Parts Ordering and Core Return Handling
Purpose
A core deposit is not a fee, it is a loan against a part the distributor expects back, on a clock that starts the day the replacement ships. Miss the window and the deposit converts to a charge on a job that has already been invoiced and closed, so the loss lands in a month where nothing explains it. Order the wrong number and the loss is worse, because a non-stock electronic control is usually non-returnable, and the customer is holding a second-visit date that now cannot be kept.
This SOP treats the core as part of the repair rather than as paperwork. The failed compressor, control board or motor leaves the customer's house in a box with a return label already generated, and the tech who removed it is the person who packs it, because that is the only moment when the part, the job and the person are in the same place.
Scope
Covers ordering a part against an open job, receiving it, tracking the core back to the distributor, and closing the deposit. Applies to any part carrying a core deposit, which in this trade is mostly compressors, sealed-system components sold as assemblies, electronic control boards, drive motors and some inverter boards.
Does not cover warranty claim filing, which is owned by the warranty claim SOP, and does not cover stock replenishment of consumables such as belts, hoses, thermal fuses and gaskets. Those run on a min and max reorder point off the truck inventory, and mixing them into this procedure is what turns a core queue into an unread list.
Roles and responsibilities
| Role | Owns | Hands off |
|---|---|---|
| Technician | Model and serial capture, part identification, removing and packing the core the same visit | Hands the office the packed core with the job number written on the box, not a verbal "core is in the van" |
| Parts / office | Lookup against serial, ordering, RMA generation, receiving, core queue | Hands the tech a received-in-hand confirmation before any second visit is booked |
| Warehouse or whoever opens the boxes | Matching received parts to jobs on the day they land | Hands the office an exception list of anything received that no open job claims |
| Owner / manager | Weekly core queue review and the write-off decision | Signs off any deposit written off, so a repeat failure is visible rather than absorbed |
The handoff that decides whether this works is the first one. A core that stays in a van becomes a core nobody can find in three weeks, and every shop that has run this badly has a shelf of unidentifiable boards.
Procedure
Look the part up against the full serial number, not against the model alone, and read the supersession chain to its current end. Acceptance: a part number the manufacturer's current lookup returns for that exact serial, with any superseded numbers noted. Wrong looks like a number pulled from an old exploded diagram or from a marketplace listing, which is how a shop ends up with a board that fits the connector and not the firmware. Stop rule: the lookup returns nothing for that serial, do not guess from the model, escalate to the brand parts line and record the case reference. Desk step, no hazard, and it is the step that prevents the two most expensive failures in this SOP.
Confirm before ordering whether the part carries a core deposit, whether it is returnable if unused, and what the return window is, and write all three into the job. Acceptance: three values recorded, taken from the distributor's terms for that order rather than from memory. Wrong looks like assuming all boards are non-returnable and all compressors carry cores, both of which are usually true and neither of which is universally true. Stop rule: terms unclear, ask before ordering, because a non-returnable part ordered on an unconfirmed diagnosis is a loss the shop absorbs in full.
Order with the job number in the distributor's reference field so the box arrives labelled with the job it belongs to. Acceptance: order confirmation attached to the job, reference field populated, promised ship date recorded separately from any promised delivery date. Wrong looks like a bulk order covering four jobs on one purchase order, which arrives as one carton and gets sorted by whoever is free. Stop rule: if the distributor cannot carry a reference, write the job number on the packing slip at receipt before the box leaves the receiving bench.
Generate the return authorization for the core at the time of ordering, not at the time of removal. Acceptance: an RMA number and a printed return label held with the job, and the return window's expiry date recorded as a date rather than as a duration. Wrong looks like an RMA requested after the tech has already removed the part, which stalls the core in a van for the days it takes to come back. Stop rule: distributor will not issue an RMA until the replacement ships, record the expiry date as unknown, flag the job, and set the queue to chase it on the ship date.
Receive the part against the job on the day it lands, and confirm the number on the part against the number ordered before anything is booked. Acceptance: part number on the item matches the order, quantity matches, and the job is stamped received with a date. Wrong looks like a partial shipment booked as complete, so a second visit is scheduled against a kit missing its harness. Stop rule: mismatch or shortfall, do not mark received, do not book the visit, and open the shortage with the distributor the same day.
