Ok, this may be a weird design, still I want to try to make it and in a way that I can send the design to batchpcb (not to a different pcb house)
I want to have a atmega168 mlf32 placed in between the 2 rows of an Atmega168 dip28. So far, I have not been able to do that with the batchpcb limitations. Everywhere I get DRC errors (mostly clearance and width errors)
I have tried many things to get it (auto) routed, but to no avail. I have tried to ‘mirror’ the smd to the bottom layer, put the mlf32 outside of the legs of the dip, you name it, I have tried it.
I haven’t looked at a layout like that but my gut says that it isn’t going to work. You might try creating your own footprint with smaller pads for the DIP. Have you thought about doing both as surface mount? that should be really easy
I finally managed to get the board 100% routed. Submitted it to the batchpcb bot and it says I have 1 error:
Checked Aperture 10 (0.0059): Failed
Error - Aperture too thin
Checked Aperture 11 (0.0480): Passed
Checked Aperture 12 (0.0070): Passed
Checked Aperture 13 (0.0380): Passed
I found some other posts about this kind of error, one said to check the ‘Sizes’ tab in DRC. Before I started the auto-router I changed ‘Minimum width’ from 8mil to 7mil. When I change it back to 8 mil, DRC comes up with 132 errors on layer 1, type ‘Width’.
Philba:
Surface mount DIP sockets?? Yes, I know how silly that sounds…
Philba,
Not really. I used some recently to mount old technology DIP
packages to a SMT PCB.
Gave me a new component density - the SMT DIPs were on
one side, while the SMT drivers for them were on the
other side. Useful as the DIPs are 1970s technology displays and
had to be placed in a pleasing pattern. I’ll post some PIX - time
permitting.
Comments Welcome!
Actually, I’ve done the same - I make my own boards and really really hate drilling so I migrated to SMT but couldn’t bear tossing a bunch of perfectly good dips so I widened the two rows of pads and bent out the pins. Viola! Surface mount. It’s a bit tricky getting the pins totally coplanar but I used a screw driver to push down the dip until the solder hardens up. Probably wouldn’t work for machined pin sockets.
I realized my cheap ways were a waste of time so I stock SMDs these days. The DIPs are good for prototyping anyway.
jandirks:
I finally managed to get the board 100% routed. Submitted it to the batchpcb bot and it says I have 1 error:
Checked Aperture 10 (0.0059): Failed
Error - Aperture too thin
Checked Aperture 11 (0.0480): Passed
Checked Aperture 12 (0.0070): Passed
Checked Aperture 13 (0.0380): Passed
I found some other posts about this kind of error, one said to check the ‘Sizes’ tab in DRC. Before I started the auto-router I changed ‘Minimum width’ from 8mil to 7mil. When I change it back to 8 mil, DRC comes up with 132 errors on layer 1, type ‘Width’.
Anyone that can help?
JD
get a gerber viewer, open up the gerber. then use a text file editor and open up the gerber file - it’s a bit gibberish at first but you should ble able to figure it out… look for Aperture 10 and find where it’s being used. that will give you the coordinates - in the gerber viewer, you should be able see where it is. From there it should be a matter of fixing…
Thanks for your help so far Philba, I really appreciate it.
I downloaded Viewmate and opened up the Gerber (cmp file). The contents didn’t make much sense to me at first, but after a while I found the aperture 10 error (0.0059).
I changed it to 0.007 and zipped up all files and submitted it. The first check done but batchpcb succeeded, no aperture to thin errors anymore.
After a few minutes I received an email stating that DRC errors (clearance) were found.
When I checked the attached images, I found that they were all on the 4 inside corners of the atmega168 mlf32 SMD pads, see image:
There is nothing I can do about this, right? I am a noob in doing this electronics stuff… I am not too much looking forward into making a new footprint…
CD
Edit: I should add that I have used a clearance of 7 mil, I think I read somewhere that user ‘busonerd’ (?) stated that allthough they promote the batchpcb service with 8mil spacing/tracing, the bot allows for down to 6.9mil. Please correct me if I am wrong
Pfff… I am close to giving up. I made the pads narrower and shorter 3 times now, to no avail. I keep getting DRC errors back from the BOT. The errors are all in the smd pads, at all 4 corners the first 2 pads.
Maybe the limitations of the batchpcb service make it impossible to use an Atmega168 MLF32 in a design?
I will start a new thread to see if anyone has ever managed to get this chip in a design:
