MLCC Capacitance Loss Under DC Bias, Finland: 420 of 526 Walsin X5R and X7R Parts From 1 uF Up Have the Same Value in a Bigger Case, 388 at a Higher Voltage, and 44 Have Neither

MLCC capacitance loss under DC bias reaches 35 to 65 percent at rated voltage on a 10 uF 6.3 V X7R 0805. We counted Walsin’s general-purpose X5R and X7R parts from 1 uF up: 420 of 526 have the same value in a bigger case, 388 at a higher voltage, and 44 have neither.

---

Venatech Limited, an authorised Walsin distributor in Hong Kong, holds the Walsin Technology (Walsin) MLCC parametric export, 00_ALL_MLCC_parts_combined.xlsx, and 26 Walsin MLCC datasheets. This post asks the question a power-supply designer in Oulu or Tampere asks when a 10 uF capacitor on a 12 V or 24 V rail measures half its label: where in the Walsin range is the same value in a bigger case or at a higher voltage? Every part number below is real and was checked against our master list.

How much capacitance does an X5R or X7R MLCC lose under DC bias?

Between a third and two thirds at rated voltage, depending on the part. An EE Power technical article reports minus 35 to minus 65 percent at rated voltage on a 10 uF 6.3 V X7R 0805 compared across three makers (EE Power, fetched 20 October 2026). It adds that the loss follows the electric field, volts per unit of dielectric thickness, so thinner layers lose more.

ROHM’s ceramic-capacitor application note makes the same point in case terms: a smaller part loses more under the same voltage, and its example 1608 part rated 10 V “can only be used for 1V” (ROHM application note, fetched 20 October 2026). Both sources send the reader to the maker’s own curve for the exact part.

Walsin publishes one typical curve per dielectric, not per part. We read page 5 of 00_MLCC_Full_Catalogue.pdf (Walsin catalogue) and found a single DC bias chart for the general-purpose line. By eye, X7R falls about 37 percent at 30 V and X5R about 57 percent at 20 V. The chart has no case size, no rated voltage and no thickness axis. The page says the curves are “for reference only”.

Does a bigger case or a higher voltage rating cut the loss?

A bigger case usually does, and a higher voltage rating sometimes does not. The EE Power article recommends moving up one case size, for example 0603 to 0805 or 0805 to 1210, to cut the loss. ROHM’s note shows a 10 uF 3216 part where the 50 V version lost more capacitance under bias than the 16 V version, so a higher rating is a place to look, not proof.

That is why this post counts twins and does not rank them. A twin is another Walsin part with the same dielectric code and the same capacitance value. It is a candidate. The bias curve, the height and the temperature rating still decide.

Move	What usually helps	What the source warns about	Match / mismatch
Same case, higher voltage rating	Thicker dielectric layers can cut the field per volt	ROHM: the 50 V 3216 example lost more than the 16 V one	Mismatch risk — ask for both curves
Next case size up, same voltage	More dielectric volume per microfarad	Footprint and height change	Match on field logic, mismatch on board area
X5R to X7R, same case	Wider temperature range (125 C against 85 C)	Walsin lists the same plus or minus 15 percent over temperature for both	Mismatch — bias loss is not in that figure

Walsin prints the temperature characteristic of X7R and X5R as plus or minus 15 percent, with an operating range of minus 55 to plus 125 C for X7R and minus 55 to plus 85 C for X5R (catalogue page 5). That figure is measured without bias. It tells you nothing about the capacitance at 12 V.

How many Walsin parts offer the same value in a bigger case or at a higher voltage?

For 526 distinct part numbers, 420 have a larger-case twin and 388 a higher-voltage twin. We parsed all 3,513 rows of the “General Purpose Caps” series, kept X5R and X7R from 0201 to 1812, dropped repeated part numbers, and kept the 526 of 1 uF and above. For each part we searched the same dielectric and capacitance in the same file.

A larger-case twin is a bigger case at the same or higher rated voltage. A higher-voltage twin is the same case at any higher rating. Only 314 of the 526 have a same-case twin at twice the rating or more, which is the usual derating target.

Case size	Parts of 1 uF and up	Same-case twin at 2x voltage or more	Bigger-case twin at same or higher voltage	Match / mismatch
0201	8	3	8	Match on value, all 8 can move up a case
0402	31	13	31	Mismatch — 18 have no 2x voltage twin
0603	86	46	86	Mismatch — 40 have no 2x voltage twin
0805	119	79	117	Match on case, 2 parts at the top
1206	143	97	120	Mismatch — 23 cannot move up
1210	116	65	58	Mismatch — only 58 reach a 1812
1812	23	11	0	Mismatch — the largest case in the series

We have not seen this counted anywhere: 44 of the 526, 8.4 percent, have neither a higher-voltage twin in their case nor a larger-case twin. They are 12 X7R 1812 parts, 11 X5R and 8 X7R 1210 parts, 10 X5R and one X7R 1206 part and two X5R 0805 parts.

What does the 10 uF ladder look like for X5R and X7R?

At 10 uF the dielectric choice moves the voltage ceiling by two grades. X7R 10 uF starts at 0805 and stops at 16 V there, while X5R 10 uF reaches 25 V in 0603 and 50 V in 0805. We grouped the 46 X5R and 34 X7R 10 uF parts by case and listed the lowest and highest rating in each.

