A Dock-Door Portal Design Worksheet: Antenna Count, Beam Footprint and Dwell Time
A dock-door portal is a geometry problem bounded by a legal power ceiling and a moving clock, and all three have to close on one page. The three pieces normally live in three places: RFID pages describe beamwidth in adjectives, generic antenna calculators do the trigonometry in the abstract, and the channel-occupancy ceilings sit inside standards documents. This worksheet puts them together in one unit system.
It turns published beamwidths into read-zone footprints in metres, footprint overlap into seconds in the zone, seconds into completed antenna cycles and milliseconds of reader-on time, then checks the answer against the two rules that govern how long a reader may illuminate one channel. Every figure is arithmetic you can redo on a calculator, and the worked 3 m × 4 m door at the end is filled in line by line.
The five inputs a portal design actually needs
These five numbers are what turn an antenna count into arithmetic. They cost an afternoon to collect and they set everything downstream.
- Clear opening. Width and clear height measured to the inside of the frame, in metres — the measured opening rather than the nominal door size.
- Load envelope. The tallest and widest thing that actually passes: a double-stacked pallet, a 2.4 m cage, a coil on a trailer bed. The envelope, not the average.
- Speed. Vehicle speed through the opening in km/h. A pedestrian with a hand pallet truck and a reach-truck driver are two acceptance cases, each with its own line.
- Tag count and tag position. How many tags on one load and where they sit: four pallet-corner labels, one tag per carton in a 60-carton stack, one tag per crate in a cage.
- What the load is made of. Metal-rich and water-rich loads reflect and absorb, and they earn their own settings pass: the configuration that clears 60 cartons of dry goods is re-tuned for 60 drums of liquid in a steel cage.
The output is one number, and it is read attempts per pass rather than read range. Range is a static measurement and a dock door is a timed event, so the figure that governs is how much transmission time a tag receives while it crosses the opening.
The worksheet runs in this order, one line of arithmetic per step: beamwidth → footprint → antenna positions → zone depth → seconds in zone → antenna cycle time → reader-on time per pass → that time checked against the population and against the channel ceiling.
Beamwidth to footprint, by trigonometry
EN 302 208 defines the term exactly, and its definition is the one to work with: the beamwidth is “the angle between the two half-power (−3 dB) points of the main lobe, when referenced to the peak effective radiated power of the main lobe” (clause 4.3.4.2). It is a −3 dB contour, so it marks where margin halves rather than where reading ends; tags outside it still read, with less margin.
Footprint width at distance d from the antenna face falls out of the triangle:
w = 2 · d · tan(θ / 2)
Worked once, at d = 3 m with a 68° panel: tan(34°) = 0.6745, so w = 2 × 3 × 0.6745 = 4.05 m. The rest of the table is the same line repeated.
| Half-power beamwidth | Footprint at 1.5 m | at 2.0 m | at 3.0 m | Antenna class with that beamwidth |
|---|---|---|---|---|
| 20° | 0.53 m | 0.71 m | 1.06 m | ground/mat antenna, narrow plane |
| 68° | 2.02 m | 2.70 m | 4.05 m | ~8.5 dB low-profile panel, both planes |
| 70° | 2.10 m | 2.80 m | 4.20 m | 9.0 dB nominal area antenna, both planes |
| 80° | 2.52 m | 3.36 m | 5.04 m | ground/mat antenna, wide plane |
| 90° | 3.00 m | 4.00 m | 6.00 m | the standard's horizontal beam-width ceiling above 1 W (see below) |
| 100° | 3.58 m | 4.77 m | 7.15 m | ~3 dB compact rugged outdoor part |
| 115° | 4.71 m | 6.28 m | 9.42 m | ~5.5 dB low-profile part, both planes |
Those beamwidths are real published figures rather than illustrations. One widely circulated fixed-reader accessory guide lists an ultra-low-profile indoor/outdoor panel at about 8.5 dB gain with 68° vertical and 68° horizontal beamwidth, a high-performance area antenna at 9.0 dB nominal with 70° in both planes, a compact outdoor part at about 3 dB with 100°, a low-profile part at about 5.5 dB with 115° in both planes, and a ground/mat antenna at about 9 dB with 20° in one plane and 80° in the other. Identium quotes portals to a schedule written the way this column is written — pattern, gain and port count — so the geometry survives a change of part.
