SmartDC is one place to see the electricity feeding your servers, from the utility connection through the UPS to the rack, and to catch what is going wrong while it is still an advisory and not an outage.
A data center depends on a long chain of electrical equipment: the utility connection, the UPS that converts power to DC and back, the batteries that carry the load through a failure, and the distribution that reaches every rack. A problem rarely starts where it finally shows up. A connection can run warm long before anything trips, and a disturbance on the incoming supply only matters if it gets through to the servers.
SmartDC measures that whole chain and brings it together in one place, on both independent power paths. It watches the electrical system, not the servers’ software or network: if you already monitor those, this is what watches the power underneath them.
Power reaches a server through a chain, and each point on it answers a different question. This is one path. A second, independent copy runs beside it, with no DC tie between them, so losing one leaves the other carrying the load.
Energized and unloaded in normal operation. It is measured anyway, because it is the way around the UPS.
What the utility is delivering, before anything of ours touches it.
What actually reaches the UPS, and how the rectifier loads the supply.
The rectifier, the battery and the inverter sections all meet here. Measured for voltage, ripple, current, heat and insulation.
What the UPS is producing for the load, whatever the supply is doing upstream.
What finally reaches the racks. This is the reading the servers live with.
Charge, health, cell voltages, temperatures and estimated runtime.
A reference, not a limit. Voltage class, grounding, ratings and sensor selection all get mapped to your installation.
It follows the power through every stage, tests readings against your limits, and shows how far each one can be trusted. All of it runs in the demo, on a simulated data center in your browser, and nothing you do there touches any equipment.
The whole site in one view: the power chain from the utility to the rack on both paths, what is protected and how much margin is left, how long the batteries would last, and every active alarm.
Open in the demo →The live state of the supply at whichever meter you choose: voltage, current, frequency, real and apparent power and power factor on each phase, with the conversion path, a phase diagram, line voltages, unbalance, neutral current and 24 hours of energy and demand.
Open in the demo →Distortion broken down order by order, from the 2nd harmonic to the 50th, for voltage or current and for each phase. Alongside it, the shape of the waveform, neutral loading, crest factor and flicker, with waveform distortion kept apart from the demand-based figure a compliance study uses.
Open in the demo →Every sag, swell, interruption and transient, with how deep it went, how long it lasted and which phases it touched. The disturbance upstream sits next to what the UPS output did at the same moment, so you can see whether it ever reached the load.
Open in the demo →Take a reading from a meter or import one from a CSV file, and check it against limits you set: voltage, frequency, distortion, unbalance, power factor and phase rotation. Each run freezes its evidence and its limits, and exports as a CSV report or a JSON record.
Open in the demo →Both 800 V buses at a glance: power, voltage, loading, a one-line diagram, demand trends and every feeder. Choose a feeder and you get its voltage, current, power and joint temperature, its breaker state, and what its load would become if its paired feed were lost.
Open in the demo →Pole-to-pole and pole-to-earth voltage, current, ripple and power for each bus, the loss across the distribution, and triggered captures of the fast dips a slow display would never catch.
Open in the demo →A map of every bus joint and tap-off, by temperature or by rise above ambient, with rate of rise, voltage drop and estimated resistance. A joint running hotter than its neighbours stands out, and a sensor that has stopped reporting shows as unknown, not as fine.
Open in the demo →Active conditions explained in plain language, insulation and residual current, protective device status, alarm thresholds and event history. Acknowledging an alarm records that someone saw it, and the condition stays active until it actually clears.
Open in the demo →Energy and rolling demand, the capacity left if one path is lost, the status of every rectifier module, and each battery’s charge, health and estimated runtime.
Open in the demo →Which sensors are reporting, how the gateways and clocks are doing, which channels have gone stale, and what is still waiting on commissioning. It is how you know how far to trust every other reading.
Open in the demo →Every measurement defined in one searchable, exportable place: its unit, its type, the instrument behind it and how to read it, with the standards and technical references behind them.
Open in the demo →A daily operations report written from the same readings: energy, peak demand, capacity, backup, the hottest joints, power quality events and the alarm log, ready to print, save as a PDF or export.
Open in the demo →See what the system does on a bad day. Take out a utility feed or a whole UPS path, overheat a joint, fail a rectifier module or fault the insulation, and every reading, diagram and alarm responds.
Open in the demo →Joint J04-A climbs to about 63 °C, around 35 K above its local ambient, while its neighbours sit in the mid-40s. The voltage drop across it puts its estimated resistance at about twice theirs, and it is rising faster than any of them.
