
A radar site and a radio site both look like "a building with an antenna," but their backup power problems are not the same problem. One is dominated by a pulsed transmitter that draws enormous peak current for microseconds and comparatively little average power; the other is dominated by a continuous DC load that must never see a switching transient. Getting the two confused is the most common reason a backup specification looks reasonable on paper and fails on site. This paper sets out the operating conditions, the load shapes, the power trains and the physics behind both, and then draws a deliberately narrow boundary around what a sealed nickel-metal-hydride (NiMH) pack can and cannot do in these stations.
An airport surveillance site and a telecom tower share a site plan, a lightning risk and usually a diesel generator. What differs is the shape of the load and the consequence of losing it.
At a radar site the consequence is a surveillance gap: the controller loses track updates, separation minima may have to be increased, and the gap is visible in the safety case. The load is dominated by a transmitter whose instantaneous demand is out of all proportion to its average demand, plus a rotating antenna whose motor behaves like a starting induction machine every time it accelerates. At a radio site the consequence is a coverage hole: traffic is carried by neighbours, calls drop, and the operator's key indicators are availability and mean time between outages. The load is a continuous DC draw — baseband, radio units, transmission, monitoring, sometimes cooling — with almost no inrush and no pulsing.
That single difference explains why the two industries specify backup power with different vocabularies. Radar specifications talk about overload class, crest factor, static bypass transfer time and voltage distortion during the generator block load. Telecom specifications talk about rectifier redundancy (N+1, N+M), float voltage windows, low-voltage disconnect and ampere-hours of autonomy.
| Site class | Typical RF band and power | What dominates the electrical load | Backup form usually specified |
|---|---|---|---|
| Primary surveillance radar (PSR) | L-band 1250–1350 MHz, pulsed power around 2.5 kW; S-band 2.7–2.9 GHz at tens of kW peak | Pulsed transmitter, antenna drive, cooling, signal processing | Online UPS plus diesel generator; dual independent feeds |
| Secondary surveillance radar (SSR) | 1030 MHz interrogation, 1090 MHz and 740 MHz replies; around 1 kW pulse power | Transmitter, receiver, monopulse processing | Same UPS/generator architecture as the PSR it is paired with |
| Precision approach radar (PAR) | X-band 9.10–9.50 GHz, solid-state around 350 W pulse | Transmitter, antenna, shelter cooling | UPS sized for a defined hold-up; generator for longer outages |
| Airport surface / weather radar | S- or C-band, solid-state, duty cycles up to the mid-teens percent | Transmitter average power and thermal management | UPS plus generator, often with a separate transmitter UPS |
| Coastal / marine radar | X-band and S-band, lower power | Transmitter and antenna motor | Local UPS, often single-source with generator |
| Telecom macro site | 700 MHz–3.5 GHz radio units | Continuous DC load: radio, transmission, monitoring | -48 V DC plant with VRLA or LiFePO4 and N+1 rectifiers |
| Microwave relay / trunk node | 6–38 GHz | Continuous DC load plus shelter cooling | -48 V DC plant, 3–8 h autonomy typical |
| VHF/HF ATC radio site | 118–137 MHz, HF 3–30 MHz | Continuous transmitter/receiver, low current | DC plant or small UPS, generator for extended outages |
A radar transmitter's nameplate is usually a peak figure. The number that matters for cables, rectifiers, UPS sizing and cooling is the average, and the number that matters for the DC link is the energy that must be stored between pulses.
The arithmetic is simple and worth doing explicitly. For a rectangular pulse of width τ repeated at a pulse repetition frequency PRF,
D = τ × PRF Pavg = Ppeak × D
A 22 kW S-band solid-state transmitter described in the published literature operates at 50 µs pulse width and 5 percent duty cycle, built from fifty 500 W high-power amplifiers combined through a low-loss radial combiner. Five percent of 22 kW is about 1.1 kW of average RF — the rest of the input power becomes heat. That is why a transmitter hall can consume several kilowatts of site power while radiating a small fraction of its peak rating.