Do the calculation twice for every antenna — azimuth against the door width, elevation against the load height. For a symmetric 68° × 68° panel the two answers match; for the 20° × 80° mat antenna they differ by nearly a factor of five (0.53 m against 2.52 m at 1.5 m); and both planes go in the worksheet, because site acceptance tests both.
Tilt moves the footprint down the approach instead of straight across it, and it is the tool for overhead positions. An antenna at height h tilted φ below horizontal puts its beam axis on the floor at horizontal distance h / tan φ, at a slant range of h / sin φ. At h = 2.6 m and φ = 20°: floor intercept 2.6 / 0.3640 = 7.14 m, slant range 2.6 / 0.3420 = 7.60 m. At 30°: 4.50 m and 5.20 m.
The cost is paid in path length, and at small angles it is genuinely small. Tilting by φ lengthens the path to a design point d metres away in the horizontal to d / cos φ. Over a 3.00 m design point: 10° gives 3.046 m, which is 20·log(3.046/3.00) = 0.13 dB outbound and, because a passive tag’s reply travels the same path back, about 0.27 dB round trip. 20° gives 3.193 m, 0.54 dB outbound and about 1.08 dB round trip. That is why 10° to 20° is the useful range: it biases coverage into the approach for roughly a decibel of round-trip budget. Steeper angles climb fast — 40° over the same design point is 3.92 m, 2.32 dB outbound and 4.64 dB round trip.
Gain and beamwidth move in opposite directions, and the standard joins them
Gain and beamwidth are one piece of physics seen twice: a higher-gain panel concentrates the same power into a narrower cone. The published figures above show it in a single line, roughly 3 dB at 100° rising to 9 dB at 70°. What turns that from a curiosity into a design constraint is that the harmonised standard ties the two together at clause 4.3.4.3, and it does so only in the horizontal orientation.
| Band | e.r.p. setting | Permitted horizontal beamwidth | Antenna class that satisfies it |
|---|---|---|---|
| 865–868 MHz (lower) | ≤ 500 mW (≤ 27 dBm) | no restriction | any, including the 100°–115° wide parts |
| 865–868 MHz (lower) | > 500 mW to ≤ 1 W (≤ 30 dBm) | ≤ 180° | any, including the 100°–115° wide parts |
| 865–868 MHz (lower) | > 1 W to 2 W (≤ 33 dBm) | ≤ 90° | the 68°–80°, 8.5–9 dB panel class |
| 915–921 MHz (upper) | ≤ 1 W | no restriction | any |
| 915–921 MHz (upper) | > 1 W to ≤ 2 W | ≤ 180° | any |
| 915–921 MHz (upper) | > 2 W to 4 W (≤ 36 dBm) | ≤ 90° | the 68°–80°, 8.5–9 dB panel class |
Two consequences are worth money. First, because the limit is specified for the horizontal orientation, the way to cover a tall door at full power is a narrow azimuth pattern with a generous elevation pattern: buy height in the antenna pattern, not in the transmitter. Second, the wide 100° and 115° parts are the unrestricted-pattern choice in the lower band at or below 500 mW e.r.p., while the 68°–80° class is what carries the full 2 W. That is a trade between pattern width and power, decided once, per antenna position — and on a 3 m door the 68° panel already over-covers the opening by 35% at the far jamb, so the narrow part is usually the comfortable choice rather than the compromise it sounds like.
An Indian installation is designed to the same ladder, by way of the standard that the rules name. Table-IV of G.S.R. 853(E) sets the legal limits directly: 2 W e.r.p. in 865–868 MHz on channels of 200 kHz or less, interrogator transmissions at that power permitted only on the four channels centred at 865.7, 866.3, 866.9 and 867.5 MHz, tags replying at −20 dBm e.r.p., and EN 302 208 named in the table as the harmonised reference standard. The beam-width ladder above is clause 4.3.4.3 of that named standard, so it belongs in a design worksheet as engineering practice under the referenced standard, while the figures Table-IV itself carries are the power, channelling and channel-access limits. Our port-to-air power budget guide works the dBm arithmetic behind those ceilings.
