MG is Walsin Technology’s “Automotive Caps Without AEC-Q200 Certification” series, 788 part numbers against MT’s 2,176. We parsed every row of both files: 656 of 788 MG tolerance cells print an unreadable negative number instead of a percentage, and at 0402, MG’s 100% real X5R parts out-range MT’s AEC-Q200 ceiling by 6.8 times.
Venatech Limited, an authorised Walsin distributor in Hong Kong, is an authorised distributor for Walsin Technology (Walsin). We hold the automotive MLCC parametric export in four files: 01_Automotive_MT.xlsx (2,176 rows), 02_Automotive_MG.xlsx (788 rows), 04_Automotive_SoftTerm_ST.xlsx (118 rows) and 03_Automotive_SoftTerm_VT.xlsx, which is a 2-row stub stating zero orderable part numbers. We opened and parsed all four files for this post, plus the 11_Automotive_MT.pdf and 14_MG_Series_High_Reliability.pdf approval sheets. Every part number below is real and came out of those files.
What exactly does “without AEC-Q200” mean on Walsin’s MG series?
It means the series name itself denies the qualification Walsin’s MT series claims. Walsin’s own ordering key reads “Automotive Caps Without AEC-Q200 Certification (MG series),” and we counted it: the exact string appears 6 times across 00_MLCC_Full_Catalogue.pdf, every one of the 6 using the word “without.” MG is still named automotive — Walsin’s own catalogue filing places it beside MT under “Automotive” — but automotive and AEC-Q200-qualified are two separate claims, and MG carries only the first. There is no partial-qualification wording anywhere in the MG approval sheet; it is a binary denial, stated once per heading and repeated on every page of the datasheet.
How many MT and MG automotive MLCC part numbers actually exist?
2,173 distinct MT part numbers and 785 distinct MG part numbers. We counted rows first — 2,176 in the MT file and 788 in the MG file — then counted distinct part numbers and found the gap is duplication, not new parts: 3 MT part numbers and 3 MG part numbers each appear on two rows. MT21B104K500, MT21B223K101 and MT21B224K500 are the MT duplicates; MG21B102K500, MG21B223K101 and MG31B471K251 are the MG duplicates. Every one of the 6 duplicate pairs carries a different published thickness on its two rows — for example MT21B104K500 prints 0.80 mm on one row and 1.25 mm on the other. We have not seen this duplication-with-conflicting-thickness pattern counted against both series together before.
| File | Rows | Distinct part numbers | Duplicated part numbers |
|---|---|---|---|
01_Automotive_MT.xlsx |
2,176 | 2,173 | 3, each at two thicknesses |
02_Automotive_MG.xlsx |
788 | 785 | 3, each at two thicknesses |
04_Automotive_SoftTerm_ST.xlsx |
118 | 118 | 0 |
03_Automotive_SoftTerm_VT.xlsx |
0 orderable | 0 | — |
So the series that claims AEC-Q200 by name (MT plus ST) totals 2,291 part numbers, and the series that denies it (MG) totals 785. Add VT’s zero and the whole automotive MLCC shelf we can quote today is 3,082 rows across three live series.
What happens to the MG series Tolerance column on 656 of 788 rows?
It stops reading as a tolerance at all. We checked every cell in the MG Tolerance column: 656 of 788 rows print a bare negative number — -0.1 on 369 rows, -0.05 on 220, -0.01 on 30, -0.2 on 21 and -0.02 on 16 — where MT’s own Tolerance column on the same fields prints +-10%, +-5%, +-1%, +-20% and +-2%. The other 132 MG rows print a readable absolute tolerance such as +-0.25pF, matching MT’s own convention for NP0 parts under 10 pF. So the defect is confined to the percentage-tolerance rows, not the whole column.
We did not stop at counting the broken cells; we decoded them. Every one of the 656 bare values matches the tolerance letter already encoded in the part number, read at face value as a negative decimal fraction of the percentage: -0.01 is the 1% (F) parts, -0.02 is 2% (G), -0.05 is 5% (J), -0.1 is 10% (K), and -0.2 is 20% (M), with zero exceptions across all 656 rows. MG15X104K100 carries the letter K in position nine of its part number and the broken cell reads -0.1; decoded, that is ±10%, exactly what K means on every MT part number in the same position. The data is recoverable, but a cross-reference tool reading the Tolerance column as a number rather than decoding the part number would either drop the row or report a negative percentage that does not exist on any datasheet. We have not seen this exact decode — the broken cell, the part-number letter, and the one-to-one match across all 656 rows — published anywhere.
| MG Tolerance cell (as exported) | Rows | Part-number letter | Decoded tolerance |
|---|---|---|---|
-0.01 |
30 | F | ±1% |
-0.02 |
16 | G | ±2% |
-0.05 |
220 | J | ±5% |
-0.1 |
369 | K | ±10% |
-0.2 |
21 | M | ±20% |
Does MT or MG cover more capacitance at the same case size?
