Oscillator NOT

[3v0]

I built a small board with a PIC and the resonator will not oscillate. After cleaning the board, resistance from the oscilator pins to all nearby traces is infinite.

I programed the PIC and tested it on a breadboard with the oscillator prior to installation. The same setup works if I bypass the board and solder the resonator to the PIC pins.

Can nearby traces add enough capacitance to cause the resonator to not work ?

I found this on the Abracon site.

LAYOUT HINTS:
When you design circuit on a PCB, please take the following
considerations:
Please design the trace length as short as possible.
Try to avoid thin line on resonator traces (< 0.010”). Design as
wide as possible.
If noise occurs, please insert EMI filter with serial mode between
XOUT pin and resonator or between capacitor and GND.

The square pads in the image are the resonator. Resonator lines are .016 with a .024 ground between them.

 

[kchriste]

One question would be; is the resonator one with the built in caps? Some need external caps and the breadboard may have just provided enough stray capacitance for it to work.
What is the part number/manufacturer of the resonator?
Which PIC are you using it with? What is your PICs _CONFIG setup?

 

[Nigel Goodwin]

You might also double (or triple) check the rest of the connections, it's suprisingly easy to do some thing silly like not connect MCLR (been there, done that!).

 

[ericgibbs]

hi 3v0,

I would suggest, as the PIC is most likely in a socket on the pcb, is to do a ohms/resistance check on the PIC socket pins ,
[with the PIC out] to the copper track.

Sometimes during assembly of multipin ic sockets, its possible to bend a pin under the socket housing, which is missed
during soldering and the solder just fills the hole.

Thats why its always a good idea to check that the pins are straight before insertion of the socket or ic, into the pcb.

Eric

EDIT:
hi 3v0
Lets know what you eventually find thats causing the problem. Good Luck.
[sounds like a case for Arther C Clarke!]

 

[3v0]

No Joy.

The processor and resonator were tested on a breadboard prior to installation on the PC. That rules out the config values.
As you can see in the attached image I tried bypassing the PCB resonator by soldering one to the PIC pins. With this the circuit worked.

The osc/resonator traces are not shorted. The connection from socket to resonator is good.

 

[Blueteeth]

Hmm, just an idea, have to doubled up the traces between the PIC and the resonator with wires? preferably as short as possible. I've had no end of trouble with a stripboard layout, where the osc traces are almost identical to yours and I still haven't found a way to make the b*stard oscillate - although, I am using a xtal instead. It seems I only get this trouble when I have a GND connection inbetween the two osc pins...god knows why that would interupt its working, if anything it *should* make it more stable.

Blueteeth

 

[Nigel Goodwin]

Hmm, just an idea, have to doubled up the traces between the PIC and the resonator with wires? preferably as short as possible. I've had no end of trouble with a stripboard layout, where the osc traces are almost identical to yours and I still haven't found a way to make the b*stard oscillate - although, I am using a xtal instead. It seems I only get this trouble when I have a GND connection inbetween the two osc pins...god knows why that would interupt its working, if anything it *should* make it more stable.

I've never had any problems at all with crystals on veroboard, have a look at the layouts for my tutorial 876 and 877 boards.

 

[3v0]

Nigel,

Your tutorial schematics for the 876 and 877 use crystals with caps.

Blueteeth said did not have the problem when he used crystals and that the problem only showed up when he ran ground between the two osc pins.

I am going to take the dremmel and remove the ground and replace it with a wire elsewhere and see what happens.

In the end this board may end up in the trash but I would like to understand the cause of the problem.

On another note. Posting is an art I do not understand. If one provides too little info you end up doing 20 questions. If one provides enough information readers skim the post and ask the questions anyway.

 

[Nigel Goodwin]

Nigel,

Your tutorial schematics for the 876 and 877 use crystals with caps.

Blueteeth said did not have the problem when he used crystals and that the problem only showed up when he ran ground between the two osc pins.

But he said "I am using a xtal instead", which made me believe he's using a crystal?.

 

[3v0]

You are correct.

 

[luisgerman]

Can nearby traces add enough capacitance to cause the resonator to not work ?

There are two important electrical differences between a resonator and a crystal. The impedance of the resonator is lower than that of the crystal and that its frequency response is wider. While a crystal’s ESR may be 20 to 1000 ohms, a typical resonator will be in the range of 20 to 150 ohms. The frequency of a typical resonator is less accurate than that of a typical crystal. At start-up, the resonator’s lower impedance develops a lower voltage across it over a wider frequency range, which can make it more difficult for the oscillator to ’find’ the fundamental frequency and start oscillation. Getting a noise signal fed back to the input from the output becomes more important.