Reading dBic against dBil, and covering the bottom of the pallet
One accessory guide prints both numbers on the same catalogue line. The overhead and ceiling star antenna kits are listed as “gain-max 9 dBic, gain linear max 6 dBil”; the localised point-of-sale kit as “gain-max 7.3 dBic, gain linear max 4.9 dBil”. Same antenna, two figures, 3.0 dB and 2.4 dB apart.
Which one to budget from is a question about the tag, not the antenna. The circular figure is what the antenna delivers to a circularly polarised counterpart. The linear figure is what it delivers to a linearly polarised dipole — which is what a passive UHF label is. When tag orientation on the load is uncontrolled, the linear figure is the one to build the budget on.
Run the range calculation twice off the same antenna and the size of the choice becomes visible. Take any baseline: suppose the forward-link budget using 9 dBic puts the activation threshold at 6.00 m. For a fixed power-at-tag threshold, range scales as 10^(ΔG/20), and 10^(3/20) = 1.41, so the same threshold budgeted from 6 dBil sits at 6.00 / 1.41 = 4.25 m. The reverse link lands on the identical factor, because antenna gain appears on both the outbound and the return path: 3 dB of gain is 6 dB of round-trip budget, and 10^(6/40) = 1.41. On the 2.4 dB pair the factor is 10^(2.4/20) = 1.32.
Where that 1.75 m lands is what decides whether the choice of figure changes a design. On the worked 3 m door it is margin: the longest slant from a jamb panel at 1.40 m to a tag on top of a 3.60 m load at the far jamb is hypot(3.00, 2.20) = 3.72 m, so both 6.00 m and 4.25 m reach it — one with 2.28 m of headroom, one with 0.53 m. On a deep overhead position the same 1.75 m changes the answer: an overhead antenna asked to activate tags 5 m down the approach is inside budget on the 9 dBic figure and outside it on the 6 dBil figure. Budgeting from the linear figure keeps the 3 m door closing and tells you the truth about the 5 m one, which is the whole argument for using it.
The bottom layer of a pallet is answered by geometry rather than by more power. A side panel at 1.40 m looking across a 3 m opening has its −3 dB contour well above the bottom carton by the time it reaches the far jamb, and raising power widens the footprint outside the door faster than it fills the floor. A floor or mat antenna rated for the traffic solves it directly: one published part measures 915 × 305 × 8 mm, about 9 dB gain, 20° in one plane and 80° in the other, and withstands payloads over 200 kg. At 1.5 m above the floor its footprint is 2 × 1.5 × tan(10°) = 0.53 m by 2 × 1.5 × tan(40°) = 2.52 m — a stripe, not a circle. Lay the 2.52 m dimension across the doorway so the load drives along the narrow axis.
Circular polarisation earns its keep exactly where orientation is unknown, which is most inbound freight and most of what a pallet tracking portal sees. Where orientation is fixed and you control label application — an outbound line, a returnable crate with a moulded tag pocket — a linear panel aligned to the tag recovers the polarisation loss and is the better buy.
Antenna count per door, and reader ports
The worked door. 3.00 m clear width, 4.00 m clear height, loads to 3.60 m, mixed cartons, 60 tags, acceptance speed 4 km/h. Antennas: the 68° × 68°, ~8.5 dB class, on both jambs facing across the opening, two per jamb at 1.40 m and 2.60 m. The jamb panels are rotated 15° in the horizontal plane toward the approach, which biases the coverage stripe upstream of the threshold while leaving each panel’s elevation boresight on the horizontal — so the elevation arithmetic below runs off the mounting heights as measured. Downtilt is kept for overhead positions, where the formulas in section two apply.
Take the far jamb, d = 3 m, footprint 4.05 m, half-height 2.02 m. The antenna at 1.40 m covers −0.62 m to 3.42 m in that plane; the one at 2.60 m covers 0.58 m to 4.62 m. Their union spans −0.62 m to 4.62 m, which contains the full 4 m opening. The 15° rotation lengthens the slant to the far jamb to 3.00 / cos 15° = 3.11 m, widening each footprint by about 4% — margin in the right direction.
