Confusion DPU, PPM & DPMO vs. Sigma Level - ASQ Six Sigma Black Belt hand book

  • Thread starter Stephane Dubreuil
  • Start date
S

Stephane Dubreuil

Hi all

In the ASQ six sigma black belt hand book,

PPM = DPU * 1,000,000
So PPM is not = to DPMO

Question, why many reference says that the sigma level of 3.4 DPMO and 3.4 PPM are both six sigma?

Which one is good?

Please help
:thanx:
 

Miner

Forum Moderator
Leader
Admin
Re: Confusion DPU, PPM & DPMO vs Sigma level

Read the thread that Jim linked. That should make the relationship between DPU and DPMO clear (and saves me from repeating myself). One item that may not have been completely clear in that thread is:

  • [*]DPU and DPMO are for defects. You may have multiple defects per part.
    [*]% and PPM are for defectives. A defective is a part that contains one or more defects.

There is no direct relationship between DPMO and PPM. If 100% of all the defectives contained exactly one defect out of one opportunity per part, then DPMO would equal PPM. That is the only time.
 
S

Stephane Dubreuil

Re: Confusion DPU, PPM & DPMO vs Sigma level

The problem is the inconsistency between forum members and between this forum and the ASQ one.

Some are saying that PPM can be for defects or defective parts. It is up to the compagny to decide. This where start the confusion.

It is look like the literature is not consistent.

Who says the right thing?

Example from the ASQ Black belt book p179:

Process produce 40,000
Three type of defects can occur
-Blurred printing:36
- wrong dimensions: 118
-Rolled end: 11
Total Number of defects: 165

DPU= 165/40,000 = 0.004125
PPM = DPU x 1,000,000 = 4125
RTY = e>-0.004125 = 0.996

DPMO
Opportunity
-Blurred printing: 1 => TOP = 40,000
-Wrong dimensions: 3 => TOP = 120,000
-Rolled end: 2 => TOP = 80,000
Grand TOP = 240,000

DPMO= DPU * 1,000,000 / 240,000 = 687.5
 
S

Stephane Dubreuil

Re: Confusion DPU, PPM & DPMO vs Sigma level

“In a quality setting, "PPM" could quite reasonably be applied to:
* Defects per million items
* Defectives per million items
* Defects per million opportunities
* Defectives per million opportunities”


Regards
 
S

Stephane Dubreuil

Re: Confusion DPU, PPM & DPMO vs Sigma level

Forgot to Say, my last reply is a response from the ASQ forum for that same subjet. Great forum and great members by the way.

Thanks
 

Miner

Forum Moderator
Leader
Admin
Stephane,

You need to make the final decision on what you will use in your company. I will just make two points:

  1. Just because you saw an answer in the ASQ forums does not mean that it is sanctioned by ASQ, nor does it mean that it has widespread acceptance. It is based solely on the knowledge of the person answering your post.
  2. In quality usage, PPM means Parts per Million. Nonconforming parts = defectives, not defects. Defects per million (opportunities) = DPMO
I have worked at multiple companies in many industries, including automotive. All have used PPM to mean parts, not defects. This is not to say that there are not companies out there that do otherwise. Obviously, there is a lack of universal agreement on this issue. Therefore, you must decide what makes sense for your company.

Does it make more sense to precisely define PPM as I have attempted to do, or leave it ambiguous?
 

Tim Folkerts

Trusted Information Resource
Re: Confusion DPU, PPM & DPMO vs Sigma level

“In a quality setting, "PPM" could quite reasonably be applied to:
* Defects per million items
* Defectives per million items
* Defects per million opportunities
* Defectives per million opportunities”
Regards

My main point when I said this (on another forum) is that "PPM" by itself is simply a ratio and can be applied to pretty much anything. It doesn't have to be a physical part(as in "our factory makes 16 different parts for medical devices") but it could also be a more abstract part (as in "rework is a part of the Cost of Poor Quality).

  • the CO2 concentration in air is about 380 PPM (there are 380 m^3 of CO2 in every 1,000,000 m^3 of air)
  • the sales tax where I live is about 60,000 PPM ($60,000 tax for every $1,000,000 spent -- more commonly thought of as the ratio $0.06 for every $1 = 6%)
  • limits on lead in water in the US are 0.015 PPM (0.015 mg Pb per 1 kg water = 15 PPB)
  • the Pareto Principle states that 800,000 PPM of effect come from 200,000 PPM of the causes.
You need to think about what you mean by "PPM" before blinding plugging it into an equation.

With all that said, I agree with Miner that the most common meaning of "PPM" in quality "defective parts per million parts".


Tim F
 

Bev D

Heretical Statistician
Leader
Super Moderator
agree with Tim & Miner on the techncial and 'loose" definition of the term ppm. If your customer has a specific definition it hsoudl be used, otherwise you really can use it in your own organization in any way that all agree on. It isn't a law of physics or mathematics.

The yield formula you cited however is a law of mathematics and the exponent is the proportion of defects per unit. Not the number of defectives. It comes from the Poisson distribution which deals with defects. (Of course at low defect rates it won't matter as the Poisson and BiNomial converge as long as the defects are independent)

DPMO is largely useless except as a means of normalizing defect rates between two or more disparately complex processes or products. A complex product or process will naturally have more defective units than a relatively simple product or process since there will be more opportunities. It can enable rational discussions among rational people when making comparisons internally. But in the end even complex products and processes must be of high quality. I've found that most attempts at calculating dpmo are simply a waste of precious time, resources and emotional energy.

I prefer to simply quantify my problems and begin improving the worst few and then move to the next 2 or 3 and so on. While this fancy accounting of defects is theoretically interesting it does nothing to assist in improving quality...
 
D

davidong

I thought I have question along this thread.
My QA engineer said that we have less 3.4 parts per Million, so he concluded that we can tell our customer that we are a six sigma company. Is this a fair statement? Because we may have most of the parts are in the spec limit, but we may not have the dimension outside the 6 sigma range.

I thought that 6 six sigma means low variant of common causes. So we can said if we have a six sigma process , then we can have 3.4 defects per Million. But can we said the other way around ?
 
Top Bottom