The same arithmetic applied to an older tube-based surveillance radar is more striking. One published description of the ASR-9 gives a klystron peak of 1.3 MW, a 1 µs pulse and a PRF between 325 and 1200 pulses per second. Multiplying through, the average RF at the antenna works out to roughly 0.42 kW at the low PRF and 1.56 kW at the high PRF. A megawatt-class transmitter, in other words, is an average-power device of the same order as a few domestic kettles — but its instantaneous current, and the energy that must be delivered within each microsecond, are nothing like a kettle's.
| Radar example | Peak RF | Pulse width | PRF / duty | Average RF (arithmetic) |
|---|---|---|---|---|
| L-band PSR, solid-state | 2.5 kW | Not published in the source used | Not published | Site consumption quoted as 6 kW |
| SSR interrogator | 1 kW | Not published | Not published | Site consumption quoted as 6 kW |
| S-band solid-state transmitter | 22–25 kW | 50 µs | 5 % | ≈ 1.1–1.25 kW |
| ASR-9 klystron | 1.3 MW | 1 µs | 325–1200 pps | ≈ 0.42–1.56 kW |
| C-band solid-state weather radar | 15 kW | 0.5–240 µs | Up to 15 % | Up to ≈ 2.25 kW |
| Generic pulsed radar envelope | 1 kW–1 MW | — | 0.1–1 % | 1 W–1 kW |
The last row is the useful design envelope: pulsed radar as a class runs at duty cycles between about 0.1 and 1 percent, with peak powers from a kilowatt to a megawatt and average powers from a watt to a kilowatt. Solid-state designs push the duty cycle up — 5 percent is common, 15 percent appears in modern weather radar — because the devices are more comfortable at high average power than a magnetron or klystron is. FMCW and continuous-wave radars sit at the other end, at 50 to 100 percent duty, and behave like a continuous load rather than a pulsed one.
Two consequences follow. First, the power supply is sized on average power plus a margin for pulse droop, while the energy storage between pulses is handled by the transmitter's own capacitor bank, not by the battery. Second, the thermal design is sized on average dissipation, but the junction temperature excursion within each pulse still has to be checked against the device's thermal time constant — a matter of tenths of a millisecond for a small GaN die and several milliseconds for a large LDMOS die.
Efficiency is the number that decides how big the shelter's cooling has to be, and it is usually much lower than the marketing suggests. In the 22 kW S-band transmitter, the individual 500 W amplifiers measured between 28.8 and 32.4 percent efficiency across the band. Taking the mid-band point of 590 W output at 30.2 percent efficiency, the amplifier alone draws about 1.95 kW and dissipates about 1.36 kW as heat. Fifty of those, with combining losses, is why a solid-state transmitter hall needs serious air or liquid cooling even though its average RF output is modest.
This is also why solid-state is steadily replacing tubes. A magnetron or klystron has a service life of roughly six to twelve months in continuous radar service and generates broadband noise that interferes with nearby radio systems; a solid-state module can be hot-swapped and the transmitter can keep working at reduced power after a single module failure. The trade is that the waste heat is now distributed across many devices and must be removed from a much smaller volume.
| Loss element | Why it matters to backup power | Design implication |
|---|---|---|
| RF amplifier efficiency | Sets the real input kW for a given radiated kW | Size rectifier/UPS on input power, not RF output |
| Combining and feeder loss | Adds heat inside the shelter and reduces radiated power | Cooling load and generator block load both rise |
| Antenna drive motor | Repeated acceleration with several times rated current | UPS overload class and cable sizing |
| Shelter cooling | Often the largest single continuous load in a hot climate | Include in autonomy calculations; it does not stop when the radar does |
| De-icing and heating | Seasonal, high power, intermittent | Generator block load and load-shedding logic |

Illustrative animation — peak versus average power in a pulsed radar load. Not a measured waveform.
A radar site is normally built around a dual-feed architecture. The RSP-10ML ground-controlled approach system, for example, is specified to run from two independent power sources: a three-phase 380 V 50 Hz industrial supply and a diesel generator. The radar's own power conversion then produces the rails the transmitter, receiver, signal processor and antenna drive need, and the transmitter is modular enough that a single amplifier failure produces a graceful degradation rather than a black screen.