At mid-door, d = 1.5 m, the footprint halves to 2.02 m, giving ±1.01 m about each boresight: 0.39 m to 2.41 m and 1.59 m to 3.61 m, a contiguous union of 0.39 m to 3.61 m. That leaves 39 cm at the floor and 39 cm at the lintel to answer, and it is the arithmetic that earns a fifth antenna on purpose rather than a hopeful one: a mat antenna in the floor for the bottom layer, or an overhead unit for the top of the load. Add both and the door wants six ports.
Design for overlap. Two footprints that overlap by half a metre cost a little duplicate reading, which software resolves in a millisecond; half a metre of gap in the middle of the door costs a tag. Where the budget forces the choice, overlap.
Ports follow the antenna count. One published fixed-reader family ships as 2-port and 4-port variants in the smaller model and 4-port and 8-port variants in the larger, so a five-antenna door calls for an 8-port reader, or a 4-port reader with an antenna hub: one published hub fans a single reader port out to eight ports, reaches 32 ports on one reader, and switches from one antenna port to another in under 25 milliseconds.
What decides between those two routes is radio time rather than port count, because a reader serves its antennas in sequence. Five antennas at a 35 ms dwell and 25 ms of switching is a 300 ms cycle; twenty antennas on that same reader is a 1.2 s cycle, which is longer than a forklift takes to cross an opening. So a bank of four five-antenna doors is four reader radios, one door per reader, each an 8-port reader or a 4-port reader with one hub — and the hub’s 32-port reach earns its keep on static inventory positions, where a slow round robin is exactly what you want. The worked door below is costed on the 4-port-plus-hub arrangement, because that is the one with a published switching figure to budget against; on a reader’s own ports, take the inter-antenna switching time from the reader datasheet and put it in line 16.
Cable is the last subtraction, and it is worth checking the specific kit: in that accessory guide the star antenna kits ship 17 ft (about 5 m) cables — two in the dual-antenna overhead/ceiling kit, one in the single-antenna overhead/ceiling and wall-mount kits, while the localised point-of-sale kit in the same guide ships a 20 ft cable. e.r.p. at the air is port power minus cable and connector loss plus antenna gain referenced to a dipole. Keep the two units straight while you do it: 33 dBm e.r.p. is 35.15 dBm e.i.r.p., the 2.15 dB being the dipole reference.
Physical dwell: zone length over speed
Convert once and reuse: 4 km/h = 1.11 m/s, 6 km/h = 1.67 m/s, 8 km/h = 2.22 m/s, 12 km/h = 3.33 m/s. Time in zone is zone depth divided by speed.
| Speed | m/s | 1.5 m zone | 2.0 m zone | 3.0 m zone |
|---|---|---|---|---|
| 4 km/h | 1.11 | 1.35 s | 1.80 s | 2.70 s |
| 6 km/h | 1.67 | 0.90 s | 1.20 s | 1.80 s |
| 8 km/h | 2.22 | 0.68 s | 0.90 s | 1.35 s |
| 12 km/h | 3.33 | 0.45 s | 0.60 s | 0.90 s |
Two cautions. The zone depth is the depth of the footprint overlap measured along the direction of travel, which is a different quantity from the depth of the door structure. On the worked door the jamb panels spread 2.02 m along travel at mid-door and 4.05 m at the far jamb, so 2.0 m is the defensible planning figure — the mid-door number, the tightest point — and 3.0 m is a figure to claim after a survey has measured it.
And the speeds are yours to set. A pedestrian pushing a hand pallet truck, an electric pallet truck and a reach truck crossing the same threshold produce times in zone that differ by a factor of three across that table. Fix the acceptance speed in the specification and the commissioning conversation is settled in advance.
Air time: what one pass actually buys
Per-antenna dwell is a setting, not a property of the hardware. The round robin is:
cycle = N × (dwell + switching time)
Five antennas, a 35 ms dwell and the published sub-25 ms switching figure give 5 × 60 ms = 300 ms. Set the dwell to 500 ms instead and the cycle becomes 5 × 525 ms = 2.63 s, so a sub-second pass is served by one or two antennas, whichever the cycle happened to be on. Long dwells suit static inventory; a door wants short dwells and many cycles.
