A German buyer cannot copy one MLCC insulation resistance limit from Walsin: 25 of 26 PDFs state one, in six different forms, and 13,138 of the 24,026 export rows (54.7%) sit in series where the floor depends on dielectric, voltage and capacitance together.
Venatech Limited, an authorised Walsin distributor in Hong Kong, holds 28 Walsin Technology (Walsin) datasheets, 451 pages, and 26 of them are MLCC documents. This post is written from the extracted text of those 26 PDFs and from the 24,026-row combined export 00_ALL_MLCC_parts_combined.xlsx, read against each other for one field. Every part number is real and every count is ours. If you write an incoming-inspection sheet for a German plant, the question is whether one gigohm figure can go on it. It cannot.
How many of Walsin’s 26 MLCC PDFs state an insulation resistance limit, and in what form?
25 of 26. We extracted the text of all 26 PDFs and searched each for an insulation resistance figure. 24 print a “gigohm or ohm-farad product, whichever is smaller” limit as an initial requirement. One, the feed-through sheet, prints megohm values part by part. One, the 12-page application guide, prints no figure at all.
| What the PDF prints | PDFs, of 26 | Which | Match / mismatch against a single gigohm floor |
|---|---|---|---|
| A gigohm term or an R×C product, “whichever is smaller” | 24 | 23 approval sheets plus 00_MLCC_Full_Catalogue.pdf |
Mismatch — the floor falls below the gigohm figure once capacitance passes a crossover |
| Megohm values, per part number | 1 | 22_Feed_Through_FT.pdf: 25 MΩ on 2 part numbers, 500 MΩ on 1, 1000 MΩ on 8 |
Mismatch — megohms, not gigohms, and set by part |
| A flat 10 GΩ with no product term, somewhere in the sheet | 4 of the 24 | 04, 10, 17, 18 |
Match on a single gigohm floor, but only for one band or one page of the sheet |
| No figure | 1 | 00_MLCC_Application_Guide.pdf: four lines name insulation resistance, none carries a number |
Not applicable |
The application guide is a mounting and derating document, not an approval sheet, so its silence is expected.
Which six insulation resistance limits do the Walsin datasheets print?
Six, counting only the initial requirement. The product term is written R×C in ohm-farads, so a larger capacitor is allowed a lower resistance. After-stress rows add at least four further looser forms (1 GΩ or 50 Ω·F, 500 MΩ or 25, 1 GΩ or 10, 500 MΩ or 5) and we leave those out.
| Form as printed | Sheets that print it, of 24 | Export series it governs | Match / mismatch against the 10 GΩ or R×C ≥ 500 Ω·F form |
|---|---|---|---|
| 100 GΩ or R×C ≥ 1000 Ω·F, whichever is smaller (NP0 or C0G) | 2 — 16, 18 |
Safety S2, 87 NP0 rows | Mismatch — ten times the gigohm term, twice the product |
| 10 GΩ or R×C ≥ 500 Ω·F | 14 | General Purpose, MT, MG, SH, SG, 01R5, HT and the X7R rows of S2 | Reference |
| 10 GΩ or R×C ≥ 100 Ω·F | 18 | Big Sizes, Middle to High Voltage, RF, UF, HH, TT, and the Class II tiers of the table-driven series | Mismatch — one fifth of the product |
| R×C ≥ 50 Ω·F, printed with no gigohm term | 8 | the lowest tier of General Purpose, MT, ST, MG, SH and SG, and X5R at 01005 | Mismatch — one tenth of the product, and no ceiling |
| A flat ≥ 10 GΩ | 4 | Middle to High Voltage at 1000 to 3000 V, HH below 200 V, and the S2-H and S3 summaries | Match on the gigohm term, mismatch on form — no product term |
| 25, 500 or 1000 MΩ per part | 1 — 22 |
Feed Through, 8 rows | Mismatch — a different unit and no formula |
The forms overlap inside sheets. Seven of the 24 (01, 11, 12, 13, 14, 19, 20) print three of the six at once, as a tier table keyed on rated voltage and capacitance. The source URL of each sheet is in our INDEX_MLCC_series.csv, so every line can be checked.
Which Walsin MLCC series have one limit we can quote, and which need a lookup?
