There's been some ad-hoc discussion here about how ThinkPad models with 20V power determine if they are connected to a 65W or 90W power block, with some speculation that there is something in the 20V DC cable wiring associated with this. In another thread beagle72 had occasion to do this.....


... which definitely confirms that the DC cable from the power block only contains two conductors and there is no other "sensing/signalling" conductor in the cable.

The original thread that beagle72's text above was copied from was mostly about other power issues, but if you want to look at it, you can find it here... http://forum.thinkpads.com/viewtopic.php?f=29&t=72437

Cheers,

Bill B.

Interesting.

So he ended up with a 90W adapter with a plug from a 65W one. If there are only two wires in both models and the plug is apparently identical in both, what's the third contact for? Future enhancement?

Are we sure the third contact isn't an additional ground? Is there continuity between the known ground on the 90W plug and this third contact?

That would be my guess. In many countries the electrical code requires that all devices are grounded by a separate wire. This is not the case in North America.

Also perhaps the "auto-detect" feature isn't based on wattage but rather voltage. When you use a 65W adapter with a T60p, perhaps the ThinkPad draws so much power that the nominal DC voltage drops below some threshold and when you use a 90W adapter with a ThinkPad model that only needs a 65W adapter the DC voltage exceeds some threshold. Circuitry inside the ThinkPad detects when these thresholds are breached. If the voltage is too low, the ThinkPad doesn't work. If it's too high something like a Zener diode limits it so as to protect the system.

Oh... a DC/DC story, well... I have to have some comment on that.
Hausman is probably right - DC/DC converters are usually done in that manner - there are a couple of ways to ensure protection on these. Good DC/DC converters are even protected from the short circuit (high current, zero voltage - standard AC/DC adapters usually end up with a fried transformer after applying short circuit on the DC side). In contrast to the standard AC/DC adapters (linear, open-loop circuit configuration), DC/DC converters use negative feedback and a totally non-linear circuit (switching) configuration, so they are actually very power efficient (70-95%).
Usually, DC/DC converters have a raw converter (power switches (diode, MOSFET, IGBT, thyristor...), transformer, inductor, capacitor) and a PWM (Pulse Width Modulation) driving circuit which, on good DC/DCs, is very smart (has sensing for the output voltage and protection in case of a short circuit), but since these are extremely non-linear circuits, and we're talking about extreme amount of power during transition periods (peak current at 15 amps, if normal current is at 4 amps), Zener diodes are not used as a protection ('cause you'll need a very big Zener diode if you don't want it to get fried first time you plug it in the AC). Besides, Zener diodes are mainly used as a component in linear voltage stabilizers, and have been abandoned for any serious use because of the noise they produce (white Gauss noise), so there's no place for Zeners in a Thinkpad power supply
So... if we exclude the possibility that IBM engineers have developed a dedicated modulation type so they could insert the adapter info into DC lines, and totally forgetting that they could only add one more wire to do so... yes, the output voltage goes down as soon as inductor hits i_max, in order not to fry anything.
Also HP uses the system where they have like standard 18.4V for 65W adapter and 19V for a 90W (0.6 volts is enough to see which adapter you're using) on the same type of plug - even for the same type of laptop (but one model has integrated video, another has dedicated) - HP nc6400 for example.