Signal repeater selection is driven by five binding gates — input/output signal class, bandwidth, gain budget, galvanic isolation, and supply — and skipping any of them produces the field failures that drive most support tickets [S1]. Across 1900 MHz indoor boosters the published envelope is 76 dB uplink / 83 dB downlink gain at 60 MHz bandwidth with a 4.5 dB uplink noise figure, against a –8 dB return-loss floor and a 30 dBm ALC ceiling [S2]. Process-control pulse repeaters live on a different axis: open-collector, voltage-pulse, current-pulse, and non-voltage-contact inputs each demand their own external-resistor discipline per Yokogawa's 2026 field guidance [S1].
China's mobile-repeater supply base, anchored by Shenzhen-volume factories with ISO 9001 / CE / RoHS / CCC certification [S4][S5], ships the bulk of the world's 800–2600 MHz quad-band boosters at sub-$70 retail on AliExpress as of July 2026 [S3]. The same supplier cluster also feeds the dedicated 800MHz/2GHz CDMA digital repeater market under YD/T 3303-2017, which covers radio, optical-fibre, and trunking repeaters plus in-building distribution amplifiers [S6]. Selecting across both worlds — industrial pulse/analog and cellular RF — needs a spec-first pass, not a price-first pass.
Gate 1 — Lock the Signal Class Before You Touch a Datasheet
The first binding decision is signal class: open-collector, voltage pulse, current pulse, non-voltage contact, 4–20 mA analog, or RF [S1]. Yokogawa's 2026 troubleshooting guide for the JUXTA D / M / VJ pulse-repeater family lists eight separate checks for "no output" complaints, and seven of those eight are signal-class checks: supply wiring, supply-voltage match, polarity, the mating instrument's pulse spec, external resistance on voltage/open-collector/non-voltage-contact inputs, accessory-resistor fit on current-pulse inputs, and verification that the output terminals have their required external power [S1].
Mis-classifying a current-pulse input as voltage pulse, or omitting the external pull-up on an open-collector output, is the dominant root cause behind dead repeaters in the field [S1]. RF repeaters do not have this failure mode — they fail on band mismatch — but industrial pulse repeaters almost always do, because the input impedance and biasing differ across contact types. A signal conditioner and a pulse repeater are not interchangeable; the conditioner sits upstream normalising 4–20 mA / 1–5 V, while the repeater reconstructs a discrete pulse train with defined voltage and current levels.
Gate 2 — Bandwidth and Frequency Plan (RF) or Pulse Rate (Industrial)
RF repeater selection lives or dies on the band plan. The 1900 MHz indoor reference unit covers 1850–1910 MHz uplink and 1930–1990 MHz downlink inside a 60 MHz window, with 76 dB ± 1 dB uplink and 83 dB ± 1 dB downlink gain [S2]. Multi-band consumer boosters on AliExpress in July 2026 list five to fifteen simultaneous bands — B1/B2/B3/B4/B5/B7/B8/B12/B13/B17/B20/B25/B26/B28/B66 — with 800/900/1800/2100/2600 MHz as the recurring core [S3]. Picking an unlisted band means dead coverage; picking too many bands means higher noise figure and lower per-band selectivity.
For industrial pulse duty, the equivalent of "bandwidth" is maximum pulse rate plus minimum pulse width. Yokogawa's JUXTA M Series plug-in signal conditioners span more than 30 I/O variants and support a wide pulse-rate range through external scaling resistors [S1]. A counter or signal calibrator downstream of the repeater will reveal whether the chosen unit preserves the edge at the highest input frequency — anything above the rated pulse rate is silently lost, not clipped. Where pulse rate exceeds the repeater's passband, the only fix is a faster model, not a software tweak.
Gate 3 — Gain, Noise Figure, and ALC Envelope
The published gain budget on a 5 W indoor 1900 MHz booster is 76 dB up / 83 dB down at ±1 dB flatness, with in-band ripple held to ≤ 8 dB and a maximum input power of 0 dBm before damage [S2]. Return loss is specified at ≤ –8 dB, and the ALC threshold is 30 dBm with 31 dB manual gain control in 1 dB steps [S2]. The uplink noise figure of 4.5 dB is the binding number for coverage uplink; the 8 dB downlink figure is acceptable for in-building use but too high for outdoor long-backhaul links.
For industrial pulse repeaters, "gain" translates to output drive capability — open-collector ratings typically land at 30 V DC / 100 mA, and voltage-pulse outputs sit at 12–24 V amplitude. Exceeding those numbers without a buffer stage is a common burnout mode. A signal isolator placed downstream of a high-gain pulse repeater can break ground loops that otherwise feed noise back into the input, and the isolator's 3-port isolation rating should be checked against the repeater's common-mode voltage rating on the same loop.
Gate 4 — Isolation, Impedance, and Cable Discipline
RF repeaters need 50 Ω impedance end-to-end with N-Female connectors as the reference fit [S2], and any mismatch on the coax run is multiplied by the gain. The same datasheet publishes 45 dBc third-order intermodulation and an MTBF above 20 000 hours, but both numbers collapse if the antenna separation is too tight. Self-oscillation is the dominant failure when indoor and outdoor antennas can see each other — the LED logic on the reference unit flags "always-green uplink" as the self-excited state, with the corrective action being physical antenna separation or barrier insertion [S2].
