On 05:23 AM 23/07/2003, Brad Velander said:
Hi all,
well I finally had to break my normal pattern of flat hierarchy this morning. I am having difficulties getting sheet symbols to connect to the subsheet ports. Everything netlists and updates to the PCB just fine except for the 4 points that connect between sheets. What am I missing here?

I have checked that the ports seem to connect to the wire at the correct point. This is not the easiest, nor am I sure of this one, because the electrical snap grid doesn't work consistently. It doesn't always show me the enlarged dot when I try to connect to the port, sometimes it does but not always.
I have 4 ports, IP1, OP1, IP2 & OP2.

OK I assume that IP1 is a signal going into that net and OP1 is an output on the subsheet? Is this correct? The stuff below assumes that IPx is an input to the *sub-sheet* and OPx is an output from the sub-sheet. You will need to change things if you IPx is an input on the higher level sheet.

In my sheet symbol they are set as input, output, input & output respectively.

Yes this is what I would do.

On the subsheet I have the ports set as output, input, output & input respectively.

I would have the sub-sheet ports set as for the sheet entries - that is an input Sheet Entry needs a driving signal and it maps over an underlying port which is also an input. I view the matching port and sheet entry as different views of the same thing.

I gather you are thinking more that they are different ends of a special component that is an input on the higher level sheet (sheet entry) and is an output on the lower level ports, which then drives the input pin (residing on the sub-sheet). Is that right?

So for an input on net XYZ I would have an input pin "driven" by an *input* port, which matches an *input* sheet entry. Somewhere in the hierarchy I should have, on the same net XYZ, a suitable *output* pin driving an *output* port which matches an *output* sheet entry. I know it may be odd having an input port driving a pin, but I prefer to think of what the port is doing on the side not connected to a wire. So I answer "No" to the question about swapping directions when you create a sheet symbol from a sheet. But this view of the world is just my preference - the system should not care (assuming the ERC matrix is set up appropriately).

When synchronizing I have ports and nets set to global.

Shouldn't this be Ports Only or Sheet Symbols/Sheet Entries?

The only thing that doesn't connect are the 4 port connections. What am I missing guys?

In P99SE, ports and sheet entries don't name nets (this is optional in DXP). Do you have net labels on the nets? I usually make sure that the net label, port name and sheet entry name all match precisely (unless it is a full hierarchical design). Even when I do nets global I will usually fully wire up ports and sheet entries and have a top level sheet showing the interconnections. At all levels I have net labels on every wire and bus, and the *same* name (identically) for any associated ports and sheet entries (the INS key is great time saver here, if you are placing a net label, port or sheet entry and you hit INS while hovering over a net label, port or sheet entry, just the text is copied, not the whole object, so it is easy to make a net label with the right name by "morphing" from a sheet entry for instance).

This scheme means that it makes no difference whether I use Nets and Ports Global or Ports Only - in fact I sometimes will update a PCB with both net scopes and confirm nothing odd happens. But this scheme won't work in multi-channel designs where you are attempting to use a full hierarchical net list (with duplicated sub-sheets). But if you are doing this DXP is light years ahead anyway - P99SE is pretty broken in this case.


* * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
* To post a message: mailto:[EMAIL PROTECTED]
* To leave this list visit:
* http://www.techservinc.com/protelusers/leave.html
* Contact the list manager:
* Forum Guidelines Rules:
* http://www.techservinc.com/protelusers/forumrules.html
* Browse or Search previous postings:
* http://www.mail-archive.com/[EMAIL PROTECTED]
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * *

Reply via email to