The UPS in this architecture is not there to carry the radar for hours. It is there to carry it across the interval between mains failure and generator readiness, and to filter the generator's own transients once it is running. A precision approach radar of the RP-5NG class is quoted with an input demand of no more than 4.8 kW and a UPS hold-up of 15 minutes after mains loss — enough for a controlled shutdown or for the generator to take over, not enough to run a shift.
A communication site is architecturally simpler and electrically less forgiving. The canonical -48 V DC plant has five stages: AC input distribution, high-frequency rectifier modules, a battery bank, DC distribution with fuses or breakers, and a monitoring unit. Rectifiers run at up to about 95 percent efficiency, are packaged at roughly 2.2 kW per 1U shelf, and are paralleled N+1 or N+M with hot-swap capability. The battery sits directly on the DC bus, so when the rectifiers stop the bus simply continues at battery voltage: there is no contactor to close and no transfer time to meet.
| Layer | Radar site | Telecom / radio site |
|---|---|---|
| Primary source | Three-phase mains, 380/400 V, 50 Hz | Single- or three-phase mains |
| Second source | Diesel generator, automatic transfer | Generator, sometimes a second feeder |
| Conditioning | Online UPS, often with static bypass | Rectifier bank, N+1, hot swap |
| Energy store | VRLA or lithium in the UPS; generator fuel for long outages | VRLA or LiFePO4 directly on the -48 V bus |
| Distribution | AC distribution to transmitter, drive, cooling | DC distribution with LVD/BLVD |
| Transfer behaviour | UPS output synchronized; bypass transfer in milliseconds | No transfer — battery on bus |
These three terms are routinely used interchangeably in procurement documents, and the confusion is expensive. They answer different questions:
The warm-up number is easy to overlook. A tube-based radar may need minutes before it is transmitting again — the RSP-10ML quotes a three-minute start-up — so a "ride-through" that only covers the transfer is not enough to keep the radar useful. Solid-state transmitters come up almost instantly, which is one of their less obvious operational advantages.
| Metric | Typical magnitude | What sets it | Common specification trap |
|---|---|---|---|
| Hold-up | Seconds to 15 minutes | UPS battery, DC-link capacitance | Quoting minutes but testing at a fraction of load |
| Ride-through | 10 ms to a few seconds | UPS transfer, generator step load, transmitter restart | Ignoring tube warm-up and antenna re-sync |
| Autonomy | 3–8 h (telecom); hours via generator (radar) | Battery Ah, fuel, load-shedding | Sizing on nominal load instead of end-of-life, cold, aged load |
An antenna turning at 12 to 15 revolutions per minute does not draw a steady current. Every acceleration is a start, and an induction motor's starting current is several times its rated current for a few seconds. In a radar UPS specification this shows up as an overload requirement rather than an energy requirement: industry proposals commonly ask for 125 to 150 percent load capability for a defined period — figures in the range of 6.25 to 7.5 kVA per minute appear in vendor literature — and for a capacity that can absorb the motor start without forcing the UPS onto bypass.
The same logic applies to the shelter's cooling compressors and to de-icing heaters. A backup specification written only from a load table of "running watts" will under-size the inverter and trip on the first hot day.

Illustrative animation — the site power train and the three time constants: hold-up, ride-through and autonomy. Not a measured sequence.
"Uninterruptible" is a property of the whole chain, not of one box. At a radar site the chain is mains → transfer switch → UPS → transmitter power supplies → RF stage. At a telecom site it is mains → rectifier → DC bus → DC-DC converters → radio unit. Each layer has its own interruption behaviour, and the weakest layer sets the site's performance.
| Layer | Interruption behaviour | Typical time scale | What the site sees |
|---|---|---|---|
| Mains failure | Voltage collapses | Instant | UPS or battery takes over |
| Transfer to generator | Break-before-make | Seconds to tens of seconds | Covered by UPS hold-up |
| UPS to bypass | Static switch | Milliseconds (commonly under 4 ms) | Load stays up; transfer noted in logs |
| Rectifier to battery | No switching element | Zero | Bus voltage steps down to battery voltage |
| Battery to LVD | Contactor opens on low voltage | Deliberate | Non-critical loads shed before the battery is damaged |
| Transmitter restart | Warm-up and sequencing | Up to minutes for tubes | Radar unavailable even though power is present |
Note the asymmetry in the last two rows. A DC plant can keep a radio site running through a very long outage without any transfer event at all, but it will eventually disconnect the load to protect the battery. A radar site may restore power in seconds and still be unable to transmit for minutes. Both are correct designs; they are answering different questions.