Two things have to be true of one pass, and they pull against each other.
Constraint one, cycles. Three complete cycles per pass is the planning minimum we use, because each cycle re-presents the tag at a different antenna position and polarisation, which is how an awkwardly oriented label eventually answers. That means cycle time no greater than a third of the time in zone: N · (dwell + switch) ≤ T/3. At T = 0.90 s with five antennas and 25 ms switching, dwell is capped at 900/15 − 25 = 35 ms.
Constraint two, reader-on time. Only the part of the pass spent transmitting counts. Reader-on time = cycles × N × dwell, which in the continuous limit is T × dwell/(dwell + switch) — the duty factor. At 35 ms dwell and 25 ms switching the duty factor is 35/60 = 58%, so a 0.90 s pass contains 525 ms of transmission and 375 ms of switching. Note which symbol is absent from that expression: N. Antenna count changes how finely the air time is sliced, not how much of it there is.
The requirement side comes from the population. Take the RAIN Alliance’s published worked figure of 100 tags per second of reader throughput: 60 tags on the load needs 60 / 100 = 600 ms of reader-on time. Four pallet-corner labels instead of 60 carton labels needs 40 ms, which is why the decision about what carries a tag at the door is the largest single lever on this page.
The worked door, at the acceptance speed. 4 km/h through the 2.0 m zone gives T = 1.80 s. Constraint one caps dwell at 1800/15 − 25 = 95 ms. Constraint two needs 1.80 · d/(d + 25) ≥ 0.600, which solves to d ≥ 12.5 ms. Any dwell between 12.5 ms and 95 ms satisfies both, and 35 ms sits comfortably inside: cycle 300 ms, 1.80 / 0.300 = six complete cycles, reader-on time 6 × 5 × 35 = 1,050 ms against the 600 ms the population needs. The door closes with 75% margin, and line 24 of the worksheet reads pass before anything has been bought.
Reader-on time stands in for tag air time here because of the footprint arithmetic, and that step is worth showing. What a tag receives is the part of the reader’s transmission spent on an antenna that illuminates it. At mid-door each jamb panel’s elevation footprint is ±1.01 m about its boresight: 0.39 m to 2.41 m from the panel at 1.40 m, 1.59 m to 3.61 m from the one at 2.60 m, on both jambs. So a tag anywhere from 0.39 m to 3.61 m on the load sits inside the −3 dB contour of all four jamb panels for the crossing, and the mat antenna covers the bottom 0.39 m. Four of the five ports illuminate the bulk of the population, so the reader-level budget is the governing one and it is carried across whole. On a wider door, where each pair of panels covers only part of the opening, the population genuinely divides between illumination volumes — and then the divisor is worked from the footprints in the same way, rather than assumed.
Faster traffic, and five ways to buy air time. Take the same door at 8 km/h through the 2.0 m zone, T = 0.90 s. Constraint one caps dwell at 35 ms while constraint two asks for at least 50 ms, so the design moves one of its inputs. In order of what they cost:
- Tags to clear at the door. Pallet identity at the portal and item detail read inside: 4 tags needs 40 ms and every constraint relaxes at once. This is a data-model decision, and it is free.
- Acceptance speed. 4 km/h doubles T against 8 km/h and reopens the window to the 12.5–95 ms dwell range above. It costs a line in the specification.
- Surveyed zone depth. 3.0 m at 8 km/h gives T = 1.35 s: dwell capped at 65 ms, needed at 20 ms, and 35 ms gives a 300 ms cycle, four complete cycles and 4 × 5 × 35 = 700 ms of reader-on time against 600 ms. That closes on a 17% margin rather than 75%, which is a number to quote as surveyed rather than as planned.
- Duty factor. A longer dwell spends less of the pass switching. At T = 0.90 s a 65 ms dwell gives a 450 ms cycle, two complete cycles and 2 × 5 × 65 = 650 ms of reader-on time: it buys the population budget at the price of the third cycle, so it is a trade to make deliberately and to test on the day.
- Power last, because it widens the footprint outside the door and buys stray reads from the neighbour along with the extra margin.
One caution about the throughput figure itself. Gen2 arbitration is framed slotted Aloha, and the formal result is that the optimum is reached by setting the frame length equal to the backlog size at each subsequent frame, with an asymptotic efficiency of e^−1. Resolving n tags therefore costs on the order of n·e slots rather than n slots: 60 tags is about 163 slots, 300 tags about 815. A published tags-per-second figure already carries that arbitration overhead inside it, which is why the budget above is built from tags per second — and it is also why a figure quoted in slots or inventory rounds is worth converting before trusting it.
The two regulatory ceilings on how long you may illuminate a channel
Two rules bound the clock, and they are different shapes.
In the United States, 47 CFR 15.247(a)(1)(i) requires that where the 20 dB bandwidth of the hopping channel is under 250 kHz, the system “shall use at least 50 hopping frequencies and the average time of occupancy on any frequency shall not be greater than 0.4 seconds within a 20 second period”. Paragraph (b)(2) pairs channel count with power: 1 watt maximum peak conducted output for systems employing at least 50 hopping channels, 0.25 watts for systems employing fewer than 50 but at least 25.
In Europe’s lower band and in India the ceiling is ten times longer and differently shaped. EN 302 208 clause 4.3.7.3 caps continuous transmission on a channel at 4 s with an off-period of at least 100 ms before returning to it, and Table-IV of G.S.R. 853(E) carries the same numbers in almost the same words: the maximum period of continuous interrogator transmission on a channel shall not exceed 4 s, and the period between consecutive transmissions on the same channel shall be at least 100 ms.
The asymmetry that changes portal design sits at the top of the band: clause 4.3.7.3 states plainly that there is no specific limit on transmission length for interrogators in the upper band, 915–921 MHz. We covered that band’s wider consequences in the 915–921 MHz guide.
This is where an exported portal configuration earns a re-engineering rather than a re-label, and the RAIN Alliance makes the point in numbers. About 20 tags in the read zone at 100 tags per second is 200 ms, comfortably inside the 400 ms FCC channel budget, so the whole population is read inside one dwell period and session S0 behaves. Raise the population until clearing it takes longer than that budget and the reader hops mid-population, S0 flags reset on the hop, and the design moves to a persistent session. The identical population inside the lower band’s 4 s allowance can stay on S0. One portal, two continents, two session choices — and one published radio-mode table underlines the split by offering backscatter data rates up to 640 kbps on FCC-band readers while topping out at 320 kbps on ETSI-band readers.
Two more clauses are easy to miss at design stage and expensive at commissioning. Lower-band interrogators are required to support trigger techniques indicating the presence or arrival of objects that may be tagged, and to stop transmitting once they have ceased to read further tags — so a photo-eye or loop trigger on an 865–868 MHz dock door is part of the design, and it belongs on the worksheet as line 26. And a reader that polls instead of waiting for a trigger is in presence-sensing mode, where each transmission is restricted to less than 1 s with at least 100 ms between successive transmissions.
On the worked door these two ceilings land on the trigger and the session rather than on the dwell. A 35 ms per-antenna dwell sits far inside both allowances with room to spare. What the ceilings do decide is that the 600 ms of reader-on time needed to clear a 60-tag load spans several channel changes under the FCC rule, so the session choice is made for you in the 902–928 MHz build and left open in the 865–868 MHz one.
The worksheet, blank and filled
Copy the middle column into your own door. The right-hand column is the 3.00 m × 4.00 m door carried end to end.