Industrial pulse repeaters do not have an antenna-isolation problem; they have a ground-loop problem. Yokogawa's JUXTA D Series is DCS-supported and modular with JIS/EIA mounting nests, terminal blocks for field wiring, and dedicated connectors for DCS comms — a layout that simplifies clean separation between field and panel grounds [S1]. Where a signal tower light or annunciator chain is fed from the same repeater output, the common-mode return path must be planned, or the pulse edges ring enough to trigger false counts downstream.
Gate 5 — Power Supply, Environment, and Compliance
The 1900 MHz reference accepts AC 110–240 V at 50/60 Hz or DC 27 V 3 A, draws 81 W, runs from –10 °C to +55 °C at under 90 % humidity, and complies with ETS 300 694-4 and GB 6993-86 [S2]. China's CDMA digital repeater market sits under YD/T 3303-2017, which defines network architecture, RF performance, O&M, power adaptability, environmental adaptability, and safety for radio, optical-fibre, and relay repeaters plus in-building distribution amplifiers [S6]. ISO 9001, CE, RoHS, and CCC are the recurring four-cert baseline on the China supply side [S4][S5].
Industrial pulse repeaters are typically 24 V DC, DIN-rail or rack-mount, with the JUXTA F Series drawing its 24 V DC from a panel supply and fitting 24 mm-wide housings [S1]. For hazardous-area panels, an explosion-proof pressure transmitter on the same loop will not be isolated by the repeater; the divider/isolator has to sit on the safe-area side. The right shortlist logic is: signal class → bandwidth → gain → isolation → supply, in that order, and never let price re-order those gates.
Comparison: Industrial Pulse Repeater vs RF Cellular Repeater vs Signal Conditioner
Three option types cover most 2026 spec sheets. An industrial pulse repeater (JUXTA D / M / VJ class) handles open-collector, voltage-pulse, current-pulse, and non-voltage-contact I/O at 24 V DC, with no RF section and no ALC — its pass criterion is correct external-resistor fit [S1]. A cellular RF repeater (1900 MHz / quad-band 800–2600 MHz class) handles 76–83 dB gain across 60 MHz windows with ALC at 30 dBm and 4.5–8 dB noise figure, and its pass criterion is band-plan completeness plus antenna isolation [S2][S3]. A signal conditioner is not a repeater — it normalises 4–20 mA / 1–5 V DC and lacks the discrete-pulse reconstruction step.
On cost, a 1900 MHz 5 W indoor booster retails around $57–$70 on AliExpress as of July 2026 with CE/RoHS, two-year warranty, and 10-piece MOQ [S2][S3]. An industrial pulse repeater module lists higher per channel but eliminates the RF antenna engineering. For projects that mix both — a process skid with a cellular gateway on the panel — the signal isolator sits between the pulse repeater and the gateway's analog input, not between the booster and the donor antenna.
Who Should NOT Pick the Mainstream RF Repeater
The mainstream consumer quad-band booster is the wrong pick where uplink noise figure must stay below 3 dB, where the deployment is outdoor long-backhaul, or where a public-safety band (TETRA, P25, 700/800 MHz LTE) is in scope. The reference 1900 MHz unit's 4.5 dB uplink NF is acceptable for in-building consumer coverage but will desensitise a donor-site link budget [S2]. For TETRA, the answer is a selective repeater with band-pass filtering at the donor antenna, not a wideband quad-band unit. For a relay module driving a control relay from a pulse output, an industrial pulse repeater — not an RF booster — is the correct architecture.
Likewise, engineers who only need 4–20 mA loop extension should not buy a repeater at all; a signal conditioner is the right block. Repeaters reconstruct a discrete or RF signal; conditioners normalise a continuous analog signal. The cheapest RF booster on AliExpress is also the wrong pick for any panel that needs functional safety certification — none of the sub-$70 units carry SIL ratings, and YD/T 3303-2017 does not substitute for IEC 61508 [S3][S6]. For SIL-rated pulse work, the path is a dedicated safety pulse isolator, not a general-purpose repeater.
Field Failure Modes and Shortlist Logic
Five failure modes recur across both worlds: wrong signal class (pulse / analog / RF) on the input, exceeded pulse rate or out-of-band RF, undersized output drive, ground loop on the analog side, and antenna self-oscillation on the RF side [S1][S2]. The reference 1900 MHz unit publishes MTBF > 20 000 hours and an auto-shut-off that triggers above the ALC threshold to prevent base-station interference — but neither spec covers an installer who mounts indoor and outdoor antennas on the same mast [S2].
The shortlist logic for a 2026 build: (1) write the signal class first — pulse, analog, or RF; (2) write the bandwidth or pulse-rate budget; (3) write the gain and noise-figure envelope; (4) confirm isolation and connector fit; (5) confirm supply, environment, and certification. Apply that to a shielded-cable run feeding a counter module and the repeater disappears as a problem block. Skip any of the five gates, and the support ticket that follows is the one Yokogawa's FAQ already answers [S1].