I've spent about 20 hours trying to find the cause of a problem in my design...and I found it...but I don't know what to do to fix it.

I am using two SX-28 chips, one running at 4MHZ (using a blue Murata 4Mhz resonator) and reads a serial line and ouputs characters to an LCD. The other SX-28 runs at 20MHZ (using a ZTT resonator) and runs a timing circuit. I developed the circuit using the PDB, then burned a PCB and wired it up only to find out it didn't work at all. After hours of working with an oscilliscope I decided to wire it all up on a separate prototype board using push connectors. To my surprise, it didn't work there either (ok right about now, half of you are saying duhhh, do that first).

It turns out the cause of the problem is the center ground wire from the oscillators. If they are close together, they mess each other up. If I move the ground wires apart, they work fine. How do I fix the design? The circuit board is small and I can't move the ground wires apart, so that isn't an option.

Is there something I need to do when using two resonators on a project?

Using two resonators, should work, in my opinion...

Just a couple of suggestions...

Have you tried shielding on the proto board yet?  (try using shielded wires, then shield the two oscillators, or any mix of the two).

Are you using a ground plane between traces?

If you can answer some of these questions it will be easier for us to help figure out a solution:

- Does everything share a common ground?

- Can you post an image of the section of the PCB where the resonators are located so we can see how the layout looks?

- Where exactly did you install a 220 ohm resistor?  Is it from ground to one of the other resonator leads? It isn't clear from your message what the other end is connected to.

- Depending upon the resonator used some people install a 10K resistor across the resonator and I have also seen some add a small cap (around 5pf) across a couple leads as well.  That may help.

As always, i appreciate the help. I'd like to find the right way to solve this problem before I make another board.

1)Everything does share a common ground.

2)I can post the bottom of the board, but not the top.

3)I installed the 220 ohm resistor from the center post of the 4Mhz resonator to ground. The resonator is not supposed to need any capacitors or resistors.

4) I'm not using sheilded wires, as this is a PC board. I did use a copper pour to give as much of a plane as possible.

Thanks again!

It is clear that you are running one SX28 at 4Mhz.  Are you running the other one at 20Mhz or 50Mhz??  It mentions 50Mhz on the image you scanned.  4Mhz is forgiving when you do your board layout.  20Mhz you have to be a lot more careful and at 50Mhz you are getting into RF range and you have to think along analog lines when you do the layout.

I can't see why attaching a 220ohm resistor from the ground on the resonator to ground would make a difference.  Something isn't right and you'll need to double check everything.

Can you tell which SX28 chip isn't working?  Do you have access to a scope?

Rather than having ground traces snake from one component to another like that, you should really either make one large rail with smaller branches going to each component or utilize that copper pour as a ground plane.  Routing traces like that is just asking for trouble...

Sorry, the original design was using a 50Mhz. then I was testing it with a 20Mhz when I wrote the original question. It is using a 50Mhz processor.

Yes, I do have an O-scope, and it is the SX-28 running at 4Mhz that is affected.

Velocita, I agree, I need to get the ground under control. I was in a rush and used the autorouter function in the PCB software.

Robot, I'm trying to fix it...or at least understand why it wasn't working.

Paul is correct.You actually have a ground trace running in between the ground plane.  Instead you should be using the ground plane.  The center pin of the resonators should be connected to the ground plane in this situation.  Unless I am missing something your flooded plane doesn't seem to connect to anything which would likely create more noise issues.

[Quoting: "DosManDan"]I installed the 220 ohm resistor from the center post of the 4Mhz resonator to ground. The resonator is not supposed to need any capacitors or resistors.

A lot of the SX-based schematics I see (and that I build) usually place a 10k or so resistor between OSC1 and OSC2 (across the outside pins of the resonator). This value seems to work for 20MHZ and 50MHZ, and doesn't seem to negatively affect circuits where it might not be required. Guenther has more to say about this if you poke around the forums.

The SX datasheet has resistor and capacitor values that are theoretically required for proper resonator startup. The small surface mount resonators do not include caps, the big dip ones usually do.

This may not be it at all, of course, but it may be worth looking at. Certainly something that could be soldered to the resonator pads on the bottom of your board.

In the newer version of the board, I used a ground plane and also the 220Ohm resistor. Things are working fine, I'm just worried a problem might creep up since I never figured out the cause of the problem.

I experienced the SAME problem when I used a stand alone prototype board (push wire). If the ground were closer than about 2 inches, the problem started. When they were farther apart, everything worked fine. Basically, I simply moved the ground wire closer or farther appart.

I wouldn't worry about it too much until you actually experience the problem again.  When you start dealing with higher frequencies, such as 50MHz, RF begins to take effect.  I'm sure the anomaly you witnessed was an RF artifact.

Thanks Zoot, I'll try that and see if it works for me. I'm just worried as heck that I didn't pick the right solution. I called Parallax today to see what they recommended. Chris was answering the question above. With the new design and the changes to the copper pour, I'm keeping my fingers crossed.

Okay, I see what you've done.you have a filled plain instead of a bunch of islands, however your ground lead still runs in between.  Please see the picture below.  This is what I was referring to when I mentioned a ground plane.  All lines that go to ground directly connect to this plane via thermal pads (you could use solid).  This reduces noise greatly.

Chris,
My appologies for not taking a better picture, but the new design (green) is using the ground plane exactly as shown in your photo.

I've included a magnified photo for you to see. I took the picture before the rosin was cleaned off, so I think it was keeping you from seeing the traces.

Dan,
  I thought the two boards you posted were the same design, but now I see they are radically different.  In any event it almost looks as though a few traces and pads are shorted to the ground plane on this photo.

It's a light reflection, they really aren't shorted. This board works perfectly, I'll take a better picture tonight. I was just trying to get it online and the lamp above it caused a lot of white spots in the photo. After I took the picture I was looking at it and I too thought, wow, did I bridge that joint? When I looked at the board it didn't appear anything like the photo.