Differential pair routing
Differential pair routing
We have added differential pair routing to version 2.9 beta. If you have any issues or suggestions about diff pair implementation please ask in this topic.
Re: Differential pair routing
Okay, I'll re-post this here...
In the Schematic Editor, the units have been set to inch. In the PCB Editor, the units have been set to mm. As shown in the Net Classes dialog of the Schematic Editor (first example) below, the units used for Differential Pairs are the same as those used in the Schematic Editor. Shouldn't the units used for Differential Pairs be the same as those used in the PCB Editor, instead? If not, could you let us choose which units to use in the Net Classes dialog (see second example)?
In the Schematic Editor, the units have been set to inch. In the PCB Editor, the units have been set to mm. As shown in the Net Classes dialog of the Schematic Editor (first example) below, the units used for Differential Pairs are the same as those used in the Schematic Editor. Shouldn't the units used for Differential Pairs be the same as those used in the PCB Editor, instead? If not, could you let us choose which units to use in the Net Classes dialog (see second example)?
- Attachments
-
- dp_units.jpg (358.42 KiB) Viewed 2665 times
-
- dp_choice.jpg (186.62 KiB) Viewed 2665 times
Tom
modular connectors
The biggest use I have for differential pairs is for ethernet. Specifically ethernet on RJ45 connectors.
The outer two pairs seem to be able to be routed ok.
The inner pairs, there isn't any way I can ascertain to get the system to permit me to escape the connector with a differential pair.
See the example .dip file attached. Try routing the four differential pairs. The pairing is correct on these pins.
Any ideas?
The outer two pairs seem to be able to be routed ok.
The inner pairs, there isn't any way I can ascertain to get the system to permit me to escape the connector with a differential pair.
See the example .dip file attached. Try routing the four differential pairs. The pairing is correct on these pins.
Any ideas?
- Attachments
-
- diff example.dip
- RJ45 differential example.
- (11.67 KiB) Downloaded 293 times
Re: Differential pair routing
Hello forrestc,
We plan to improve the Free Edit Trace feature. Now it does not work in beta version.
It can solve the problem.
Serg Luts
DipTrace Team
We plan to improve the Free Edit Trace feature. Now it does not work in beta version.
It can solve the problem.
Serg Luts
DipTrace Team
-
- Posts: 22
- Joined: 11 Sep 2015, 22:20
Re: Differential pair routing
Hi,
I would like to say overall well done on the diff pair routing tools, its great to see the high speed features being added to Diptrace.
I did notice a couple of things that are probably worth mentioning:
1. In the differential pair manager there doesn't seem to be a way to compare the routed lengths from one complete diff pair to another, sometimes in design a group of diff pairs need to be routed the same length, there needs to be a way to view the routed lengths simultaneously.
2. Starting to route a diff pair in PCB from a set of pads seems to work as expected, although ending the diff pair onto another set of pads sometimes does not work as expected, it wont connect. This is possibly due to the grid setting though?
Also will we get the feature to length match / tune / meander single ended signals as well as diff pairs in version 3.0?
Keep up the good work, overall DipTrace is a great product
Kind Regards,
Martin Johnson.
I would like to say overall well done on the diff pair routing tools, its great to see the high speed features being added to Diptrace.
I did notice a couple of things that are probably worth mentioning:
1. In the differential pair manager there doesn't seem to be a way to compare the routed lengths from one complete diff pair to another, sometimes in design a group of diff pairs need to be routed the same length, there needs to be a way to view the routed lengths simultaneously.
2. Starting to route a diff pair in PCB from a set of pads seems to work as expected, although ending the diff pair onto another set of pads sometimes does not work as expected, it wont connect. This is possibly due to the grid setting though?
Also will we get the feature to length match / tune / meander single ended signals as well as diff pairs in version 3.0?
Keep up the good work, overall DipTrace is a great product
Kind Regards,
Martin Johnson.
Re: Differential pair routing
Hello,
1) We will think about it.
2) The grid settings do not to affect on that. Usually in this situation the diff pair has a little place to completed the routing. Try to do the undo (on 2-3 steps) for diff pair. After that try to route repeatedly.
