Testing RFID tags on metal and liquid goods before a full count

RFID looks effortless in demonstrations: a reader is pointed at a pallet and the screen fills with numbers. Real warehouses are less cooperative. Steel parts, cans of fluid, bottles of coolant and foil-wrapped electronics all interfere with radio signals, and a tag that reads perfectly on a cardboard box can fail completely when stuck on a metal housing. The remedy is not pessimism but a pilot test, done before the whole warehouse is committed to RFID, and short enough to run in half a day.

Why metal and liquids cause trouble

An RFID tag works by receiving energy from the reader and answering with a small signal. Several things can disturb that exchange:

• Metal reflects radio waves and can detune a tag that sits flat against it, so that it no longer responds at the expected frequency or distance.
• Liquids absorb energy. A tag placed on a full bottle or a tank can lose most of its range.
• Dense stacks shield the tags inside. A pallet of metal-cased parts may read well on the outside layer and poorly in the middle.
• Neighbouring tags can disturb each other if they sit too close together.

None of this means RFID cannot be used with metal or liquids. It means that tag type, tag position and stacking pattern are decisive, and that they are found by trying.

The same item can also behave differently from one batch to the next: a thicker can wall, a different cap or a foil layer inside the carton can change the result. That is one more reason to test the actual goods in your warehouse, and not a sample picked from a catalogue.

What to test

Test the factors that you can actually change:

• Tag type: a standard paper-faced tag against a tag designed for metal or one with a spacer.
• Position: top, side, or the corner of the carton; flat against the item or lifted off it.
• Orientation: relative to the reader and to the item.
• Reader settings: power and antenna direction.
• Distance: the real distance at which workers will stand, not the best distance in a demonstration.

A half-day test plan

1. Choose five representative item types: for example a steel casting, a boxed metal part, a bottle or can of liquid, a foil-wrapped electronic item and a plain carton as a reference.
2. For each type, prepare 20 units and test three tag positions. Mark each tag with a number.
3. Read each group from 1, 3 and 5 metres, three times each, and note how many of the 20 are read each time.
4. Then repeat in the real arrangement: stacked on a rack or a pallet, in the aisle where the count will happen, with neighbouring goods present.
5. Record the read rate for every combination, the tags that were missed and the tags that were read twice.

Photograph every tag position and write it on the test sheet. The people who attach the tags later must be able to reproduce exactly what worked, and not what they remember of it.

Agree in advance what counts as a pass. For illustration, you might require that at least 99 of every 100 tags are read within three passes at the working distance. Items that fall short do not have to be abandoned: they move to a different tag position, a different tag type or simply to barcode scanning.

What to do when a group fails

• Try a spacer or a tag designed for metal, and compare the cost with the benefit.
• Put the tag on the packaging side that faces away from the metal, or on a non-metal handle or plate.
• Tag the carton instead of the item, if the carton holds a known quantity.
• Use barcodes for that group and keep RFID for the groups where it works. A mixed method is normal.

Summary

A half-day pilot costs little compared with a count that must be redone because half the tags did not read. Smart Inventory tests tag placement on metal and liquid goods first, records the read rates and then decides, group by group, between RFID and barcode scanning, using UHF readers where they perform and handheld barcode scanners where they do not.