A UHF RFID reader can detect a sample tag across a room and still miss tagged items on a nearby shelf. Before replacing the reader, check the tag on the actual item, its position, the surrounding stock and the reader configuration. Read range depends on the combination of reader, antenna, tag and environment. Passive RAIN RFID tags obtain their operating energy from the reader’s radio field. Impinj’s system overview explains how these components work together.
For inventory, the useful question is: can an operator identify the intended items within the available time, while keeping nearby stock out of the result? This guide explains how to investigate that question for a handheld UHF RFID workflow.
Define what a successful read must achieve
Start with the physical area and business task. Counting one shelf, checking a trolley and searching for a particular asset require different outcomes. Write down which items belong in the task, where the operator can stand, whether items can be moved and how much time is available.
Separate three measurements:
- Maximum observed distance: the farthest point at which a particular tag was detected under recorded conditions.
- Target coverage: how many intended tagged items were detected within the test window.
- Unwanted reads: which tags outside the intended task were also detected.
A demonstration of the first measurement does not establish the other two. Ask for the tag, item material, orientation, reader settings and test method behind a quoted distance.
If you are still deciding what to tag and how to identify each item, begin with UHF RFID Tag and Reader Selection Basics.
Why the same reader performs differently across a shelf
The tag must suit the item it is attached to
Test the complete tagged item, including its contents and packaging. A label tested loose on a desk does not reproduce its final mounting conditions.
Metal can alter the radio environment, while water and other liquids can absorb energy and detune tags. Tags designed for mounting on metal or for challenging materials provide options to evaluate; an ordinary label should not be assumed suitable for every surface. GS1 explains these material effects and the role of dedicated tag designs.
For a filled container, compare candidate placements on the filled container. For a metal asset, follow the tag supplier’s mounting instructions. Record the exact tag construction and placement, so a successful sample can be reproduced.
Orientation can change the result
A tag that reads when facing the reader may behave differently when rotated or packed among other items. Impinj’s antenna technology note illustrates how tag designs can have different responses at different angles.
During a test, rotate the item through positions that occur in the real workflow. Record whether another viewing angle or an additional sweep recovers a missed tag. If success depends on manually turning every item, include that handling time in the comparison.
Power and read-zone control need to be considered together
Where the reader permits adjustment, compare supported power settings under the same conditions. Record both the intended items recovered and the unwanted tags detected at each setting. Increasing power is worth evaluating only against both outcomes.
Power should be configured for the deployed reader, antenna and region. Impinj’s guidance on reducing RF emissions describes power and antenna considerations for its equipment. Its particular settings should not be copied to an unrelated device.
If the task requires one deliberately presented item rather than the longest possible range, compare a proximity UHF RFID read pad with the proposed tags and neighbouring items; lower transmit power alone does not establish that selection behaviour.
Reader mode and time in the field also matter
RF modes can trade data rate against tolerance to interference. A mode suited to repeated reads of an easy tag population may not be the best choice for finding difficult tags. Impinj documents this distinction in Reader Modes Made Easy.
Ask which modes the candidate reader actually supports. Compare them using the same population and scan time, then repeat at the operator’s normal pace. A result obtained by holding the reader still indefinitely does not establish performance during a working sweep.
Locate the failure before changing hardware
Treat these observations as diagnostic starting points, not confirmed causes. Where supported, compare the reader’s tag report with what the application displays or records.
| Observation | Possible explanation to investigate | Useful next comparison |
|---|---|---|
| A tag reads loose but fails after attachment | Tag design or placement does not suit the item | Compare approved placements on the complete item |
| A missed tag appears when the item is rotated | Orientation affects the radio link | Repeat the same route with recorded item angles |
| Tags read individually but some disappear in a packed group | Packing, orientation or inventory settings change the result | Compare isolated items with the actual loaded shelf |
| The count rises when nearby stock is added | Tags outside the intended task are being included | Separate known target IDs from known neighboring IDs |
| A reader report contains an ID that the application omits | A filter, mapping or data-delivery step needs checking | Follow that ID from reader output to the application record |
| Report totals are high but some expected IDs are missing | Repeated reports may be masking incomplete coverage | Compare unique expected IDs, not the total number of reports |
| Results vary greatly between operators | Movement, angle, route or scan time may differ | Repeat with a defined route, then test normal working practice |
Filtering also needs an explicit definition. Application filtering can hide a tag from the displayed result without preventing its radio response; reader or protocol filtering works at a different stage. The Impinj emissions guidance linked above describes this distinction. Record where a filter operates before interpreting the resulting count.
For the wider device-to-application path, use the AIDC Device Integration Checklist.
Run a controlled shelf test
The following is a suggested evaluation method. Agree the sample and acceptance criteria with the project team before treating the result as a purchasing decision.
- Establish the expected population. Prepare a verified list of the items physically present and their tag IDs. For an item count, confirm the relationship between each item and its identifier. A unique-ID count cannot distinguish two items that have accidentally received the same identifier.
