Same Length Wiring in Ultiboard!!

[liquidzell]

Hey all, I must take this last step since I have many more boards to lay out in the next year or so...and if I can accomplish this "dream"

will save me hours and hour of tedious work....my situation is described in the link below, please see it AFTERyou read my question ( I took a print screen)

The thing is I am wiring several SMA input to a clock....many clocks in my boards...and each of these wires have to be the sam length..so what you do obviosly to achieve

this is to have some curves in some of the wires since the space is limited...I had asked if there is anyway the autorouter ( am now using the professional version which has more

autorouter options)could do this for you or if there is anything I can do to make my life simpler and not have to go wire by wire playing with until I get the 2.79 length I need for each.

A member of this board a while back gave me  agood suggestion which was to edit the net list and make a max and min length for each of the wires I want to be the same...and this

seemed resonable of course but if you look at the picture, even though the max and min length that I set (2.79) was applied...the autorouter still routed a wire with 1.71 length!

I tried this several times and for several wires and nothing...so i guess my question is  if there is any possible way to achieve this type of feat automatically or eve easier on Ultiboard

...considering all the future clock drivers I am going to have to do..this will same me sooooo much work!!! Thanks for any help or tips!!!!

http://homepages.udayton.edu/~arroyoea/autowriring.JPG

Message Edited by liquidzell on 10-19-2007 05:57 PM

[kittmaster]

Its typical of ultiroute to completely ignore values that you set, I'm betting this is one of those options. They are supposed to fix a lot of little things like this. Not to say its happening here, but seems since the version 10 integration, lots of strange things like this have been happening.

[liquidzell]

...Anybody else?

[lacy]

Hi there again,

I am sorry that didn't work for you the way you wanted but if ,like Kittmaster says, Ultiboard ignores your settings like this then I don't know of any other solution to help you. Maybe the NI Team has a better option that I couldn't find or didn't see.

[lacy]

It appears to me that the curves in the traces from SMA1 don't match. The top curves are shorter that the bottom curves. This may be what is throwing the bottom one off. It appears, according to what I see from your picture, that Ultiboard is going ahead and wiring it even if maximum length can't be achieved which in my opinion it should not do. If the length can't be achieved it should give you a warning message or soemthing.

Does the bottom trace curves have to be longer that the top traces curves? My thinking is that unless both curly traces are equal in length then the connecting traces are going to be unequal.

I hope I understand this correctly as this is the only think that makes sense to me.

[kittmaster]

Lacy, he is trying to do impedance matching traces, the shapes and angles are unimportant, but the matching differential traces length are what are most important to his design.

Ultiroute is NOT part of ultiboard, its a seperate "plugin" that has been merged into Ultiboard in this revision (10). They still haven't mastered variable passing between UB and UR or else this issue would not even be discussed because it would follow his rules.......just one of those things that still needs to be fixed.....just like palletization.....trace width importing and a whole slew of other minor things.

Did anyone ever get a timeline for the next rev release?

[lacy]

Thanks, Chris. I had an idea that the way he had it was necessary for some reason and you have clued me in on that reason. I am not concerned about the shape or angles, just the length of those shapes and angles.  I will try to explain. In the first line his trace has what I would call 3 curves and the second one has 4 curves. This makes the second one longer that the first. Therefore Ultiboard/Ultiroute is shortenting the ltrace from the IC  to mate with the lengthened second set of curves  Reguardless, like you say, Version 10 is not following his rules. Even if it couldn't make the trace the way he wanted it shouldn't just arbitrarily do what it wants and shorten it. The least it should do is through up a warning flag or something.

I guesss this is another issue for the NI Team to look at and try to solve. Maybe they can find a temporary workaround for him as I don't know of anything that could help in this situation.

As far as updates go, Nestor explained to me in a previous post as to why they couldn't give information about what may in the next update or even a timeline to when it will be available. He said he was going to bring this up at their next meeting and see if they could give out any information. I haven't heard how that meeting went so I don't have any clue as to whether this idea was accepted or rejected. You can read the post for yourself to see his exact wording in his response to my question. Just look for the post labeled "Updates?" with me as the aurthor.

[lacy]

Sorry everyone, I am wrong (I hate when that happens). I was under the false assumption that the curvy traces were created manually. I re-read the post and I now understand that these were created by the rouuter in order to get the trace lengths to his specs and still failed to do so. Sorry about that misunderstanding. I guess I am going to have to read the post a little more carefully from now on.

Message Edited by lacy on 10-20-2007 04:28 PM