DC Link & Pre-Charge Design for Motor Drives
← Back to: Motor & Motion Control
This page is where I organise everything needed to bring a DC link up safely. Whenever a drive or cabinet needs controlled pre-charge, inrush limiting, weld detection and a clear checklist for sizing parts and picking ICs, I use this page as the reference instead of guessing.
What this page solves
This page explains how to bring a motor-drive DC link up smoothly instead of slamming large capacitors directly across the supply. The goal is to limit inrush current, avoid welded relays or MOSFET overstress and prevent nuisance trips on upstream breakers or fuses.
It focuses on the pre-charge path, main relay or contactor drive and basic weld-detection hooks around the DC link. Detailed gate-driver design, full STO implementations and front-end AC PFC topologies are handled on dedicated pages and only referenced here as interfaces.
After walking through this page, a designer can place the DC link pre-charge block correctly in the drive power tree, understand why a limited inrush path is mandatory and decide which inrush-limiting technique is appropriate for a given bus voltage and capacitance range.
DC link structure and pre-charge principle
A motor-drive DC link sits between the front-end supply and the inverter power stages. It typically consists of a rectifier or DC source, an optional pre-charge branch, a main relay or contactor and a bank of DC link capacitors that stiffen the bus for one or more drives.
When the main relay is closed with an empty capacitor bank, the DC link behaves like a short circuit and the current is limited mostly by wiring, transformer and source impedance. A dedicated pre-charge path inserts a resistor, NTC or MOSFET-based limiter so that the capacitors charge along a controlled voltage ramp before the main relay closes under a small remaining voltage difference.
Typical DC link capacitance in motor drives ranges from tens of microfarads up to several millifarads, while bus voltages span 24 V, 48 V and 80 V low-voltage systems to 400 V and 750–800 V industrial and traction drives. As bus voltage and stored energy increase, the pre-charge branch moves from simple resistor or NTC schemes toward MOSFET-based soft-start solutions with dedicated control and diagnostics.
In multi-drive cabinets, a single front-end DC link can feed several inverters. The pre-charge function is usually placed at the shared DC link entrance, while additional contactors or breakers segment the bus into zones. This page treats those segmentation points as interfaces and concentrates on how the pre-charge and main relay work together to deliver a predictable, repeatable DC link start-up.
Pre-charge design flow
The DC link pre-charge function can be implemented as a simple state machine that supervises the front-end supply, brings the capacitor bank up through a limited inrush path, hands over to the main relay and then validates that the contactor can both close and open as expected. Each step is tied to measurable quantities such as bus voltage, inrush current and relay feedback contacts.
The design starts with a small set of inputs: the DC bus voltage range, total capacitance, allowable inrush current, target pre-charge time and the available sensing channels for voltage, current and relay position. These parameters define the resistor, NTC or MOSFET sizing and the thresholds used when comparing DC link voltage to the input supply before the main relay is closed.
A typical sequence begins by checking that the front-end supply and low-voltage control rails are in a valid window. The controller then enables the pre-charge branch while keeping the main relay open and monitors the DC link voltage ramp. Once the DC link reaches a defined percentage of the input voltage, the main relay closes and the pre-charge path is turned off after a short overlap period.
After handover to the main relay, weld and open-circuit conditions are checked by observing how the DC link behaves during commanded on and off events. The resulting status is reported to a safety microcontroller or PLC, which can then apply the appropriate fault reaction and system-level interlocks if the pre-charge or contactor path does not behave as intended.
Inrush limiting technique selection
Inrush limiting on a DC link is usually implemented with one of three approaches: a fixed resistor that is bypassed by a relay, an NTC thermistor that presents a high resistance when cold, or a MOSFET-based soft-start using a linear or controlled current profile. Each method trades simplicity, cost, thermal behaviour and diagnostic capability in different ways.
A fixed resistor plus bypass contactor offers straightforward operation and low component count, but the resistor must be sized for significant short-term power dissipation and can be bulky. An NTC thermistor automatically reduces its resistance as it heats, which lowers steady-state losses, but its limiting behaviour degrades when the device is already hot and it is not suited to frequent restarts or high ambient temperatures.
