• Discrete Manufacturing
  • Warehouse Operations


  • Spotlight Server 2.0.1, Windows Svr, ASP.NET, Microsoft SQL Server Cluster
  • Handheld UHF RFID readers
  • Fixed UHF RFID readers at dock door portals and selected locations on the production line
  • Windows touch PC vehicle mount computers on fork lifts
  • UHF RFID tags on air conditioner units, coil carts, and fork lifts


  • Specification, design, implementation and deployment services
  • Annual support


Carrier has a 900,000 square foot manufacturing facility in Collierville, Tennessee that produces air conditioner and heating units for commercial and residential customers. This facility undertook a project with S3Edge to use RFID technology to streamline and improve its finished goods shipping processes.

Carrier uses just in time supply chain and streamlined manufacturing processes to receive sheet metal and other component parts at one end of the facility and deliver finished air conditioner units at the other end.  The finished air conditioners are immediately loaded on trucks and shipped to Carriers’ finished goods warehouses.  The facility produces more than one million air conditioning and heating units per year.

There are 9 separate production lines which start with a base plate being placed on the production line and the completed air conditioner is built up as it passes through a number of stations where parts are added, it is tested and boxed and ultimately stacked as finished goods at the end of the production line.  The completed units are then moved through one of 18 dock doors and loaded into trucks for shipment to the warehouse.

The focus of the RFID tracking tracking system initiative was to:

    • improve quality, efficiency and productivity during the finished goods shipping process
    • improve reliability of unit identification during the assembly process
    • reduce errors associated with attaching wrong components during the assembly process

S3Edge’s Spotlight software was customized and deployed with a network of fixed RFID scanners, mobile hand-held scanners, and fork-lift vehicle mount computers to address Carrier’s business needs. As up-time and reliability was a critical part of this application, the new system was developed as a High-Availability cluster-aware solution.


The benefits of the system can be grouped into three major categories:

    • Labor savings – Prior to deploying the RFID tracking system, a dedicated Carrier employee was stationed at each pair of dock doors (9 full time employees for each of potentially 3 shifts).  The sole purpose of this employee was to barcode scan each air conditioner as it was placed on the truck for shipment to the warehouse. The RFID system eliminated the need for these manual barcode scans and therefore roughly 9 full time staff equivalents per shift.
    • Improved reliability of unit identification during the assembly process – Units sometimes fail the final test and need to be temporarily pulled off the production line for rework. The RFID system allows these units to be easily identified and tracked.  It also prevents these units from erroneously being shipped to the finish goods warehouse until the problem is corrected.
    • Reduction of errors associated with attaching wrong components during assembly – The RFID system was used to error-proof the production line at two key stations, the coil forming station and the  boxing station.  Prior to having the RFID tracking system, it was common that errors were made at these two stations when a production line transitioned from producing one model to the next.  The wrong coil or the wrong box would sometimes be used at the transition point were the new models appeared on the production line.


The network architecture of the turnkey solution developed is shown in the figure and includes:

    • A central Spotlight server that controls a network of fixed and handheld UHF RFID readers and vehicle mount computers. The Spotlight server is also connected to and exchanges information with three existing backend manufacturing execution systems of record.
    • Fixed RFID readers are installed at the base lay in station, the coil forming station and the boxing station on each production line. Fixed readers are also installed at each of the 18 dock doors in a direction sensing portal configuration.
    • WiFi connected handheld RFID readers are used to scan, view and update the state of work in process and units being reworked.
    • WiFi connected touch-screen vehicle mount PCs are installed on each fork lift along with a UHF RFID tag that identifies the fork lift.
    • Each of the shop carts used to move heat exchange coils to the coil forming station also have a RFID tag attached that uniquely identify them.
    • Finely, a three way (UHF RFID, barcode, and eye-readable) label is commissioned, printed and attached to the base-plate of each unit as it is initially placed on a production line.  The label contains the ID the backend manufacturing execution systems assigned to the unit that is used to track the unit through out the manufacturing process.
Carrier solution overview


A 3-way label is printed by the manufacturing execution system and attached to the base plate of each unit as it is placed on the production line.  The 3-way label includes an RFID tag, a bar code, and an eye-readable unit ID.  The manufacturing execution system tracks the model type and other information about the unit including the bar code and RFID tag ID assigned to the unit.  From this point forward, the unit is tracked using the identification numbers associated with the label attached to the base plate.


