This has nothing to do with gerber file generation, these are just regular DRC violations.
My attitude towards thermal reliefs is that it’s nice if KiCad is content with them, but whenever the automated algorithm gets into a pinch, I either change the pad connection for such a pad to “solid”, or I manually draw a track.
There is a screenshot of this PCB in the thread below:
It’s nearly all THT, so it would not matter much whether the GND pour is on the top of the bottom layer. How much room for a GND plane is left after the tracks have been layed is another matter.
So this error I’m getting is because the ground pour is not covering the whole surface? can I ignore that?
I change the copper zone to bottom layer and now I’m getting only two errors.
No, the DRC violation is: Error: Thermal relief connection to zone incomplete. And that’s a pretty accurate description. I also already gave you some possible solutions in my first post. But for the bigger picture, the root cause is that the routing of the tracks on your PCB is quite bad. From the small snippets I can see it looks like just some tracks being drawn between pads without putting much thought into it. Designing a “good” PCB needs much more attention to details, and that is also why projects like Freerouter are barely useful. Such autorouters do not have enough intelligence to route the tracks properly. The easy way out is to make it a 4 layer design, and that is a tradeoff between cost of the PCB and design time.
Is it not possible to route it manually on two layers only?
What could be the downside of using freerouting plug for this board? Can it reallycause it not to work?
It’s a relatively simple PCB, so sure it’s possible to do this properly on a 2 layer PCB. But each PCB is always a mix of goals. You seem to prioritize aesthetics (nice row of resistors) and compactness. That are two goals that make it much more difficult to do the routing properly. EMC compliance is probably not high on your list. It also looks like you made the classic beginners mistake of forgetting to reserve PCB area for the actual routing. And from there you glide into a grey area of what you consider a “working” pcb.
So do you recommend to make the board slightky bigger in order for being able to manually route?
At the end of the day im coming from the diy world of synthesizers made on stripboard so i can guess moving to pcb, even if route by a plugin, is much better the stripboard and wire routing(?)
You could either re-place all the components to make it easier to route, or rip up all the autoplaced traces and see if you can do a better job manually.
It’s always a mix of design goals, and compromises. I definitely recommend to ditch the auto router and route the PCB manually. (Autorouters can be a useful tool, but they are definitely not a magic tool that “simply does all the wiring” for you.) Footprint placement is a very important part of the PCB design, and learning what works (and what does not work) is something you only learn by experience.
When I design a PCB, I usually start without expectations of PCB size, and do the part placement and the routing for small sections of footprints that connect to each other. And from there I go to bigger sections, while also keeping an eye on secondary goals such as connector placement (acessibility) and switches / buttons (for example for a front panel) that have specific location requirements. But it’s an interactive process. On a PCB like this, it is not a good way to first fully do the footprint placement, and only then start routing the tracks.
Reserve some time to do this. It probably takes much more time then you would expect (My guess is 2 days of more just to route this PCB) It’s not just the routing, but from research to what are important factors for designing a “proper” PCB, to trying out things, then ripping them apart and doing things in a different way. It’s not the design process that is important here. It’s the experience in routing PCB’s that is the most valuable you will get out of it.