Lean Manufacturing

[drawoh]

Could somebody out there please recommend a good book on Lean Manufacturing?

I posted this to the Cost Analysis forum, but I deleted it. It was not getting any traffic.

Thanks.

--
JHG

 

[tygerdawg]

Gosh, what a tough question.

Lean mfg is a philosophical approach to mfg that focuses IMHO on a single goal: eliminate waste in all its forms. Ultimate success requires a mindset change in the organization, and my experience is that the Top Dawg must be fully committed. Doing this covers so many different facets, topics, and irritating buzzwords, with attendant "experts" who publish a plethora of books (some/most of dubious usefulness). I've been steeped in this since the early '80s when "World Class Manufacturing" was all the rage at that time. It's good, it works, I'm a believer and a practical fanatic about it. It's a joy to see a well-designed lean manufacturing system cranking out the product.

A collection of topics/buzzwords that fit into the basket of Lean Manufacturing:
5S, Toyota Production System, Just In Time, Value Stream Mapping, Continuous Improvement, 7 Forms of Waste.

Before you purchase anything, I'd go to your local library and start reading anything they have there on Lean Manufacturing and Toyota Production System. I've recently purchased my own copy of Toyota Production System by Monden (find it on Amazon) and puts into quantitative perspective some of the things I've done over the years. I've also used the Learning To See workbook about Value Stream Mapping (lean.org, Amazon). But most of my knowledge on the subject came from reading the philosophical mind-set of eliminating waste, reading the case histories and example, and then doing it in my own job (usually successful).

TygerDawg
Blue Technik LLC
Virtuoso Robotics Engineering

http://www.bluetechnik.com

 

[MikeHalloran]

How NOT to do 5S:

Get everyone lathered up about eliminating waste. Fine idea on its own.

Gather up all the short leftover pipes, shapes, angles, and sheet left over from the prior half-infinity of jobs that involved making skids of all sizes and shapes. Sell them to a scrap dealer.

Since the stock racks are now empty, sell them to the scrap dealer, too.

Sure enough, after a short time, in comes an order that generates a need for a couple of skids.
The major structural members are no problem; we order them, and they come in on a truck.
As the design evolves, there is a need for short bits of angle and tubular products.
None are in stock, so we order full lengths, and pay a little extra or wait for them to fall off a truck going in our direction with someone else's order, or send someone to pick them up.
All of this involves extra delay, compounded every time a need for a new shape falls out of the design process.

There are no stock racks in which to store the cutoffs, so we end up with a bigger housekeeping problem than we had before, and the project is a little late in part due to the compounded single-day delays.

The company now wastes a lot less than it did before, because it's now a lot smaller, and produces a lot less product, and a lot less revenue.



Mike Halloran
Pembroke Pines, FL, USA

 

[drawoh]

tygerdawg and MikeHalloran,

I am writing up some notes on DFMA, and the stuff I am finding on Lean Manufacturing seems to provide a useful context. DFMA is another thing your "Top Dawg" commit to. I am looking for some respected literature.

So much of this stuff is a mode of thinking. So many people out there are unimaginative and uncreative, and they are looking for a set of procedures to follow rigidly and without question. Any system will fail if it is implemented stupidly enough.

--
JHG

 

[tygerdawg]

How true. Everybody wants a simplistic recipe to follow. I don't believe there are any. There are some generally-successful approaches & guidelines, though. But again, without Management commitment to force workers/Middle-Managers/SiloDwellers out of their comfortable little rut, any attempt is all in vain.

If I had to produce a "recipe" it would sort of be like this (generally taken from Monden's TPS book + my own experience).[ul]
[li]Get Management commitment[/li]
[li]Start small with focused projects, but seek the big picture[/li]
[li]Fix the stupid, obvious problems first[/li]
[li]Do the 5S thing to clean everything up & make it more efficient[/li]
[li]Use data-driven improvement (like six-sigma tools) to fix processes (center the mean, reduce the variation)[/li]
[li]Use data-driven process analysis to re-configure functional layout production systems into cellular production systems (where appropriate)[/li]
[li]Deploy continuous improvement methods to pursue elimination of waste in all its forms: setup reduction, transport, scrap, supporting processes[/li]
[li]Implement Just In Time & KanBan systems for inventory control / inventory reduction[/li]
[li]Rinse & Repeat[/li]
[/ul]

Yeah, all of this should only take you a couple weeks. Hop to it.

