MLCC Capacitance Code Against the Capacitance Column: 28 of 24,026 Walsin Rows Disagree, Audited for United States Buyers

We decoded the MLCC capacitance code embedded in all 24,026 part numbers in Walsin’s MLCC export and compared each one against the export’s own Capacitance column. 28 rows disagree. Nineteen of those 28 sit in the two AEC-Q200 automotive series. We have not seen this counted anywhere.

---

Venatech Limited, an authorised Walsin distributor in Hong Kong, holds the full parametric export for Walsin Technology (Walsin) MLCC — 00_ALL_MLCC_parts_combined.xlsx, 24,026 rows — plus 26 MLCC datasheet PDFs and the series index INDEX_MLCC_series.csv. For this post we parsed every part number, decoded its three-character capacitance code ourselves, and set the result beside the Capacitance field in the same row. Every part number below is real and appears in that export.

How many Walsin MLCC rows have a capacitance code that disagrees with the Capacitance column?

28 of 24,018 comparable rows. We parsed all 24,026 part numbers in 00_ALL_MLCC_parts_combined.xlsx and every one carried a decodable three-character code: 17,338 in plain three-digit form, 6,504 using R as the decimal point, 184 leading with R. Eight rows have no capacitance figure we could compare. Of the remaining 24,018, 23,990 agree and 28 do not.

What we measured in 00_ALL_MLCC_parts_combined.xlsx	Count
Rows in the export	24,026
Part numbers whose capacitance code decodes cleanly	24,026
Rows with no comparable Capacitance value	8
Rows where code and column agree	23,990
Rows where code and column disagree	28
Disagreements that are an exact factor of ten	12
Disagreements that are a factor of one thousand	4

The decoding rule is the ordinary EIA one and you can repeat it in a spreadsheet in five minutes. Characters six, seven and eight of the part number are the capacitance code. Three digits means two significant figures and a power-of-ten multiplier in picofarads, so 223 is 22 nF and 106 is 10 µF. An R is a decimal point, so 0R5 is 0.5 pF and 5R1 is 5.1 pF.

Which MLCC series carry the disagreements, and which carry none at all?

Six of the sixteen series in the export carry all 28. Ten carry zero. The rate is not even: the two AEC-Q200 automotive series — MT and ST — hold 2,290 of the 24,018 comparable rows, 9.5% of the file, and 19 of the 28 disagreements, 68% of them. General Purpose holds 8,466 rows, more than a third of the export, and three.

Series	Comparable rows	Rows where code and column disagree	Rate
Automotive Caps Qualified to AEC-Q200 (MT series)	2,172	18	0.83%
Automotive Soft term Caps Qualified to AEC-Q200 (ST series)	118	1	0.85%
Microwave Caps (RF series)	3,615	4	0.11%
High Q & Low ESR Caps (HH series)	1,709	2	0.12%
General Purpose Caps	8,466	3	0.04%
The other ten series, Middle & High Voltage down to FT	7,938	0	0

The concentration matters more than the total, because the MT series is the line a United States buyer reaches for when the part has to go through AEC-Q200 paperwork. The MT approval sheet is WTC_MLCC_Automotive_MT.pdf and the ST sheet is WTC_MLCC_Soft_term_Auto_AECQ200_ST.pdf. AEC-Q200 is Walsin’s own word, compliant, and the qualification report has to be requested against your exact part number — the export does not carry a qualification column at all.

By case size the 28 land at 1210 (9 rows), 0402 (9), 0201 (5), 0603 (4) and 0805 (1). By dielectric: NP0 15, X7R 11, X5R 1, X7S 1.

Is the MLCC capacitance code or the Capacitance column the field to trust?

The code, on this evidence. We grouped the export by the 13 capacitance codes involved in the 28 disagreements and counted how the other holders of each code read. In all 13 cases the majority agrees with the decode. Code 223 appears on 306 rows: 298 read 22 nF, 8 read 2200 pF. Code 106 appears on 162 rows: 160 read 10 µF, 2 read 4.7 µF.

The 106 group is the cleanest to look at, because 4.7 µF already has its own code in Walsin’s scheme — 475 — and the export uses it correctly elsewhere.

