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

Printed Circuit Board Assembly & PCB Design Forum

SMT electronics assembly manufacturing forum.


Acceptable standards

K

#34512

Acceptable standards | 26 May, 2005

Hi Folks,

My boss has me chasing mischuck figures and I need some stats for a report. All of my machines are humming at over 99% efficiency so theres no major loss there, but still there is a huge descrepancy between stock after its been through our process. Typical to change all feeders on all machines up to 3 or 4 times a day so with all the handling theres a potential for loss. Problem is when these parts are counted back on their return to stores, we've lost about 10-20% of the components, (mostly discrete). Mainly down to operators loosing parts setting up/tearing down the feeders etc. My boss is looking for industry standards to compare our process to others. Any of you care to give input on your average component loss?

K

reply »


RDR

#34513

Acceptable standards | 26 May, 2005

This 10 - 20% loss, is it 20% of 10 parts or 10,000? How many parts are lost from a "Perfect feeder teardown"? For example our Samsung 8mm feeders lose 8 components on 4mm pitch every time the cover is released. So if we had an 8 piece run that required 1 per we would have a 100% attrition (8 parts placed, 8 parts lost in the teardown). Your numbers do seem very high however. I would take a serious look at your operators and see what they are doing along with stockroom personel. People may be cutting the ends off of the tape so you have to strip them back to load the feeder etc... < 1% machine fallout is entirtely acceptable from everywhere I have been we use 2% for 0603 and larger and 5% for 0402 and smaller. Stick and tray parts we leave at 0% percent although that doesn't always happen.

reply »

#34517

Acceptable standards | 26 May, 2005

How are you counting parts back into stock? If it is an estimation then that could be your problem. One place I worked, the stockroom estimates were usually double what the lead hand would estimate. If you are using a counter then your rate is probably high because you are using small quantities per run.

We think effiencieny is a dangerous word. Always always always ask what it means for each specific case and how it is arrived at. It lulls people into a false sense of security. Trust you eyes more than the machine efficiency numbers, are there loose parts in or around the machine?

Are you keeping track of losses for the month? Is it possible you keep "losing" the same parts. I.E. you are not really losing 10-20% overall but are using a tracking method that leads to exageration on a single run.

If you are an OEM or a CM using your own part numbers you might want to look at dedicated feeders. You are probably loading and unloading 100N 0603 3 or 4 times a day.

reply »

PeteP

#34525

Acceptable standards | 26 May, 2005

An elusive cause can be Bill of Material Quantity errors vs. actual placements. Check the programs to the Bills and designators on the BOM to Qty. if the postings above do not net results.

reply »

#34536

Acceptable standards | 26 May, 2005

I have never seen a published "industry standard". What I've experienced with Fuji CP and IP is less than 1%. If feeders are changed over 3-4 times per day that will certainly add to this number. Not to mention sloppy operators. Some machines consume less parts in a feeder changeover than others. Your material planner(s) should be factoring in what we call "shrinkage" when they release material to production, this is where component loss is accounted for. But 10-20% is a massive amount for standard qtys, even with 3-4 changeovers per feeder per day, unless you are talking about using only hundreds of compnents per change, because some feeders waste up to 5 components per change.

reply »

K

#34576

Acceptable standards | 27 May, 2005

Thanks for your replies, to go a little deeper..... The sample run I chose was a very small run.... 3 PCB's, 340 parts in total required for the 3 boards. Machine mischuck info gave a total of 360 parts used, 20 lost to various reasons, recognition/no pick which would be our typical loss for a first board thru (i'e feeders settling in). Our stores people count the reels on a "county-S" physical part counting machine which is pretty accurate. The sample was counted before it came out of stores and again on return. 560 parts were used from the kit. telling me there was about 200 parts "lost" outside of the machines scope. 45 feeders used in the set up, so about 5 parts lost per feeder (not machine related). I did check inside the machine and there were no parts dropped in there. I think I'm fooling myself here though as the sample was a really small run of 3 PCB's. If I built 30 of the same product, the percentage of lost parts against qty of parts used would be far less. I'm having problems trying to come up with a way of quantifying this though. If I was running the same product 24/7 life would be easy but no such luck. We run a wide range which changes both in parts used and qty's produced from week to week. Need to run a larger batch.

reply »

#34581

Acceptable standards | 27 May, 2005

How much does each part costs? balance that against the cost of counting and recounting and the time you are spending on this. Unless each parts costs dozens of $$$ you are wasting dollars to save pennies.