TygerDawg
Blue Technik LLC
Virtuoso Robotics Engineering

http://www.bluetechnik.com

 

[drawoh]

tygerdawg,

I don't want to do it. I want to read about it!

--
JHG

 

[MikeHalloran]

I got a few days of formal training in DFMA by an outfit called Boothroyd-Dewhurst.
We didn't get anything to take home and keep.
I think the paper instructional material was pricey,
and they were really trying to sell us the software tool, which was stratospherically pricey.

Toward the end of the training, one of the senior MEs posited that they were talking about stuff that we already did, or tried to do, but were frustrated at every turn.

Small example: The designer who was apparently held in highest regard by management hardly ever erased anything. If you asked him to add a feature to something, he would add a bracket and a bunch of fasteners instead of reconsidering the root part.



Mike Halloran
Pembroke Pines, FL, USA

 

[SkipVought]

No doubt union thinking enters into the prevailing bog-down status quo culture.

Skip,

_{Just traded in my OLD subtlety...
for a NUance!}

 

[drawoh]

MikeHalloran,

I have the Boothroyd, Dewhurst and Knight textbook. I am a graduate of a three year engineering technology course, and I took a lot of stuff on manufacturing. This has been very useful over the years. My college now is a four year university. The graduates are eligible to become professional engineers. They don't teach manufacturing anymore. The Boothroyd, Dewhurst and Knight textbook book fills in all sorts of blanks for a newly minted EIT who wants to do mechanical design.

I am not familiar with their software, and I cannot conceive of how software would help me, or help an engineering graduate with no knowledge of manufacturing processes. Their work tables seems to be for existing designs that require improvement. I don't want my stuff to require improvement.

In over thirty years, I cannot think of a single, major component I have designed where I did not know how it would be manufactured. How about you?

--
JHG

 

[MikeHalloran]

Same here; I've always had at least some idea of how my stuff would be made.

My friend Frank Z, our shop supervisor at one place, also operated a machine shop on the side. He would bid on all sorts of stuff, basically using classical processes, bidding at a bare profit, and winning a few percent of the jobs he bid on. For the ones he got, he then went through and improved the process to drive his cost down without changing his bid price; that was how he made his profit. I learned a lot from him.

The B-D software basically helped you map out a process for making or assembling stuff, and assigned costs for features and fasteners, kept track of the aggregate cost as you went along, and nagged you about minimizing fastener counts and simplifying stuff. I think we rented one chair for one year, and hardly anyone used it, especially the ones who could have benefited most.




Mike Halloran
Pembroke Pines, FL, USA

 

[drawoh]

No doubt union thinking enters into the prevailing bog-down status quo culture.

I don't see where union thinking enters into this. How many engineers and designers are unionized?

People don't like changing the way they do things.

--
JHG

 

[KENAT]

Drawoh, what volumes are you making stuff at?

Sadly an awful lot of the material out there is really targeted ah high volumes, the book I thought I found & ordered turned out to be more focused toward artists and industrial designers and all but useless.

thread404-316744

Posting guidelines faq731-376

http://eng-tips.com/market.cfm?

(probably not aimed specifically at you)
What is Engineering anyway: faq1088-1484​

 

[JNieman]

If you're looking for standards specific to accommodating LEAN'ing practices, I would seek section R311.7 of the 2012 IBC, or OSHA standard 1910.23, depending on context.

 

[IRstuff]

As an example waste elimination, I cite what the Japanese were doing at their heyday. We got to second-source a Hitachi CRT controller chip, that Hitachi was second sourcing from Motorola. The mask set arrived with no test chips, which was completely counter to our internal products which had 5 test chips per wafer; that meant that 5 possible product sites were non-productive. We asked Hitachi about that and they said their process didn't require test chips!! The mask set came without a process flow and processing parameters like oxide thicknesses and ion implant densities/energies, so we asked for those, and they said, "don't worry about it, just run whatever your standard NMOS process is.

Disbelief... We ran the first lot; it yielded better than ANY chip of our own, and the second lot yielded even better. So, a couple of takeaways:
> lean manufacturing is about building robust product that survives the vicissitudes of a varying process, which means,
> designing a robust product that can tolerate process variations and provide product that fully meets its requirements without having to cherry-pick
> you must know and fully understand your production process, to the point where you can understand exactly every variability in the process and be able to design around them -- this means that you're looking beyond SPC and mean/tolerance range, because those alone don't tell the whole story. -- where humans are involved in machining, there's always been a lackadaisical attitude that trying to hit the center of the range was too much work, and that anything that came in under the limit was acceptable -- I think this was the root of many of the problems of US manufacturing in the 70s and 80s.

