Main Site Documentation

EMX Clock speed


#1

Hi,

What speed is the main oscillator on the NXP chip set to, and what PLL multiplier do you use? I am having issues CE marking a product, as my device is emitting too much noise at certain frequencies - trying to eliminate any clock sources as the culprit.

Thanks


#2

You can read the crystal. It says 12mhz. The processor runs at 72mhz as shown on product page.

I do not have PLL settings but you can read them from registers.


#3

Thanks. What FCC / CE certificates does the product hold?


#4

EMX module does not require FCC/CE certification since it is not an end-user product and it does nto deal with radio frequencies. These certifications can be applied to the final product that uses EMX.


#5

You may not need the certification (though I’d keep a close eye on the Raspberry pi situation if I were you, it is akin to the your spider/cobra products in many ways). but you are not ultimately going to sell many EMX units if people can’t get products through EMC tests easily. I must confess I had hoped that with your 8 layer board and very tight design and not too high clock frequencies we might have been OK!!! … but it seems not.

D Proctor, I would really appreciate it if you could share any experiences you may have in getting your product through the EMC regs. I’m just at the start of that process at the moment.

Thanks very much,

David Cowan


#6

Guys, I would think it would be worth taking this kind of commercial conversations offline with GHI directly (at least the bits that are less than conversational)


#7

Sorry for the rant. Used to do EMC consultancy and this sort of issue is so often a problem. On a sample of one I don’t know that there is a problem in this case obviously… but if there is it will come back again and again and ultimately using the ‘component’ defense will not help.

I do think it would be immensly helpful and engender confidence if anybody with any actual experience in these matters could share them with the forum.

My tupence (sorry 2 cents) worth. Won’t post again on this


#8

Sorry, don’t mean to come off sounding like a forum admin, because I’m not, so let me clarify.

Soliciting other people who have experiences like you’re seeking is fine, although the commercial work that is likely to have done this is less likely to hang out on the forums :slight_smile: And I’d trty to keep the conversations light and not critical. I also suspect that GHI are better situated to know what other customers may have done in this area, and while they are unlikely to disclose that here you may have a different conversation in a 1:1 setting.


#9

Yes, of course you are right. Bit of a touchy issue for me and I shouldn’t have jumped!


#10

We are taking about a 72mhz chip here. I would be really surprised if it is causing any significant noise. And if it did, then NXP will be in trouble with hundreds of customers.

The PLL setup values are the same ones you see in any example code from NXP.

Let us know if we can help further.


#11

I know I said I wouldn’t respond… so sorry.

You’d be amazed…l I’ve seen little 8bit 2MHz micros hummmm… 30dBs over… and my customers were likewise surprised… It’s all about the layout, but 72MHz is fast in EMC terms and quite capable of causing severe problems unless every precaution is taken. Limits are actually set stupidly low, I think mostly to create non tarrif trade barriers.

As I said the multilayer board gives confidence… but not too much. Must confess I had a bit of a ‘cold sweat moment’ when somebody reported an EMC issue. I really don’t want to have to get my spectrum analyser working again!!!

BTW could you sort my post counter it seems to have been inadvertantly reset.

Thanks,


#12

Assuming there is a problem, this will be a problem for NXP more than GHI. They sold millions of these chips so I am assuming things went well.

Finally, EMX has been out for years and have never heard of such problem so another assumption there is no problem.

You can always order a spider/cobra and run any tests you like before you start any design.


#13

About the counter, it looks correct http://www.tinyclr.com/forum/user/7337

Where are the other posts?


#14

It really is not anything to do with NXP chip, rather the design of the clock circuitry - Typically the copper tracks that act as the antennas for emitting spurious noise.

The EMC issues, for me, are in the 10 - 100MHz range (only up to 6dB over), so the 12Mhz and its harmonics could easily be the culprit. Like I said, im trying to identify the possible sources, so that I can start eliminating them, and the EMX is just one of few possibles at this stage.


#15

Great, hope you are right!!

Thanks,


#16

About the number of posts. I think i made one post as user 6376 and then the next one was as user 7337 on this thread. I didnt re-register or anything.


#17

6dB doesn’t sound too bad. Again I’d appreciate it if you could let me know how you get on!!.


#18

Then these are 2 users not one :slight_smile:


#19

No at 6dB over we were quite happy due to the standard operating conditions it would be used in. I will post an update as to how we get on.


#20

OK somehow I seem to have two. Can you delete the 7337 one.