Since the “Q” of ceramic resonators is generally lower than quartz, they are more easily pulled off frequency by variations in circuit or parasitic reactances. This is desirable if a circuit is designed with a variable element, as greater tuning range is realized. It is not desirable if the highest possible stability is the design goal, because the resonator will be more susceptible to variation in parasitic reactances, such as capacitors formed by circuit board etch, and temperature variations of intended circuit reactances. These variances will add to the already substantial deviation over temperature of the resonator itself.

 

[Blueteeth]

Yes, don't wanna hijack the post...I am using a crystal, and its the first time I've had trouble with the osc on a *>single<* stripboard layout...isolated incident.

Nigel, you have recommended your layouts several times, and I'm sure they're just dandy, but my board is slightly different. The crystal is quite close to the PIC, maximum trace length 7mm....via a jumper wire that is. However, due to the generous size of your layout...your xtal/caps circuit seems well isolated from anything else, so I'm guessing its a 'coupled electric noise' thing, with yours being less suseptable.

It's related to 3v0's problem, because the layout of this board is the same his. I had a GND trace inbetween the two osc pins so that I could use a resonator when needed (removing the caps). I too will probably trash this board, I'm completely confused, short traces, clean connections (no cold solder joints), checked with a DMM, caps...its all good.

3v0, could post a larger pic of your layout, so we can see what traces are near the osc pins? Although many insist 'stripboard is fine' is isn't the most reliable for oscillators, are you using SRBP? This issue is bugging me something chronic, lets hope we can sort it out.

Blueteeth

 

[3v0]

3v0, could post a larger pic of your layout, so we can see what traces are near the osc pins? Although many insist 'stripboard is fine' is isn't the most reliable for oscillators, are you using SRBP? This issue is bugging me something chronic, lets hope we can sort it out

Yes maybe in a few hours.

I am using a board designed with EagleCad and etched/drilled.

 

[Nigel Goodwin]

Nigel, you have recommended your layouts several times, and I'm sure they're just dandy, but my board is slightly different.

Mine work!

But to be fair, it's NOT a particularly critical thing.

 

[Blueteeth]

But to be fair, it's NOT a particularly critical thing.

It shouldn't be I think the god of parasitic capacitance may have farted in my face. I'll replace the caps on mine tomorrow.

3v0, let us know your results. Oo just a random thought, if you've got I/O connections around the resonator, try shunting these to VCC or GND with a high value resistor, so that they only carry a tiny current, regardless of whether you're using them or they're inputs or outputs....just a thought. I'm toying with the idea of a parallel resistor.

Blueteeth

 

[donniedj]

Umm.... Are those with the resonator issues using a scope to monitor the oscillation frequency on both OSC1 and OSC2 pins? I suspect some other PIC pin has a cold solder joint, resistor with micro fracture, or a micro open in a trace.

Remember the bigger and harder the problem the smaller the root cause.

 

[3v0]

Umm.... Are those with the resonator issues using a scope to monitor the oscillation frequency on both OSC1 and OSC2 pins? I suspect some other PIC pin has a cold solder joint, resistor with micro fracture, or a micro open in a trace.

Remember the bigger and harder the problem the smaller the root cause.

I have removed and replaced (with new) the processor socket. I have removed and replaced the osc. I have checked all power, gnds, and osc pins at the pic socket with ohm meter.

 

[Nigel Goodwin]

This may, or may not, be helpful? - but I have occasionally had oscillator problems, never on my stripboard layouts, but on bought PCB's - in all cases (after LOT'S and LOT'S of messing about), it turned out to be something stupid (like config fuses wrong, pin bent under a chip, mclr not connected, chip plugged in wrong way round - stupid things like that). In no case ever has changing the crystal or capacitors helped - yet I did it every time!

 

[luisgerman]

Umm.... Are those with the resonator issues using a scope to monitor the oscillation frequency on both OSC1 and OSC2 pins?

To measure the oscillator, it should be tested with an oscilloscope connected to OSC2 (Remember that a scope probe adds its own capacitance to the circuit, so this may have to be accounted for in your design).

 

[3v0]

If you probe osc1 it will stop the osc even with a 10x probe. So you probe osc2. Got that.

It can not be the config value because the chip works on the breadboard.

I have checked the solder joints with an Ohm meter many time. Old hat.

Lets put this on hold for a bit. I will try to post a up close and personal image of the underside of the board this afternoon (US central).