Case	X5R 10 uF parts, voltage range	X7R 10 uF parts, voltage range	Match / mismatch
0402	3, 4 to 10 V	none	Mismatch — X7R does not exist here
0603	8, 4 to 25 V	none	Mismatch — X7R does not exist here
0805	10, 6.3 to 50 V	6, 6.3 to 16 V	Mismatch — X7R ceiling 16 V
1206	12, 6.3 to 50 V	13, 6.3 to 35 V	Match from 6.3 to 35 V
1210	13, 6.3 to 50 V	14, 6.3 to 50 V	Match across the range

Take a 10 uF X7R 0805 on a 12 V rail. The 16 V part, 0805B106M160, is 1.33 times the rail, the thinnest margin in the list. The next X7R rung is 1206B106M350 at 35 V, which gives 2.9 times the rail and costs about 2.05 times the board area. The table below sets the realistic candidates beside that incumbent.

Candidate	Dielectric and case	Rated voltage	Height (T)	Match / mismatch with a 10 uF X7R 0805 16 V incumbent
0805B106M160	X7R 0805	16 V	1.25 plus or minus 0.20 mm	Incumbent, the reference
0805X106K500	X5R 0805	50 V	1.25 plus or minus 0.20 mm	Match footprint and height; mismatch X5R, 85 C ceiling
1206B106M350	X7R 1206	35 V	1.60 plus 0.30 minus 0.10 mm	Match dielectric; mismatch area 2.05x, height up to 1.90 mm
1210B106K500	X7R 1210	50 V	2.50 plus or minus 0.30 mm	Match dielectric; mismatch area 3.2x, height up to 2.80 mm
0603X106M250	X5R 0603	25 V	0.80 plus or minus 0.20 mm	Mismatch — smaller case, expect more bias loss

Areas are nominal length times width of the case code. The last row is the one the ROHM note warns about: a smaller case at a higher rating. The Walsin general-purpose sheet, WTC_MLCC_General_Purpose.pdf, gives the ordering code and dimensions but no per-part bias curve.

Which Walsin parts sit at the top of both ladders?

Forty-four parts, and 13 of them are 100 V. The 44 split by rating as 13 at 100 V, 12 at 25 V, 11 at 50 V, three at 16 V, two each at 10 V and 35 V and one at 6.3 V. Typical members are 1206X106K500 (X5R 10 uF 50 V), 1210B106K500 (X7R 10 uF 50 V) and 1210B476K100 (X7R 47 uF 10 V).

For these there is no same-value part in the series with more headroom. The levers left are a lower working voltage, two parts in parallel at half the value each, or a different dielectric. Ask us for the Walsin curve first, because the loss may already be inside your budget.

Our post on DC bias and voltage derating sets out the derating rule, and the four empty DC bias columns shows what Walsin’s enriched export does not hold.

What does the Hong Kong shelf hold for a DC bias swap?

As at 23 September 2026 we hold 51 of these 526 part numbers in Hong Kong stock, and 38 of the 51 have a same-case twin at twice the rating somewhere in the series. Only 9 of the 51 have a same-case twin at twice the rating that is also on the shelf, 33 have a bigger-case twin on the shelf, and 16 have neither. We joined the stock snapshot to the 526 and repeated the twin test inside the stocked lines.

Stocked part	Dielectric, case, value	Twin on the shelf	Match / mismatch
0805B106K100	X7R 0805 10 uF 10 V	bigger case, for example 1210B106K160	Mismatch — 1210 height 2.00 mm nominal
0805X106K160	X5R 0805 10 uF 16 V	same case at 2x: 0805X106K500, 50 V	Match on dielectric, case and height; mismatch only in rating
1206X106K500	X5R 1206 10 uF 50 V	none	Mismatch — top of both ladders

Stock is a snapshot and moves. Check ready stock and send the part number to the enquiry form.

How should a Finnish power design pick the replacement?

By working voltage, not rated voltage. State the real DC at the capacitor, the capacitance the loop needs at that voltage and the highest temperature, then keep the twins that clear all three. Our count says the first filter, a bigger case or a higher rating, leaves a candidate for 482 of the 526 parts; the second filter, Walsin’s curve for that exact part, is not published and has to be requested.

Run the twin list through our free cross-reference tool, which takes a competitor part and returns Walsin parts with the differences stated. The MLCC range lists the general-purpose series. Qualification stays with you: confirm against the Walsin datasheet and the per-part bias curve, and qualify in your own circuit. We have not requested the bias curves from Walsin in writing yet; they are on our request list.

The short version

DC bias removes 35 to 65 percent of an X5R or X7R capacitor at rated voltage, and Walsin publishes one typical curve per dielectric, not per part. Of 526 general-purpose parts from 1 uF up, 420 have a bigger-case twin, 388 a higher-voltage twin, and 44 have neither. At 10 uF, X7R tops out at 16 V in 0805 while X5R reaches 50 V.

Send us the capacitor that measures low on your rail, with working voltage and temperature, and we will return the Walsin twins with the case, height and dielectric differences beside each. And a question back: do you size your capacitors from the label or from the bias curve?

---

Frequently asked questions

---