It crosses the 60 °C advisory limit. The advisory names the joint, the feeder and the reason, next to the current, the rate of rise and what its paired feed would carry if this one were lost. The system itself trips nothing.
Acknowledging it records that someone saw it, and the condition stays active. The answer is an inspection arranged under the site’s own electrical work procedures, while it is still a maintenance job.
Open Thermal health, choose J04-A, and follow it from the map to the feeder, the advisory and what its paired feed would carry.
The fault simulator takes out a utility feed or a whole UPS path, overheats a joint, fails a rectifier module or faults the insulation. Every reading, diagram and alarm responds, and nothing touches any equipment.
An alarm that resets as soon as a reading dips is noise, so each limit comes with a delay and a separate reset point. These are starting points for monitoring, not protective relay or breaker settings, and they get reviewed against your equipment.
Ninety-nine definitions across fifteen disciplines. Which ones apply depends on the sensors fitted at your site, and the cadences are targets. Open a group for the detail.
V01 · V DC
MeasuredMean voltage between positive and negative conductors at each main bus and feeder. The nominal system value is 800 V pole-to-pole.
V02 · V DC
MeasuredSigned positive-conductor potential relative to protective earth. On a floating IT system this may drift; it is not a regulated +400 V rail.
V03 · V DC
MeasuredSigned negative-conductor potential relative to protective earth. Together with positive-to-PE voltage, reveals displacement of the floating system.
V04 · V DC · time
MeasuredLowest and highest pole-to-pole voltage in the selected period, with timestamps and the meter aggregation interval. Separate from waveform event extrema.
V05 · V · %
DerivedUpstream voltage minus feeder-end voltage; percent drop uses upstream voltage. Compare time-aligned readings under stable load and account for meter uncertainty.
V06 · V DC
DerivedCommon-mode potential = (Vpositive-to-PE + Vnegative-to-PE) / 2. Diagnostic only; a balanced value does not prove healthy insulation.
I01 · A DC
MeasuredCurrent through each independent bus. Positive denotes source-to-load flow; negative denotes reverse flow. Do not sum outgoing and returning conductor magnitudes as load current.
I02 · A DC
MeasuredBranch current for each rack or zone. Retain the circuit rating and polarity convention alongside the measurement.
I03 · A DC
MeasuredCurrent in each parallel positive or negative conductor where sensors are fitted. Supports identification of uneven sharing and overloaded individual paths.
I04 · %
Derived100 × maximum absolute deviation from the mean conductor current / absolute mean current, calculated for parallel conductors of the same pole. Suppress near zero load.
I05 · A · time
MeasuredHighest current with capture interval and timestamp. High-speed event peaks are distinct from peaks of slow display aggregates.
P01 · kW
MeasuredMean of simultaneous instantaneous voltage × current samples. Positive is import to the load. Mean voltage × mean current is only a steady-state approximation when ripple correlation is negligible.
P02 · kW
DerivedAverage imported power over a rolling 15-minute window ending at each 5-minute sample boundary. The dashboard uses rolling demand rather than fixed demand blocks.
P03 · kW · time
DerivedLargest rolling 15-minute demand in the selected period, evaluated at 5-minute sample boundaries, with interval start and end timestamps. Independent per-path maxima must not be added.
P04 · kWh
MeasuredTime integral of positive active power. Keep cumulative, interval and daily totals distinct, with reset/rollover handling.
P05 · kWh
MeasuredTime integral of negative active power, reported as a positive exported-energy quantity. Do not hide reverse energy inside a net-only register.
P06 · % · A spare
DerivedAbsolute current divided by the applicable continuous operating limit. A configured nameplate rating is not automatically the allowable installed operating limit.
P07 · kW · %
DerivedLoss = input power minus summed downstream power, adjusted for storage and other branches. Efficiency = output/input. Flag results below combined meter uncertainty.
Q01 · V RMS
MeasuredRMS of v(t) minus its DC mean. Always retain the measurement bandwidth, window and sample rate; example design band is 10 Hz–20 kHz.
Q02 · V p-p
MeasuredMaximum minus minimum AC-component voltage over the stated capture window and bandwidth. Not interchangeable with RMS ripple.
Q03 · %
Derived100 × AC-component voltage RMS / absolute DC mean voltage, using the same observation interval. State the bandwidth and suppress near zero DC voltage.
