How do I minimise electrical noise/signal interferance inside a projector?

[kitatit]

Hey guys,

I've got a question about good design practice to minimise noise and interference inside my projector.


How much of a potential issue is this?

here are a couple of pic showing the approximate direction I'm heading in with my projector design.

I wanted to shift the mains AC in socket closer to the ILDA board and ILDA in out sockets to allow more space for mounting holes to screw the rear panel to the heat sink.


Are there any rules of thum about how close mains voltage (240V) should be kept from the various types of signals running around inside a projector?

I'm aware that there are a wide range of frequencies af varying amplitudes bouncing around inside here but I'm not sure how much of an issue it is.

Are there any no-no I should avoid?

Thanks
kit

[mixedgas]

Follow the recomendations in Bill Benner's grounding PDF. There is not much you can do. The good news is dry air needs 2.2 Kv per mm to break down, unless it is between needlepoints.

The spec you want is refered to as "creepage voltage" and should not be a problem unless your routing the 250V on a home made PC board or other nonstandard connection.

So 15 mm is not a bad start, using heavily insulated mains lead.

The specs for creepage are here:

https://b2b.harting.com/Webhelp/EGds..._distances.htm

Your looking at 4 mm for 250VAC, so 15 mm should be a nice distance.

Small signal leads ( Galvo inputs, Galvo Feedback signals) near the 50 Hz should be differential and shielded.

Only tie the shields to ground at one end of the shielded wire.

After that it comes down to routing for least hum, and a filtered, shielded IEC C14 as the socket, if you have room. consult BBs paper if your using one of the IECs that grounds the case for you. I like the ones with the built in filters, as they provide a path to return the spikey noise on the line from INSIDE the case, as well as keeping the incoming noise down.

The paper on star/home run wiring is some place on the Pangolin web site.

Ferrite beads can also help, place them on the small signal stuff, or wrap the signal leads a few turns through a ferrite torroid if your picking up RF gack from the switching psus.

This can require a bit of art if there is HF noise pick up.

I dont think you'll have much of a issue to worry about.

http://www.pangolin.com/resguide12.htm


Steve

[kitatit]

Thanks Steve! thats great info.

The IEC socket I have is a basic fused,switch but not filtered. I was looking at the filterd ones when I bought it.
The Vicor supply I'm using has the "FARM" front end input brick. I didn't have a close look at what filtering it does but I assume it is doing a similar job of filtering as the filtered IEC socket.
http://www.vicorpower.com/cms/home/products/brick/FARM

Is it still worth using a filtered socket or would this be a case of too many cooks in the kitchen?

[mixedgas]

It looks like "FARM" will help a great deal. Keeping things out of the enclosure from the outside is the main reason for a filtered C14. I doubt you would see much difference unless your using it at a gig with large amounts of metal halide or arc lights. I went to filtering after a gig next to a electric arc steel mill in Pittsburgh.

I'm partial to the inlet filter, but "FARM" will by its nature protect the nearby parts.

Steve

[kitatit]

Thank for your insights Steve, it's much appreciated. Sounds like I'm on the right track.

Cheers
Kit

[dnar]

A few of my general rules of thumb.

1. Don't run low voltage and high voltage wires parallel to each other. Doing so results in cross-talk. If wires must cross, do so at 90 degrees.
2. Keep low voltage and high voltage wires separated. The more distance, the less the interference.
3. Follow Bill's ILDA wiring guide, it works and is not dissimilar to good audio amplifier wiring and grounding practice.

I'll add to Steve's comments:

Only tie the shields to ground at one end of the shielded wire.

This is good advice and results in what we used to call "Telescopic" shielding. My audio experience has shown bonding the shield at the input side (receiver) is best.

A good single point "star" ground is best as prescribed in Bill's wiring guide. Watch out for phantom grounds...

[kitatit]

Thanks Wayne, I remember having the 90deg recommendation for midi leads and mains power leads.

Thanks man!

[dnar]

Thanks Wayne, I remember having the 90deg recommendation for midi leads and mains power leads.

Thanks man!

It's an old trick, minimises cross-talk as the "receiving" wire runs 90 degrees to the "transmitting" wire, the EM field is not coupled to the receiving wire. Most old school motor mechanics know this from running high tension spark plug leads parallel on high compression propane fueled engines and the misfires that result...

Another good one to know, avoid sharp 90 degree bends in wires carrying large currents, as these bends produce more concentrated EM fields.

All these concepts apply equally to PCB layout also, which is why you don't see many 90 degree traces on boards, they are 2 x 45 degree traces or even curves in RF applications.

[mixedgas]

[QUOTE=dnar;220972]It's an old trick, minimises cross-talk as the "receiving" wire runs 90 degrees to the "transmitting" wire, the EM field is not coupled to the receiving wire. Most old school motor mechanics know this from running high tension spark plug leads parallel on high compression propane fueled engines and the misfires that result...

What the heck had high compression propane engines?

Steve

[dnar]

What the heck had high compression propane engines?

Steve

My old hand built "ride" for a start! 10.5:1 compression, 5.1 litres, running propane. If you want to see and hear her, let me know I'll start another thread.

Plenty of US V8's ran propane for Baja racing etc.