Eagle ground plane issue

[electrosonic]

I accidentally sent OSH Park a .brd file that had isolate=0 (the default value for some reason) for a ground plane. The board passed the DRC test, but my power rails are shorted to ground somewhere (testing continuity on an unpopulated board).

Anyone seen this before?

Andrew.

[pappasmurfsharem]

I accidentally sent OSH Park a .brd file that had isolate=0 (the default value for some reason) for a ground plane. The board passed the DRC test, but my power rails are shorted to ground somewhere (testing continuity on an unpopulated board).

Anyone seen this before?

Andrew.

Did you use OSHs design rules they have the file on their site

[electrosonic]

Yeah I used Laen's DRC file. I have asked the same thing on the OSH support page, so I'll see what response I get. The boards look OK, there appears to be space between the traces and the ground plane, smaller than usual I guess.

I think I am working through every possible mistake in learning EAGLE.

Andrew.

[pappasmurfsharem]

Yeah I used Laen's DRC file. I have asked the same thing on the OSH support page, so I'll see what response I get. The boards look OK, there appears to be space between the traces and the ground plane, smaller than usual I guess.

I think I am working through every possible mistake in learning EAGLE.

Andrew.

If you want to send my your BRD file I can take a look, no promises though.

[electrosonic]

Thanks for the offer, they are small boards I think I will just correct the issue (and a few other minor things) and move on to the next revision. Just curious as I had never seen this before.

Since it takes about 3 weeks for me to get boards from OSH Park I usually have 3 or 4 projects on the go between various revisions and parts orders.

I will shelve this one (Attiny85 VCA using a vactrol) for a few weeks and move on to my LM3046 fuzz face / super fuzz hybrid that I have been working on.

Andrew,

[pappasmurfsharem]

Thanks for the offer, they are small boards I think I will just correct the issue (and a few other minor things) and move on to the next revision. Just curious as I had never seen this before.

Since it takes about 3 weeks for me to get boards from OSH Park I usually have 3 or 4 projects on the go between various revisions and parts orders.

I will shelve this one (Attiny85 VCA using a vactrol) for a few weeks and move on to my LM3046 fuzz face / super fuzz hybrid that I have been working on.

Andrew,

Are all 3 boards having issues? I have only made a few orders from OSH but they all seem to work just fine.

It's possible a copper pour went awry.

Did you use a V+ pour as well? If not maybe you can just exacto and cut along your V+ lines and pads. It'll be fun!.... not really.

[electrosonic]

I semi populated one and fired it up. Smelt the regulator getting hot. Checked the regulator pinout, removed the power supply caps, removed the regulator - still a short.

Decided to check the other two unpopulated boards, both measure a short between the supply rails and ground.

Rechecked my Eagle files and realized I had done the isolation=0 for the ground pour. There is visible isolation between the pads and the ground, but maybe I am below the tolerances of the fabrication.

I am surprised that Laen's DRC would pass a board it couldn't make. Or maybe I have have overlooked something else. I have probably used OSH Park 20 times and never had any issues that weren't my own fault.

Andrew.

[pappasmurfsharem]

I semi populated one and fired it up. Smelt the regulator getting hot. Checked the regulator pinout, removed the power supply caps, removed the regulator - still a short.

Decided to check the other two unpopulated boards, both measure a short between the supply rails and ground.

Rechecked my Eagle files and realized I had done the isolation=0 for the ground pour. There is visible isolation between the pads and the ground, but maybe I am below the tolerances of the fabrication.

I am surprised that Laen's DRC would pass a board it couldn't make. Or maybe I have have overlooked something else. I have probably used OSH Park 20 times and never had any issues that weren't my own fault.

Andrew.

Looks like Laen's DR has an isolation of 1.5011mm and a Thermal of 10mil

Somehow your isolation must have changed.

Either way that sucks, to wait so long and be disappointed. It's especially bad for me with my problem with patience.

[electrosonic]

I heard back from OSH Park, the isolation value defaults to 6 mil if you try to go lower than that.

I used my kids $8 plastic microscope from Lee Valley and my iPhone and managed to get a picture of the problem area.



Something weird is going on in the ground plane and you can see the pad (V+) is attached to ground by a thin trace.

Andrew.

[electrosonic]

I used an exacto knife to cut that small trace - it must be 4 or 5 mils thick and I couldn't see it without magnification.  Problem solved, though I am not sure what I did wrong. It looks like I put writing in the wrong layer somehow.

Andrew.