| Electronics > RF, Microwave, Ham Radio |
| Where to connect the shields of wiring and connector advice |
| (1/3) > >> |
| daqq:
Hi guys, Where should I connect the shields of a shielded differential pair cable in the following situation: There are two metal boxes. Both are grounded locally, there is less than 1 meter between them. In the whole space around them there is a nasty pulsed RF field at around 10GHz that we can't do anything about (existing installation). There are 3 devices - one master (C), two slaves (A, B). There will be one connector board on the boundary of the first main enclosure. All of the signals are floating (transformer coupling on both sides). The crossed circles represent connectors. The signals are low power. The signal path will be as follows: Source device -> Transformer -> Connector -> Shielded cable -> Connector -> Connector PCB -> Connector -> Shielded cable -> Connector -> Transformer -> Target device My question is: At which points should I connect the shields of the cables to the enclosures? Should It be one single, undisturbed connection, or should I connect part of the signal path to one enclosure, the other part to another enclosure? Using what way would I achieve the least RF interference getting into either enclosure and onto the transmitted signal? My guess would be that the best place to connect the shields would be at the connector board (3, 4, 5, 6) and no where else. Is this a reasonable assumption? Also, I was looking into using SATA cables - they are relatively cheap, well shielded, with a well defined impedance. Would you recommend their usage in this kind of situation? Thanks, David |
| T3sl4co1l:
First preference: can you do this? (pic) If there's ANY distance between the enclosure entry and the PCB, you'll have radiation from those inside lengths. The distance between "B" and the grounding point, or any other point where a cable enters the PCB to the grounding point, also allows radiation. At low frequencies, we'd be talking stray inductances, but these will be full random fractions of wavelengths, even for tight construction, so you really need shielded connectors, coaxial cables (or fully shielded/screened), EMI spring stock stuff, and all that. Any kind of transformer isolation is simply a short circuit at those frequencies, and the immunity of a given device will be down to how well the CM filtering and rejection is, and what the susceptibility of the devices are. Tim |
| Rerouter:
Generally the shield is connected at the side with the lowest impedance, as hard tying it to a ground return can cause a loop picking up more noise rather than lessening it, The is obviously different if the shield is your ground return, but i thought i would give both cases, For both these cases my knowledge runs out somewhere close to the 10's of Mhz range, so for higher, you need an RF wizard to hop in, |
| KJDS:
I've used SATA connectors successfully for high speed data. I'd then move one box next to the other rather than leave it a metre away. |
| Ian.M:
At 10GHz, a joint between two plates with a 1.5cm section not making electrical contact looks like a resonant 1/2 wave slot antennae. :scared: I'd look at the possibility of running metal conduit (rigid or metal flexible with outer metal braid) between the two enclosures, with threaded or conductive compression end fittings to make them effectively a single Faraday cage. Depending on the possible earth fault currents, it may need a bonding strap between the cabinets. Then use screened cable within the conduit, with an anti-EMI ferrite ring right at the conduit end fitting in both cabinets, with its screen grounded at the source end to the source PCB. An extra ferrite right up against the connectors at each end would also help. I doubt anything less will keep out high energy local X band signals. What is being done to stop the EMI coming in on the other wiring? |
| Navigation |
| Message Index |
| Next page |