Remove the failed part at the machine and pack it into the box the new part arrived in, at the machine, before the appliance is reassembled. Acceptance: core in the box, job number written on the outside, return label attached, box in the van in a designated core position rather than loose. Wrong looks like a compressor sitting on a shop floor with no marking. Stop rule: if the box is damaged or absent, pack the core anyway in equivalent packaging and note it, because a core rejected for packaging is still a rejected core. Hazards, and they are specific per part: a removed compressor holds residual refrigerant and oil, so cap both service stubs before it goes in the box and recover the charge under 40 CFR part 82 subpart F rather than venting it, and treat an R-600a isobutane charge as flammable, with the area ventilated and ignition sources controlled before the system is opened. A removed control board can retain charge in its supply capacitors, so allow the manufacturer's stated dwell after disconnection before handling it and hold it by its edges in an antistatic bag, because a board damaged by handling is a board the distributor can refuse.
Get the core out of the van and into the outbound carrier flow within one working day of removal. Acceptance: a carrier tracking number recorded against the job. Wrong looks like a full core shelf, which is the physical form of the same failure. Stop rule: a core still in a van at the end of the second day gets escalated to the manager, not left to the tech's next quiet morning.
Close the deposit when the credit lands, and reconcile the credit amount against the deposit charged rather than assuming they match. Acceptance: credit posted, job closed, core queue entry cleared. Wrong looks like a partial credit accepted silently because the distributor downgraded the core for damage or missing components. Stop rule: partial or absent credit, dispute it against the tracking number and the packing photograph while the reference is fresh, and escalate to the manager if the distributor holds.
Review the core queue weekly against expiry dates and act on anything inside the last third of its window. Acceptance: no open core past its expiry date, and every core inside the last third of its window either shipped or explained. Wrong looks like a queue sorted by date opened rather than by date expiring, which buries the urgent ones under the old ones. Stop rule: a deposit that will expire before the core can physically get there is written off deliberately by the manager with a reason recorded, rather than discovered later.
The record this produces
One core ledger line per deposit, living on the job and visible in a single queue:
- Part identity: ordered number, superseded number where one applies, the serial the lookup was run against.
- Terms: deposit applies yes or no, returnable if unused yes or no, return window as an expiry date.
- Movement: order date, ship date, received date, removal date, carrier tracking number, credit date.
- Evidence: photograph of the packed core with the job number visible, and the packing slip.
The parts clerk reads the queue weekly. The manager reads the write-off line monthly, because a rising write-off count is almost never a distributor problem and almost always a van problem. The bookkeeper reads the credit dates when a supplier statement does not reconcile, and without the tracking numbers that reconciliation is guesswork.
Worked pass: a compressor core that missed its first window and was recovered
A no-cool call on a top-mount refrigerator diagnosed as a failed compressor, sealed-system parts covered by the brand's longer term, labor customer-pay. Step 1 returned a current compressor number against the serial with one supersession behind it. Step 2 confirmed a core deposit applied and the distributor's stated return window was 30 days from the ship date, recorded as a hard expiry date rather than as "30 days". Step 3 ordered with the job number in the reference field. Step 4 could not produce an RMA at order time because that distributor issues them only on shipment, so the job was flagged and the queue set to chase on the ship date. That is the exception path in step 4 doing its job rather than being skipped.
The part shipped two days later. The RMA and label came with it and were printed into the job, and the expiry date was set from the ship date. Step 5 received it clean, number matched, and the second visit was booked. Step 6 ran correctly at the machine: charge recovered by a Type I certified tech, stubs capped, old compressor into the shipping box, job number on the outside, label attached.
Step 7 is where the pass fails. The core went into the van and stayed there. The tech worked four days of calls, the core shelf position was occupied by another job's box, and the compressor rode around under a set of drain hoses. The step 7 stop rule fired at the end of the second day: the queue flagged a core with a tracking number field still empty. The manager escalated rather than waiting, the tech offloaded that afternoon, and the core shipped on day three of a window measured from a ship date that was already four days old. It arrived inside the window with room to spare, and the credit posted in full.
Two things worth taking from that. The deposit was recovered not because the tech remembered but because the queue was sorted on an expiry date recorded as a date, which is why step 2 insists on converting the duration into one. And the failure was physical, not clerical: a designated core position in the van that another box was occupying. The corrective action was a second position, not a reminder.
What would change the answer. On a non-stock electronic board rather than a compressor, step 2's returnable field usually reads no, which moves the whole risk forward to step 1: an unconfirmed diagnosis on a non-returnable part is the one case where the correct move is a second diagnostic visit rather than an order. And on a warranty repair, the core often has to go back to the manufacturer rather than to the distributor, on the manufacturer's own reference rather than an RMA, which changes step 4 and nothing else.
References
- 40 CFR part 82 subpart F (refrigerant recovery and technician certification; Type I covers small appliances)
- Distributor terms of sale for each supplier you buy from, which is where core deposit, return window and restocking terms are actually defined
- Manufacturer parts lookup for the brand, run against the serial number rather than the model
- See related: warranty-authorized repair claim process SOP; second-visit scheduling when parts are ordered SOP