Q04 · A RMS · %
MeasuredRMS of current after removing its DC mean, with declared bandwidth. Relative ripple uses absolute mean current and is invalid near zero current.
Q05 · V · ms · count
MeasuredEvent type, threshold, start time, duration, minimum/maximum and pre/post-trigger waveform. Event duration comes from local fast acquisition, never inferred from 2-second dashboard polling.
Q06 · V/ms
DerivedMaximum and minimum dV/dt over a specified event window. Derivative results must include sample interval and filtering to avoid presenting noise as a transient.
Q07 · A/ms
DerivedMaximum and minimum dI/dt with capture interval, sensor bandwidth and filter settings. Useful for dynamic load and fault-event diagnostics.
T01 · °C
MeasuredPositive and negative bar joints, terminations and conductor temperatures with location identifiers. IR measurements require valid emissivity and line-of-sight assumptions.
T02 · °C
MeasuredLocal air temperature paired with each thermal zone. Use it to distinguish high room temperature from local connection heating.
T03 · °C rise
DerivedJoint temperature minus local ambient temperature using aligned readings. Operating limits depend on the approved equipment, insulation and installation.
T04 · °C/min
DerivedSlope over a specified recent window, such as five minutes, with concurrent current/loading trend. A fast rise can matter before an absolute limit is reached.
T05 · mV
MeasuredSmall differential voltage across a defined connection under load. Measurement must tolerate the bus common-mode voltage and use suitable isolation.
T06 · µΩ
DerivedR = 1000 × joint voltage drop in mV / current in A. Valid only at sufficient, stable current after offset compensation; this is an online estimate, not an offline micro-ohmmeter test.
T07 · W
DerivedApproximate resistive loss = I RMS squared × R. With R in µΩ divide by 1,000,000. Distinguish this local estimate from metered whole-path losses.
G01 · MΩ
MeasuredEffective resistance of the connected active system to PE. Trend each independent bus and retain response time, self-test status and lowest recorded value.
G02 · MΩ
DerivedPositive-to-PE and negative-to-PE estimates only if the selected IMD provides valid pole-specific data. Never infer an exact insulation resistance from pole voltage alone.
G03 · µF
MeasuredConnected system capacitance to PE. It affects insulation-monitor response and fault behavior; retain the IMD permitted capacitance and response characteristics.
G04 · mA DC / RMS
MeasuredResidual current around all active conductors, with bandwidth and DC/RMS basis stated. On a floating system this complements the IMD; it does not replace insulation-resistance monitoring.
G05 · state · time
StatusDevice healthy/fault, test result and measuring-lead continuity. If buses are coupled, coordinate IMDs according to the approved architecture; do not assume independent monitors can operate on one connected system.
G06 · circuit · state
StatusIdentified feeder, locator confidence/status and event time when a fault-locator is installed. Show Not instrumented if unavailable; residual current alone does not prove the location of an IT insulation fault.
R01 · state
StatusClosed, open, tripped or unknown for every main and branch device. Distinguish position feedback from commands and from current flowing through the circuit.
R02 · cause · time
StatusLatched cause, trip timestamp, operation sequence and captured fault data where supported. Preserve acknowledged versus cleared states and source timestamp quality.
R03 · state
StatusTrip-unit healthy, auxiliary supply healthy and trip-circuit continuity where supervised. No status may be labelled healthy when the channel is stale or absent.
R04 · V DC · A · kA
ConfiguredRated DC voltage/current and applicable interrupting/withstand ratings, with required pole configuration and conditions. These are engineering metadata, not measured fault current or a live safety margin.
R05 · A · V · ms · °C · MΩ
ConfiguredPickup, time delay, hysteresis, latching and reset rules, separately identifying hardware protection and dashboard alarms. Example dashboard thresholds are not approved equipment settings.
R06 · state
StatusShow each fitted fuse alarm, contactor position, precharge ready/fault and surge protective device condition separately. Use Not instrumented for devices without feedback.
R07 · state · time
StatusDetector health and alarm record only where an appropriate detector is installed. An absent alarm, normal bus telemetry or dashboard status cannot establish absence of arcing.
S01 · online / total · state
StatusAvailable, failed, standby and current-limited modules, including controller health. Available output capacity must reflect failed modules and operating derating.
S02 · A / module · %
DerivedIndividual module output currents and maximum deviation from the mean among active modules. Evaluate at sufficient load and account for differing module ratings.