Part number	Code	Code decodes to	Capacitance column reads	Size	Rated voltage	Match / mismatch
ST32B106K500	106	10 µF	4.7uF	1210	50Vdc	Mismatch — column disagrees with its own code and with four siblings
ST32B106J500	106	10 µF	10uF	1210	50Vdc	Match — same code, same size, same voltage
ST32B106K250	106	10 µF	10uF	1210	25Vdc	Match on capacitance; mismatch on rated voltage, 25Vdc not 50Vdc
ST32B106M500	106	10 µF	10uF	1210	50Vdc	Match on capacitance and voltage; mismatch on tolerance, M not K
MT32B106K500	106	10 µF	4.7uF	1210	50Vdc	Mismatch — the identical defect in the MT series
MT32B106J500	106	10 µF	10uF	1210	50Vdc	Match — same code, same size, same voltage
MT32B475K500	475	4.7 µF	4.7uF	1210	50Vdc	Match — this is what a real 4.7 µF part number looks like

We will not tell you a part is equivalent. We will show you where it differs — and here the difference is between two fields inside one row of the manufacturer’s own file. ST32B106K500 and ST32B106J500 are two part numbers of the same code, the same 1210 footprint and the same 50Vdc rating, and the export gives them capacitances a factor of 2.1 apart.

What does the widest disagreement look like on a real part number?

A factor of one thousand, on four 0201 microwave parts. RF03N180J6R3, RF03N180G6R3, RF03N180F6R3 and RF03N180K6R3 all carry the code 180, which decodes to 18 pF. All four read 18nF in the Capacitance column. We checked all 273 rows in the export that carry the 180 code: 269 read 18 pF and these four do not.

Part number	Code decodes to	Capacitance column reads	Size	Dielectric	What the rest of the export supports	Match / mismatch
RF03N180J6R3	18 pF	18nF	0201	NP0	269 of 273 180 rows read 18 pF	Mismatch, 1,000-fold
RF03N180F6R3	18 pF	18nF	0201	NP0	same code, same conflict	Mismatch, 1,000-fold
MT32N333J101	33 nF	3300pF	1210	NP0	232 of 235 333 rows read 33 nF	Mismatch, 10-fold
MT15B823J500	82 nF	68nF	0402	X7R	103 of 106 823 rows read 82 nF	Mismatch, not a decade shift
0402X683M500	68 nF	680nF	0402	X5R	167 of 168 683 rows read 68 nF	Mismatch, 10-fold the other way
HH03N3R0A101	3.0 pF	100pF	0201	NP0	174 of 175 3R0 rows read 3.0 pF	Mismatch, 33-fold
MT18N4R3B500	4.3 pF	2.7pF	0603	NP0	86 of 87 4R3 rows read 4.3 pF	Mismatch, no decimal explanation

An 0201 NP0 part reading 18 nF is the one that would survive a casual review, because the number is plausible for an 0201 X5R and nothing in the row flags the dielectric conflict. The RF series sheet is WTC_MLCC_Microwave_RF.pdf and INDEX_MLCC_series.csv records its published capacitance range as 0.1 to 1000 pF — which puts 18 nF outside the series envelope. MT18N5R1C500, the only one of 97 N5R1 part numbers reading 5.0 pF rather than 5.1 pF, is our own previously published finding and it sits inside the 28.

Which rows carry no usable capacitance in the column at all?

Eight. We read every non-numeric value in the Capacitance column of 00_ALL_MLCC_parts_combined.xlsx and sorted them into four kinds of defect. Four rows hold a datasheet path, two hold a letter glued to or standing in for the number, one holds the right digits with a capital letter O, and one is simply blank.

Part number	Capacitance column holds	Code decodes to	Defect	Match / mismatch
MT31B474J101	/wp-content/uploads/datasheet/WTC_MLCC_Automotive_MT.pdf	470 nF	datasheet path where the value belongs	Mismatch — no capacitance to compare
MT31B474K101	the same path	470 nF	same	Mismatch — no capacitance to compare
MT31B474M101	the same path	470 nF	same	Mismatch — no capacitance to compare
0201N0R1D250	/wp-content/uploads/datasheet/WTC_MLCC_General_Purpose.pdf	0.1 pF	same, and 25Vdc repeats into the L column	Mismatch — two fields lost
MT31B822J631	820OpF	8.2 nF	capital letter O for a zero	Mismatch on the string, match on intent
1808B681M302	680pFG	680 pF	a stray G welded to the unit	Mismatch on the string, match on intent
SH15X105K100	E	1 µF	a single letter where 1 µF should be	Mismatch — no capacitance at all
0603N102J251	blank	1 nF	empty, with L, W and T also empty	Mismatch — no capacitance at all

Four of those eight we have published before and they are reproduced here because the audit has to account for them. Three are new as far as we can establish: 1808B681M302 reading 680pFG, SH15X105K100 reading the bare letter E, and 0603N102J251 with the field empty. We have not seen any of the three reported anywhere. SH15X105K100 is the one to watch, because a 1 µF 0402 X5R soft-termination part is a high-running position and a selector that filters on capacitance cannot see it.