While there's some built-in cost to make Design for Manufacturing (DFM); the savings are tremendous:
> no test equipment and test chips dedicated to characterizing the process on EVERY wafer; that amounted to about $15 extra revenue per wafer and the elimination of about $500k of test equipment, AND elimination of test time
> there was always jokes in the manufacturing line about yield crashes every fall and summer, because the plant water supply changed because of snow/rain -- but this product was immune the those sorts of yield crashes
> even though the chip was larger than needed, it yielded better than chips that were tailor-designed for our own process, to the tune of about 30% higher yield. THAT translated into $150 extra revenue per wafer

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

 

[KENAT]

"where humans are involved in machining, there's always been a lackadaisical attitude that trying to hit the center of the range was too much work, and that anything that came in under the limit was acceptable -- I think this was the root of many of the problems of US manufacturing in the 70s and 80s."

Hmm, one might argue that if the tolerances are set correctly then indeed coming in anywhere under the limit is indeed acceptable assuming the design takes those tolerances into account. In fact by by achieving an in tolerance value at less effort one might argue they are being lean within their process.

You're starting to touch on using worst case tolerancing or statistical methods where simple statistical methods typically assume a distribution around mean, but many of the potential causes of deviation from tolerance are not random but skew one way. Hence to really take advantage of statistical methods you need to keep the distribution centered near the mean - i.e. you actually need to spend more effort on one sub-process to reap benefits over the entire process.

Posting guidelines faq731-376

http://eng-tips.com/market.cfm?

(probably not aimed specifically at you)
What is Engineering anyway: faq1088-1484​

 

[drawoh]

KENAT,

We usually run batches of five systems at a time. After twenty batches, we have made a hundred units. If the hour I save in design wastes twenty minutes per manufactured system, the company winds up way behind. DFMA applies in limited production. The techniques are different. Maybe the way you think is different. You need to understand the fabrication and assembly process.

--
JHG

 

[MadMango]

The Machine That Changed The World is the only book I've ever read about Lean Manufacturing, from what I remember it was pretty insightful.

"Art without engineering is dreaming; Engineering without art is calculating."

Have you read faq731-376 to make the best use of these Forums?

 

[KENAT]

drawoh, certainly not saying DFMA doesn't apply even to lower volumes. However if the 'dfx' fix takes a week of engineering to save $100 per and you sell a few a year then it may not be worth it when you allow for not just direct ROI but also that the Engineering time could be coming up with additional products to grow your revenue etc. - I think the MBA term may be opportunity cost.

Also, many of the rule of thumbs that come up for DFx are based on high volume i.e. "you need injection molded snap fits..."

Posting guidelines faq731-376

http://eng-tips.com/market.cfm?

(probably not aimed specifically at you)
What is Engineering anyway: faq1088-1484​

 

[IRstuff]

" However if the 'dfx' fix takes a week of engineering to save $100 per and you sell a few a year then it may not be worth it"

If your production is truly that random, then yes. But typically, you use the same CNC, so once that's characterized, you KNOW explicitly what it does and it should not really be as if you were designing for a completely new and unknown tool.

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

 

[KENAT]

IRstuff, we don't do any in house CNC, all job shopped out. We have preferred vendors that may remake same part over and over again in batches of, 5, 10, 15 ... every few months but that's it. So while they can control and characterize their tools to make sure they meat our stated tolerances, makes it harder to be holistic as the part may not always go to the same shop and mating parts are very likely to go elsewhere. So while we can design for a general capability, designing for a specific tool would not be appropriate.

Our production is truly that random, our electronic controller is the highest volume part and might hit around 300 a year with a life of maybe 10 years.

Most other stuff is significantly lower, next highest being around 100 pa possibly with shorter life etc.

Plenty of accessories in single digits per year.

Plus I do a lot of custom '1 off' stuff where it's even more of an issue that design time cost outweighs production time cost in many situations.

Posting guidelines faq731-376

http://eng-tips.com/market.cfm?

(probably not aimed specifically at you)
What is Engineering anyway: faq1088-1484​