S03 · %
Derived100 × DC output active power / AC input active power over the same interval. AC power quality belongs to the upstream meter view; it is not a DC bus measurement.
S04 · %
DerivedBMS-reported remaining charge fraction with data age and BMS estimation quality. State of charge is an estimate, not a direct guarantee of available runtime.
S05 · %
DerivedBMS-reported health using the manufacturer definition and reference capacity. Track test date and estimator availability.
S06 · kW · A
MeasuredSigned charge/discharge power and current, permitted discharge power/current, temperature derating and contactor readiness. Show polarity convention explicitly.
S07 · kWh · min
DerivedEstimated runtime = bus-side usable energy / protected load × 60. Apply conversion efficiency only if energy is specified upstream of the converter. BMS discharge limits and actual contingency load still apply.
S08 · V · mV · °C
MeasuredLowest/highest cell voltage, cell spread, hottest/coldest temperature and sensor location. Include BMS warning and fault states.
S09 · A/B state
StatusAvailability of each complete source-to-load path, including module, breaker and dual-input load readiness. A closed main breaker alone does not prove redundancy.
S10 · MW · %
DerivedEvaluate surviving source, bus, feeder and load-input limits after one path is lost. For two 800 V / 4000 A buses, nominal per-path power is 3.2 MW; 6.4 MW installed is not 6.4 MW redundant capacity.
D01 · s · quality
StatusGood, stale, missing, out-of-range or substituted flags per measurement. Suppress derived healthy conclusions when any required input is invalid.
D02 · ms · state
MeasuredClock source, offset and synchronization state for meters, gateways and event recorders. Unsynchronized readings cannot support precise event ordering or small loss calculations.
D03 · date · % / absolute
ConfiguredSensor/meter identification, range, accuracy, calibration date and due date. Include voltage-divider and current-sensor uncertainty when judging small changes or calculated losses.
D04 · S/s · Hz · s
ConfiguredAcquisition sample rate, analog bandwidth, filters, averaging window, dashboard refresh and retention. These are distinct; slow polling cannot reconstruct fast electrical transients.
D05 · online / total · ms
StatusExpected versus reporting channels, link latency, packet loss and gateway state. Coverage must be calculated against the configured inventory, not only devices that happen to report.
AC-V01 · V RMS
MeasuredTotal true-RMS L1-N, L2-N and L3-N voltages for a configured three-phase four-wire wye system. Profiles use 480/√3 V at 60 Hz or 400/√3 V at 50 Hz; line-to-neutral values are not assumed identical.
AC-V02 · V RMS
MeasuredRMS of L1-L2, L2-L3 and L3-L1 differential waveforms, including modeled harmonics. Under unbalance, calculate each line voltage from its actual phase waveforms rather than multiplying an average by √3.
AC-V03 · Hz · time
MeasuredMeasured fundamental frequency with timestamped minimum and maximum over an explicitly defined interval. Frequency and 50/60 Hz belong to the AC-side measurement points, not the internal 800 VDC link.
AC-V04 · sequence
StatusPositive sequence L1 → L2 → L3, reverse sequence or unknown. Keep phase labels and sensor polarity consistent through commissioning; a missing or invalid phase must invalidate the sequence conclusion.
AC-V05 · %
Derived100 × magnitude of negative-sequence voltage / magnitude of positive-sequence voltage. This differs from maximum RMS magnitude deviation from the three-phase mean; label both definitions separately.
AC-V06 · V RMS
MeasuredNeutral-to-protective-earth potential at the monitored AC point, with connection and bandwidth stated. This requires its own measurement channel and is not inferred from phase-voltage unbalance.
AC-I01 · A RMS
MeasuredTotal true-RMS current per phase, including harmonic components. Current-sensor ratio, polarity, bandwidth, burden and saturation performance must match the application.
AC-I02 · A RMS
MeasuredNeutral current includes phase unbalance and triplen harmonics. It can be measured directly or derived by summing the three instantaneous phase currents, and the display should say which.
AC-I03 · %
Derived100 × negative-sequence current magnitude / positive-sequence current magnitude. Suppress ratios at insufficient current and distinguish sequence unbalance from simple phase-current magnitude spread.
AC-I04 · ratio
DerivedAbsolute waveform peak divided by total RMS over the same capture interval. A distorted waveform may exceed the sinusoidal √2 ratio. Current crest factor is undefined on the unloaded bypass.
