I am laying out my first PCB. I seem to be a little nervous when I lay down a trace and some of the traces have some extra bends or may be a little close to something else. I have been deleting the entire trace and making another attempt at it. Is there an easier way to correct an existing trace without completely erasing it and starting over. Thanks Mike
You can delete the last track segment with a āaltā+ādelā - and the whole trace with ādelā. You can alter the track posture with the ā/ā key.
Make sure you are using the interactive āpush and shoveā tool - this will push your traces around and lay them out nicely.
Track width and how close tracks are placed to other features is determined by the clearances set in the Board Setup. These are usually OK if you are going to use a cheap PCB fabricator. If you are using an alternative manufacturing process - for instance toner transfer or PCB milling, you might want to adjust those - possibly quite significantly.
1 Like
Assuming version 5.1.x with modern toolset:
You can use interactive drag (shortcut d or reachable via the right click context menu of the trace) to drag parts of a trace.
You can also simply draw a new trace and kicad will replace the old one (This option can be turned of in the preferences.
Wellā¦ this looks interesting, but I can not seem to get it work. I have the trace minās set, I found the Interactive router settings, yet the shove or replace doesnāt seem to work. Is this a plugin or something I have to add? Mike
I think I found one reason why I was having trouble. I need to zoom in more. Now the replace seems to work. Yet the shove and other functions are sill eluding me. Mike
Are you in the modern toolset? (press F11 to change into it) the routing settings can be chosen when you click in the button to draw tracks and then do right click.
Yes I have the modern toolset. Yet I seem to be getting small knobs at the spots were the trace changes direction. I try and delete just that section and replace it, but the same knob apprears. The trace doesnāt seem smooth. Mike
Seems another part of my problem is that somehow I have placed traces right on top of others while attempting to correct. Iāll have to delete these and try again. Mike
Seems after I delete these extra traces, the shove seems to work. I think I need to clear everything and start again. I only have a dozen or so traces so far anyway. Thanks for the help Mike
I found it VERY helpful to deploy the āshow tracks in outline modeā function to find any fragments and/or unwanted leftovers inside tracks.
Click either left hand side menu button or just K to change the visual state of tracks.
The clean up tracks function should actually remove those little unnecessary extra track segments.
edit -> cleanup tracks and vias
Wellā¦ after cleaning out my first errant attempt stabs at trace making and doing a little experimenting, I have become much better at this. Iāve decided to move some of the IC around so that the traces do not cross so much. I need to experiment with both sides of the board more and try the Vias. I found that if I use paper and pencil I can layout just a few ICās (for example the chip select logic) much easier than on the computer. This reduces the clutter and I can see what is going on much better. Then once I have an idea of what it could be, then the traces are added to the PCB. So far what I have done is to place the interconnection of a few IC on one side of the board and then the group interconnections on the other. Any outliers, I plan on using both sides and the vias. This is my first attempt at some kind of logical layout, Iām sure there must be a method to this madness. Thanks Mike
This ability to place components is very important for creating effective board layouts. Iād rate it as even more important than the ability to efficiently route traces. As you develop and master this ability you will come to appreciate that āeffectiveā isnāt defined in the same way for all boards. Sometimes itās related to spatial efficiency - getting all the components into the smallest possible acreage. Sometimes itās related to exploiting a specified outline - perhaps a very irregular outline - usually constrained by a chosen enclosure, or the placement of other components in a system. āEffectivenessā could be measured by how well your board interfaces (mechanically) with connectors, controls, front panels, etc. Sometimes āeffectivenessā is measured by thermal performance - how much heat is produced at what locations - and sometimes effectiveness is rated by EMI measurements.
At any rate, youāll get better at placing components as you do it more, and as you work under different constraints imposed by different designs. Be patient; Rome wasnāt built in a day.
