Printed Circuit Board Assembly & PCB Design SMT Electronics Assembly Manufacturing Forum

Printed Circuit Board Assembly & PCB Design Forum

SMT electronics assembly manufacturing forum.


50 PPM

#21255

50 PPM | 21 August, 2002

Hi all,

Our management set target 50 ppm for overall process like printing,mounting & reflow soldering.

Our m/c cpk for mounter is 1.0 & for printing is 1.33.

So, pls comment !

reply »

Ken Bliss

#21256

50 PPM | 21 August, 2002

Hi Joseph

It is not clear what you want us to comment on, however I will take a shot.

1. If the boards you are assembling are expensive it may make sense to strive for 50ppm.

2. If the boards are not very expensive, in my experience you would be better off improving throughput by working on the plant bottlenecks and therefore profits than spending a lot of time and money trying to get to 50ppm as that is a very aggressive goal, although I am sure marketing would like to sell that quality level, I think improved profits and therefore be able to lower prices to be more competitive would work better overall with marketing.

3. What is your current rate ppm? If it is in the 100�s of which it probably is or you would not be making this post, a little work in both areas would be excellent.

Hope that helps some

Ken Bliss www.blissindustries.com

reply »

#21257

50 PPM | 21 August, 2002

Hi Ken, Our printing is 49ppm, mounting is 83ppm & reflow is 150ppm.

RDGS

reply »

dragonslayr

#21262

50 PPM | 21 August, 2002

I would like to address several issues. First, your reflow is 150 PPM for the same exact assemblies that have lower PPM's? or are you giving a 150 PPM rating for all products produced through the reflow process in a specific time period?

Reflow is a benign process i.e. the only thing added is time and heat. How do you derive that the reflow process is causing more defects than print, pick and place? The few things that come to mind that can cause reflow defects include an out of balance machine (jerky chain/belt, profile not optimized, convection blowers are blasting parts around, etc.) Otherwise, the root cause of defects must be attributable to the prior processes. Second- What are your defined attributes to calculate PPM? Do you include: quantity of assemblies produced x((total component count + total placements (components - PCB) + total terminations)/ 1,000,000 ? My observation regarding 50 PPM target- easily reproduced with an assembly with total component counts greater than 800. (motherboards are a good example) In this range, it takes about 10 solder defects to be equal to 1 misplaced component, affecting your PPM by a value of 1 (going from 50 PPM up to 51 PPM) (OK y'all - don't challenge the actuals here- this is just a generalized comment)

On the other hand, assemblies with low total component counts (and terminations) tend to trend much higher (200 PPM and greater). Why? A single defect with a low parts and termination count has an effectively higher influence to the PPM. 2 solder defects can cause the PPM to rise 30-40 points.

Finally - what can you do? 1. Stabilize your process ( get rid of all the obvious mistakes). 2. Accept whatever PPM (per assembly) you get as a baseline reference. 3. Apply 6 sigma rules to the daily PPM results. A big hint - there's no such thing as a Lower Control Limit when Process Improvement efforts are driving the PPM rate down. Just make sure that you react to upward trends. Periodically adjust the target value down (with good statistical practices employed)

Compare apples to apples - monitor and track PPM rates per each individual assembly! To do otherwise would be creating an average that smooths out the high and lows, giving you false PPM values.

I fail to find the significance of the dollar value of the assembly to be a determining factor as to why one would be proactive in reducing defects, therefore reducing costs.

I could produce 10 very expensive boards and have a lower rate of return on my efforts than to fix process problems for a cheap assembly that is produced in the hundreds/thousands quantities on a daily basis.

A defect is a defect, regardless of where found, all causing greater cost in rework and lost time/materials.

Good luck

reply »

Ken Bliss

#21263

50 PPM | 21 August, 2002

Hey dragonslayr

Just a follow up comment from your comments that you fail to see the significance to the value of the assembly.