MOSFET soft-start architectures use a controlled gate drive together with voltage or current sensing to shape the DC link ramp. This allows precise control of inrush current, accurate timing of the pre-charge phase and built-in fault detection, but requires a gate driver, measurement circuitry and careful safe operating area design, especially at 400–800 V and large DC link capacitance values.
For low-voltage 24–48 V systems with modest capacitance and infrequent start-up events, a resistor plus contactor often remains adequate. For medium-power AC/DC or DC bus systems where start-up intervals are long and ambient temperature is controlled, an NTC can be acceptable but must be derated for hot conditions. For high-voltage or high-energy DC links, multi-axis servo cabinets and drives that start and stop frequently, MOSFET-based soft-start or dedicated hot-swap and pre-charge controllers are typically the preferred choice.
| Method | How it works | Advantages | Limitations | Typical use range |
|---|---|---|---|---|
| Fixed resistor + contactor | Series resistor limits inrush during pre-charge, then a relay bypasses the resistor for normal operation. | Simple, easy to analyse, low silicon complexity and low cost for modest energy levels. | High short-term dissipation, bulky resistor, thermal stress under repeated start-up cycles. | 24–48 V buses, small servo drives, compact AGV controllers with moderate DC link capacitance. |
| NTC thermistor | Cold resistance is high to limit inrush, then resistance drops as the device heats during normal operation. | Self-adjusting behaviour, few components, low steady-state losses once warmed. | When the NTC is already hot, resistance can be too low to limit inrush, making frequent restarts and hot ambient operation risky. | Medium-power AC/DC front ends and DC links with long intervals between start-ups. |
| MOSFET soft-start | Controlled gate drive keeps the MOSFET in a defined region, limiting current or shaping the DC link voltage ramp. | Programmable inrush profile, strong diagnostics, suitable for high voltage and large energy storage. | Requires gate driver, sensing, control logic and careful SOA design, higher design complexity. | 400–800 V DC links, large capacitor banks, multi-axis servo and high-duty industrial or traction drives. |
An important constraint for NTC-based solutions is the cool-down time. If the drive is restarted quickly after a previous run, the thermistor may remain hot and its resistance low, so the DC link sees almost no inrush limiting. Designs with frequent on–off cycles, fast brown-out recovery or high ambient temperatures should therefore avoid using an NTC as the primary pre-charge element and move toward resistor plus bypass or MOSFET soft-start approaches.
Relay-weld detection
Relay-weld detection around the DC link ensures that contactors and pre-charge relays can both close and open as commanded. The objective is to detect welded-closed contacts that prevent the DC link from being fully de-energised and open-circuit failures that stop the bus from charging, then report these conditions to the safety controller or PLC for further action.
A common approach uses two series contactors and two-point voltage sensing. By driving K1 and K2 together and measuring voltage at the supply side, the DC link side and the midpoint between the contactors, it becomes possible to distinguish between successful opening, welded contacts and contacts that refuse to close. Auxiliary contact feedback further improves confidence in the actual contact position relative to the commanded state.
Current-based weld detection complements voltage sensing. After a commanded open event, the DC link current is expected to decay below a defined threshold within a specified time. If significant current persists despite an open command and a falling DC link voltage, a welded contact or unintended current path is suspected. Voltage-difference checks and current-decay checks are often combined to provide robust coverage across a wide set of operating conditions.
When MOSFETs or smart high-side switches act as the main DC link switch, weld detection shifts toward electronic feedback. Gate commands are compared with the measured drain-source voltage and bus current, and driver IC fault outputs are monitored for short-circuit or stuck-on conditions. From the DC link perspective, a shorted MOSFET closely resembles a welded relay and should be treated with the same level of caution in diagnostics and fault handling.
Weld and open-circuit results are summarised as status bits and diagnostic codes and passed to a safety microcontroller or PLC. These results feed higher-level safety logic, which can inhibit drive enable, command controlled braking or latch a fault that requires manual intervention. For networked drives, the same information can be exposed over CAN or time-sensitive Ethernet links so that a plant controller has a clear view of DC link disconnect integrity.
Interfaces to other system blocks
The DC link and pre-charge function sits between the front-end supply and the inverter power stages and interacts with several other subsystems. Understanding these interfaces helps keep responsibilities clear and avoids duplicating protection or state-machine logic on other pages in the motor and motion control hierarchy.
