The inductance and leakage inductance from the transformer I wound are both nearly twice those from the blogger’s winding. Is that normal? Do I need to grind an air gap?
I explained it very clearly in my blog post: after winding, it usually ends up much too high, so you need to grind an air gap or add a spacer gap; otherwise it will saturate easily.
That’s totally normal. The first one I wound came out at 3 times the target value—even worse than yours. Don’t panic.
Hand-grinding method: Find a glass plate or a tile as a platform, lay down 800~1200 grit sandpaper, add a little water, hold the center post face down, and grind in a figure-8 pattern. Grind a bit, measure, repeat; stop when you reach 150~160µH—it’s better to grind a little less and check several times; grinding too much makes the inductance too low, which is more troublesome—peak current goes up, and both the MOSFET and output transistor suffer.
If you don’t want to grind, you can also shim the air gap: put a few layers of tape or thin paper on the two outer legs—the blogger also mentioned this method. The downside is that leakage inductance will be a bit higher than grinding the center post; the upside is there’s no concentrated fringing flux, so EMI and localized heating of the windings near the gap are actually a bit better. For the lazy, shimming is recommended.
Two common misconceptions:
-
Don’t try to adjust inductance by reducing the number of turns—the turns ratio determines the reflected voltage and duty cycle, and once you change the turns ratio, the entire design has to be redone. Inductance can only be adjusted via the air gap.
-
Don’t expect grinding the air gap to bring leakage inductance down too. After grinding, the inductance drops by half, but the leakage inductance basically doesn’t budge.
For leakage inductance, first confirm the measurement method is correct: Is the secondary shorted? Is the auxiliary winding shorted? Is the LCR meter set to 150kHz? Have you scraped the pin varnish off clean? (High contact resistance will make the reading drift all over the place.) If you’ve confirmed all that and it’s still too high, then it’s a winding problem: interlayer tape too thick, margin tape too wide, wire not pulled tight, sandwich winding not done tightly, lead wires left too long. Honestly, rewinding one is more worth it than fighting it—wind two extra at once and compare measurements; that’s the fastest way to improve.
After grinding, remember to blow the dust off clean before putting the core back together.
Okay, thanks, everyone!
Does it matter that the old LCR maxes out at 100K? Will the error be large?
Bro, will the error be very large on an LCR digital bridge with a max of 100K?
No big deal.