Just my 2 cents

reply »

#34609

Acceptable standards | 28 May, 2005

We also face same problem . It is difficult to make understand pruchase & store department. They order exact qty & issue also exact qty , later stage we have to take as process loss. They always worry about small cost but never calculte production cost. Some time lot of Production loss for counting issue only. We do have lot of changeover every day. On an average we loose 1~2 % ( for all small size of component ).

reply »

#34611

Acceptable standards | 28 May, 2005

I agree with Claude's comment here. As the good old 80/20 rule suggests, 80% of your lost $$ may be coming from 20% of the parts on your list.

So, I would spend 80% of my efforts to sort the list by "loss in $$" and try to reduce the loss of such.

And to me, even if you find an industry standard percentage, that'll translate to different $ figures for everyone.

Erhan

reply »

Stephen

#34620

Acceptable standards | 29 May, 2005

Get'em on the floor. But invite themn nicely. Most people are interested in what people in the other parts of the company do. On the other hand if you say something like, you have to come and take a look so you numbskulls can know what you are doing wrong. That probably won't work. And maybe invite someone from accounting as well. They would see both sides better. And make sure they understand that you understand, that if a job is for 200 but you use 205 BGA's that cost $1,500.00 each there better be a good explaination.

A pet peeve of mine is the supply chain people prefer serialized part numbers but people on the floor prefer coded part numbers. And guess who usually wins.

reply »


URL

#34650

Acceptable standards | 31 May, 2005

Don't forget that loading a feeder will waste parts, depending on the feeder. A lot of bean counters over look this. But then again, they'll spen $100 chasing pennies around.

reply »

K

#34664

Acceptable standards | 31 May, 2005

All good points and thank you for your replies. Most of the components lost are discretes. The problem I have is my operators gather any SOIC "lost components" and return them to stores in a small metallised bag. This in turn goes into a 16" X 12" X 12" cardboard box. My stores people have 3 years worth of components which is now highlighted as a "problem" based on the fact that we have a box nearly full of lost components. If the operators swept them into a bin instead of keeping them, the loss clearly would not be as dramatic. I personally would expect a certain amount of loss due to handling. I need facts and figures though for a second report.... Do any of you recover lost valuable parts and return them into stock? Or do you have scrap factors / shrinkage to deal with the loss..... and discard the components? If you do, what would your scrap factor typically be?

I have already put a good bit of time into crunching numbers and what not trying to quantify this issue. I'll bet my time (hourly rate) has already far exceeded the value of the components in the box........

reply »


URL

#34676

Acceptable standards | 1 June, 2005

We return materials to stock all the time. Have also noticed some repair people like to pilver the reject bin every so often so they have a stock pile of their own. This too adds up on discretes.

reply »

PW

#34686

Acceptable standards | 2 June, 2005

K, aim to prevent return of rejected components back into stores. Consider potential damage caused through handling post pick and the effect of your product quality in using these components in the future.

The best place to address this problem is at the point of rejection and this means the operator and or the line technician. Get the mindset sorted with a touch of operator accountability and sell to your boss a balance of reduced attrition rates and productivity/machine efficiency without compromise to the finished product assembly quality. Ensure any engineering issues are resolved in the most expedient manner. Your relationship with the operator can make a difference in their ownership of this as an issue.

reply »

 Reflow System

Plasma prior Conformal Coating