Star-delta: the interlock and the dead time

A big motor started straight across the line pulls six to eight times its full-load current for the first few seconds. The motor survives it. The building often doesn't — lights dip, the breaker gets nervous, and the utility charges you for the peak. The star-delta starter is the oldest fix in the trade: start the motor in a weaker connection, then switch it to the strong one once it's turning.

Three contactors do it, and the whole lesson is about the moment they hand off to each other.

Why two connections

The same three windings can be wired two ways. In star each winding sees the line voltage divided by √3, which means it pulls about a third of the current — and makes about a third of the torque. In delta each winding sees full line voltage: full current, full torque.

So star is the starting connection and delta is the running one. Start weak, let the rotor come up to speed, then switch. The trade is real: a third of the torque means star-delta only works on a load that will accelerate under it — a fan or a pump, not a loaded conveyor.

The three contactors

KM1 stays in. KM2 and KM3 are the pair that hands off. And here is the rule the entire circuit is built around:

KM2 and KM3 must never be closed at the same time. Star ties the winding ends to a common point; delta ties them to each other's phases. Both at once is a phase-to-phase short across the supply, through contactors that were never asked to break a fault current.

The interlock, in the logic this time

   RUN     /T1.DN    /KM3
──[ ]───────[/]───────[/]────────( KM2 star )

   RUN     T2.DN     /KM2
──[ ]───────[ ]───────[/]────────( KM3 delta )

Each contactor's rung examines the other with an XIO. It's the same pattern from the forward/reverse lesson — but notice where it lives this time. The forward/reverse print left the interlock to the panel, because a reversing starter comes with a mechanical latch that makes "both closed" physically impossible.

A star-delta assembly has no such latch between star and delta. Nothing upstream protects you. So the interlock has to be in the program, and it is.

Same catalog, opposite decision — and both are right for their own panel. That is the judgment the print is asking you to make: what already protects this, and what's left for me?

The dead time: the part everyone forgets

Look at the second rung again. Delta doesn't wait for the transition timer T1. It waits for a second timer, T2, set to 100 ms.

Why? Because the XIO interlock only protects you from the logic commanding both at once. It cannot protect you from physics. When the scan drops KM2, the coil de-energizes — but the contactor's moving contacts take milliseconds to actually part, and there's an arc across the gap while they do. The very next scan is 10 ms later. If delta closed then, it would be closing into a star contactor that is still, electrically, in the circuit.

So the sequence is:

1. T1 finishes → the star rung breaks → KM2 is commanded open.
2. T2 runs 100 ms — the contactor is given time to physically open.
3. T2.DN → the delta rung closes KM3.

Two protections, two different failure modes:

• The XIO interlock stops the program from asking for both.
• The dead time stops the program from asking too soon.

Delete either one and the circuit still looks fine on the screen. This is the difference between a rung that passes review and a starter that survives a thousand transitions.

Where people get bitten

The dead time is the classic interview question, and the wrong answer is "to let the motor slow down." It doesn't — 100 ms is far too short to matter to a spinning rotor, and slowing down is the last thing you want during a transition. The dead time exists for the contactor, not the motor.

Second trap: the star rung examines /T1.DN, not T1.TT. It holds star while the clock is still running and drops it the instant the clock is done. Using the timing bit instead would work here — but it stops working the moment anyone resets or re-arms that timer, because TT is only true while it's actively counting.

Run the real print

Load the star-delta starter in the builder. Press START and watch the sequence: KM1 and KM2 close together, the 6-second clock runs, star drops, and 100 ms later delta picks up.

Then break it on purpose, twice:

1. Delete the /KM3 contact from the star rung. The circuit still runs, and it still looks right — that's the point. The interlock is insurance you only collect on when something else has already gone wrong.
2. Wire delta to T1.DN instead of T2.DN. Now star drops and delta picks up in the same scan. On the screen: a clean transition. In a panel: both contactors bridging for the milliseconds it takes the star contacts to part.

A simulator will happily run both of those. That's exactly why you read the print instead of watching the lights.