• You are here:
  • Knowledge »
  • Blog »
  • Horizontal integration. Vertical integration. Multidimensional integration – Part 2
Blog

Horizontal integration. Vertical integration. Multidimensional integration – Part 2

  • 4. February 2019

Horizontal integration. Vertical integration. Multidimensional integration
Episode 2: vertical

Now that we know what we mean by horizontal integration (in case you don’t then you may read about it here, in the first episode of this series), let’s talk about vertical integration.

Our product is the final result of several engineering disciplines: mechanical, electrical, electronic, software, optical etc. The mechanical guys design a box – the enclosure which will fit the device. The electrical engineers design the wires and connectors. The optics refer to the lenses of the device (in case it’s a camera or a smartphone).

All these disciplines must converge and each brings its own requirements into the game. The wires cannot be bent at more than a certain angle. The lenses must be at a certain distance to each other or have a certain radius. The PCB cannot be smaller than a certain size in order to route all signals.

The requirements of a discipline become constraints to the other disciplines. Some constraints have higher priority than others and a compromise must be reached in order to finish the design.

As we design the electronics, the sooner we know about all the constraints coming from the other disciplines, the easier we can adapt. Late changes are most expensive. The commonly-used example is that finding out, after layout, that we cannot fit our board into the case will cause us to re-do the layout, the routing and all the post-layout verifications.

Maybe the layout has higher priority and our mechanical colleagues must re-design their case. Even in this case it’s much easier for them if they know to know how large our board is before they start designing. Or where the large connectors are. Or where the power-hungry chip is located so that the fan or the venting holes are placed accordingly.

Ideally, any part of a design must have a model with the characteristics which are relevant in the other “worlds”. The board must be a 3D model in the MCAD software, the power-hungry chips must have a thermal model for the thermal engineer so any changes made in a “world” are reflected into “the other world” and thus impose constraints.

As in the case of the horizontal integration, if information flows in an intelligent format, then nothing will be lost on the way. If the mechanical engineer only gets a drawing of the board on a piece of paper, then maybe he will miss a digit when entering it into his software and thus use the wrong data.

Any data should be entered only once and then transmitted automatically and in an intelligent format anywhere it must be used.

Change management should also be used. We get the full board model from mechanical CAD but we send back only to them only a proposal with the holes which we need in the board. The mechanical engineer accepts some proposal, rejects others or sends counter-proposal. If we sent the whole board back and forth it would be hard to notice what has actually changed.

More to follow…..