| Line | Unit or formula | Your door | Worked door |
|---|---|---|---|
| 1. Clear width | m | — | 3.00 |
| 2. Clear height | m | — | 4.00 |
| 3. Tallest load | m | — | 3.60 |
| 4. Acceptance speed | km/h → m/s | — | 4.0 → 1.11 |
| 5. Tags to clear at the door | count | — | 60 |
| 6. Antenna class, azimuth / elevation beamwidth | degrees, gain | — | 68 / 68, ~8.5 dB panel |
| 7. Mounting heights | m | — | 1.40 and 2.60, both jambs |
| 8. Rotation toward approach (horizontal plane) | degrees | — | 15 |
| 9. Footprint at far jamb | w = 2d·tan(θ/2), d = 3.00 m | — | 4.05 m |
| 10. Footprint at mid-door | same, d = 1.50 m | — | 2.02 m |
| 11. Elevation union at far jamb | m above floor | — | −0.62 to 4.62 (covers 0–4.00) |
| 12. Elevation union at mid-door | m above floor | — | 0.39 to 3.61 |
| 13. Extra antenna for load bottom / top | count and class | — | 1 mat antenna, 20° × 80° |
| 14. Ports consumed | count | — | 5 |
| 15. Reader arrangement | ports | — | 4-port reader + 1 antenna hub (5 of 8 hub ports); one reader per door |
| 16. Switching per antenna change | ms, from the datasheet | — | < 25 (hub, published) |
| 17. Per-antenna dwell | ms | — | 35 |
| 18. Cycle time | N × (dwell + switch) | — | 300 ms |
| 19. Zone depth along travel | m (= line 10 at the tightest point) | — | 2.00 |
| 20. Time in zone T | depth ÷ speed | — | 1.80 s |
| 21. Complete cycles per pass | line 20 ÷ line 18, rounded down | — | 6 |
| 22. Reader-on time per pass | cycles × N × dwell | — | 1,050 ms |
| 23. Reader-on time needed | tags ÷ throughput | — | 600 ms |
| 24. Verdict | pass / tune | — | pass, 1.75× margin |
| 25. Channel ceiling in force | s / ms | — | 4 s on, 100 ms off (865–868); 0.4 s avg in 20 s (902–928) |
| 26. Trigger | type | — | photo-eye or loop (required in 865–868) |
A site survey confirms this, and it is worth knowing what the survey actually measures: the real −3 dB contour in the presence of the building, the reflection pattern off the racking and the dock leveller, the read rate with the actual load rather than a test pallet, the true overlap depth along travel for line 19, and the stray read rate from the neighbouring door.
Then write the acceptance test into the tender as a timed event. A usable clause states a first-pass read rate over a stated number of passes, at a stated speed, with a stated tag and load — for example: 60 tagged cartons on a shrink-wrapped pallet, 30 passes at 4 km/h, 100% first-pass read on pallet identity and at least 99% on item-level tags, with fewer than one stray read per pass attributable to the adjacent door. That is testable on the day, which is what an acceptance clause has to be. Our 22-clause tender template has the surrounding clauses, and the warehouse management and logistics platforms are where the portal data lands.
Once the geometry closes, what remains is settings: session, Q, target flag, transmit power and antenna dwell, each interacting with the others and with the tag population. That is the next worksheet.
Frequently asked questions
How many antennas does an RFID dock door need?
Work it from the opening rather than from a rule of thumb. On a 3.00 m × 4.00 m door, two 68° panels per jamb at 1.40 m and 2.60 m cover the full height at the far jamb (union −0.62 m to 4.62 m) but reach only 0.39 m to 3.61 m at mid-door, leaving 39 cm at the floor and 39 cm at the lintel. A fifth antenna answers it — a floor mat for the bottom layer, or an overhead unit for the load top. Five ports is the realistic figure for a fully covered door of that size, six if you want both top and bottom.
How high should an RFID portal antenna be mounted, and at what tilt?
Mount so the two elevation footprints overlap across the whole opening. With a 68° panel on a 4 m door, 1.40 m and 2.60 m works: at the far jamb, 3 m away, the footprint is 4.05 m, giving ±2.02 m about each boresight and a union that spans the opening. For jamb panels, rotate 10° to 20° toward the approach in the horizontal plane and keep the elevation boresight level, so the height arithmetic stays valid. For overhead positions, downtilt is the tool and its cost is small at small angles: over a 3.00 m design point, 10° lengthens the path to 3.046 m for 0.13 dB outbound and about 0.27 dB round trip, and 20° to 3.193 m for 0.54 dB and about 1.08 dB. An overhead antenna at 2.6 m tilted 20° meets the floor 7.14 m out at a 7.60 m slant range.
How do I calculate an antenna's read-zone footprint from its beamwidth?
w = 2 · d · tan(θ/2), where θ is the half-power beamwidth in the plane you care about and d is the distance from the antenna face. At 3 m: a 68° panel gives 4.05 m, a 70° panel 4.20 m, an 80° pattern 5.04 m, a 100° part 7.15 m. Do it twice per antenna — azimuth against the door width, elevation against the load height — because asymmetric parts such as a 20° × 80° mat antenna give answers nearly five times apart (0.53 m against 2.52 m at 1.5 m).
