Beginner SMD design

[Hughieandrolf]

Hi.
I'm a first-timer here. please be kind...
I'm designing a PCB for a home built clock project around a 44 pin TQF PIC16F887 using Eagle PCB design software.
I'm hoping to self etch prototype PCBs, (I've done this a few times using through-hole components)  and if all goes well, have 20 or so PCBs professionally made.
I want to keep the PCB as small as possible and I have a few questions:
 
Is it a good idea to route tracks via the pads of  N/C pins of components to gain access to other pads from the inside of the chip footprint? 
Can tracks attach to any edge or corner of a pad or should it always be 'axial' to the component?
Is it good/bad practice to route tracks between pads and under components as I would do with say a axial lead resistor or between rows of DIP pins?
Is it OK to lay components at 45 degrees to others?

Cheers.

[crispy_tofu]

Welcome to the forum!

Is it a good idea to route tracks via the pads of  N/C pins of components to gain access to other pads from the inside of the chip footprint? 

For this one, it's generally not a good idea because the N/C pins could be grounded, diagnostics, etc.
As this stackexchange question states:
(http://electronics.stackexchange.com/questions/8496/can-i-run-traces-through-no-connection-pins)

It's a grey area. Some manufacturers tell you that's used for calibrating. Some manufacturers even will tell that you that certain pin has a function, only used by them for calibrating. Some tell you only not to connect it, or just say it's an unconnected pin.. You can't know for sure. The datasheet is information the manufacturer want to tell to you about using the device, but it might not be everything.

I recommended you do not connect them. If you get some generic IC from a different manufacturer or even batch the behaviour might be different. If you're engineering a project, you don't want to throw in unpredictability. You would have to test every single batch before you're going to use that particular batch. It depends on whatever you want to do that.

Is it good/bad practice to route tracks between pads and under components as I would do with say a axial lead resistor or between rows of DIP pins?

That should be fine.

This video might be able to help you:


Sorry, I can't answer the other questions, hopefully a more experienced member can inform you. 

[Simon]

no solder resist on a TQF ? good luck, what are the pin pitches ?

[crispy_tofu]

no solder resist on a TQF ? good luck, what are the pin pitches ?

Looks like 0.65mm between centres... eek

[Simon]

I just get cheap prototype PCB's made these days, no point in prototyping something that will be made in a certain way in production in a way that is near impossible to prove the prototype.

[tggzzz]

I'm a first-timer here. please be kind...
I'm designing a PCB for a home built clock project around a 44 pin TQF PIC16F887 using Eagle PCB design software.
I'm hoping to self etch prototype PCBs, (I've done this a few times using through-hole components)  and if all goes well, have 20 or so PCBs professionally made.

Welcome, and don't be afraid to make mistakes - but try to ensure they are new mistakes!

I've recently returned to fabbing PCBs after a 20 year hiatus - so I've been in your position recently. My experiences are at https://entertaininghacks.wordpress.com/category/homebrew-pcbs/

Toner-transfer is just about good enough for 0.1" PTH double sided boards. I wouldn't try it for fine-pitch SMD components. You will spend a long time debuggin your manufacturing technique.

If you are making 20, then you might as well use the same implementation for the prototype - otherwise you will do debugging twice.

Cheap'n'cheerful PCBs (double sided, solder mask, silk screen, 0.5mm pitch SMD components) can cost £10 for 10 delivered in 10 days; see the above reference for the source I used.

[exmadscientist]

Is it a good idea to route tracks via the pads of  N/C pins of components to gain access to other pads from the inside of the chip footprint?

Completely depends on the component. Most components just say "NC", "no connect", and I leave those alone. Some say "DNC", "do not connect", and you have to leave those alone. Some say "NIC", "no internal connection", and you can do whatever you want with them (within reason, of course; no component not rated for it is going to like 1000V appearing on a random pin). And some components flat out don't tell you the whole story in their datasheets. For example, the AD9257 has two "DNC" pins, which should mean they have to be left alone. But ADI engineers say it's OK if they're grounded, as they were on the earlier, otherwise-pin-compatible AD9222. So sometimes you can ground NC pins, sometimes you can't.

I just leave the damned things alone, and it's never served me wrong.

Can tracks attach to any edge or corner of a pad or should it always be 'axial' to the component?

Traces coming out straight from the center of a pad are best, but there's no hard and fast rules. SMT is pretty forgiving when it comes to this, at least at larger pad pitches. The smaller things get, the more careful you must be. I have had no trouble with all sorts of crazy trace routing on 0603 component pads.

Is it good/bad practice to route tracks between pads and under components as I would do with say a axial lead resistor or between rows of DIP pins?

This is OK (in general) if and only if you have solder mask. With solder mask, you can run a trace between the pads of an 0603 resistor with no trouble. Without solder mask, I think trying to do that would likely end in frustration. As others have said, this is a very good incentive to get professionally fabricated prototype PCBs. It's so cheap these days that there's little reason to bother doing it yourself.

Is it OK to lay components at 45 degrees to others?

Sure. Modern pick and place machines have no problem with that (as long as the usual mechanical clearances are maintained, of course).

[KL27x]

no solder resist on a TQF ? good luck, what are the pin pitches ?

No luck needed. 0.65mm pitch is no problem. I have done this aplenty.

None of my production flex boards have soldermask between the pins. I've soldered hundreds of 0.65mm pitch IC's on them by hand. And that was just today.

Done hundreds of toner transfer boards with no mask, same pitch, too. I've also proto'd a few 0.5mm pitch pcb's with toner transfer. No problem, at all.

Soldermask is important to cover copper that isn't supposed to get tinned. It isn't particularly important to have where there's no copper.

Toner transfer can get down to maybe 10 mils with all-day-reliability, if you do it right. Smaller in a pinch, if you care to do more examination/testing and throw away some boards. Down to 6 mils is maybe the practical limit where unevenness in the etching starts to become nearly as wide as your trace - and at this level, you should probably be using 1/2 oz copper, and your etching technique better be good.

OP, no problem routing traces thru pins you KNOW are unused and high impedance. If you're not using all your pins, for instance, you might intentionally route a special function pin through the pad for an unused pin and just make sure that pin is left in the digital input state. The only traces I would be wary of doing this are ICSP traces, if you're using them. (Obviously can't do this with the high voltage programming line, but I would avoid doing it with any of them; weird stuff can happen to the unused pin during programming).

No problem routing traces under SMD parts. I have routed traces between 0603 resistors (custom footprint to widen the gap) and thru SOT 23 pads (standard footprint), using toner transfer with no problem. The disclaimer is I have been doing toner transfer and etching with a custom CuCl2 tank for quite a while.

[Christopher]

no solder resist on a TQF ? good luck, what are the pin pitches ?

we have an lpkf milling machine, qfp and even qfn is possible with lots of strong flux

but yeah i much rather get the boards sentout

[KL27x]

Milling leaves a rough surface that might be stickier to solder. I have found essentially no difference in soldering TQFP on regular etched board with or without mask. You always need enough flux.