and tricks are welcome, or referals to similar threads.

Thanks!

Guy 

[#2]

It will take some JSL to do what you want.  I am assuming that your data table has columns for each of the limits you want to graph, since you indicated they can change over time.  Questions I have are; are the rows in time order, once a limit is different on a new row, is that value repeated on all subsequent rows until the next change, is row your X value?

If you could supply a mock up of the graph you want, and the data table you have, it would be helpful

[#3]

Actually, I would like to store the limits in a separate table, like this:

Parameter Name	LSL	USL	LACL	UACL	LAL	UAL	LCL	UCL	Effective Date
Parameter x	50	80			65	70	60	75	1-jan-22
Parameter x	55	80			65	70	60	75	1-jul-22
Parameter y				100			40	80	1-jan-22
Parameter y			10	100			40	80	12-sep-24
Parameter y		120	10	100			40	80	5-dec-25

The actual data I'd like to trend is in a separate, big table, like this. The parameter names in both tables correspond.

Batch	Manufacturing Date	Parameter x	Parameter y
A	1-jan-21	61	56
B	10-jan-21	62	102
C	21-jan-21	53	57
D	31-jan-21	74	88
...	...	...	...

[#4]

You will need to use JSL to accomplish what you want given your data table and limit table you specified.  I envision that one would create the given chart and then loop through the limits in the limit table and create  matrices that represent the start and stop points for the line for each limit.  Finally, then using Add Graphics Script, use the Line() function for each matrix to display each of the lines. 

[#5]

I'm not sure I understand what you want to do? What questions are you trying to answer? Are you trying to assess if there is a trend in your process? Plotting the actual data is what you should do to assess this. I can't understand why spec limits would change over time (they are typically not a function of process variation)? Control limits are always a function of how you collected the data and subsequent subgrouping strategies. 

[#6]

spec limits could change if a user requires a tighter spec.
rare - but not 100% excluded.

Maybe so rare that it's not worth to implement the case in the software.