Hello guys I hope you can help me, I had worked with machines that usually tell you how many components are dropped off, now I have some very old machines without such information, any idea how to measure if attrition rates are normal? I expect to be arround 3-5%.
PS.-We are a very low volume & high mix, many changeovers a day.
If 3-5% of the attempted parts placed are dropped, you have a serious problem...that's 1 part dropped for every 20 or 30 placements. Does it mean that it drops the part in the scrap bin and then successfully places the same part or would the board come out of the machine with missing parts?
We have an 11 year old Quad and we see a missed part out of the machine once every few months. You are undoubtedly running many more boards than we are but 3-5% is totally unacceptable even if it were 3-5% of the assemblies had a missing part. Do all of your machines have this problem or just one? You need to find the problem or get better equipment. There is no sense in creating assemblies with this many defects.
It means the machine places another component, no missing PCBs are built. As an update I have measured this % in the newest machine and it is below 1%, but in our worst machine (old quad) it is as high as 9% for a small batch of 10 boards (~15 comp per board), I know it is high but I had to count each component the machine drops, also the operator does not even know when the machine is dropping such quantity of components, what controls you may suggest to implement? I can not count every component and tell the operator to stop the machine... thanks!
Now knowing that its a Quad, I'm assuming it has QuadAlign which spins the part it just picked in a beam of light and a sensor measures the length, width and height of each part. If it falls outside of a defined range, it will throw it in the dump bin. On the parts it tends to throw out, I'd first make sure the tolerance is wide enough to cover part variation. If the part dimensions were inputted from simply measuring the part with a pair of calipers and the tolerance is zero or a very low value, then a slightly larger part will get rejected even though it meets the manufacturer's spec. If it picks another part and places it successfully this might be the reason for the rejects. If your old Quad does not have QuadAlign, then something else might be happening which I have little advice on and it may be time to upgrade (I don't think I'd want to run a Quad without QuadAlign).
Thanks for the info, I will check tolerances and the quadaling option. As an update I ran this past week and machine was arround 5% for a very small lot size, for a larger size lot it was less than 2%, Now I know it still higher than our newest machine (assembleon ~0.3%) the expectation is to bring the old quads to such level.
Attrition... how to measure it?| 15 February, 2012
The only way you will be able to measure it is to count the reel before you put it on the machine, count them when they come off and deduct the amount expected to be placed. you will then have the result of attrition on the machine.
Pick the parts from the drop tray/bucket and hover the inside of the machine/between feeders and count it manually.
Both not really credible methods - with this kit your using i would just get the operators to monitor the amount of parts found in the machine after each run and bag em (low value)then weigh it. That can be you KPI method!?
Even machines that count (I have 7 Europlacers and one Samsung Quad) and use the PROMON software to monitor waste and it's very good for measuring KPI's and monitoring machine faults, but even this is not absolute.
My attrition rate per machine is usually 0.78% - it makes the owners happy to see that rate! =)