At 0402, MG does — by 6.8 times. We filtered both files to Size 402 (0402) and took the maximum capacitance in each: MT’s 0402 ceiling is 100 nF, reached only in X7R; MG’s 0402 ceiling is 680 nF, reached only in X5R, a dielectric MT does not offer at 0402 at all. The gap exists because MG carries 11 X5R part numbers at 0402 that have no counterpart anywhere in the AEC-Q200-qualified MT series. Moving an MG 0402 X5R line to an AEC-Q200-qualified part is not a drop-in — it is a dielectric change, because the capacitance headroom MG offers at that case size does not exist on the qualified shelf.
| Part number | Series | Dielectric | Capacitance | Rated voltage | Match / mismatch vs MT15B104K160 (0402, X7R, 100 nF, 16 V) |
|---|---|---|---|---|---|
MT15B104K160 |
MT, AEC-Q200 | X7R | 100 nF | 16 Vdc | reference part |
MG15X104K100 |
MG, no AEC-Q200 | X5R | 100 nF | 10 Vdc | Match on size and capacitance; mismatch — not qualified, X5R not X7R, 10 V not 16 V |
MG15X684K100 |
MG, no AEC-Q200 | X5R | 680 nF | 10 Vdc | Mismatch — 6.8× the capacitance at the same case size; no MT part reaches it |
ST15B104K160 |
ST, AEC-Q200 | X7R | 100 nF | 16 Vdc | Match on every electrical field and on qualification; mismatch — wider mechanical tolerance (±0.20 mm vs ±0.05 mm) from the soft-termination layer |
At 0201 the picture reverses to a tie: both MT and MG stop at exactly 10 nF, and neither offers X5R at that case size, so the qualification gap costs nothing in capacitance headroom at 0201 — it only shows up once the case size reaches 0402.
Do MT and MG match on size and voltage when they differ on qualification?
Mostly on size, rarely at the top of the voltage range. We compared the full Size distribution of both series and both top out across the same six case codes, 0201 through 1210, plus one 1812 code MT does not carry.
| Case size | MT rows | MG rows |
|---|---|---|
0201 (201) |
204 | 36 |
0402 (402) |
485 | 199 |
0603 (603) |
585 | 264 |
0805 (805) |
512 | 198 |
1206 (1206) |
312 | 67 |
1210 (1210) |
78 | 18 |
1812 (1812) |
0 | 6 |
Voltage is where the two series pull apart at the extremes. MT’s single highest-voltage row is MT32N103J12E at 1250 Vdc; MG’s highest is a group of three rows at 1000 Vdc (MG31N102J102, MG31N680J102, MG31N101J102). Below those outliers, both series sit on the same voltage ladder — 10, 16, 25, 50, 100, 200, 250, 500, 630 Vdc — in the same case sizes, so the voltage gap is a single part number on each side, not a structural difference across the shelf.
What do you give up moving from MT to MG, and what do you gain?
You give up the AEC-Q200 claim and the top 1,250 Vdc rating, and you gain X5R capacitance headroom at 0402 and six more part numbers at 1812. Neither direction is free, and neither series is the same as the other — they are two different shelves that happen to share a footprint on most of their overlap.
- If the line item requires AEC-Q200, MT (or ST, where soft termination is also needed) is the only shelf that carries the claim. MG cannot satisfy that requirement regardless of how closely its electricals match.
- If the line item does not require AEC-Q200 and needs X5R capacitance above 100 nF at 0402, MG is the only one of the two that offers it — MT has no X5R option at that case size at all.
- Either way, read the Tolerance column as the part number, not the cell, on any MG row — 656 of 788 cells need the decode above before they mean anything.
- Cross the whole automotive section of your BOM in one pass. Paste up to 500 part numbers, or drop your Excel or CSV, into the free cross-reference tool, filter to automotive grade, and the AEC-Q200 version is listed beneath the commercial one where one exists. Then check ready stock — a dated snapshot as at 23 September 2026, not a live feed, and automotive MLCC does not currently sit on it.
For the exact documents to request once you have picked a series, see the AEC-Q200 qualification report: 1 of 3 documents a German Tier 1 will ask for. For the earlier dielectric and voltage breakdown of these two series, see AEC-Q200 MLCC: Walsin MT vs MG series. The full automotive and general-purpose range sits on the Walsin MLCC products page.
The short version
MT and MG are both named automotive by Walsin, and only MT’s name — and ST’s — carries the AEC-Q200 claim. We parsed all 2,176 MT and 788 MG rows: 2,173 and 785 distinct part numbers after removing 6 duplicate codes, each duplicated at two conflicting thicknesses. 656 of 788 MG tolerance cells print an unreadable negative number that decodes cleanly, without exception, to the percentage already encoded in the part number. At 0402, MG’s X5R parts reach 680 nF against MT’s 100 nF ceiling — the unqualified shelf, not the qualified one, carries the larger capacitance at that case size. We will not tell you a part is equivalent. We will show you where it differs.