AC-I05 · V · A · degrees
DerivedPer-phase fundamental RMS magnitude and phase angle on a common reference. Current displacement from its own voltage establishes fundamental lead/lag; phasors do not depict total harmonic distortion.
AC-I06 · mA RMS · bandwidth
StatusResidual current measured around all live conductors is separate from load neutral current. Converter loads require a sensor suited to the possible AC/DC components. Preserve sensor type and bandwidth.
AC-P01 · kW
MeasuredMean instantaneous voltage × current per phase, summed for total power. Includes real harmonic power where both voltage and current harmonics exist. Positive denotes import from source toward the load.
AC-P02 · kVA
DerivedThis dashboard defines total apparent power as the arithmetic sum of per-phase Vrms × Irms. Other effective-apparent-power conventions exist; preserve the meter definition when integrating real equipment.
AC-P03 · kvar
DerivedSum of per-phase V1 × I1 × sin(displacement angle). Positive is lagging. Q1 excludes harmonic and distortion contributions; do not assume kVA² = kW² + Q1² under distorted conditions.
AC-P04 · ratio
DerivedTrue PF = active power / apparent power using the stated total-kVA convention. Includes waveform distortion effects and is undefined when the monitored bypass carries zero apparent power.
AC-P05 · ratio · lead/lag
DerivedCosine of the fundamental voltage-current displacement angle per phase. Total DPF here is total fundamental active power / summed fundamental apparent power. It is distinct from true PF on nonlinear loads.
AC-P06 · kWh
MeasuredAccumulate positive and negative active energy separately; net energy is import minus export. Totals derived from other meters are not billing-grade registers.
AC-P07 · kW · time
DerivedRolling 15-minute average active power evaluated at 5-minute sample boundaries, plus its maximum over the stated history window. Per-path peaks are not added unless their demand intervals coincide.
AC-H01 · %
Derived100 × root-sum-square voltage RMS in harmonic orders 2–50 / fundamental voltage RMS, separately by phase. This excludes interharmonics and components above order 50; it is not a universal compliance verdict.
AC-H02 · %
Derived100 × root-sum-square current RMS in orders 2–50 / fundamental current RMS. High THD at light load needs context. THD is undefined on the unloaded bypass and must not be substituted for TDD.
AC-H03 · % fundamental · V · A
MeasuredPer-phase magnitude and phase for each harmonic order, with percent values referenced to that phase’s fundamental, across orders 2 through 50.
AC-H04 · %
Derived100 × harmonic current RMS / maximum-demand fundamental current IL. Initially unavailable until a valid IL is configured. A configured IL, a snapshot value or an arbitrary current-THD threshold does not establish IEEE 519 compliance.
AC-H05 · V / A · Hz
StatusComponents between integer harmonics require suitable spectral grouping, windowing and instrument capability. A zero placeholder must not imply that interharmonics are absent.
AC-H06 · V / A · kHz
StatusHigh-frequency conducted components require defined bandwidth, grouping, transducers and measurement methods. An orders-2–50 spectrum does not establish coverage of the 2–150 kHz band.
AC-Q01 · % nominal · ms · phase
MeasuredStore residual RMS voltage, affected phases, start/end time, duration and threshold basis.
AC-Q02 · % nominal · ms · phase
MeasuredStore elevated RMS voltage, event duration, affected phases and synchronized waveform/envelope. Compare to the configured nominal system voltage and documented operational threshold.
AC-Q03 · % nominal · ms · location
MeasuredRecord residual voltage, affected measurement point and duration. An interruption on an idle bypass source is distinct from loss of the inverter output or a load outage.
AC-Q04 · V peak · µs / ms
MeasuredInstantaneous peak voltage, polarity, ringing frequency and duration from an adequately sampled capture. A transient peak is not an RMS residual percentage.
AC-Q05 · V · A · ms
MeasuredCompare disturbance timing at the UPS input, bypass and inverter output. Retain mode, pre/post-trigger windows and clock uncertainty. A maintained output in one event does not prove equipment ride-through or transfer performance.
AC-Q06 · index · 10 min
MeasuredShort-term flicker severity from a completed 10-minute instrument interval. Pst cannot be calculated from a short snapshot waveform or a simple voltage-RMS spread.
AC-Q07 · index · 2 h
DerivedCube root of the mean of the cubes of 12 consecutive 10-minute Pst values. Retain all interval timestamps and validity.
