next topic »

[on:]

hi,
im building a project where my 328 comes into contact with lots emf noise from 2,  3 phase frequency inverters im just wondering which would be best to use for this environment really not sure which would be better?

my first thought would be a crystal but then its incased in metal so would  would act like an airel and suck all the emf into it?

currently im using a resonator upside down so its as fay away from the scourse as possible really unsure, just trying to give my 328 all the help it can get

thanks

[Reply #1 on:]

First the proper term is crystal resonator Vs ceramic resonators. And the ceramic resonators come in either 2 pin versions requiring external padding capacitors just like the 2 pin crystal resonators, or there are 3 pin ceramic resonators the have the padding caps built into the package.

3 pin type Ceramic type resonator:
https://www.sparkfun.com/products/9420

Crystal type resonator:

https://www.sparkfun.com/products/536?

As far as EMI effecting either one, I don't think either is better or worst then the other. What is more important for these resonators is they are mounted and wire as close as possible to the clock pins of the microcontroller chip.

Lefty
 

[Reply #2 on:]

If you've large EMI problems why not enclose the Arduino in a metal box?

[Reply #3 on:]

The can that the crystal is encased in is a shield. Helps keep the noise out (and the noise form the crystal in...) An enclosed can does not function as a very good antenna, otherwise it would be pretty poor practice to use something like that as a shield, and as you ought to know, encasing electronic component inside a metal box, and grounding that box is about as good as it gets for shielding.

[Reply #4 on:]

yes ive mounted my Resinator right next to 9 and 10 pins, im actually using 2 328's cant really mount it in a metal box ive tryed all sorts of metal shielding which seems to make interference worse for some reason think it acts like an airel, the emf travels straight through double sided copper, does some real strange things

thats got me thinking we used "mew" metal years ago for something we made,  might give that a try see if that does anything.

what ive made works, just trying to help it further if i can
 

[Reply #5 on:]

yes ive mounted my Resinator right next to 9 and 10 pins, im actually using 2 328's cant really mount it in a metal box ive tryed all sorts of metal shielding which seems to make interference worse for some reason think it acts like an airel, the emf travels straight through double sided copper, does some real strange things

thats got me thinking we used "mew" metal years ago for something we made years ago,  might give that a try see if that helps.

what ive made works, just trying to help it further if i can
 

I believe the term is Mu metals and google will show you suppliers and info.

Lefty

[Reply #6 on:]

Mumetal is for shielding devices from constant magnetic fields. It won't help protect against EMI any better than a metal box will, unless the EMI takes the form of low frequency magnetic fields.

To protect against intense EMI, you need to enclose the circuit in a metal box - typically with a conductive gasket between the two halves - and also prevent EMI entering via the inputs and outputs. This is done by a combination of designing the input and output wiring so as to minimize EMI pickup, any by using RC or LC networks in the input and output circuits.

If you use a crystal, you should connect the metal case to ground if you expect EMI to be a problem.

[Reply #7 on:]

The crystal enclosure is a metal box, but I don't think it's grounded.  Maybe it's possible to solder a lead from the case to ground?  Also, make sure your enclosure is grounded.  It will pick up EMF, and then it will shunt them to ground.  This is what you want.

[Reply #8 on:]

When it comes to EMI pickup, I would think the xtal is the last thing to worry about.
The thing to worry about is all those long traces on the board that have somewhat
high impedances w.r.t. ground, as well as all those even longer wires connecting the
board to the outside world. Even if you shield the board, you still have those wires to
deal with. That's where you need to concentrate on EMI rejection.

[Reply #9 on:]

on my first board design i had top bottom 100% "gnd" fill on my board which didnt work too well

i did think of maby trying 100% top/bottom "Earth" fill to reduce noise, really unsure of this though, not sure what noise will come up the earth,  making the gap between the fill and tracks pads objects ect, quite big

the board is inside a plastic box so cant really do anything with this.

[Reply #10 on:]

The fill, as well as true ground planes, will help prevent EMI pickup on the traces, as
well as spurious propagation of high-speed signals from the traces, but won't do much
for noise being led into the board via the external connection wires.

[Reply #11 on:]

the board is inside a plastic box so cant really do anything with this.

Sure you can.  You can put it in a different box.  Physics doesn't bend to the whims of designers.  Either you're concerned about shielding (grounded metal case, proper PCB layout, and input filtering) or you're not (plastic case and bare wire).  Picking arbitrary things to optimize won't necessarily net you anything other than wasted parts.

[Reply #12 on:]

As I remember mu metal was used (as said above) as a magnetic shield around CRTs in scopes etc. Talking of which, rather than guess, get a decent scope, it'll pay for itself over and over if you learn how to use it.

And you can metallize the inside of plastic cases Check out your mobile phone.

If you look at the data sheet for any device that requires a high frequency clock, you will find very specific information about ground planes, guard rings, shielding etc.

If you are producing a commercial product, you have to comply with FCC rules!

[Reply #13 on:]

many thanks for the help and advice