Toward the power stage, the pre-charge block delivers a stabilised DC link and a “bus ready” indication. The servo power stage, BLDC or ACIM drive then takes over PWM generation, gate driving and phase-current protection. Detailed gate-driver selection, inverter topology and short-circuit protection are handled in the dedicated servo and power stage pages and are only referenced here through enable and status signals.
Toward the MCU and safety controller, the pre-charge logic receives commands to start or stop pre-charge, drive the main contactor and run periodic self-tests. In return, it reports pre-charge completion, timeout or failure states and any suspected weld or open-failure conditions. These signals feed higher-level STO, HMI and safety monitor functions, which supervise overall drive readiness and decide whether power stages are allowed to run.
The PMIC and point-of-load regulators mainly provide the low-voltage rails needed for controllers, comparators and drivers. Their power-good outputs are part of the pre-charge start conditions, but the DC link block normally does not influence PMIC behaviour. In the opposite direction, the EMC and front-end PSU stage influences the allowed inrush profile, because its rectifier, PFC and filter design define the source impedance and the tolerable pre-charge current envelope.
Additional interfaces include brake choppers and dynamic braking modules hanging on the same DC link, as well as data logging and predictive maintenance functions that consume pre-charge and weld-detection diagnostics. Detailed design of braking, EMC and safety monitoring is covered on their respective pages, while this page concentrates on the DC link and pre-charge behaviour at their shared interfaces.
| Interface block | Relationship | Detailed page |
|---|---|---|
| Power stage / drives | Receives DC link voltage and “bus ready” status, returns drive enable and braking requests. | Servo power stage, BLDC / PMSM driver, ACIM V/F–FOC inverter |
| MCU / safety controller | Issues pre-charge and main contactor commands, receives pre-charge and weld-detection diagnostics. | STO, HMI and safety monitor pages |
| PMIC / PoL regulators | Supply control and sensing rails; provide power-good signals used as pre-charge start conditions. | Multi-rail PoL / PMIC topic |
| Front-end EMC / PSU | Defines source impedance and DC_OK signals, constrains allowable inrush current and pre-charge profile. | Front-end PSU for drives, EMC subsystem |
| Brake chopper, logging / PdM | Share the DC link and consume pre-charge and weld-detection status as part of braking and health monitoring. | Brake chopper & dynamic braking, data logging & predictive maintenance |
DC link & pre-charge design checklist
A structured checklist helps capture all key inputs for DC link and pre-charge design before sizing resistors, NTCs, MOSFETs and contactors. The entries below summarise bus voltage, stored energy, allowable inrush, timing constraints and contactor capabilities so that component selection, safety analysis and RFQs are based on a consistent set of numbers.
The DC link voltage range and total capacitance define the stored energy that the pre-charge path must handle. The allowed inrush peak and target pre-charge time then determine whether a fixed resistor, NTC or MOSFET soft-start is realistic, and whether the front-end PSU and contactors can tolerate the resulting transient. Recording these values in one place avoids underestimating stress when multiple drives share the same bus or when cabinet-level capacitors are added later.
Contactor lifetime and contact current ratings link the pre-charge profile to real-world endurance. Specifying the expected number of start–stop cycles and the typical inrush current at each cycle helps avoid selecting devices that pass nameplate current but wear out too quickly in the field. The checklist also records whether redundant weld detection has been implemented, which influences how the DC link can be integrated into a wider safety concept.
Finally, the reporting interface and any combination of soft-start with pre-charge are noted. This makes it clear whether pre-charge faults are exposed on CANopen, local I/O, HMI alarms or simple LEDs and whether the same MOSFET stage also enforces motor soft-start limits. The result is a compact template that can be reused across projects and stored alongside design reviews, RFQs and maintenance documentation.
| DC link voltage range | ________ V (min / nominal / max, e.g. 650–750 V) |
| Capacitance total | ________ mF (cabinet + drive DC link, effective total) |
| Inrush peak allowed | ________ A (limited by rectifier, PFC, contactor and fuse ratings) |
| Target pre-charge time | ________ ms (to reach e.g. 90–95 % of bus voltage) |
| Contactor lifetime and contact ratings | Mechanical: ________ ops, electrical: ________ ops at ________ V / ________ A |
| Redundant weld detection | Yes / No (dual contactors, two-point sensing, current decay check) |
| Fault reporting interface | CANopen / other fieldbus / digital I/O / HMI / LED (select all that apply) |
| Soft-start combined with pre-charge | Yes / No (same MOSFET stage used for DC link pre-charge and ramped start-up) |
This checklist can be turned into a reusable template in a design-review form, spreadsheet or RFQ attachment so that every new cabinet or drive platform starts from a documented set of DC link design assumptions.
