The MLCC Capacitance Cliff: Walsin’s Ceiling Falls 2.2 Million-Fold from 4 Vdc to 6,000 Vdc

We measured the highest capacitance available at every one of Walsin’s 25 MLCC voltage classes. It falls from 220 µF at 4 Vdc to 100 pF at 6,000 Vdc — a 2.2 million-fold drop — and it is not a straight line down. High voltage MLCC capacitance limits are a staircase with four steps that go the wrong way.

---

Venatech Limited, an authorised Walsin distributor in Hong Kong, measured this from Walsin Technology (Walsin)’s own 00_ALL_MLCC_parts_combined.xlsx, 24,026 rows. We parsed the Capacitance and Rated Voltage fields on every row, resolved 24,018 of them to a numeric pair, and grouped the result by voltage class to find the single highest capacitance published at each. Every part number below is real and every figure is read from the export, not a datasheet summary.

How does Walsin’s maximum MLCC capacitance actually fall as rated voltage climbs?

It falls by roughly four orders of magnitude between the bottom and the top of the range, across 25 distinct rated-voltage classes we found in the Rated Voltage column. The ceiling opens at 220 µF (4 Vdc and 6.3 Vdc) and closes at 100 pF (6,000 Vdc) — a ratio of 2,200,000 to 1. We have not seen this full ladder assembled in one sweep anywhere, including in our own prior posts on either end of it separately.

Rated voltage	Rows	Capacitance ceiling	Holding part number
4 Vdc	43	220 µF	1210X227M4R0
25 Vdc	3,205	47 µF	2220B476K250
100 Vdc	3,106	22 µF	2220B226J101
500 Vdc	1,143	1 µF	2220B105K501
1,000 Vdc	861	220 nF	2225B224K102
2,000 Vdc	511	56 nF	2220B563K202
4,000 Vdc	59	1.8 nF	2220B182K402
6,000 Vdc	10	100 pF	S252N101J602

Eight of 24,026 rows carry a capacitance or voltage value we could not resolve to a number — a datasheet URL where the capacitance belongs on three rows, a letter in place of a digit on another, and similar defects we have already flagged elsewhere — so this table is built on 24,018 clean pairs, not 24,026.

Is the fall a smooth curve, or does it move in steps?

It moves in uneven steps, and four of the 25 classes exist specifically because Walsin’s General Purpose, Big Sizes and Middle-to-High-Voltage series overlap rather than hand off cleanly. We read every one of the 25 rungs in order and the ceiling does not halve per step or follow any single curve — it drops hard in some places and barely moves in others.

Voltage step	Ceiling before	Ceiling after	Change
100 → 200 Vdc	22 µF	2.2 µF	10x drop
630 → 1,000 Vdc	1 µF	220 nF	4.5x drop
1,000 → 1,250 Vdc	220 nF	10 nF	22x drop
3,000 → 3,500 Vdc	18 nF	10 nF	1.8x drop

The steepest single-step drop in the table is 1,000 to 1,250 Vdc, where the ceiling falls 22-fold in one rung — sharper than the far more frequently discussed 1,000-to-2,000 Vdc span, where the fall is a comparatively gentle 4x (220 nF to 56 nF). A designer budgeting margin by “the next voltage class up” needs to know which boundary they are actually crossing, because the cost is not uniform.

Does the ceiling always fall when voltage rises, or does it ever climb back up?

No — it rises at four of the 24 steps in the ladder, which is the fact in this post we have not seen counted anywhere. We checked every adjacent pair of the 25 voltage classes in order and found the ceiling increase, not decrease, moving from 400 to 450 Vdc, from 1,250 to 1,500 Vdc, from 2,500 to 3,000 Vdc, and from 4,000 to 5,000 Vdc.

Rebound	Ceiling at lower voltage	Ceiling at higher voltage	Multiple
400 Vdc → 450 Vdc	330 nF	1 µF	3x up
1,250 Vdc → 1,500 Vdc	10 nF	100 nF	10x up
2,500 Vdc → 3,000 Vdc	10 nF	18 nF	1.8x up
4,000 Vdc → 5,000 Vdc	1.8 nF	10 nF	5.6x up

Each rebound has the same cause: a thinly populated class (1,250 Vdc has 2 rows, 2,500 Vdc has 29, 4,000 Vdc has 59) sits next to a deeper class served by a different series — the Big Sizes shelf reappears at 1,500 and 3,000 Vdc, and the S2 safety series anchors 5,000 Vdc. The lesson for anyone reading this table top to bottom is that “higher voltage always costs capacitance” is false on four of the 24 boundaries in Walsin’s own data.

Where is the single steepest drop in the whole ladder?

1,000 to 1,250 Vdc, a 22-fold collapse from 220 nF to 10 nF, and it is steeper than it looks because the 1,250 Vdc class itself is almost empty. We isolated the rows at exactly 1,250 Vdc and found two, both AEC-Q200 automotive parts — MT32N103J12E at 10 nF and ST32N103J12E, the ST soft-termination part, also at 10 nF. A designer who rounds up from 1,000 Vdc to “the next standard class” and lands on 1,250 Vdc is choosing the thinnest, lowest-capacitance rung in the entire high-voltage half of the table, not a safe intermediate step.

Field	2225B224K102 (1,000 Vdc ceiling)	MT32N103J12E (1,250 Vdc ceiling)	Match / mismatch vs a 1,250 Vdc requirement
Rated voltage	1,000 Vdc	1,250 Vdc	Mismatch — the part at 1,000 Vdc is not rated for 1,250 Vdc
Capacitance	220 nF	10 nF	Mismatch — 22x less capacitance at the higher voltage
Case size	2225	1210	Mismatch — smaller footprint, but driven by capacitance, not voltage headroom
Dielectric	X7R	NP0	Mismatch — different dielectric family entirely
AEC-Q200	No	Yes	Mismatch — the 1,250 Vdc rung is automotive-qualified; the 1,000 Vdc ceiling part is not

What holds the ceiling at the very top of the voltage range?

