In "The Algorithm of Inventing" Altshuller defined technical contradictions as follows: "to improve one parameter (quality, part) of a system in the known ways one has to pay with the worsening of another parameter (quality, part) of the system." What "improve" and "worsening" meant was not specified. For example, if increasing the speed of a ship in the known ways requires increasing the power of its engine, is it a worsening or an improvement of the parameter "power"? While the accompanied increase in the engine's size can be viewed as a worsening of this parameter, one has difficulty to accept that an increase in the engine's power is also a worsening of anything, be it parameter "power" or anything else. The common sense dictates that an increase in the power of an engine is always good and hence this parameter is thereby improved. (See Alex Acosta's letter to that effect.)
In practice, however, Alsthuller did not adhere to his own definition. For example, in his text problem of icebreaker he applied contradiction matrix to parameters "speed" (of icebreaker) and "the power" (of its engine). He did not explain the logics behind his choice and many people (like Alex) do not understand it. They don't not see a technical contradiction between icebreaker's speed and engine's power here, but rather a technical contradiction between icebreaker's speed and engine's size/weight. It is because of how Altshuller defined technical contradictions.
But if technical contradiction is defined as any unwanted relationship between parameters then speed and power indeed become contradictory here. It is because an increase in icebreaker's speed in the known ways requires an increase in the engine's power, which is unwanted under the terms of the problem. This is probably how Altshuller understood technical contradictions but could not properly define.
So, let's correct the definition:
Technical contradiction is an unwanted coupling between parameters/qualities of a system.
This definition is wider than that, which Altshuller gave, and consistent with how he used them.