New large caps are being manufactured below specified value. Request for measurements
- Thread starter Blue Shadow
- Start date
EchoForge
New Member
I’ve been noticing the same trend for a while. Here are some of my own measurements, mostly from parts purchased/tested in early 2025. Values shown are averages of identical new parts.
What catches my attention is not simply that many parts measure below nominal. It is how tightly grouped parts from the same batch usually are.
Whether I measure 2, 10, 50 or 100 pieces, I often find very little part-to-part variation, yet the whole group tends to be on the same side of nominal, increasingly toward the lower end of the tolerance range.
I’m seeing this with film capacitors as well as electrolytics. For example, in a recent bag of 100x Kemet R82 2.2 nF ±5% film caps, a large portion measures around 2.108 nF, or about -4.2%, very close to the lower limit while still being tightly grouped.
The table above is mostly from early 2025, and at that time I still occasionally had batches such as the FC or YXM that were at or above nominal. What I find different in 2026 is that I almost never seem to receive a batch centered anywhere near the upper tolerance limit anymore.
Since it has already been mentioned here by Blue Shadow that Nichicon confirmed deliberately targeting below nominal on some production while remaining within tolerance, I’m wondering how widespread this practice has become.
Has anyone here recently received a sizeable batch 10, 20, 50 or 100 parts whose average was clearly closer to the upper tolerance limit than to the lower one?
I’m not talking about one capacitor measuring high. I mean an actual production lot statistically centered toward the upper end.
It would be interesting to know whether low-side centering has effectively become an economic manufacturing standard across most of the major capacitor manufacturers.
| Series | Rated | Qty | Average | Deviation |
|---|---|---|---|---|
| Nichicon LLS | 8,200/50 | 2 | 7,285 µF | -11.2% |
| CDE 380LX | 10,000/50 | 2 | 8,635 µF | -13.7% |
| Nichicon LLS | 10,000/80 | 2 | 8,440 µF | -15.6% |
| Chemi-Con SMH | 8,200/80 | 2 | 7,300 µF | -11.0% |
| Chemi-Con SMH | 10,000/100 | 2 | 8,630 µF | -13.7% |
| CDE 380LX | 10,000/100 | 2 | 9,055 µF | -9.5% |
| Rubycon TXW | 100/100 | 4 | 82.55 µF | -17.5% |
| Nichicon UKW | 100/100 | 4 | 89.73 µF | -10.3% |
| Panasonic FR-A | 47/50 | 4 | 43.00 µF | -8.5% |
| Nichicon UPM | 10/100 | 2 | 9.572 µF | -4.3% |
| Panasonic FC | 10/100 | 2 | 10.108 µF | +1.1% |
| Rubycon YXM | 1/100 | 2 | 1.023 µF | +2.3% |
| Nichicon UDB BP | 4.7/50 | 2 | 4.320 µF | -8.1% |
What catches my attention is not simply that many parts measure below nominal. It is how tightly grouped parts from the same batch usually are.
Whether I measure 2, 10, 50 or 100 pieces, I often find very little part-to-part variation, yet the whole group tends to be on the same side of nominal, increasingly toward the lower end of the tolerance range.
I’m seeing this with film capacitors as well as electrolytics. For example, in a recent bag of 100x Kemet R82 2.2 nF ±5% film caps, a large portion measures around 2.108 nF, or about -4.2%, very close to the lower limit while still being tightly grouped.
The table above is mostly from early 2025, and at that time I still occasionally had batches such as the FC or YXM that were at or above nominal. What I find different in 2026 is that I almost never seem to receive a batch centered anywhere near the upper tolerance limit anymore.
Since it has already been mentioned here by Blue Shadow that Nichicon confirmed deliberately targeting below nominal on some production while remaining within tolerance, I’m wondering how widespread this practice has become.
Has anyone here recently received a sizeable batch 10, 20, 50 or 100 parts whose average was clearly closer to the upper tolerance limit than to the lower one?
I’m not talking about one capacitor measuring high. I mean an actual production lot statistically centered toward the upper end.
It would be interesting to know whether low-side centering has effectively become an economic manufacturing standard across most of the major capacitor manufacturers.
EchoForge
New Member
Nichicon UKL1V100KDD1TA, 10 µF / 35 V
Measured 5 out of 50:
9.708, 9.734, 9.696, 9.676, 9.749 µF
Average: 9.713 µF (-2.87%)
Spread within the measured sample: 0.073 µF, or about 0.75%.
Rubycon 50TWL3R3MEFC5X11, 3.3 µF / 50 V
Measured 7 out of 50:
2.929, 2.925, 2.900, 2.915, 2.933, 2.921, 2.929 µF
Average: 2.922 µF (-11.46%)
Spread within the measured sample: 0.033 µF, or about 1.13%.
Both samples are tightly grouped internally, even though the Rubycon lot is centered substantially farther below nominal value.
Measured 5 out of 50:
9.708, 9.734, 9.696, 9.676, 9.749 µF
Average: 9.713 µF (-2.87%)
Spread within the measured sample: 0.073 µF, or about 0.75%.
Rubycon 50TWL3R3MEFC5X11, 3.3 µF / 50 V
Measured 7 out of 50:
2.929, 2.925, 2.900, 2.915, 2.933, 2.921, 2.929 µF
Average: 2.922 µF (-11.46%)
Spread within the measured sample: 0.033 µF, or about 1.13%.
Both samples are tightly grouped internally, even though the Rubycon lot is centered substantially farther below nominal value.