AMS1117/Why you need one/02. How a regulator holds still
Why you need one · 02 of 11

How a regulator holds still

Inside the chip, a transistor sits in the supply line and a comparator watches the output. When the output moves, the transistor opens or closes until it comes back. Whatever the transistor drops, it turns into heat.

A valve in the supply line

Between VIN and OUT there is one transistor. Think of it as a valve: fully open, it still drops about a volt; partly open, it drops more. The chip sets how far open it is, all the time, from one measurement.

The measurement

Two resistors inside the chip divide the output down, and a comparator checks that fraction against a fixed 1.25 V reference. For the 3.3 V version the resistors are chosen so that the tap is 1.25 V exactly when OUT is 3.3 V. (Engineers call this comparator an error amplifier: it answers "how far off", not just "yes or no", but the idea is the same.)

Too low, and the comparator opens the valve. Too high, and it closes it. Press play to see both:

Measure the output, correct the transistor, repeat
the load jumps
What changes
OUT
3.30 V
Across the transistor
1.70 V
Heat
0.06 W
Steady: 5 V in, 3.30 V out, the part drawing 20 mA.

The loop never stops. A part that suddenly draws more pulls OUT down, and the chip opens up to match. A supply that rises pushes OUT up, and the chip closes down. The output capacitors on the board hold OUT steady for the instant before the chip catches up.

Every volt the valve drops is heat

The part gets 3.3 V. The supply gives more. The difference is dropped across the transistor, and a voltage across something with current through it is power: volts times amps, and all of it heat.

From 5 V at 240 mA that is 1.7 V × 0.24 A, about 0.4 W — the chip runs warm. From 9 V at the same current it is about 1.4 W, and the chip shuts itself down. Where the extra volts go is the whole of that sum.

The adjustable version

The AMS1117-ADJ is the same chip with the two resistors left out. Its first pin, ADJ, is the comparator's input, and the board supplies the resistors — one of them a trimmer you turn. The chip still holds 1.25 V, now between OUT and ADJ. How the ADJ board sets its voltage takes it from there.

When it does not work

The output reads 3.28 V, not 3.30 V

That is inside the tolerance. The datasheet allows 3.234 V to 3.366 V for the 3.3 V version, and a cheap meter adds an error of its own. Anything in that window is a working regulator.

The output is fine with no load and drops when I connect the part

The regulator can only correct while it has headroom. If VIN falls under load too — a thin USB cable, a tired battery — the output follows it down. Measure VIN with the part running. The volt it needs has the numbers.

Does a higher input voltage make it work better?

No. Anything above the output plus about a volt is simply dropped across the transistor as heat. The part gets 3.3 V either way; the chip gets hotter.

Where this goes next

Eighteen fixed at 3.3 V, six you set yourself, and every part on each.

The two boards

Edit this page — content/books/ams1117/how-a-regulator-holds-still.mdx

Community

Questions about this product

See what other owners have asked, and read their solutions.

Ask a question ↗

AMS1117 LDO Regulator Kit, 24 Boards (18 × 3.3 V, 6 × Adjustable)

Loading discussions…

Discuss this article

Ask about this page. The answer stays here, on the page it belongs to, for whoever hits the same wall next.

Browse Modules and blocks on the forum