A RFID reader is installed at the coil forming station with three antenna is shown in the figure to the right. As each air conditioner arrives at the coil forming station, its RFID tag is scanned along with the RFID tags on the two coil carts from which heat exchange coils are currently being pulled.  The Spotlight server queries the manufacturing execution system to confirm that the coil about to be used is compatible with the model type of the air conditioner base plate just scanned.  A stack light alerts the coil forming station operators if a mismatch exists and direct them swap the current coil carts with coils that match the unit being processed. This prevents an incorrect coil from being installed on a unit which otherwise requires significant rework or possibly sapping the entire unit.

Coil forming station error proofing

A similar error proofing is implemented later on the production line at the boxing station.  An RFID reader is installed at each boxing station.  As each air conditioner arrives at the boxing station, its RFID tag is scanned.  The Spotlight server queries the manufacturing execution system to determine the type of unit being processed and displays model information on a monitor to ensure the boxing station operator places the correct box over the unit.  Again this prevents errors which if not caught immediately require significant exception handling processing if discovered later in the supply change (e.g, at an end-user customer site).


As completed air conditioners arrive at the end of a production line, they’re moved through one of the 18 dock doors and loaded onto a waiting truck.  The units are moved using an RFID tagged “clamper” fork lift.  Each “clamper” forklift is capable of picking up and moving multiple RFID tagged air conditioner units at the same time.

As the fork lift is about to enter the truck, the operator pauses momentarily to confirm the correct units are being loaded. A direction sensing RFID reader at the dock door scans the RFID tags on all the air conditioners being moved as well as the RFID tag attached to the fork lift.  The Spotlight server queries the manufacturing execution system with the IDs of the units scanned to confirm they’re all ready to be shipped.  Spotlight updates the vehicle mount computer display on the fork lift (identified by the fork lift RFID tag just scanned) to display the number of air conditioners detected and a confirmation all the scanned units are ready to be shipped.

If a unit requiring rework is detected, a red screen alert is displayed indicating the operator has picked up an incorrect unit and needs to correct the error before proceeding.  Similarly, if the number of units detected does not match the actual number of units currently on the fork lift, the operator knows there is a problem scanning one or more of the units that must be corrected prior to proceeding to load the truck.

If all the units scanned are ready to ship and the number scanned matches the actual number on the fork lift (indicated by a green screen and a matching unit count), the operator confirms the load operation on the vehicle mount computer touch screen.  This confirmation serves as a change-of-custody transfer point from the production factory into finished goods.  The Spotlight server reports this confirmation event to the manufacturing execution system which allows the manufacturing execution system to automatically generate a shipping manifest including the unit IDs of all units that have been loaded on the truck.

A short video of a fork lift in action loading finished goods at one of the dock doors is included here.

One of the challenges in this system was its high metal content environment which makes it difficult to reliably scan the RFID tags of interested and at the same time not read nearby unrelated tags.  Tuning was required at the dock doors ensure only the desired units were scanned while filtering out stray reads of nearby units.

WiFi coverage is also not guaranteed in all locations in the factory so the vehicle mount and handheld computers need to cache a local copy of the information where possible to allow them to continue operation even when connectivity to the server is not available for a period of time.

Finely, one of the more interesting aspects of this deployment is that the RFID system is a mission critical application that must be available 24 X 7 since the operation of production facility is dependent on it being available.  This motivated having the software be “cluster-aware” so that it can be installed on a server configured with redundant hardware to automatically recovered from hardware failures.