Lot Number Part Life Traceability

Hello all,

First time posting here, have a bit of a complex issue:

We are in the aerospace industry, and we require full traceability for the entire life of any given part. I realize this can be achieved by serializing all of our parts (including purchased parts) and using the serial matching functionality.

That said, we would like to avoid using full serialization if possible. Has anyone using Epicor needed the same tracebility and achieved this thru means other than serialization?

Specifically, we would need to know which LOT#s (purchased parts or assemblies) were used to make which parent part (piece specific). My apologies if this is not a clear enough explanation, any help is greatly appreciated!

Nate

We make Food Packaging and traceability by Lot is critical for meeting customer and food safety requirements. For this purpose we use Lot control of our raw materials and finished goods. This helps in trouble shooting also for customer complaints. If serializing is not required lot tracking provides full traceability.

In our case we have made a simple lot number. We use the date as the lot number. So all products made on a date have the same lot number say for today it will be 190527. The date as lot also helps the customers maintain FIFO stock where needed, and identifying old stock is very intuitive.

For traceability we use Lot tracker and Part/Lot Where used report.

Thanks

Vinay Kamboj

I strongly recommend using the supplierā€™s lot number - usually specified on their cert of conformance - for purchased parts.

This adds the benefit of automatically tracing your product back to the vendors lot.

A downside to assigning your own lot number based on the the date you received it, is that a supplier could have made enough for multiple of your PO releases.

If you a assigned them lots upon receipt date, you could have two different lot numbers that really segregate a single lot. Vendor lot ā€˜101ā€™ for 100 pcs received in 2 lots of 50 gives you your lot ā€˜190526ā€™ (50 pcs) and lot ā€˜190530ā€™ (50 pcs)

If you did analysis and found an issue with products made with your lot ā€˜Aā€™, you might only recall 50 of those 100 products.

We use supplierā€™s lot for purchases. When I mentioned date it is only for what we manufacture.

Vinay Kamboj

Thanks for the suggestions guys - much appreciated!

We currently generate an internal lot number upon receipt, and also record the supplierā€™s lot or ā€˜heatā€™ number.

Our issue is when more than one material lot number is used to fill job requirements, we lose traceability for an individual piece on the job.

For example:
Job qty 10pcs
5 pcs of material with lot number 123
5 pcs of material with lot number 456

When the job is complete, any of my 10 finished pcs could have been made with material from either of these 2 lot numbers. As far as I can tell, the only way for a direct match up is thru serialization. We were just hoping to avoid that route if possibleā€¦

Thatā€™s the down side of lot numbers.

You have to weigh the risk-reward of using lots vs s/n.

The reward of using lot #'s is less overhead for tracking. The risk is youā€™d have to recall all of any produced lot that used even just one of the affected component lot.

Sorry if I focus on recalls. We make a product used in explosive atmospheres, and if a supplier notified us that a specific lot of a component was not to spec, weā€™d have to notify the customer (i.e. recall). That said, we still use lot #'s as the likelihood is low, and weā€™d be okay with recalling more than might actually be affected.

Thanks Calvin, that is exactly the issue we are trying to circumvent.

Currently, our certs are put together manually (by tracing the life of the part via job traveler) We would like to automate this process when we go live with Epicor. Seems like the only way to do so is with serial tracking.

I really appreciate your help sir!

What you can try in such cases is to report quantity two times of 5 each. The sequence and timing of transactions should clarify which lot is used. Another could by writing the same in labor notes or using a UD field in labor entry.

Vinay Kamboj

My former employer used 100% Lot control for tracing. Purchased items used the Receiver number as the lot number. Manufactured parts used the Datecode as the lot number. For high-end customers we had rules requiring ā€œno split lotsā€ meaning that all materials had to be from the same lot. This sometimes meant that we had to have smaller jobs, OR we had to purchase/make more parts to make a larger lot.
One reason we chose to use our own lot number instead of the supplier was better tracabilityā€¦ for example, we had Epoxy that was purchased. Each can would get its own lot number, even though they may have the same mfg lot. This was because each can has its own life. Each can is opened at a different time, and may have had its lid off longer.
Same rule for Wire. We made very fine wire devices. Each spool of wire had its own lot number. This was in case someone dropped the spool (thus damaging the wire). This may go undetected until final inspection of the finished product, but now we needed to trace back to all other lots made with the same spool.
In E10, you CAN track the MFG lot number in addition to your own lot number. This way you could have lots A, B, and C, that all point to the MFG Lot number Xā€¦ so if your suppleir does a recall of their lot, you can still trace back to your lots and where they were used.