IC and brands mapping for DC link and pre-charge
Once the DC link and pre-charge requirements are captured in a checklist, the next step is to translate those numbers into concrete IC categories and sourcing terms. The mapping below links each functional block in the pre-charge path to typical IC types, example brands and search keywords that work well in distributor parametric tools and RFQs.
Pre-charge MOSFET stages are usually built around hot-swap, eFuse or ideal-diode controllers that supervise current, drain-source voltage and thermal limits. Relay driver ICs handle contactor and pre-charge relay coils with integrated diagnostics, while NTC-based current limiting may be coordinated by simple controllers that monitor temperature and restart intervals. Dedicated diagnostic ICs or safety monitors observe contact voltages and auxiliary contacts to detect welded or open contacts and feed this information into the safety controller.
The same brands often supply companion parts such as bus-voltage monitors and current-sense amplifiers used in inrush limiting and weld detection. Recording function, IC category and sourcing keywords in a single table makes it easier to brief suppliers, align internal designs, and avoid overlooking suitable device families when migrating between voltage levels or safety targets.
| Function | IC category | Typical brands | Sourcing keywords | Notes |
|---|---|---|---|---|
| Pre-charge MOSFET stage | Linear hot-swap / eFuse / ideal-diode controller | TI, ADI, NXP | “high-side switch MOSFET precharge”, “48 V hot swap controller”, “DC link pre-charge controller” | Check SOA, current-limit mode, fault handling and maximum bus voltage rating. |
| Relay / contactor driver | Automotive relay driver / high-side or low-side driver with diagnostics | Infineon, ST, Renesas | “automotive relay driver”, “high-side relay driver with diagnostics”, “contactor driver IC” | Focus on coil current, channel count, open-load and short-circuit diagnostic coverage. |
| NTC-based inrush control | Temperature-based current limiter / soft-start controller with NTC interface | Microchip and selected power-control vendors | “NTC soft-start IC”, “inrush current limiter controller” | Ensure that restart intervals and ambient temperature keep NTC devices cool enough to limit inrush effectively. |
| Relay-weld diagnostics | Safety monitor / relay weld detection IC | TI, ADI | “relay weld detection IC”, “contactor diagnostics IC”, “safety monitor for contactors” | Look for multi-point sensing, built-in self-test and documentation suitable for safety analysis. |
| Bus-voltage monitoring | Window supervisor / high-voltage monitor | TI, ADI, Microchip, NXP | “DC bus window supervisor”, “high-voltage monitor” | Used for pre-charge completion thresholds and over- or under-voltage detection on the DC link. |
| Bus current sensing / inrush feedback | High-side current-sense amplifier / shunt monitor | TI, ADI, ST, Infineon | “bus current-sense amplifier”, “high-side shunt monitor” | Supports inrush limiting, current decay checks and long-term health or predictive maintenance logging. |
When preparing RFQs or internal BOMs, it is helpful to include bus voltage, inrush limits, safety requirements and preferred IC categories alongside these mapping entries so that suppliers can converge quickly on suitable device families and voltage ratings.
Request a Quote
FAQs about DC link and pre-charge design
This FAQ collects the most common questions that come up when sizing DC link capacitance, choosing a pre-charge method, planning weld detection and mapping the design to concrete IC choices. Each answer stays compact enough to reuse in design reviews, supplier discussions and FAQ structured data.
1. How do I decide whether a DC link actually needs a dedicated pre-charge stage?
I start by estimating how much energy the DC link stores and how stiff the supply is. If a direct connection would create inrush currents above what rectifiers, fuses or contactors can tolerate, a pre-charge stage becomes mandatory. I also add pre-charge when several drives share one bus or when safety requires controlled de-energisation.
