Crystal + caps under the socket?

[MrDEB]

Etching a board today for an LED matrix project and need more PCboard landscape so I am putting the crystal and associated caps under the 40 pin socket and hopefully no issues.
Wonder if this has been tried. Looks like it will fit with clearance to spare?

 

[Sceadwian]

Bad idea. Clearance isn't the issue it's how the 40 pin chip being that close will bleed parasitic signals into the oscillator circuit, not to mention localized heating

 

[throbscottle]

And /that/ is why I love this forum. I thought it was probably okay, now I know better

 

[ericgibbs]

Etching a board today for an LED matrix project and need more PCboard landscape so I am putting the crystal and associated caps under the 40 pin socket and hopefully no issues.
Wonder if this has been tried. Looks like it will fit with clearance to spare?

hi DEB,
If your 40 pin IC sockets are the open frame type ;ie have free space between the IC and PCB, mounting the crystal under the PIC is common practice and should not cause any problems.

E.

 

[MrDEB]

YES it is an open frame socket. The board is to be etched today after I remove the silk screen layer from the printed on copper image. I just plain forgot to disable the silk screen layer.

 

[ericgibbs]

YES it is an open frame socket. The board is to be etched today after I remove the silk screen layer from the printed on copper image. I just plain forgot to disable the silk screen layer.

hi,
You can/could also buy a moulded part which has crystal and caps , together with an inbuilt Lithium battery, which plugs into your IC socket and the IC then plugs into that socket, piggy back.

E.

 

[Sceadwian]

eric, how can that be common practice? Having a crystal+caps in such a enclosed environment so close to components that can cause cross talk can't be good.

 

[Mr RB]

Scaedwian, this is done often in commercial appliances and generally would not cause a problem.

The silicon chip in the PIC is maybe 2mm square, right in the centre, so a xtal inside the socket is probably "remote" to the silicon chip anyway (5mm? 10mm?), similar to a tiny QFN package that might have the xtal 5 or 10mm away. And even if it right under the silicon chip xtals etc are usually in a metal can and operate at a micropower level so it's unlikely the field would affect the silicon.

Having said that I'm agreeing with you on it being a good design rule to get the xtal away from the silicon a bit, even if 10mm away I would "feel" happier with a design.

 

[MrDEB]

I used the wrong cap pattern for the 22pf caps I thought I had (could use the 47pf?). This is just a test board for an LED matrix so going with internal oscillator seeing how no timing issues are to bother with. Just hope counting pulses from a transmitter won't be any trouble. Only need to count say 10hz and 40hz.in a one second interval.
link to the transmitter section. Was advised it may not be reliable but then maybe it will be???

https://www.electro-tech-online.com/threads/wireless-communication.125363/

 

[Sceadwian]

I stand corrected on the practicality, it just bugs me though it is efficient space engineering so if the practical folks say it's okay I'll get behind it =) Even with many crystal applications timing isn't as critical as isolating the crystal that much generally.

 

[MrDEB]

I almost went with the internal oscillator but have had issues with that and just figured always go with external but that may change with this project.
Looking through the latest Jameco catalog and see a SMD oscillator part # 1592263. Am assuming its different than a crystal. The data sheet for the 18F4520 mentions an oscilliator along with crystals etc. This project will evolve into a SMD w/ a 5x7 LED matrix w/ 10mm dots.
NEED to compact the board as small as feasible.

 

[Sceadwian]

If it needs to be as small as possible then use the Internal oscillator. What problems have you had with the internal oscillator before? Unless you need precise timing they're generally not needed, they can usually be synchronized to other chips without too much difficulty.

 

[MrDEB]

I have used the internal oscillator numerous times but my last project went arway when I attempted to use it. WHY??
That project is on the back burner for now .
Have present project almost ready to insert PIC and the ULN2803's. Spent an hour tracking down a short on the board. Think I corrected it. Onward.

 

[MrDEB]

pictures of crystal inside socket frame

I powered up the PIC and flashed some LEDs all is a go. The crystal appears to not be affecting the PIC.
The 30G wires are to connect top layer to bottom layer. Have yet to trim.

 

[3v0]

I have used the internal oscillator numerous times but my last project went arway when I attempted to use it. WHY??

Should not have. Without more info nobody can tell. Expect it was user error

Good work on the PCB.

 

[Mr RB]

Yeah nice PCB, a good result for a double sided home toner+etch job! It's great to see someone making projects and not just students asking homework questions.

So what does the "LED display matrix" project do?

 

[MrDEB]

This is just a test board to test the code I am working on. The matrix is a bicycle turn signal that will be controlled wireless. The board pictured has a 5 pin header other than the ICSP header for imputing the wireless signal.
Pretty simple.

 

[Pigskin Tony]

Something like this? ( wrong link on last post - moderator please deflate)

https://www.kickstarter.com/projects/1058059772/bike-luminance-lighting-and-turn-signals-that-will

 

[MrDEB]

similar but not the same. What he is doing in the video could be done with just a 7555 timer. VERY easy to do.

 

[Pigskin Tony]

Just a 7555 timer? Wow, that guy seems to think more parts are needed. How's a 7555 going to respond to head tilt and switch the three groups of LEDs.

You're way smarter than me on this stuff!