3 Likes

Thanks Tim. The issue is the exact traceability for a specific manufactured ā€˜pieceā€™. So if we have a non conformance against 1pc of a batch of 10pcs, and multiple lots of the same material were used on the job, we canā€™t tell which lot was used to make that 1 specific piece.

and THAT is why we had rules about ā€œno split lotsā€ for manufacturing. If that is important, then you need to apply the same rule.
FYI, I had a customer who wanted to ā€œserial trackā€ wireā€¦ the problem with that concept is that you can only serial track items that are zero decimal UOMs (ieā€¦ EACH)ā€¦ and you cannot issue or consume a partial spool, or partial length of an EACH. Serial tracking only works for discrete units.
But Serial tracking COULD work if you are using discrete items in your manufacturingā€¦ just be careful what you wish for hereā€¦ if you have a job for 10 pieces, and all the materials are serial tracked, you will then need to do ā€œLower Level Serial Matchingā€ on each component. And the parents would also all need to be serial tracked as wellā€¦ there is alot to be done to get lower level serial tracking.

I think we will need to apply the same rule. We are planning on inventorying our extrusion in inches so obviously this will not work.

Really appreciate everyoneā€™s help. Iā€™ll post back when we come up with a solution

@timshuwy
Just to complicate things (because thatā€™s the kind of guy I am), do you require only using the specific resources for lot controlled jobs too?

Itā€™s straightforward to think about the lineage of the components that go into a final assembly. But there are cases where you might have multiple resources (like two different CNC stations) and want to use both to get the job out faster (or to not have one sitting idle). The ultra anal retentive person would make that as two separate jobs, with two different resulting lot codes.

And that doesnā€™t even take into account other resources like tools (drill bits, end mills, punch dies, etcā€¦) or the QC equipment (gauges, go-no-go gauges, calipers, DVMā€™s, etcā€¦)

We report in multiple packages on jobs using report quantity for every package on that job. Each time they go into report quantity a unique UD Table record is created and then they scan the source purchased lot and that is stamped on the UD record also.

If they send that unique item to another process then a new record can be created and the original source puchased lot is held along with the UD record source.

I also store vendor tag, heat, chemical properties at receipt time into another UD table.
When a pack slip prints they get complete material certifications for every lot on the pack.

funny you mention thatā€¦ we had one product that was manufactured on 4 different winding machines. These machines ā€œWet woundā€ some flat wire into a slinky size/shape coilā€¦ wet winding meant that the wire went through epoxy before being wound around a mandrel under pressure. Then the part was torqued, and baked.
Each part was stamped with a lot number that told us:

  1. which Wire Spool lot was used
  2. which winding machine was used
  3. which operator was processing the work
  4. The Epoxy lot #
    One spool of wire only made about 10-12 finished coilsā€¦ we could make about 100 parts per day per machine, so we had about 40 lots per day.
    OHā€¦ just ā€œfor funā€ the specifications required us to do a destructive pull test on one unit per wire spoolā€¦ meaning we had at least a 10% loss. If it failed, we had to destroy a second from that same lot. (what a mess).
    THEN we gathered all the data, showing pull strength correlation over time. We could monitor the performance of the machines and people. If a machine started producing bad parts, we knew it within hours when the parts from that machine started failing pull strength.

You hit at least 4 of the 6 Mā€™s of Six Sigmaā€¦

  • Materials (wire and epoxy)
  • Machine (which winder)
  • Man (which operator)
  • Method (inherently captured by the Job that lot was built to.
1 Like

2 others:

  • Mother Nature: We had a cleanroom environmentā€¦ so this was a constant.
  • Measurement: we were constantly gathering dataā€¦ used SPC extensively to monitor output. There were tools/gauges/visual inspections, etc all along the wayā€¦ at the end, we even had random final products each day that were pulled from production for further testing and destruction.
    One might ask ā€œWhat was this unobtanium Object you are makingā€ā€¦ it was a voice coil for the Head Disc Actuator (HDA) on a very high end disk drive manufactured by IBM. It was the ā€œworlds largest and fastestā€ at the time with (if I remember right) 14 gigabytes of storageā€¦ There were two HDAs per drive. The pull strength had to exceed 150 pounds minimum. some of ours exceeded the pull capacity of the tester (1000 pounds).

Clean rooms are intended to be constant - but that doesnā€™t always mean they are. :wink:

We had logs of our cleanrooms (temp, humidity - surprisingly no particle counts). So we could kind of work backwards to derive the ā€œMother Natureā€ variables, if he had to.

1 Like

Hey- itā€™s been couple of months for now and Iā€™m facing similar issue. Did you figured out a way to track from a single job the ā€œspecific pcsā€ comes from ā€œspecific raw material lotā€?

Thanks

Unfortunately this is not doable without splitting your jobs based on material lots. Thats what we have decided to do