- Include realistic neighbors. Place a separately identified group of tagged items where adjacent stock would normally sit. Record their positions. This makes unwanted reads visible in the test instead of discovering them during rollout.
- Record the configuration. Note the reader and antenna, tag model, placement, region setting, power, available RF mode, firmware, application version and filters. Photograph the arrangement without exposing confidential item data.
- Define one trial. Specify the starting point, route, scan window and whether the operator can reposition items. Clear the trial’s result set before each new run. Keep repeated reports in a separate log if the software exposes them.
- Change one variable at a time. Compare placement, angle, supported settings or route individually. Preserve the previous result so an improvement can be attributed to a particular change.
- Repeat under working conditions. Include representative operators and realistic packing variations. Record each run separately, including missing and unwanted IDs. Do not combine several attempts into one apparently complete first-pass result.
Once the read behavior is understood, extend the evaluation to interruptions, exceptions and application acceptance using the AIDC Pilot Validation Checklist.
Interpret the result: 105 IDs can still mean an incomplete count
Illustrative example only: these figures are invented to explain the calculation. They are not AIDC GO test results, a customer case or a performance target.
Suppose a shelf contains 100 items, each with one verified, distinct tag ID. Nearby stock has a separate list of IDs. During a 30-second trial, the reader produces 2,000 tag reports. After removing repeated IDs, the result contains 105 distinct IDs: 98 from the target shelf and seven from neighboring stock.
The useful findings are:
- Target coverage is 98 / 100 = 98% for that trial.
- Two target items remain unread. Identify them for investigation.
- Seven unwanted IDs were detected. Check whether the application excluded them correctly.
- The 2,000 reports are radio-read observations; they do not represent 2,000 items.
An interface showing “105 items found” would obscure both problems. If an application filter displays only the 98 expected IDs, that improves the displayed result but does not establish that the physical read zone excluded the neighboring tags.
If a second pass recovers the missing items, report the first-pass and final results separately, together with the extra time. Set the acceptable result according to the business task and cost of exceptions. This example does not establish a universal pass percentage.
Use the evidence to choose the next change
If the same items repeatedly fail, investigate those items and their tags first. Compare placement, packing and orientation before concluding that the reader needs replacing.
If the target population reads but neighboring stock is also included, review the physical read area and the application’s rules for accepting IDs. Decide whether excluding unwanted IDs in software meets the task, or whether the workflow requires tighter control of which tags are interrogated.
If reader reports are complete but the application result is not, follow the data path. Replacing the reader will not by itself repair an incorrect item mapping or application filter.
When comparing a sled with an integrated handheld, use the same tagged population, permitted operating conditions and timing method. Include the host application and connection in the test. Our RFID Sled vs Integrated RFID Mobile Computer guide addresses that architecture decision.
What to include in a supplier discussion
A useful brief gives the supplier something reproducible to evaluate:
- The task, target area and positions of nearby tagged stock.
- Actual item materials, packaging and contents, with representative photos or samples.
- The tag model and placement, or a clear statement that tag selection remains open.
- The required working distance, operator movement and available scan time.
- Expected IDs, missing IDs and unwanted IDs from each recorded trial.
- Reader configuration, destination country, application requirements and unresolved questions.
Ask for any reported result to identify these conditions. A range figure without its test setup is insufficient for comparing two proposals.
The AIDC Hardware Project Brief Template provides a structure for organizing the wider requirement.
Frequently asked questions
How far can a handheld UHF RFID reader read?
There is no single distance that applies to every handheld, tag and item. Request the conditions behind the quoted range, then test at the distance and pace required by your workflow. Include both target coverage and unwanted reads.
Will turning up the reader power fix missed tags?
It is one variable to test, where adjustment is supported. Compare the identities of recovered tags and any additional unwanted tags. A change that raises the total count without recovering the missing target items has not resolved the original problem.
Does RFID work on metal or liquid containers?
Suitable tag designs and placements can make these applications possible. Evaluate the complete item using the tag supplier’s guidance; do not assume the result from an ordinary label on an empty package will carry over to the filled or metal item.
Is one successful inventory pass enough to approve a reader?
One pass documents one result. Repeat with the variations that matter in the intended operation, and preserve each run’s exceptions. Use the results to agree a supported operating method and an exception process before rollout.
Discuss your RFID inventory requirement
Bring the item materials, tag information, required working distance and the missed or unwanted reads you need to resolve. Contact AIDC GO to discuss the hardware requirement and what should be confirmed for the proposed configuration.
This guide covers passive UHF RFID inventory evaluation. The cited GS1 and Impinj materials support the technical explanations; the shelf-test method and numerical example are editorial guidance. References to another manufacturer’s documentation do not establish that an AIDC GO model supports the same settings or features. Model capabilities and project performance require their own evidence.