Accounting for Rework in Production Schedule

[RBCSCI]

I have a very small volume, high complexity, long cycle time product. A typical build takes roughly 6 months. There are test points in the build process with low yields and long rework cycles. I'm trying to account for those failures/reworks in my production plan. Here is my current approach:

Test Point 1 - 50% yield with a 14 week rework cycle
Test Point 2 - 80% yield with a 20 week rework cycle

My approach is to incorporate an extra 7 weeks of cycle time (.5 x 14) following Test Point 1 and 4 weeks of cycle time (.2 x 20) following Test Point 2 into the standard cycle time for each unit. I expect that I will end up with some units ahead and some behind, but the aggregate should stay close to nominal.

My alternate approach might be to build back-up units in parallel, which could be used as line fill for failures, but my unit cost is $200K each. That seems cost prohibitive.

I'm looking for ideas on how to maintain a stable production schedule that won't have my customer panicking every couple weeks.

Thanks!

 

[truckandbus]

would all units (backup or otherwise) be saleable upon completion/rework?

 

[RBCSCI]

No. The contract is only for a specific number of units. There's no potential to sell extras on this contract. There is a possibility of pushing partials/extras to a follow-on contract if one is awarded, but they may not meet the revision/pedigree requirements for the next contract. There are often changes to the Technical Data Package (TDP) from contract to contract.

 

[truckandbus]

Ok - so building only what you need looks like the way to go. Plan for worse case timing and either deliver early or sit on the inventory?

 

[RBCSCI]

Agreed. Just checking to see if I might be missing an option. I'm constrained on Period of Performance and trained resources so it's a real challenge to fit into the box on this one. Worst case cycle times put me outside my PoP and I have limited opportunities for schedule improvement without additional heads for parallel builds. I guess that's the manufacturing business.

 

[Agent666]

I'm left wondering if you can't put more test points in place to check earlier. I mean waiting that long to find 50% of you units need weeks and weeks of rework seems like there's an opportunity to identify significant non-conformance much earlier in the process?

 

[CWB1]

Sounds to me like you need more in-process quality checks as your rework time is ridiculously high. Force the work to slow down and be done correctly the first time.

 

[IRstuff]

Sorry, but your yield is atrocious. Just for the sake of argument, let's say the variable cost is $50K, which is then roughly $2k/week of fab. That means your 14-week rework is throwing away $28k and your 20-week rework costs $40k. Since your yield for the former is 50% and the latter is 80%, you're essentially adding $22k/unit of additional cost to each unit shipped, not counting the material costs. That could be the difference between a nice profit and a barely break-even profit; if you make only 50 units, that's almost $1.1 million of lost profit.

Shouldn't you be trying to figure out why the yield is so miserable? Spending, say, $100k to fix problems that could recoup $1 million; isn't that better than planning and scheduling for failures?

TTFN (ta ta for now)
I can do absolutely anything. I'm an expert!

https://www.youtube.com/watch?v=BKorP55Aqvg

faq731-376 forum1529 Entire Forum list

http://www.eng-tips.com/forumlist.cfm