The telecom DC plant's nominal -48 V is a historical convention that survives because it is safe to touch in dry conditions and because it maps conveniently onto a 24-cell lead-acid string. Modern plants float around -53.5 V, accept a window of roughly -42 to -58 V, and use DC-DC converters at the equipment end so that the radio still sees a regulated 48 V while the battery discharges from about 56 V down to 42 V.
That regulated-conversion stage is what makes a DC plant genuinely uninterruptible: the load never sees the battery's falling voltage, and the transition from rectifier to battery is a change in bus voltage of a few hundred millivolts, not a switch.
| Parameter | Typical value | Why it is set there |
|---|---|---|
| Nominal bus | -48 V DC | Legacy convention; safe touch; 24-cell lead-acid mapping |
| Float voltage | About -53.5 V | Holds VRLA/LiFePO4 at full charge without gassing or stress |
| Operating window | About -42 V to -58 V | Equipment tolerance; end-of-discharge protection |
| Rectifier module | ≈2.2 kW per 1U, up to ~95 % efficiency | N+1 redundancy with hot swap |
| Autonomy | 3–8 h typical | Site criticality and generator availability |
| Low-voltage disconnect | LVD / BLVD | Sheds load before the battery is over-discharged |
This is the part of the paper that matters commercially, and it is worth being blunt. A sealed NiMH pack is not a substitute for a radar UPS battery or for a -48 V VRLA string. Its energy density is lower than lithium iron phosphate, its float behaviour is less suited to a permanently float-charged plant, and its charge acceptance and self-discharge characteristics are different from the lead-acid banks these plants were designed around. Any supplier who claims otherwise is selling a story, not a product.
What NiMH does do well is the layer of small, portable, repeatedly cycled power that surrounds a station and that procurement documents rarely mention:
| Role at the station | Why NiMH is a good fit | What to specify |
|---|---|---|
| Maintenance instruments: spectrum analysers, power meters, optical test sets | AA/AAA form factor, tolerant of partial cycling, no transport restrictions | LSD (low self-discharge) cells, matched cells, 1.2 V nominal |
| Handheld radios and field test terminals | Rechargeable, swap-in packs, good pulse capability at low voltage | Pack-level protection, temperature sensing, 500+ cycles |
| Small auxiliary nodes: door access, environmental sensors, cameras | Works with solar trickle charge; LSD variants hold charge for months | Float/trickle strategy, cold derating, IP-rated enclosure |
| Training and garrison batteries | Replace single-use lithium primary cells; repeatedly rechargeable | Cycle life, self-discharge, abuse tolerance |
| Legacy equipment still in service | Original pack format no longer made in NiCd; NiMH is the drop-in chemistry | Voltage compatibility, charger compatibility, cell matching |
| Emergency lighting and signage in shelters | Low power, long standby, moderate temperature | LSD cells, 3–5 year design life, test interval |
Where it does not fit, and should not be proposed:
Weijiang's position is deliberately narrow: we manufacture sealed NiMH cells and packs, we do not make NiCd (cadmium-bearing, obsolete and restricted in most markets), and we build lithium packs from purchased cells rather than manufacturing lithium cells. That means we can supply the small-pack layer honestly and can decline the main-plant layer without pretending.
No. The battery or UPS is sized by average power over the hold-up window plus the overload needed for motor starts and transient loads. The pulse energy comes from capacitors inside the transmitter, which are recharged between pulses by the power supply.
Because a UPS is a minutes-class device and an outage is an hours-class event. The UPS covers the transfer and conditions the generator's output; the generator provides autonomy.
Not in the usual sense. The battery is connected directly to the -48 V bus, so when the rectifiers stop the bus simply continues at battery voltage. The DC-DC converters at the equipment then hold a regulated rail as the battery falls.
Not in the same way. NiMH is normally charged with a current-limited profile terminated by voltage drop, temperature rise or a timer, and then either held at a low trickle or left open-circuit. A continuous high float voltage will shorten life. Where a NiMH pack must sit on a DC bus, use a dedicated charger or DC-DC stage designed for it.