3) Yes, we plan to add this feature in version 3.0.
Serg Luts
DipTrace Team
1) We will think about it.
2) The grid settings do not to affect on that. Usually in this situation the diff pair has a little place to completed the routing. Try to do the undo (on 2-3 steps) for diff pair. After that try to route repeatedly.
3) Yes, we plan to add this feature in version 3.0.
Serg Luts
DipTrace Team
-
- Posts: 1
- Joined: 16 Sep 2015, 11:51
Re: Differential pair routing
Good evening,
I too am having issues with differential pairs on Ethernet connectors, the pair that connects to pin 3 and pin 6 will not route correctly.
The route always goes over other pins and there seems to be know way to successfully do it other than not to set it as a differential pair.
Has anyone else run into this ?
Any suggestions on getting this working ?
Regards,
Craig
I too am having issues with differential pairs on Ethernet connectors, the pair that connects to pin 3 and pin 6 will not route correctly.
The route always goes over other pins and there seems to be know way to successfully do it other than not to set it as a differential pair.
Has anyone else run into this ?
Any suggestions on getting this working ?
Regards,
Craig
-
- Posts: 22
- Joined: 11 Sep 2015, 22:20
Re: Differential pair routing
Hi,
Took a look at the Ethernet example posted above, one reason I can immediately see is that the pad and hole sizes don't provide enough space for a single diff pair trace, however I am still not convinced that the tool will work as expected even if that is corrected, other PCB software has similar issues with internal traces on Ethernet connectors.
Usually (other software) I tend to route the internal pairs out manually and then finish them off using the differential pair tool, on the beta of Diptrace 2.9 it doesn't yet seem possible, the reason being that there is no manual route ability of any part of the diff pair trace.
I think there needs to be some difference between differential pair routing and route single ended manual, so its possible to allow the user to route part of a diff pair manually (if they choose) and then join up the remainder using a specific diff pair route tool, rather than just defaulting without any choice to route diff pair based on the net class setting. It should then be possible to overcome the issue.
One other thing I have also noticed is that the diff pair tools do not yet support curved traces, they can be useful in some cases for very high speed designs to reduce the signal loss.
Kind Regards,
Martin Johnson
Took a look at the Ethernet example posted above, one reason I can immediately see is that the pad and hole sizes don't provide enough space for a single diff pair trace, however I am still not convinced that the tool will work as expected even if that is corrected, other PCB software has similar issues with internal traces on Ethernet connectors.
Usually (other software) I tend to route the internal pairs out manually and then finish them off using the differential pair tool, on the beta of Diptrace 2.9 it doesn't yet seem possible, the reason being that there is no manual route ability of any part of the diff pair trace.
I think there needs to be some difference between differential pair routing and route single ended manual, so its possible to allow the user to route part of a diff pair manually (if they choose) and then join up the remainder using a specific diff pair route tool, rather than just defaulting without any choice to route diff pair based on the net class setting. It should then be possible to overcome the issue.
One other thing I have also noticed is that the diff pair tools do not yet support curved traces, they can be useful in some cases for very high speed designs to reduce the signal loss.
Kind Regards,
Martin Johnson
Re: Differential pair routing
Hello All,
The DipTrace Beta version has the opportunity to edit a diff pair parts manually after the routing.
If you have the routed diff pair you can try to delete this diff pair from main menu "High Speed / Differential Pair Manager" or
from the diff pair context menu "Remove Diff Pair".
After that edit the traces manually.
Serg Luts
DipTrace Team
The DipTrace Beta version has the opportunity to edit a diff pair parts manually after the routing.
If you have the routed diff pair you can try to delete this diff pair from main menu "High Speed / Differential Pair Manager" or
from the diff pair context menu "Remove Diff Pair".
After that edit the traces manually.
Serg Luts
DipTrace Team
Re: Differential pair routing
One possible scenario would be to allow the user to right-click on the end segment of the differential pair and convert it to a standard trace. This would permit editing of the newly-orphaned segment as a standard trace and, at the same time, retain what remains of the differential pair class for future editing as such...
- Attachments
-
- dp_seg.jpg (200.74 KiB) Viewed 2575 times
Tom