Richard,

Thanks for your thoughts. We do have exact spec levels, which are
detailed in documents called 'VIDG E-marking' and another more rigorous
Home Office test called 'Spec-5'. These EMC specs are too lengthy to
post to the list, but I can tell you that they are quite hard targets to
hit. We have to test our equipment at approved testing houses, at a cost
of about UKP 1000 per day. Therefore, I wish to ensure that there is
some chance of this new PC design actually passing before I burn the
companies money!

Peter Moreton




> -----Original Message-----
> From: pic microcontroller discussion list
> [mailto:PICLIST@MITVMA.MIT.EDU] On Behalf Of
> Richard.Prosser@POWERWARE.COM
> Sent: 25 February 2004 19:19
> To: PICLIST@MITVMA.MIT.EDU
> Subject: Re: [EE] EMC compliant design
>
>
> You may be heading for an uphill battle.
> It will mostly depend on how good the motherboards etc are to
> start with.
> The main problem is going to be how do you know when it is
> satisfactory? Do
> you have spec levels for conducted & radiated emissions or is
> it just going
> to be a case of "No, not good enough yet, try again". Or even
> worse - "We
> have a set in XXXX that is interfering with his locally made
> radio system.
> Your equipment needs to be fixed. Now. " (XXXX being a place
> very far away,
> hard to get to and uncomfortable to live in)
>
> If you do have spec levels to meet, you are going to have to
> get the beast
> tested. This can be an expensive proposition, particularly as you are
> likely to need several trys.
>
> And what about susceptability to the RF transmissions from
> the vehicle? You
> don't want it to reset every time they transmit. Again, more expensive
> testing.
>
>
> On the plus side, computers are manufactured to meet emission
> standards and
> hence the addition of ferrite cores to video cables etc over
> recent years.
> Some motherboards offer the ability to dither the clock to produce a
> spread-spectrum noise - rather than a single frequency. I
> would suggest
> that you check if this is available as it can improve
> (measured) output
> noise by ~10dB or so. (The noise is still there but the
> average over the
> measurement bandwidth is lower).
>
> Add-in filters to the power cables, video, keyboard etc lines may be a
> proposition but require careful layout and well specified
> components. And
> more testing.
>
> EMC is more of an art than a science and what works in one
> situation can be
> useless in others. My advice would be to talk to  someone with as much
> experiance in the field as possible. Then design,  build,
> test, get more
> advice and repeat. It will cost but not as much as trying to
> start from
> scratch.
>
> As a starting point, try to get a firm specification sorted.
> This needs to
> include  frequency ranges.  "Radiating almost no RF " will trip you up
> sooner or later.
>
> Sorry I can't be more encouraging - Best of Luck!
>
> Richard P
>
>
>
>
> This is not really a PIC question, but I really need some
> quick pointers
> so apologies for a slightly OT posting...
>
> I have just been asked to put together a design for an In-Car
> PC system
> to go into emergency service vehicles, and which must radiate
> almost no
> RF. I'm planning to go for a Mini-ITX form factor motherboard (the
> CL1000 mobo) and a battery backed 12v PSU / UPS controlled by a PIC.
>
> The whole caboodle will go into a custom made metal enclosure
> with a lot
> of attention to RF shielding, to ensure that no RF leaks out
> of the vent
> holes etc. I'm worried about how to prevent RF from escaping
> out of the
> Kbd, VGA, Serial etc cables and was considering a
> 'filter-pcb' to filter
> all communications to the outside world.
>
> I know absolutely zilch about RF / EMC issues, so if anyone
> can point me
> towards any URL's where I can learn the tricks of the trade, or just
> advise what works and what doesn't, that would be great.
>
> Thanks, Peter Moreton
>
> --
> http://www.piclist.com hint: To leave the PICList
> mailto:piclist-unsubscribe-request@mitvma.mit.edu
>
> --
> http://www.piclist.com hint: To leave the PICList
> mailto:piclist-unsubscribe-request@mitvma.mit.edu
>

--
http://www.piclist.com hint: To leave the PICList
mailto:piclist-unsubscribe-request@mitvma.mit.edu