The MLCC capacitance change limit after a reliability test is not one number. We read every X7R limit in 11 Walsin Technology (Walsin) datasheets and found six: 7.5%, 10%, 12.5%, 15%, 20% and 25%. Four sheets also print limits for 12 dielectric codes that their own series holds no part in.
Venatech Limited, an authorised Walsin distributor in Hong Kong, holds 28 Walsin datasheets. This post is written from the second half of them alphabetically, the 11 MLCC sheets that state an X7R limit: MG, SH, SG, TT, VT, S2, S2-H, S3, Mega Cap M, Mega Cap V and FT. We converted each PDF to text (the MG, SH, SG, TT, HT and FT sheets are public on Walsin’s site), read the “Requirements” column of every reliability row, and set the dielectric codes the sheets name against the 24,026-row parametric export. Every part number below is real. The question is the one a qualification engineer asks the day a part drifts: how much drift did the sheet allow?
How many different capacitance change limits does Walsin print for X7R after reliability tests?
Six, from plus or minus 7.5% to plus or minus 25%. We listed every “cap change” limit that the 11 sheets print for an X7R part after a reliability test. The spread is 3.3 to 1, and it is set by the test, not by the part. Soldering heat is the only test where all 11 sheets agree.
| Test | Lowest X7R limit and where | Highest X7R limit and where | Match / mismatch across the 11 sheets |
|---|---|---|---|
| Resistance to soldering heat | ±7.5%, all 11 sheets | ±7.5% | Match — one value everywhere |
| Temperature cycle | ±7.5% (MG, SH, SG, TT, S2, S3, M, V, and FT21B) | ±12.5% (FT15X); ±10.0% on VT and S2-H, both at 1,000 cycles | Mismatch — three values |
| Damp heat, steady state or biased | ±10% (VT moisture resistance) | ±25% (TT; MG, SH and SG at 6.3 V and below); ±15% on S2 and S3 | Mismatch — four values |
| Damp heat load and endurance | ±12.5% (MG, SH, SG at 10 V and above; M; V; FT; VT) | ±25% (TT; MG, SH and SG at 6.3 V and below); ±20% on S2, S2-H and S3 endurance | Mismatch — three values |
| Board flex or bend | ±7.5% (FT, 2 mm) | ±12.5% (MG, SH, SG, TT, VT, S2, S3, M, V) | Mismatch — and S2-H states only “within the specified tolerance” |
The 7.5% soldering-heat limit and the 25% TT humidity limit are 3.3 times apart for the same dielectric code. Neither is wrong. They answer different tests, and a qualification report that quotes “X7R within ±12.5%” has quoted one of six. The cycle-count side of this is in our temperature cycling post, and the load-life side in our load life post.
Which limit applies to a real part number?
It depends on rated voltage, capacitance and sometimes the prefix. The MG, SH and SG sheets print a voltage rule under the Class II limit: 10 V and above within ±12.5%, 6.3 V and below within ±25%, with named exceptions at 10 V. We parsed all 788 MG rows against that rule: 8 of 444 Class II rows fall in the ±25% band by voltage alone, and 47 of 444 (10.6%) if the “C ≥ 1 µF” line printed beside it is read as applying to every Class II part. The SH sheet gives 7 of 1,074 by voltage and 125 of 1,074 (11.6%) on the wider reading.
| Part | Series and values | Test examined | Limit that applies | Match / mismatch against ±12.5% |
|---|---|---|---|---|
MG15X104K100 |
MG, X5R, 10 V, 100 nF | Damp heat load | ±12.5% | Match |
MG15B104K6R3 |
MG, X7R, 6.3 V, 100 nF | Damp heat load | ±25% | Mismatch — the 6.3 V rule |
MG21B105K250 |
MG, X7R, 25 V, 1 µF | Damp heat load | ±12.5% by voltage, ±25% if the C ≥ 1 µF line applies | Mismatch under the wider reading |
SH31X106K160 |
SH, X5R, 16 V, 10 µF | Damp heat steady state and load | ±25% on the wider reading (the SH steady-state row carries the C ≥ 1 µF line) | Mismatch |
TT15X105K6R3 |
TT, X5R, 6.3 V, 1 µF | Damp heat, load and endurance | ±25% on all three | Mismatch |
FT21B103M500 |
FT21B, X7R, 50 V, 10 nF | Temperature cycle, 10 cycles | ±7.5% | Mismatch — tighter |
FT15X435M4R0 |
FT15X, X5R, 4 V, 4.3 µF | Temperature cycle, 10 cycles | ±12.5% | Match |
The SH sheet prints the C ≥ 1 µF line in its damp-heat steady-state row. The MG and SG sheets print it only in the load and endurance rows. Same series family, same test name, one line different. The FT sheet keys the cycling limit to the prefix: 7 of the 8 FT export rows are FT21B at ±7.5% and 1 is FT15X at ±12.5%.
Does the temperature cycle limit depend on the number of cycles?
Not in the way you would expect. The cycle count runs 5, 10, 100 and 1,000 across these sheets, and the limit does not rise with it. The Mega Cap M and V sheets run 100 cycles at ±7.5%; the VT and S2-H sheets run 1,000 cycles at ±10.0%; the MG sheet runs 5 cycles at ±7.5%. So 200 times the cycles (5 against 1,000) buys 2.5 points of drift, and 20 times the cycles (5 against 100) buys none.
That is a statement about the sheets, not about the capacitors. Only the VT and S2-H sheets pair the longer test with the looser limit. The M and V rows show that the limit was not derived from the cycle count at all. If your specification quotes a drift limit and a cycle count together, take both from the same sheet.