My point was that if the board is worth say 100.00 cost and you built 100 per day and 2 per day require manually rework and you spend days or weeks tweaking the process to get to 1 or .5 defective boards per day, the same effort applied to throughput will have significantly more impact on the goal of the plant which is to make money. Basically at the defect rate mentioned you are possibly chasing pennies at the cost of dollars.

The only reason management would direct anyone to reduce the defect rate to a certain amount would be to increase profits. If the value of the board makes sense to do this then he should focus on that, especially if the defective parts are non reworkable or scrap and or coming out of a bottlenecked machine when they had the defect prior to entering the bottlenecked machine. It also depends on one product ongoing or small batches.

Do not get me wrong, removing defects is always a good thing and you never want them to get out of the plant. But, keep the costs and efforts focused on why the plant is there in the first place. To make money.

Just my thoughts

reply »

#21269

50 PPM | 21 August, 2002

First, it�s GREAT that your defect rate of 282 is very close to your Boss� goal of 50 ppm.

Second, your Cpk and your ppm numbers don�t tie out with our expectations. You say Cpk for: * mounter is 1.0 and �your mounting is 83ppm�. * printing is 1.33 and �your printing is 49ppm�.

� but when we use our handy, dandy conversion chart � Sigma||ppm||Cpk 6.0||3.4||2.0 5.5||32||1.83 5.0||233||1.66 4.5||1350||1.5 4.0||6210||1.33 3.5||22750||1.16

� we go �hmmmm�.

reply »

dragonslayr

#21271

50 PPM | 21 August, 2002

..and the choir is humming in 5 part harmonies. Personnally, I never gave much truck to Cpk stuff. Seems to be a high falutin way of showing the lowerlings that academics can manipulate numbers. Dazzle 'em with brilliance or baffle 'em with bull pukey. I prefer the straight dope- easier to get a 6th grade graduate to understand. They understand "High PPM - bad", "Low PPM - good"

KISS principle applies.

reply »

dragonslayr

#21273

50 PPM | 21 August, 2002

surely, a big picture perspective is in order to keep from charging off, tilting at windmills and waning Quixotic.

Just to tic ya off a bit more (come on now, smile), would those 2 boards in rework have just one defect each or are there multiple defects per unit? My point is still apples to apples. I agree that chasing a single solder bridge for unusual lengths of time is counter productive. But what happens with recurring defects across all assemblies. If you find the fix for one, betcha that fix works for most the others. Focusing on the highest rate recurring defect affects the random, less occuring defects.

...and it still don't matter if its a $100 board or a $10 board. Saving money is still saving money. Consider the 10X rule: cost of component placement = $0.04 cost of pre-reflow repair = $0.40 cost of post reflow repair = $4.00 cost of post test repair = $40.00 cost of post Final QA repair = $400.00

fixing defects early (regardless of resale value of the PCA) saves you time and money. Keep the bottle neck and bone pile out of Test. RMA's can get real expensive when you lose a customer or worse yet gain a bad reputation. Then it don't matter what the initial value of the PCA was, it ain't worth doodly if ya can't sell it.

just my humble opinion

reply »

#21309

50 PPM | 23 August, 2002

thats the problem with numbers... how accurate are they anyway? Can someone tell me how to collect data and assure me that its not someones biased opinion?

reply »

dragonslayr

#21396

50 PPM | 3 September, 2002

Jersbo- I will refer you to the late, Great Mark Twain, who once wrote:

"There are 3 kinds of lies in this world -

Lies

Damn Lies and

Statistics"

reply »

Melo_Guy

#21407

50 PPM | 4 September, 2002

I like that "Lies, dam lies and statistics." thats a riot, only because its SO TRUE. Ill remember that i'm sure ill be able to use it sometime. My company does a lot of those graphs and charts that all relate to things that make no sense. I'm in customer service not QC (quality control)

reply »

#21408

50 PPM | 4 September, 2002

We fondly refer to those charts as "wall art".

reply »

Plasma Prior to Conformal Coating

Heller Industries Mark 5获得了 2015年中国SMT远见奖 - 回焊炉领域