Dale
3 Likes
Thanks, Dale. I have some patients, but as an old man my time is running out. This is the 1st board I have attempted. In the past I have made many wire wrap boards. These also require some layout experience, but not as much as the PCB. Iām not a electronics guy, I was a steam power plant guy during my working days. Most of my electronics has been self taught. I built my 8080 CP/M 2.2 computer in the late 1970ās, using wire wrap. So this is a step up for me, Iām getting a little closer to the 21st century. Thanks for the help. Iām sure there will be more questions. Mike
1 Like
As another member of the Medicare set, I know that feeling! Thatās why I havenāt done more than a cursory introduction to 3-D modeling.
You can be proud of being self-taught. One of the most respected analog designers of our lifetime, Jim Williams ( and HERE ), was an autodidact. If I recall correctly, he started but never completed a degree quite unrelated to electronics - literature, or sociology, or something like that.
What would motivate a steam power-plant guy to build a CP/M computer? I have the highest admiration for ANYBODY who could complete that project!
Keep asking questions! Somebody will respond. You seem to be very teachable. And you put some effort into finding an answer before you post a question here - āInitiativeā like that is highly regarded around here.
Dale
2 Likes
Dale, thanks for the reply. Wellā¦ back in the 1960ās when I toke up the challenge of learning how to burn coal, most of the control circuits used at the power plant were relay logic. Once you know the process, logic is logic, whether it is relays, vacuum tubes, transistors or what ever. Then in the early 1970ās I got involved with some new power pant computerized control, using a DEC PDP8E. I was impressed with that the DEC could do. So in between battling air heater fires and electric motor failures, I enjoyed reading about computerized control. Later the micro-processor became popular. The plant purchased a Motorola D2 kit, using a 6800. We built a test jig for coal feeders. This helped trouble shoot the many feeders that the plant had. I got the bug and wanted to have my own computer. Purchasing a machine was too expensive. So I started building my own 8080. I spent maybe 10 years working the machine up from 1k memory and machine language programs to a CP/M 2.2, with 2 8" disk drive, 64k memory machine. Then my family started to grow and then I had to invest my time there. The 8080 was put away. I retired about 20 years ago and as my physical strength started to wane, I remembered the 8080 and dragged it out, cleaned it up and got it to work again. Iām still interested in control theory. I picked up a FlexoWriter and connected it up to the 8080. The FlexoWrtier needed some special I/O, which I made using wire wrap, the only method of circuit boards know how to make. I used Orcad to make schematics for my boards, but the very old software would not work with my newer Windows computer. A younger friend suggested looking into KiCad. I started using KiCad about 2 years ago and it seems to be working. My goal is to make this new I/O board a PCB. So, if it works great and if not, it is fun making the attempt and keeps me busy. Iām sure Iāll have more questions. Thanks Mike.
4 Likes
Hi Mike, also useful if you are placing the ICs, enable the ratsnest. This will give you an idea of where the higher density traces are and the ideal placement of ICs to ensure easiest routing afterwards. Also useful, to me at least, are the shortcut keys to select a part of a track (U) and entire track (I). An excellent resource on YouTube is Chris Gammellās Contextual Electronics videos here: https://www.youtube.com/user/contextualelectronic/featured.
1 Like
Iām learning to use KiCad. I have never used āUā and āIā. Selecting is an introduction to make something else.
What is (what are) that else what you do after selecting part of track or track?
Piotr, if you click on part of a track an press āUā it will select the single track up to a node or where the track splits. This is useful to select all the segments including the small <1mm pieces that sometimes form part of a track. If you want to delete the entire net then āIā works well. I will try to add pictures to show the difference.
Single segment selected
Pressed āUā to select single track
Pressed āIā to select all connected tracks
The aim of the above selections is to make it easy to delete an entire trace if you want to try different routing which the original poster tried to do
I understand that you are using U to selectively delete part of track.
I donāt have yet a lot of complicated tracks (mainly one to one connection). I supposed that Backspace deletes one segment and Del deletes whole track. I have just checked at one my track connecting in serie a many foortprints. Del deleted one such connection only. When I used U kay the same was selected as Del (without previous selection) deletes.
Later added:
As Del donāt deletes whole track (as I supposed) that means I can be usefull.