Which dielectric codes does a sheet print limits for that its series does not contain?
Twelve codes in four series. For each export series that has a sheet, we listed the dielectric codes named in the reliability section and compared them with the codes in the series’ export rows. We counted 12 codes named in the MG, SG, SH and TT sheets that appear on no export row of the series, and 26 HT rows carrying a code the HT sheet prints no limit for. We have not seen this counted anywhere.
| Series | Export rows | Dielectrics in the export | Codes named in the sheet’s reliability section but absent from the series | Match / mismatch |
|---|---|---|---|---|
| MG | 788 | X7R 414, NPO 344, X5R 30 | Y5V, X6S, X7S (3) | Mismatch |
| SG | 65 | X7R 55, NPO 10 | X5R, Y5V, X6S, X7S (4) | Mismatch |
| SH | 1,524 | X7R 1,045, NPO 450, X5R 27, X7T 2 | X6S, X7S (2) | Mismatch |
| TT | 57 | X5R 52, X7R 5 | NP0, X7E, X6S (3) | Mismatch |
| HT | 128 | X8G 102, X8R 26 | none; the reverse — X8R appears once in the sheet, in a construction table | Mismatch — 26 rows, no printed limit |
| S2 | 173 | NPO 87, X7R 86 | none | Match |
Some of this is spelling. The sheets write NP0 with a zero and the export writes NPO with a letter, and we treated them as one code. Most of it is not. The MG and SG sheets print five numeric Y5V limits: ±20% after soldering heat and after temperature cycling on both, and ±30% after bending on SG. Neither series holds a Y5V part. The only two Y5V rows in the whole export are General Purpose parts, 1206F474M500 and 1206F474Z500.
The HT sheet runs the other way. Its cover line reads X8G, its ordering key lists only G for the dielectric digit, and its limits are written for X8G alone. But 26 of 128 HT export rows carry R in that position, such as HT18R101K500, against HT15G2R2B500 for X8G. The sheet’s own wording makes the X8R parts a gap, not an exception. Our HT series post covers the X8G range itself.
What should you ask for before accepting a drift limit?
Ask which sheet, which test row, and which rule line within that row applies to the exact part number. Three facts decide it: rated voltage against the 6.3 V and 10 V thresholds, capacitance against 1 µF, and for FT the prefix. The free cross-reference tool shows the Walsin parts that match your specification and states the differences from your original, but it does not select a drift limit for you, and 11 sheets are not one table.
We hold 28 Walsin datasheets, 451 pages, and send the one covering your part on request. VT, S2-H, S3 and Mega Cap M and V have no export rows, so the tool cannot offer them. Stock in Hong Kong is on the ready stock page: 1,125 part numbers as at 23 September 2026, a dated snapshot and not a live feed. Capacitors sit on the MLCC page. Qualification stays with you. Confirm against both datasheets and qualify in your own circuit.
The short version
After a reliability test an X7R part is allowed ±7.5%, ±10%, ±12.5%, ±15%, ±20% or ±25% depending on the test and the sheet, a spread of 3.3 to 1. Only soldering heat is the same on all 11 sheets. Of 444 MG Class II rows, 8 take ±25% on voltage alone and 47 do if a C ≥ 1 µF line is read widely. A 100-cycle test on the M sheet carries the same ±7.5% as a 5-cycle test on MG. Twelve dielectric codes are named in four sheets whose series hold none, five of them with numeric Y5V limits, and 26 HT rows carry X8R where the sheet limits X8G.
Send us the drift limit in your specification and the part number it governs, and we will tell you which Walsin sheet row it came from. And a question back: when you wrote “within 12.5%” into the qualification plan, which of the six did you mean?
Frequently asked questions
How many different capacitance change limits does Walsin print for X7R?
Six: plus or minus 7.5, 10, 12.5, 15, 20 and 25 percent. We read every X7R capacitance change limit after a reliability test in 11 Walsin datasheets. The limit is set by the test and the sheet, not by the dielectric. Resistance to soldering heat is the only test with one value, 7.5 percent, on all 11 sheets.
Why does the same X7R part get 25 percent on one sheet and 12.5 on another?
Because the humidity and endurance rows key the limit to rated voltage and capacitance. The MG, SH and SG sheets print 12.5 percent at 10 volts and above and 25 percent at 6.3 volts and below, and the TT sheet prints 25 percent on all three humidity and endurance rows. Read the rule line, not just the dielectric.
Does a longer temperature cycling test allow more capacitance drift?
Not consistently. The Mega Cap M and V sheets run 100 cycles at plus or minus 7.5 percent, the MG sheet runs 5 cycles at the same limit, and the VT and S2-H sheets run 1,000 cycles at plus or minus 10 percent. Take the drift limit and the cycle count from the same sheet.
Do Walsin sheets print limits for dielectrics the series does not hold?
Yes. Four sheets name 12 dielectric codes that appear on no export row of their own series: 3 in MG, 4 in SG, 2 in SH and 3 in TT. The MG and SG sheets print five numeric limits for Y5V, a code that neither series holds. The only two Y5V rows in the export are General Purpose parts.
Which HT export rows have no printed capacitance change limit?
The 26 X8R rows out of 128 in the HT export. The HT sheet cover line reads X8G, its ordering key lists only G for the dielectric digit, and every limit is written for X8G. X8R appears once in the sheet, in a construction table. The X8G rows are covered. Confirm the X8R limit with Walsin.
How do I find the right limit for my Walsin part number?
Identify the sheet that governs the series, then the row for the test, then the rule line for your rated voltage and capacitance. For the FT series the prefix decides, FT21B or FT15X. We hold 28 datasheets, 451 pages, and send the one for your part on request.