In the PCB layout, I have global Solder mask min width as 0.2mm and Solder mask clearance as 0.2mm as this is the tolerance of the board manufacturer.
When I don’t add any tracks to these pads
I get DRC errors PAD near pad error on the pads (left column vs right coloumn)- where does the pad to pad clearance specified? is it derived from solder mask clearance?
In this case Pad 1 (Col1 first row) and Pad 6 (Col 2 first row) have a space of .23mm
I assumed that the error is in reference to solder mask being too close to the adjacent pad as it is 0.2mm wide and this leaves only .03mm. In real world scenario both the pads solder mask merge in the gap.
Should I worry about this error?
I changed the Net Pad clearance for each of these individual pads under Local clearance and settings to 0.03mm - with this DRC passed.
@krishna_kumar_c Can you attach the footprint in question here (kicad_mod file) so we can examine it?
Also that screenshot you got there… can you switch pads and zones to outline mode for another screenshot please?
And Silkscreen and Fab layers OFF please. It’s hard to see what the problem is when you got the important bit’s not visible/too much unneeded information shown.
I notice just 2 arrows there ? If this was a footprint space issue, you should have more ?
Maybe some NETs have special rules, eg Check the GND rules.
Of course, PcbNew could do with smarter reporting around errors, it should state what the actual failing distance is, and which exact rule it violated.
I have figured out the issue, the issue is with the track clearance which results in pad near pad error. I managed to change the track width on pad 5 to 0.21mm earlier it was 0.26mm with this I got rid of error between pad 2 and 5
The track on Pad 6 and Pad 3 belongs to RFDummy net which starts thinly at 0.21mm and becomes .366mm but I guess the clearance for the thin part is still retained from the appropriate net list.
I have tried moving the thicker track away from the footprint and connecting with a thinner trace (albeit with the same clearance) - this doesn’t get rid of the errors.
Is there a way to reduce clearance for the thinner part of this net.
The mod file doesn’t get uploaded it keeps on showing uploading 100% but nothing happens, waited for a pretty long time.
** Found 3 DRC errors **
ErrType(19): Pad near pad
@ (147.140 mm,111.000 mm): Pad 1 on F.Cu, Non-copper of U1
@ (148.070 mm,111.000 mm): Pad 6 on F.Cu, Non-copper of U1
ErrType(19): Pad near pad
@ (147.140 mm,111.500 mm): Pad 2 on F.Cu, Non-copper of U1
@ (148.070 mm,111.500 mm): Pad 5 on F.Cu, Non-copper of U1
ErrType(19): Pad near pad
@ (147.140 mm,112.000 mm): Pad 3 on F.Cu, Non-copper of U1
@ (148.070 mm,112.000 mm): Pad 4 on F.Cu, Non-copper of U1
I think this is a newbie error mainly attributed to the use of OPENGL canvas which doesn’t show the clearance of the tracks. On top of this the PAD near PAD error code confused me to go and fix things incorrectly.
What I don’t understand is that you set the track clearance to 0.21, when your fab house is able to do 0.2?
Loaded your footprint… Why did you put it off-center?
You also set local pad clearances… I hope you got rid of them now and those were just for ‘bughunting’.
One also doesn’t usually use text to mark pin 1 on silkscreen
That’s how I would make it (KiCAD standard would put REF field onto silk as you had it, but I don’t have them there):
Your placement outline is also pretty faint… sure the fab can do 0.05 mm?
Reference and value field are being populated by the schematic… the footprint should have them as **REF and **VAL.
The device itself is actual really really tiny. You’re going to reflow that? And you’re a beginner? Hat’s off to you sir and good luck
PS: I can’t upload other files than images either… just tested. The big boys of the forums got a message from me… we should sort this in due course I guess.
Thanks for the tips, the reflow will be done by the fabhouse itself, they offer PCBA as well - I needed to test a PA and LNA and the evaluation boards were not available hence just trying my hand on making them - either way even if I do few iterations they will be cheaper than buying the EVB’s and the added gain is PCB design knowledge.
> The fabhouse states
> Minimum PCB track 6mil (Recommended >8mil)
> Minimum Track/Vias Space 6mil (Recommended >8mil)
> Minimum pads Space 8mil
`Minimum silkscreen text size is 32mil Regarding silk screen thickness - I have to correct it
I didn’t want to test their limits on track width and clearance hence used my own higher limits to warn me.
I notice that you have references on the Fabrication layer - what is the signification of this compared to having it on Silk screen layer? Other than visual cleanliness is there any fabrication significance?
Is there a quick way to move all the component references on the populated board to Fabrication layer instead of messing with individual components.
You have two REF** in the footprint one on Fab layer and another on Eco1.user in this footprint - is it just to show me that you can move it to any of these layers or is there more to it?
Reference on silkscreen and Fab layer is there to help assembly/repair. Silkscreen will be placed where it can, but Fab layer reference is usually centered on the component and is more clear. To me, the Fab layer is useless without references.
As @jwpartain1 says, I don’t do references on silkscreen as the devices are so small, there seldom is space for it (or I just don’t want it to be there in the first place).
There are two ways to get values and references from the schematic into the layout… one is via **REF/**VAL and the other is %R and %V.
The former react to the render tab settings and can be used only once per footprint, while the latter are bound to the layer and can appear as often as you want (different layers).
Images following show same area as example…
Now, when I do create documentations I want the Ref+Val to be visible together with the Fab outline layer (and readable = relatively big) but don’t have them distract me while layouting…
(device values aren’t cleaned up yet, thus they’re so content rich as the current library conventions force me to do… for example that C106 would just be ‘100n’ instead of ‘C_100n_50V_X7R_kemet_0805’ if done properly)
During layout I don’t want to see them though (but the fab outline) so I put **REF/**VAL onto the fab layer and switch them OFF in the render tab (out of the way so to speak).
Now… how do I get a ref designation during layout to help me? (that is small + not in the way + sticks out if I look for it) - right, by using %R on a custom purpose layer like Eco1 with a nice color :-).
PS: I only use my own libraries with atomic parts (=each part I use is defined as a symbol and has a footprint pre-attached, I don’t use cvpcb to assign footprints to symbols). That’s why the part names need to be so detail rich, as otherwise the library algorithm fights me…
