Positive Vs. Negative Planes? or both?

Positve planes have the follow advantages

1) WYSIWG What you see is what you get.
2) Thermals/antipads can be idone to the DRC , the padstacks values defined, or a value added to the drc value or antipad
3) Handles mix planed plane and wiring better
 
Positive plane have the following disadvantages.

1) Large Gerber files

Negative Planes have the following  advantages

1) smaller data files

Negative Planes have the following disdavantages

1) Difficult to make WYSIWIG and in some instance where certain switches are used  in gerber creation impossiible. Such as full contact thermal reliefs.
2) DRC is not accurate and needs to be verified thru a 3rd party tool suc as ADI, CAM350 or VALOR
3) I'm not sure if this problem still exists but the SI tool wouldn't check impedance properly with negative planes.

Positve and Negative planes can be dynamic or static.

Planes can be set negative or positive independent of padstacks provided you do define your thermal flashes and antipads.

Second opinion:

I've never had problems with NEG planes, but I learned to use them, and understand what can go wrong. If you go with NEG planes, please learn the process of using them WELL. Yes, they can cause problems, but it's generally because the user tries something quirky. They've worked very well for me, and I've done 4-5 planes also on the same layer. One final issue is with gerber viewers. Some DO NOT bode well with Allegro negative layering.

Now, I've also used PADS alot! and I never learned their NEG plane creation. Everything I've done in PADS was POS. It works well too.

File size doesn't seem to be an issue now-a-daz to me. :)

Good day.
Mitch

Additionally, positive planes tend to have less copper on the plane than does a negative plane. This is because the positive plane will abide by the DRC settings and void around pins an extra amount than does the negative plane that has no DRC involved in it. This can be a problem with impedance controlled signals at EHF frequencies and bit-rates. My current employer uses negative plane structures to its advantage for minimal discontinuity signal transitions and other similar applications which can [u]not[/u] be duplicated with positive planes.

As has been mentioned, you need to know [b]exactly[/b] what you are doing when using negative planes, so I would probably suggest positive planes unless there was a need for negative planes, even though I personally prefer negatives.

Mitch good point about the gerber viewers, I'd like to add this though it is with Cadences 274X which is a full implementation of the GERBER 274X spec the problem comes when you use a viewer that only incorprated the aperture part of the 274X spec. I would reccommend going with a viewer that supports the complete 274X spec.

The file size does matter in the time it takes the photplotter to process.