An on-board charger looks like any other power supply until you notice what it drives. Its load is a battery, and that decides how its capacitors get chosen. The pack voltage climbs as the cells fill and drops as they discharge, and the current reverses when the car feeds power back out. The whole assembly sits in a sealed box that sees every season. Numbers that look generous on a bench can be marginal in that environment.
Our earlier piece covered where aluminum electrolytic capacitors sit inside an 800 V charger. This one is about the conditions that decide the part.
A Charger’s DC Link Never Sits at One Voltage
A data center PSU can be designed around a bus that stays put. A charger cannot. The pack voltage climbs as the cells fill and drops again as they discharge, and the DC link follows it, so a capacitor sized around a single operating point is really working across a range. The PFC stage on the grid side holds its own rail at a different level again.
That range makes the voltage question harder here than in a rack. The part has to clear the highest point the link reaches, overshoot on a load step included, while the capacitance you need is set by the lowest point the converter still has to regulate at.
The Load Is a Battery, and Batteries Do Not Like Ripple
On the output side the capacitor sits between a switching converter and a battery rather than ahead of a downstream load. Any ripple that reaches the pack is current heating the cells without doing useful work. Designers who care about pack life push output ripple down, and that pushes the burden back onto the capacitor: less ripple at the battery means more absorbed in the can, so the output position needs a higher ripple current rating and a lower ESR than a general-purpose part offers.
ESR matters at the end of life as much as at the start. It rises as the electrolyte dries, so the ripple passed to the battery is lowest when the capacitor is new and grows from there. Sizing on a fresh-part ESR figure leaves the pack more exposed later in the vehicle’s life, when the same current produces more heat and more voltage ripple.
When the Current Reverses
Bidirectional chargers, now common on vehicles offering V2L or V2G, run the same hardware in both directions. Power electronics do not care about the arrow on the diagram, but the capacitor duty does change. In discharge mode the converter feeds a load or the grid rather than a pack, the operating point moves, and the ripple seen by the bulk and DC-link capacitors is no longer the ripple of the charging profile.
A capacitor checked only against charging can end up outside its ripple budget during vehicle-to-load operation. That is often the harder case, because the car is stationary with no airflow while the enclosure is already at its hottest.
| Position in the charger | Charging mode | Discharge mode (V2L / V2G) |
|---|---|---|
| PFC bulk | Grid-side ripple at the line frequency, plus switching ripple | Feeds the inverter side; ripple spectrum follows the load being driven |
| Intermediate DC link | Rides up and down with the pack voltage as it fills | Follows the load, which can step harder than a charging curve |
| Battery-side output | Ripple here heats the cells, so it is pushed down | Becomes the input; ripple direction reverses |
Where the Heat Actually Goes
Data center capacitors live in airflow. An OBC lives in a sealed enclosure under the hood or under the floor, with no fan aimed at it, and it sees the full annual temperature swing on top of its own losses. That has two consequences for the capacitor you order.
The ambient you design against is the worst case inside that box rather than the outside air, and it is reached after a long drive, not at startup. The thermal path from core to outside world is then the limit, so a capacitor with the same ripple rating as a rack-mounted type can still run hotter here. Long-life series exist for this kind of duty, and the gap between a 2,000 h and a 5,000 h part shows up as field life rather than as anything printed on the sheet.
The environment adds vibration and thermal cycling that a rack never imposes. Terminations and mounting have to survive both, which is part of why screw-terminal and snap-in formats dominate the higher-power positions instead of small radial cells.
What This Changes in the Order of Work
Because the DC link moves and the load is a battery, selection does not start from one bus figure the way a fixed-rail design does. Set the voltage rating from the highest point the link reaches rather than from a nominal bus, since the top of the pack voltage and the switching overshoot both land there. Size the output capacitor against the ripple you are willing to put into the battery rather than against a generic output-ripple target. Then check bulk and DC-link ripple at the worst point of the charging curve, which is not always full load.
The method for reading ripple ratings and working out life is the one we set out for data center power stages, where the same 10 °C rule and the same frequency correction apply. What changes on a charger is which condition is worst.
How the Power Class Changes the Bank
On-board chargers come in a few standard power classes, and the class decides whether you are looking at one bulk capacitor or a bank of them. A single-phase unit at the low end of the range can often be served by one or two snap-in cans on the DC link. The classes that run from a three-phase supply push both the ripple and the stored energy up, and the answer becomes several cans in parallel, sometimes split across a series string for the voltage.
Parallel cans divide the ripple between them, which helps when no single part in the range carries the current you need, but they also want matched ESR so the share stays even. Mismatched parts in a parallel bank let the lowest-ESR can do more than its share of the work and run hotter than the rest.
Where to Look in the Range
Snap-in capacitors carry most of the bulk and DC-link duty in an OBC, while screw-terminal types take the positions where the current per connection is large. Radial long-life series cover the auxiliary rails. Both of the high-power ranges include 5,000 h and 8,000 h series built for hot, sealed enclosures.
Questions That Come Up in Charger Projects
Can a capacitor specified for a data center PSU be used in a charger?
The ripple and voltage logic carries over, but the duty does not. The charger’s DC link swings with the pack, the output sees a battery rather than a resistive load, and the enclosure runs hotter with no forced air. A part that meets the electrical numbers can still be the wrong call on life.
Does bidirectional operation need a different capacitor?
Usually a different starting assumption rather than a different family. The same series can serve both directions, but the ripple and thermal check has to cover discharge mode as well, since that is often the hotter case for a stationary vehicle.
How much ripple can reach the battery?
That is a battery specification as much as a capacitor one, and it varies with cell chemistry and with how the BMS is set up. The practical route is to take the pack maker’s ripple limit as a design input and work back to the ESR and ripple rating it implies.
Where Xuansn fits
Xuansn manufactures long-life aluminum electrolytic capacitors in snap-in, screw-terminal, and radial formats, including 5,000 h and 8,000 h series for high-ripple, high-temperature duty. If you are working on an 800 V charger or its DC-DC stage, send the pack voltage range, the ripple limit at the battery, and the ambient inside the enclosure, and our engineers will match a series to the application. Reach us through the contact page or at coco@xuanxcapacitors.com.