Add the 28 disagreements to the 8 unusable rows and 36 of 24,026 rows cannot be filtered on capacitance without reading the part number. That is 0.15% of the file, which is small until it is your line.

What should a United States engineer do before pasting a Walsin MLCC code into a BOM?

Decode the code and make it the primary field. Three steps, and the first is a spreadsheet formula. Take characters six to eight of the part number, decode them with the EIA rule above, and compare the result against whatever your selector showed you. If the two disagree, the part number is the field we would trust, and we will say so against your specific line rather than send you to a general note.

1. Decode before you order. The code lives in the part number and travels with it through every system; the Capacitance column does not.
2. Cross the whole list in one pass. Paste up to 500 part numbers or drop your Excel or CSV into the free cross-reference tool, filter by commercial or automotive grade, and read every stated difference rather than a claimed match.
3. Check what is physically in Hong Kong on ready stock before you commit a build slot — a dated snapshot as at 23 September 2026, not a live feed.

Related reading: how to read a Walsin MLCC part number takes the whole 12-character code apart field by field, and why Walsin part numbers do not paste into your BOM covers the formatting faults that break a join. The line itself is on the Walsin MLCC page.

The short version

We decoded the MLCC capacitance code on all 24,026 rows of Walsin’s MLCC export and compared it against the export’s own Capacitance column. 23,990 rows agree. 28 disagree, and 19 of those 28 are in the two AEC-Q200 automotive series that hold only 9.5% of the file. Twelve of the 28 are an exact decade shift; four are a factor of 1,000. For all 13 codes involved, the majority of rows carrying that code back the decode, so the part number is the field we trust. A further 8 rows carry no usable capacitance at all — three of them, 1808B681M302, SH15X105K100 and 0603N102J251, we believe have not been reported before. Qualification stays with you.

Send us the Walsin MLCC part numbers off your United States bill of materials and we will decode each capacitance code against the export row ourselves, tell you which of the two fields we would trust on each line, and tell you what was physically in Hong Kong on 23 September 2026. If your selector showed you a capacitance that your part number does not decode to, which one did your purchasing system send to the fab?

---

Frequently asked questions

<!– wp:rank-math/faq-block {"questions":[{"id":"faq-question-01","title":"How do you decode the capacitance code in a Walsin MLCC part number?","content":"Take characters six, seven and eight. Three digits means two significant figures and a power-of-ten multiplier in picofarads, so 223 is 22 nF and 106 is 10 µF. An R acts as a decimal point: 0R5 is 0.5 pF, 5R1 is 5.1 pF. We decoded all 24,026 rows this way.”,”visible”:true},{“id”:”faq-question-02″,”title”:”How many Walsin MLCC rows have a capacitance code that disagrees with the Capacitance column?”,”content”:”28 of the 24,018 comparable rows in 00_ALL_MLCC_parts_combined.xlsx. A further eight rows hold no usable capacitance value, so 36 of 24,026 cannot be filtered on capacitance without reading the part number. We counted every row ourselves rather than sampling.”,”visible”:true},{“id”:”faq-question-03″,”title”:”Which MLCC series are worst affected?”,”content”:”The two AEC-Q200 automotive series. MT shows 18 disagreements in 2,172 comparable rows, ST shows 1 in 118, and between them they hold 19 of the 28 while making up 9.5% of the export. Ten of the sixteen series show none at all.”,”visible”:true},{“id”:”faq-question-04″,”title”:”Should I trust the part number or the Capacitance column?”,”content”:”The part number, on our measurement. For each of the 13 capacitance codes involved in a disagreement, the majority of rows carrying that code agree with the decode. Code 223 reads 22 nF on 298 of 306 rows. Confirm against the series datasheet before design-in, and qualify in your own circuit.”,”visible”:true},{“id”:”faq-question-05″,”title”:”What is the biggest single disagreement you found?”,”content”:”A factor of one thousand, on four 0201 microwave parts. RF03N180J6R3 and its three siblings carry the 180 code, which decodes to 18 pF, and all four read 18 nF in the column. 269 of the 273 rows carrying that code read 18 pF.”,”visible”:true},{“id”:”faq-question-06″,”title”:”Does Walsin’s MLCC export carry a column for qualification status or operating temperature?”,”content”:”No. We checked the column headers of the export and neither field exists, nor does lifecycle, lead time, RoHS or REACH. Walsin’s datasheets say AEC-Q200 compliant, never certified. Request the qualification report for your exact part number; Venatech makes no qualification claim of its own.”,”visible”:true}],”listStyle”:””,”titleWrapper”:”h3″,”titleCssClasses”:””,”contentCssClasses”:””,”sizeSlug”:””} –>

---