In general it is not recommended to put windings for transformers in parallel. If there is an imbalance in the windings, there can be relatively high fault current. At least, that is what somebody told me when I was still young, somewhere in the '80-ies.
Maybe production quality of the “average” transformer has improved to the point that this is unlikely, but still, if you have an easy choice and can use another model transformer and put the windings in series, there is no possibility anymore for this fault.
Hi Peter,
Sorry - after i posted i did see that the thread ran on and on and that you had provided much information and in fact made many changes.
The quick exchange below though, highlights what I think is likely ht error of your confusion: KiCAD’s work mode is not “place pad and lines”. Rather it is “create a legitimate schematic that may be rules tested, and then go into assisted layout mode for that schematic. It will help you find a solid result snad prevent ones that break the rules (opens, shorts…)
So that AC input (what I in fact guessed) must be a component with a footprint and pads. Find the riser/header footprint that matches your hardware and place it. Now you would’ve tow legitimate pads, and KiCAD would not only allow you to connect, but it would draw airlines telling you to, and it would;d flag it in the error check if unconnected. Its more work up from but believe me, it pays for itself many times over. I hope this helps.
G
Thanks for your comments.
The trafo is 1.5VA per winding and the manufacturer lists parallel as OK, but I understand your caution.
Yes the help I have had here has made me realise that KiCad starts with a ‘functioning’ Schematic and then works towards a pcb design.
The AC ±15V pads are to connect to another pcb. They could be flying lead connections but to save space I am piggy backing that board on top of this design. Perhaps I am bending the rules but I am not concerned if KiCad complains about them being unconnected. 
in theory maybe. In practice, no. And many highly respected manufacturers list that as a suggested option. BEL/Signal; Thorston; Torroid of MD, Avel-Linberg - all of them. And i do it all the time.
If you are designing for a volume run that makes a custom wound transformer economically viable - by all means go for it. For the rest of us in “off the shelf” land, its the best way to achieve the current and voltage specs we need.
When i parallel windings i will note that i never have noise or heat rise. And without one or the other there’s likely no fault / parasitic current to speak of.
Thanks both for the technical input, I can always learn!
My little project is definitely a one off, a ±15V PSU for a Record Preamp (RIAA). The configuration has been running in a breadboard form (Veroboard) for over a year so I am happy to take the risk.
Next step is to get some prices for my little project, hope I can find a UK supplier.
Hi dabbler,
The problem is the schematic is not matching with the footprint. If you want to retain the ground plane, you need to add a zone there. You can draw a polygon by clicking on the board.
When adding polygon for a particular net, make sure that polygon is connected to the net wherever the pin is. In your case, the pin is at the first connector. Use add zones option to add a copper plane on the ground. Select the net and draw the polygon on your board. 31.496 is the clearance that needs to be followed.
Maintain proper distance from other nets as per the specification for AC (whatever clearance you require from any other net). Use it when it is absolutely necessary.
Thanks for the information.
I decided in the end not to have ‘unused’ copper surface.
Now waiting for my chosen supplier to deliver!
This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.