I’m trying to troubleshoot an issue where KiCAD merges two separate global labels into a single net (in the netlist). The schematic is depicted below. I’m using Ver 5.0.
Here’s the relevant snippet of the .net file and relevant schematic. I searched all schematics and couldn’t find where the two labels become connected directly or indirectly.
Thanks for your help. NETLIST
(net (code 270) (name AFE_OUT_A)
(node (ref U11) (pin 3))
(node (ref U11) (pin 16))
(node (ref U11) (pin 15))
(node (ref U11) (pin 11))
(node (ref U11) (pin 5))
(node (ref U8) (pin 10)) — I expected this to be a separate net e.g., V_TC_IN_A not AFE_OUT_A
(node (ref D5) (pin 3))
(node (ref L3) (pin 2)))
It seems that KiCAD has the two nets connected as listed in the ERC report. But I do not see where in the schematic. I attached the schematic that is being referenced in the ERC report.
ERC report
ErrType(4): Conflict problem between pins. Severity: warning
@ (64.77 mm,114.94 mm): Pin 11 (Unspecified) of component U11 is connected to
@ (64.77 mm,120.02 mm): pin 15 (Unspecified) of component U11 (net 270).
ErrType(4): Conflict problem between pins. Severity: warning
@ (115.57 mm,122.55 mm): Pin 16 (Unspecified) of component U11 is connected to
@ (115.57 mm,104.77 mm): pin 3 (Unspecified) of component U11 (net 270).
ErrType(4): Conflict problem between pins. Severity: warning
@ (115.57 mm,104.77 mm): Pin 3 (Unspecified) of component U11 is connected to
@ (115.57 mm,107.31 mm): pin 5 (Unspecified) of component U11 (net 270).
Check connections of other labels that are directly connected to V_TC_IN_A and AFE_OUT_A.
For example if SWL_S7 and SW1_D1 are connected somewhere it will result in all of the mentioned nets being merged.
I did search for all SW1_S* and SW1_D* earlier and there is only two instances of each label in the schematic; one at U11 (top diagram) and the other is where they are being used at U8 (bottom diagram).
I also tried to inspect indirect connections, but couldn’t find yet.
Could this connection be related a bad symbol of U8 or U11? I reviewed them but I can post the text here.
I don’t know how a bad symbol can cause this.
My remaining suggestions are:
Upgrade to 5.1. It’s fully compatible with 5.0 and has fewer bugs. I don’t remember if there were any netlist bugs in 5.0 but who knows, you will lose nothing by upgrading.
Use highlight net tool and try to spot if net is highlighted where it shouldn’t be
But I’m not sure how DAC_OUT_A is connected to V_TC_IN_A in the figure above (SW1_S1 and SW1_D1 are two different pins of an open switch on U11, a source/S and a destination/D pins).
It seems you have made the schematic too complicated for anyone to understand, including yourself.
As a rule, stick to one net label per net. If you need to explain how pins are connected, that should be obvious from the connections, if not, add a text table to describe how pins are used.
I followed the advice of Qu1ck to use the net highlight tool. Here’s the result (all the relevant nets are highlighted in purple by the tool). This is matching the generated netlist exactly, but I’m still not sure why KiCAD would tie V_TC_IN_A to DAC_OUT_A?
Is it perhaps a spurious warning and the issue is that the pins are of type Unspecified? Some symbols you download come like that. First thing I do is assign them the right type in the symbol editor. Using Unspecified defeats some checks so it’s better to be more precise in the symbol data.
But V_TC_IN_A should not be the same name as AFE_OUT_A, according to my understanding of the schematic above. But KiCAD netlist is suggesting they are. I wish they are same, because unfortunately we have boards assembled based on this assumption and I need to fix them if necessary.
Yes, I’m using hierarchical sheets (U8 and U11 are on different sheets - I just placed them in the same image since I could submit only one image to the forum). I’ve been using hierarchical labels at many other places, but only when there are few of them to avoid cluttering the main sheet.