Send us the MT or MG line items off your BOM and tell us whether AEC-Q200 is a hard requirement on that line. We will come back with the matching candidates field by field, the decoded tolerance where the export needs it, and a plain statement of which fields do not transfer.
Frequently asked questions
<!– wp:rank-math/faq-block {"questions":[{"id":"faq-question-01","title":"What does \"AEC-Q200\" mean on Walsin's MG series name?","content":"Nothing positive. Walsin's own heading reads \"Automotive Caps Without AEC-Q200 Certification (MG series),\" and we counted the exact string: it appears 6 times across the full MLCC catalogue, every one using the word \"without.\" MG is automotive by Walsin's own filing, but it does not carry the AEC-Q200 claim that MT and ST carry.","visible":true},{"id":"faq-question-02","title":"How many Walsin MT and MG automotive MLCC part numbers exist?","content":"2,173 distinct MT part numbers and 785 distinct MG part numbers, after removing 6 duplicate codes we found on both sides — 3 in each file, each duplicated across two rows with a different published thickness. The row counts before de-duplication are 2,176 and 788.","visible":true},{"id":"faq-question-03","title":"Why is the MG series Tolerance column unreadable on so many rows?","content":"656 of 788 rows print a bare negative number such as-0.1 instead of a percentage like +-10%. We decoded all 656 against the tolerance letter already in the part number and found a clean match with zero exceptions: -0.1 is always the K (±10%) parts, -0.05 is always J (±5%), and so on down to -0.01 for F (±1%).”,”visible”:true},{“id”:”faq-question-04″,”title”:”Does MG ever offer more capacitance than MT at the same case size?”,”content”:”Yes, at 0402. MG’s ceiling there is 680 nF, reached only by 11 X5R part numbers, against MT’s 100 nF ceiling in X7R. MT carries no X5R dielectric at 0402 at all, so that capacitance headroom does not exist on the AEC-Q200-qualified shelf.”,”visible”:true},{“id”:”faq-question-05″,”title”:”Can an MG part be used where AEC-Q200 is specified?”,”content”:”Not on the strength of the documentation. The series name is the only field that states the qualification status — the electrical fields can match exactly between an MT and an MG part at the same case size and the qualification claim still will not transfer.”,”visible”:true},{“id”:”faq-question-06″,”title”:”Is there an MT part for every MG part, or an MG part for every MT part?”,”content”:”No. The two series are not a matched pair. MT reaches 1,250 Vdc on one part number that MG has no counterpart for; MG reaches 680 nF at 0402 in X5R that MT has no counterpart for. Moving a line between the two series is a selection exercise against the destination series’ own range, not a substitution.”,”visible”:true}],”listStyle”:””,”titleWrapper”:”h3″,”titleCssClasses”:””,”contentCssClasses”:””,”sizeSlug”:””} –>
What does “AEC-Q200” mean on Walsin’s MG series name?
Nothing positive. Walsin’s own heading reads “Automotive Caps Without AEC-Q200 Certification (MG series),” and we counted the exact string: it appears 6 times across the full MLCC catalogue, every one using the word “without.” MG is automotive by Walsin’s own filing, but it does not carry the AEC-Q200 claim that MT and ST carry.
How many Walsin MT and MG automotive MLCC part numbers exist?
2,173 distinct MT part numbers and 785 distinct MG part numbers, after removing 6 duplicate codes we found on both sides — 3 in each file, each duplicated across two rows with a different published thickness. The row counts before de-duplication are 2,176 and 788.
Why is the MG series Tolerance column unreadable on so many rows?
656 of 788 rows print a bare negative number such as -0.1 instead of a percentage like +-10%. We decoded all 656 against the tolerance letter already in the part number and found a clean match with zero exceptions: -0.1 is always the K (±10%) parts, -0.05 is always J (±5%), and so on down to -0.01 for F (±1%).
Does MG ever offer more capacitance than MT at the same case size?
Yes, at 0402. MG’s ceiling there is 680 nF, reached only by 11 X5R part numbers, against MT’s 100 nF ceiling in X7R. MT carries no X5R dielectric at 0402 at all, so that capacitance headroom does not exist on the AEC-Q200-qualified shelf.
Can an MG part be used where AEC-Q200 is specified?
Not on the strength of the documentation. The series name is the only field that states the qualification status — the electrical fields can match exactly between an MT and an MG part at the same case size and the qualification claim still will not transfer.
Is there an MT part for every MG part, or an MG part for every MT part?
No. The two series are not a matched pair. MT reaches 1,250 Vdc on one part number that MG has no counterpart for; MG reaches 680 nF at 0402 in X5R that MT has no counterpart for. Moving a line between the two series is a selection exercise against the destination series’ own range, not a substitution.