2. How do I size the pre-charge resistor for a given DC link voltage, capacitance and allowed inrush time?
I treat the DC link as a simple RC charge and pick a resistor that lets the bus reach the target percentage of the final voltage within the allowed time, while keeping current below the inrush limit. Then I check resistor power and energy during charging and verify temperature rise under worst-case ambient conditions.
3. What is a practical rule for choosing the voltage threshold to switch from the pre-charge path to the main relay?
A common rule is to switch the main relay when the DC link reaches about 90–95 percent of the input voltage. At that point the remaining inrush current through the main path is small enough to stay within contactor and rectifier limits. I still confirm the threshold against worst-case mains tolerance and capacitor tolerance.
4. When is MOSFET-based pre-charge mandatory instead of using an NTC or a fixed resistor?
MOSFET-based pre-charge becomes mandatory when inrush energy is high, voltage is in the hundreds of volts or the system needs tight control and frequent restarts. In those cases, fixed resistors and NTCs waste too much power, heat up excessively or recover too slowly. A controlled MOSFET stage also gives better diagnostics and protection options.
5. What limitations does an NTC-based inrush limiter create for frequent start–stop or brown-out conditions?
An NTC inrush limiter depends on cooling down between starts, so frequent power cycling or brown-outs leave it hot and with low resistance. That means almost no inrush limiting on the next start. I also need to accept extra steady-state losses and verify that worst-case mains dips do not repeatedly stress the NTC and rectifier.
6. How is relay-weld detection actually implemented in hardware using voltage and current feedback?
I usually combine voltage difference and current decay checks. Voltage sensing at the source, midpoint and DC link shows whether contact gaps are really open. After an open command, the DC link current should fall below a threshold within a defined window. If voltage or current behaviour disagrees with the command and auxiliary contacts, weld is suspected.
7. When should redundant weld detection with dual contactors and two-point sensing be planned instead of a single diagnostic path?
Redundant weld detection with dual contactors is worth the extra cost when the DC link feeds high-energy drives, when safety integrity targets are demanding or when maintenance access relies on guaranteed isolation. In those projects, two independent contact gaps with separate sensing points reduce common-mode failures and give the safety controller better evidence about real contact status.
8. How should pre-charge and weld-detection status be reported on CANopen or other fieldbuses so that a PLC can act on it?
I map pre-charge and weld-detection results into clear status bits and fault codes, not just generic error flags. Typical signals include “pre-charge in progress”, “bus ready”, “pre-charge timeout” and “weld suspected”. The PLC then links these bits to interlocks, controlled braking and operator messages, making the DC link behaviour visible at system level.
9. Which parameters must appear in a DC link and pre-charge checklist before sending an RFQ for contactors or MOSFET stages?
Before sending an RFQ, I include bus voltage range, total capacitance, allowed inrush peak, target pre-charge time and expected start–stop frequency. I also add contactor lifetime expectations, weld-detection requirements, fault-reporting interfaces and any safety constraints. Suppliers then have enough context to propose suitable contactors, MOSFET controllers and margin on ratings.
10. How can the same pre-charge design be adapted when multiple drives share one DC link bus or when cabinet capacitance changes?
When several drives share a bus or cabinet capacitance changes, I first recalculate the effective total capacitance and stored energy. Then I recheck inrush limits, pre-charge time and resistor or MOSFET sizing against the new values. It is important to verify contactor ratings, fuse curves and weld-detection thresholds for the updated bus capacity.
11. What IC categories and search keywords are most useful when sourcing pre-charge MOSFET controllers for 24–80 V industrial drives?
For 24–80 V drives, I usually search for hot-swap controllers, eFuses and ideal-diode controllers that support the required voltage and current. Keywords such as “48 V hot swap controller”, “high-side switch MOSFET precharge” and “industrial eFuse” highlight suitable families. I then filter by SOA, diagnostics, fault modes and package options.
12. What changes in IC choice and protection strategy are needed when moving from low-voltage (48 V) to high-voltage (400–800 V) DC links?
Moving to 400–800 V DC links, I switch to high-voltage hot-swap or gate-driver solutions and pay much closer attention to creepage, clearance and isolation. SOA margins and surge immunity become critical, and weld detection usually gains redundancy. Fuses, contactors and layout also need upgrading so that the whole protection chain matches the higher energy level.
