Hi Lee, we use a pic18F26K22 and a PIC18F67K22 using RS232 comms (admitedly only at 9,600 and 19,200 baud) and also I2c comms to an NXP RTC using the internal RC oscillator.=20 We've had no issues in 3million units, it's worth looking at before you dismiss it entirely. as alway, YMMV! -Jim -----Original Message----- From: piclist-bounces@mit.edu [mailto:piclist-bounces@mit.edu] On Behalf Of lee@baudalign.com Sent: 07 February 2014 22:43 To: piclist@mit.edu Subject: Re: [PIC] crystal physical size? (a few replies rolled into one) > From: Jim > Hi Lee, > Have you considered using the internal oscillator to save PCB space,=20 > is this an option for you? > -Jim Thanks Jim, yeah I looked into it a while ago but the consensus seemed to b= e that it wasn't reliable enough for the likes of RS232 comms and I2C etc. Also, 8MHz is a bit slow for me, I think I'm going to need all the cycles 20Mz can give. Space isn't a HUGE deal, I've got room for a crystal of some sort, but was investigating whether there were smaller options to be had for cheap. > From: David VanHorn > Not critical. The parts that I see as critical are getting the=20 > loading caps right, and returning their ground ends to the PICs ground=20 > pin and nowhere else, not dumping them into a plane, and definitely=20 > not putting them into a track that eventually gets back to the PIC ground pin. Thanks David, so I don't need to rearrange my placement too much. Not entirely too sure what you mean by the last comment? After the ring I connect to a thick ground trace (not plane) that connects straight back to the battery (about 25mm), but this trace has branches and vias that connect to the other PIC grounds, and also two buck/boost converter grounds, along the way? Just not sure whether you mean I should be having a dedicated trace all the way back to the battery negative, where it meets with the other ground connections anyway, not far from where it would have originall= y met them? > From: Spehro Pefhany > Load caps should be correct for the xtal. They're calculated the same=20 > way as for any crystal. > From: David VanHorn > Hopefullly that would be (Cl * 2) - Cp where Cl is the specified load ca= p > value, and Cp is the measured or estimated parasitic capacitance? Ex fo= r > a 22pF crystal: (22 * 2) - 6 =3D 38pF so use 36 or 39pF > From: Spehro Pefhany > Hopefulllly. I usually use 5pF for the stray, where the frequency=20 > doesn't matter enough to worry about (for a crystal). That might be=20 > high for a small SMT crystal, or depending on the ground plane and=20 > layer stackup, maybe not. Yeah I'm not, um, sure where I got my current capacitance figure from...=20 I think I copied it from a similar design without any clue whatsoever about the crystal's requirements... well, at least I know what to look for now! = ; ) > From: Martin Klingensmith > These are a couple years old and I've used them in several designs. > There very well could be something smaller now: > Abracon ABS07 for 32768 Hz (3.2 x 1.5 mm) NDK NX2520 for 16-48 MHz=20 > (2.5 x 2.0 mm) A problem is that none of the crystals I've found ("cheap" small orders, like $8 for 20pieces) have given any data whatsoever as to the model number= , and definitely no Cl figures. Thanks for those numbers Martin, gives me some specific models to look for, which also _may_ give some clues as to what kind of Cl the unnamed crystals may have... (also might be able to spy numbers on their casing pics) Thanks all Lee -- http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive View/chang= e your membership options at http://mailman.mit.edu/mailman/listinfo/piclist --=20 http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive View/change your membership options at http://mailman.mit.edu/mailman/listinfo/piclist .