All 16 export series have a printed limit, but only 4,594 of 24,026 rows (19.1%) can be quoted without looking at the part. We joined each of the 16 series in the export to its governing sheet and classified the sheet’s insulation resistance statement. Five series carry one figure, four vary by dielectric or voltage, six need dielectric, voltage and capacitance together, and one is per part.
| Class of statement | Export series (rows) | Rows | Share of 24,026 | Match / mismatch — can you quote one figure? |
|---|---|---|---|---|
| One floor for the whole series | Big Sizes (501), RF (3,615), UF (285), HT (128), TT (57) | 4,586 | 19.1% | Match — one line per series |
| Banded by dielectric or voltage only | Middle to High Voltage (4,095), HH (1,709), 01R5 (317), S2 (173) | 6,294 | 26.2% | Partial — need the dielectric or the voltage |
| Tier table keyed on dielectric, voltage and capacitance | General Purpose (8,467), MT (2,176), SH (1,524), MG (788), ST (118), SG (65) | 13,138 | 54.7% | Mismatch — a per-part lookup |
| Per part number | Feed Through (8) | 8 | 0.03% | Match — read the part’s own line |
13,138 of 24,026 rows, 54.7%, sit in six series whose sheet prints three different floors in one table. We have not seen this counted anywhere. It includes the largest series, General Purpose, at 8,467 rows. Six further series have a sheet and no export rows (RH, RT, VT, S3, Mega Cap M and V), so a limit can be quoted for them and no part number can. The export itself has no insulation resistance column: it holds 16 columns, and the 173-row enriched sample carries the field on 18 rows, which we published in MLCC insulation resistance and dissipation factor: 18 of 173 rows.
When does the ohm-farad term, not the 10 GΩ floor, set the limit?
Above a crossover capacitance of 50 nF for the 500 form, 10 nF for the 100 form and 5 nF for the 50 form. The limit is the smaller of 10 GΩ and N ÷ C, where N is the printed product. At 50 nF with N = 500 both terms equal 10 GΩ. We computed it for every row: 3,335 of 24,018 rows with a readable capacitance (13.9%) sit above it in every tier their sheet prints.
The arithmetic below is ours, not a figure Walsin prints. It shows what the printed product implies for real parts. Which tier a given part falls in is a lookup in the sheet, so the tiers are shown side by side.
| Part number | Series | Capacitance | Implied floor under N = 500 / 100 / 50 | Match / mismatch against a 10 GΩ inspection limit |
|---|---|---|---|---|
0402X104K6R3 |
General Purpose, X5R, 6.3 V | 100 nF | 5 GΩ / 1 GΩ / 0.5 GΩ | Mismatch at every tier |
1210X227M6R3 |
General Purpose, X5R, 6.3 V | 220 µF | 2.3 MΩ / 0.45 MΩ / 0.23 MΩ | Mismatch — 4,400 to 44,000 times below 10 GΩ |
MG32B476K100 |
MG, X7R, 10 V | 47 µF | 10.6 MΩ / 2.1 MΩ / 1.1 MΩ | Mismatch |
TT31X476M6R3 |
TT, X5R, 6.3 V | 47 µF | 2.1 MΩ (one printed tier, N = 100) | Mismatch |
2220B476K250 |
Big Sizes, X7R, 25 V | 47 µF | 2.1 MΩ (one printed tier, N = 100) | Mismatch |
S242N100J502 |
Safety S2, NP0 | 10 pF | 100 GΩ (the product term is 100 TΩ) | Match — the gigohm term governs |
Add your own incumbent part in a further row before you compare, because its sheet may print a different N. A 10 GΩ limit copied onto an incoming-inspection sheet for 1210X227M6R3 would reject parts that are inside the datasheet limit, which permits a reading 4,400 to 44,000 times lower.
Do the Walsin datasheets disagree with themselves on insulation resistance?
Five of the 24 series sheets print the initial floor in two forms for the same series: 09 (RT), 10 (HH), 04 (Middle to High Voltage), 17 (S2-H) and 18 (S3). We compared each sheet’s summary page with its test table. For two of the five the effective floor differs for a part inside the sheet’s range.
On 18_Safety_X2_S3.pdf the summary states a flat ≥ 10 GΩ while the test table states 100 GΩ or R×C ≥ 1000 Ω·F for NP0 — a tenfold difference on the same dielectric. On 04_Middle_and_High_Voltage.pdf the summary states a flat ≥ 10 GΩ at 1000 to 3000 V while the test table states 10 GΩ or R×C ≥ 100 Ω·F for every rating from 200 V up. 56 of the 1,238 export rows rated 1000 to 3000 V have a capacitance above 10 nF, for example 1812B123K102 at 12 nF and 1000 V, where the table implies 8.3 GΩ and the summary says 10 GΩ. The other three differ in form but not in effective value across their capacitance range. On 09_Microwave_Automotive_RT.pdf the summary states R×C ≥ 100 and the AEC-Q200 table states R×C ≥ 500, yet no RT capacitance, 100 pF at most, reaches the crossover. The 23 export rows rated 3500 to 4000 V sit outside the summary’s printed voltage bands altogether.
What should a German buyer ask for before accepting an insulation resistance figure?
Ask for the sheet page, the tier and the test condition, in that order, and refuse a bare gigohm number. For 19,432 of the 24,026 export rows (80.9%) the limit is a function of the part, not of the series. We will send the page that governs your part number. We hold 28 Walsin datasheets, 451 pages, and for thin film, inductor, varistor, disc capacitor, RF and antenna parts we hold none, so on those lines no limit can be quoted from us at all.
- Give us the part number and the grade. With dielectric, rated voltage and capacitance the tier is a lookup, and we show you the line.
- Cross the whole list. Paste up to 500 part numbers, or drop your Excel or CSV, into the free cross-reference tool. It states differences from your original rather than hiding them behind a tick.
- Check what is in Hong Kong. Ready stock lists 1,125 part numbers as at 23 September 2026, a dated snapshot and not a live feed. The line sits on the Walsin MLCC page.
Venatech Limited: Unit 207, 2/F, Sun Ling Plaza, 30 On Kui Street, Fanling, N.T., Hong Kong, (852) 2628 4777. Confirm our authorised status with Walsin. For how the capacitance and dissipation-factor conditions are stated, see MLCC capacitance test conditions: 23 of 24 datasheets.
The short version
25 of 26 Walsin MLCC PDFs state an insulation resistance limit, and the series do not share one: we found six forms, from 100 GΩ or R×C ≥ 1000 Ω·F down to R×C ≥ 50 Ω·F with no gigohm term. 13,138 of 24,026 export rows (54.7%) sit in six series where the floor is a table lookup, and 3,335 rows have a capacitance high enough that the printed product, not 10 GΩ, is the operative floor. 1210X227M6R3 implies 2.3 MΩ to 0.23 MΩ against a 10 GΩ headline. Five sheets print two forms of the floor and two of them differ for a real part. We will not tell you a part matches on insulation resistance. We will show you the line that governs it. Confirm against both datasheets and qualify in your own circuit.
Send us the insulation resistance limit from your incoming-inspection sheet and the part numbers it covers, and we will tell you which tier of which sheet each one falls in. And a question back: when your inspection rejected a large-capacitance part at 10 GΩ, did you check the rule it was held to?
Frequently asked questions
How many Walsin MLCC datasheets state an insulation resistance limit?
25 of the 26 MLCC PDFs we hold. 24 print a gigohm-or-product limit, whichever is smaller, as an initial requirement. The feed-through sheet prints 25, 500 or 1000 megohms part by part. Only the 12-page application guide prints no figure. We read the insulation resistance row of every sheet to count this.
What is the usual insulation resistance limit on a Walsin MLCC datasheet?
The commonest printed form is 10 gigohms or an R times C product of at least 500 ohm-farads, whichever is smaller, on 14 of 24 sheets. 18 sheets print the 100 ohm-farad form, usually as a Class II tier. The safety sheets for NP0 print 100 gigohms or 1000 ohm-farads instead.
Why is the limit lower for large-capacitance Walsin MLCC?
Because the product form lets resistance fall as capacitance rises. The floor is the smaller of 10 gigohms and N divided by capacitance, where N is 500, 100 or 50 depending on the tier. A 220 microfarad part such as 1210X227M6R3 implies between 0.23 and 2.3 megohms, not 10 gigohms.
Does the Walsin MLCC export carry an insulation resistance column?
No. The 24,026-row combined export has 16 columns and none is insulation resistance. The separate 173-row enriched sample carries the field on 18 rows, which we published earlier. For the other 24,008 rows the limit comes only from the datasheet, and it depends on the series and the part.
Which Walsin MLCC series have a single insulation resistance limit?
Five series, covering 4,586 of 24,026 export rows: Big Sizes, RF, UF, HT and TT each carry one printed floor. Four more vary by dielectric or voltage only, covering 6,294 rows. Six series, covering 13,138 rows, need dielectric, voltage and capacitance together, and Feed Through is set part by part.
Do Walsin datasheets disagree on the insulation resistance limit?
Five of 24 series sheets print the initial floor in two forms for the same series. Two differ for a real part: the S3 sheet gives 10 gigohms on its summary and 100 gigohms for NP0 in its table, and the Middle to High Voltage sheet differs on 56 export rows between 1000 and 3000 volts.
What should a German incoming inspection ask for?
The sheet page, the tier and the test condition for the exact part number, not a bare gigohm figure. We hold 28 Walsin datasheets, 451 pages, and will send the page covering your part number. Confirm against both datasheets and qualify in your own circuit.