How fast can a forklift pass through an RFID portal?
As fast as the arithmetic allows. Time in zone is zone depth divided by speed: a 2.0 m zone gives 1.80 s at 4 km/h, 0.90 s at 8 km/h and 0.60 s at 12 km/h. Then check it against the cycle and the population. A five-antenna round robin at 35 ms dwell plus sub-25 ms switching is a 300 ms cycle, and reader-on time per pass is cycles × antennas × dwell. With 60 tags at the RAIN Alliance’s published worked figure of 100 tags per second, 600 ms of reader-on time is needed: the worked door closes at 4 km/h through 2.0 m (1.80 s, six cycles, 1,050 ms — 75% margin) and at 8 km/h through a surveyed 3.0 m zone (1.35 s, four cycles, 700 ms — 17% margin).
How long is an RFID reader allowed to transmit on one channel?
It depends on the band. Under 47 CFR 15.247(a)(1)(i) the average time of occupancy on any frequency must not exceed 0.4 seconds within a 20 second period, with at least 50 hopping frequencies where the channel's 20 dB bandwidth is under 250 kHz. In 865–868 MHz, EN 302 208 clause 4.3.7.3 and Table-IV of India's G.S.R. 853(E) both cap continuous transmission on a channel at 4 s with at least 100 ms before returning to that channel. In the 915–921 MHz upper band, EN 302 208 sets no specific limit on transmission length.
Do I need a floor antenna to read the bottom layer of a pallet?
Usually yes, because it is a geometry question rather than a power question: a side panel at 1.40 m has its −3 dB contour above the bottom carton by the time it crosses a 3 m opening, and raising power widens the footprint outside the door faster than it fills the floor. A mat antenna rated for the traffic answers it directly — one published part is 915 × 305 × 8 mm, about 9 dB, 20° by 80°, rated over 200 kg, giving a 0.53 m × 2.52 m stripe at 1.5 m height. Lay the long axis across the doorway.
Should I buy a 4-port or an 8-port reader for a bank of dock doors?
Count ports per door first, then count radios. A fully covered 3 m × 4 m door consumes five ports, so it calls for an 8-port reader or a 4-port reader with an antenna hub — one published hub fans a single port out to eight and reaches 32 ports on one reader, at under 25 ms per antenna switch. The deciding factor is radio time: a reader serves its antennas in sequence, so five antennas at 35 ms dwell is a 300 ms cycle while twenty on one reader is 1.2 s, longer than a crossing. Four five-antenna doors therefore means four reader radios, one per door, and the hub's 32-port reach is best spent on static inventory positions where a slow round robin suits the job.
Sources
- ETSI EN 302 208 V3.3.1 (2020-08), clauses 4.2.1, 4.3.4.2, 4.3.4.3 and 4.3.7.3
- 47 CFR 15.247(a)(1)(i) and (b)(2), frequency hopping requirements and power limits
- Gazette of India, G.S.R. 853(E), New Delhi, the 10th December 2021, Table-IV (865-868 MHz Short Range Devices Exemption from Licence Rules, 2021)
- Zebra FX9500/FX9600 Technical Accessory Guide (antenna gain in dBic and dBil, beamwidths, ground/mat antenna dimensions and load rating, kit cable lengths)
- Zebra FX Series RFID Fixed Reader Integration Guide (antenna port counts; Radio Modes for FCC Readers and for ETSI Readers)
- Impinj R700 Antenna Hub User Guide v7.6 (one port to eight, 32 ports on one reader, sub-25 ms antenna port switching)
- Barletta, Borgonovo and Cesana, A formal proof of the optimal frame setting for Dynamic-Frame Aloha with known population size, arXiv:1202.3914v2, 10 March 2013 (frame length equal to the backlog size at each subsequent frame; asymptotic efficiency e^-1)
- RAIN RFID System Design Guidelines V2, section 3.1.2 (20 tags in read zone, 100 tags per second, 200 ms against the FCC 400 ms dwell limit)