The S2 safety-certified series, and the number on its Rated Voltage column is not what it looks like. At 5,000 and 6,000 Vdc, the only rows in the entire 24,026-row export are the 173 rows of 15_Safety_X1Y2_S2.xlsx, and the ceiling we measured there — 10 nF at 5,000 Vdc on S242B103K502, 100 pF at 6,000 Vdc on S252N101J602 — is read against a figure that is the impulse voltage, not the working voltage, a defect we have already published in full. We are not re-deriving that defect here; we are flagging that the top two rungs of this entire capacitance ladder stand on a column that needs a datasheet to interpret correctly.

What should a designer actually do with a 25-rung ceiling table?

Read the two rungs either side of your target voltage before committing to a class, because the step between them is not predictable from the voltage ratio alone.

1. Do not assume the next standard voltage class costs a fixed multiple in capacitance. We measured drops from 1.8x to 22x and rebounds up to 10x across 24 adjacent steps — look up your specific boundary rather than applying a rule of thumb.
2. Check the row count at your target voltage, not just the capacitance ceiling. 1,250 Vdc holds 2 rows and 3,500 Vdc holds 1; a ceiling figure from a one- or two-part shelf is a single-source risk, not a stocked option.
3. Cross the whole BOM in one pass. Paste up to 500 part numbers or drop your Excel or CSV into the free cross-reference tool. It decodes 18 manufacturers’ part numbers and matches against 800,000+ crossable Walsin part numbers, with every difference from your original stated rather than hidden.
4. Check what is physically in Hong Kong on ready stock before committing a build slot — a dated snapshot as at 23 September 2026, not a live feed.

For the 48 V to 100 V end of this same ladder, including which case sizes actually populate each class, see 48 V bus MLCC: zero of Walsin’s 24,026 rows are rated 63 V or 80 V. For the 800 V to 2,000 Vdc span and what AEC-Q200 costs there, see EV traction inverter passives for Germany. The full range is on the Walsin MLCC range page.

The short version

We mapped the maximum capacitance Walsin publishes at all 25 rated-voltage classes across 24,026 MLCC rows. The ceiling runs from 220 µF at 4 Vdc to 100 pF at 6,000 Vdc, a 2.2 million-fold fall, and it does not fall monotonically — it rises at four of the 24 steps, most sharply 10x between 1,250 and 1,500 Vdc. The single steepest drop going the expected direction is 1,000 to 1,250 Vdc, a 22-fold collapse onto a two-part shelf. The top two rungs, 5,000 and 6,000 Vdc, stand entirely on the S2 safety series, whose voltage column is impulse rating rather than working voltage.

Send us your target voltage and capacitance, and we will tell you which rung of this ladder you are actually standing on — whether it is a deep shelf or a two-part class — with the part number and the datasheet page behind it. If your BOM rounds 1,000 Vdc up to “1,250 Vdc, the next standard class,” have you checked how thin that shelf actually is?

---

Frequently asked questions

<!– wp:rank-math/faq-block {"questions":[{"id":"faq-question-01","title":"How much does Walsin's maximum MLCC capacitance fall from low voltage to high voltage?","content":"2.2 million-fold. We measured the ceiling at all 25 rated-voltage classes across 24,026 rows: 220 µF at 4 Vdc and 6.3 Vdc, falling to 100 pF at 6,000 Vdc.","visible":true},{"id":"faq-question-02","title":"Does MLCC capacitance always fall as the rated voltage class rises?","content":"No. Checking every adjacent pair among the 25 classes shows the ceiling rises at four boundaries: 400 to 450 Vdc, 1,250 to 1,500 Vdc, 2,500 to 3,000 Vdc, and 4,000 to 5,000 Vdc. Each rebound happens where a thin class sits next to a deeper shelf served by a different series.","visible":true},{"id":"faq-question-03","title":"Where is the steepest capacitance drop in Walsin's MLCC voltage ladder?","content":"Between 1,000 and 1,250 Vdc, a 22-fold fall from 220 nF to 10 nF. The 1,250 Vdc class holds only 2 rows in the whole export, both AEC-Q200 automotive parts, so the ceiling there is a single-part-number figure, not a shelf.","visible":true},{"id":"faq-question-04","title":"What is the maximum MLCC capacitance available at 1,000 Vdc?","content":"220 nF, on part number 2225B224K102 in the Big Sizes series. That is the highest capacitance anywhere in the 24,026-row export at 1,000 Vdc or above; no row at that voltage or higher reaches 1 µF.”,”visible”:true},{“id”:”faq-question-05″,”title”:”Why does the 5,000 Vdc and 6,000 Vdc capacitance figure need a datasheet to interpret?”,”content”:”Because those two rungs are served entirely by the S2 safety-certified series, whose Rated Voltage column holds the impulse voltage, not the working voltage, on all 173 rows. The working figures are in the datasheet, not the export.”,”visible”:true},{“id”:”faq-question-06″,”title”:”How many of Walsin’s 24,026 MLCC rows could not be resolved to a numeric capacitance and voltage pair?”,”content”:”Eight. Three rows carry a datasheet URL where the capacitance value belongs, one carries a capital letter in place of a digit, and the remainder carry other unreadable values. All eight are excluded from the figures in this post.”,”visible”:true}],”listStyle”:””,”titleWrapper”:”h3″,”titleCssClasses”:””,”contentCssClasses”:””,”sizeSlug”:””} –>

---