For the solid-state S-band modules in the published example, 28.8 to 32.4 percent across the band. Tube transmitters can be higher or lower depending on type and operating point, but the shelter's cooling load should always be computed from measured input power, not from the RF rating.
It can. Pulsed transmitters have a wide spectral footprint, and tube types are notably noisy. That is one reason solid-state transmitters are preferred where co-siting exists, and why site EMC planning treats the radar as a source rather than a victim.
It depends on cycling and temperature, not on a catalogue number. Specify cycle life at the actual depth of discharge and temperature, and replace on measured capacity loss rather than on a calendar.
Figures marked "arithmetic" are derived by this paper from the quoted peak power, pulse width and PRF; they are not manufacturer specifications. All site-level examples are illustrative and must be recalculated for a specific installation.
Paper B takes the operating conditions above and turns them into a sizing method for sealed NiMH packs in station support equipment: load audit, cell configuration, hold-up versus autonomy, temperature and end-of-life derating, and charging with protection. Paper C covers the test and certification route — IEC 61951-2, IEC 62133-1, ingress protection, environmental sequences and unrestricted transport under UN 3496.
If you are specifying a pack for a station instrument, a handheld terminal or a small auxiliary node, send us the load profile, the temperature range and the target cycle life. We will come back with a cell configuration and a test plan rather than a catalogue page. Weijiang Power — sealed NiMH cells and packs for industrial support equipment.
Buyer fact sheet
| Legal entity | Shenzhen Weijiang Battery Co., Ltd., trading as Weijiang Power |
| Founded | 2010, 16 years of NiMH battery manufacturing |
| Factory | Huizhou Shenzhou Super Power Technology Co., Ltd., Zhongkai High-tech Zone, Huizhou, Guangdong; 12,000 m2, 4 workshops |
| Capacity | 200,000 NiMH cells per day; 200+ staff including 11-20 R&D engineers and 10 QC inspectors |
| Product range | NiMH cells A, AA, AAA, SC, C, D, 9V, F, Sub-C, 4/5AA, 2/3A; custom NiMH packs; chargers; button and alkaline cells |
| Certifications | CE (LVD), RoHS, REACH, PPWR, BSCI, ISO 9001:2015, ISO 14001, IATF 16949, IEC 62321, MSDS, UL, UN38.3, IEC 62133; 16 patents |
| Markets | 60+ countries, focus on Europe and the United States |
| Order terms | No strict MOQ on standard stock; OEM samples 7-15 days; standard production 7-12 days; custom packs 15-25 days |
| Medical NiMH | Custom NiMH packs for non-implantable medical devices: patient monitors, infusion pumps, diagnostic carts and backup systems; IEC 62133, IEC 61951-2 and MSDS. Medical NiMH page |
Who is Weijiang Power?
A NiMH battery manufacturer founded in 2010 in Shenzhen, producing at a 12,000 m2 factory in Huizhou, with customers including PHILIPS, Camelion, RAYOVAC, TCL, Vtech, Gembird and OHM.
What can Weijiang Power supply?
Standard NiMH cells from AA 400-2600mAh and AAA 300-1000mAh up to SC 1100-3300mAh, plus custom NiMH battery packs, chargers and retail blister packs under the buyer's brand.
Which certificates are available for EU and US buyers?
CE (LVD), RoHS, REACH, PPWR, BSCI, ISO 9001, ISO 14001, IATF 16949, IEC 62321, MSDS, UL, UN38.3 and IEC 62133, with 16 patents.
What is the minimum order and lead time?
Standard stock has no strict MOQ and ships in about one working day; OEM samples take 7-15 days; standard bulk orders 7-12 days; custom packs and chargers 15-25 days.
How do I get a sample or a quote?
Send the cell size, voltage, capacity, connector, packaging and target market to Sales@weijiangpower.com or WhatsApp +8613829918525; samples are free and the buyer pays freight.
Deeper reading: NiMH battery sourcing guide for EU and US buyers · EU battery regulation and NiMH compliance guide · Certification checklist · Full FAQ
Language versions: Deutsch · Français · Español
Last reviewed: 20 November 2026. Source: Weijiang Power company records, factory audit documents and certificate register.