AC-Q08 · % nominal · time
StatusIdentify rapid changes in RMS voltage using the adopted measurement method, thresholds, hysteresis and event flags. Sag/swell examples do not by themselves provide RVC coverage.
AC-S01 · state · time
StatusNormal double conversion, static bypass, battery operation or maintenance state from the installed system. Read-only: the monitoring system provides no switching commands.
AC-S02 · kW · %
DerivedRectifier AC input, 800 VDC link, DC-fed inverter sections, inverter AC output and downstream PDU output. Calculate losses using aligned meters; do not count the same energy at successive boundaries as separate facility load.
AC-S03 · V · A · kA · curve
ConfiguredContinuous limits, CT/PT ratios, short-circuit interrupting/withstand ratings and protection settings. Prospective fault current cannot be inferred from normal operating current.
AC-S04 · quality · time · date
StatusSource timestamp, freshness, missing channels, synchronization, calibration, polarity and instrument-class evidence. A mathematically valid snapshot does not establish real meter availability or calibration.
AC-S05 · configuration · evidence
ConfiguredIdentify the actual point of common coupling, demand-current IL basis, fault-level basis and applicable measurement duration/statistics. Operational thresholds provide advisories only; no IEEE 519 or IEC compliance pass is implied.
AC-S06 · V · Hz · degrees · state
StatusRetain voltage, frequency and phase-angle differences plus manufacturer-reported transfer permitted/inhibited. An energized bypass alone does not prove availability for a safe transfer.
A monitoring system is only useful if you can believe it, so most of the design is about being honest with what it knows and what it does not.
There are no switching or trip controls anywhere in the system. It can show you what the equipment is doing, and it cannot tell the equipment what to do.
Breakers, relays and the installed control system keep doing the protecting, and any trip is theirs. Monitoring sits beside them and does not replace them.
A last-known value never poses as a current one. A stale channel is labelled and left out of the totals, and an instrument that is not fitted says “Not instrumented”, not zero.
Each value carries its source timestamp, its quality flag and how it was acquired, and a calculated result inherits the age and quality of what went into it.
A display refreshing every 2 seconds cannot see a 12 ms dip. Fast events come from recorders at the equipment, with their own sampling, and never from the dashboard’s polling.
Measured, derived, status and configured values are kept apart. A live joint resistance is an estimate under load, not a four-wire micro-ohmmeter test on a de-energized joint.
Checks run against limits you set. IEEE 519 and IEC 61000-4-30 need a defined point of common coupling, a demand basis, documented instruments and proper observation periods, so a spot reading earns no pass or fail.
Two 3.2 MW paths are 6.4 MW installed and 3.2 MW redundant. The number that matters is what one surviving path can carry, so that is the number the system reports.
Power-quality meters at the incomer, rectifier input, bypass, UPS output and PDU, plus isolated DC meters, thermal sensors, insulation monitors and the BMS. Acquisition stays local to each device.
Read compatible Modbus RTU/TCP registers, then normalize units, device identity, timestamps and quality flags.
Publish through an approved MQTT, OPC UA or HTTPS integration into the historian and the facility’s monitoring system.
Read-only measurements, trends and an alarm workflow, showing the freshness and coverage of every channel.
Software alone cannot make a reading trustworthy, so commissioning builds these records with you before it goes live on your equipment. The workspace evaluates records; it does not create disturbances, run a load bank or switch your UPS.
Meter and sensor IDs, serial numbers, firmware, locations, ratios, ranges and polarity.
CT and VT ratios, polarity, the neutral channel and phase mapping, checked through to the display.
Approved thresholds, delays, hysteresis and reset policy, reviewed for the actual equipment.
Clock source and accuracy, calibration and due dates, and transducer uncertainty.
Which points are expected, how quality is flagged, and how data is acquired and stored.
Load steps, battery and bypass transfers, with synchronized recordings of input and output.
Most of what is on this page has to be wired into the electrical system: current sensors on the bus and feeders, isolated voltage inputs, thermal sensors on the joints, insulation monitors on each bus. Fitted afterwards, that work competes with live equipment for space and access. Planned from the start, it goes on the same drawings as everything else.
That is why it belongs with our work on the TPI partnership. We already plan the switchboards, the switchgear and the crew that installs them as one job from the first drawing. Monitoring belongs in that same conversation.
Still unsure about something? Call +1 (305) 205-6409 and talk it through with someone who will be on the job.
Planning or upgrading a data center anywhere in South Florida? Tell us what you are building, and we will show you how it fits.
