Cat5 max current?

I can only offer you how I feel about it - based on the rest of my experience. The 1A should work fine in most cases - although there may be connectors which will not be good even for that. The 2A look too much for the connector, but my guess is you will have trouble only in a limited number of cases. If you do this on a large volume thing, you are likely to hear from it - but my guess is in a manageable manner. If it is about just a few devices, you'll manage it one way or the other - but then I doubt you would have asked if this were the case :-). If you will not use the RJ45 connectors and can control that part, you are unlikely to have any problems out of the cable itself, some measurable voltage drop, so what. It won't get too hot.

Dimiter

------------------------------------------------------ Dimiter Popoff Transgalactic Instruments

------------------------------------------------------

larwe wrote:

The Power over Ethernet folks use 48VDC at 400mA max., apparently. Presumably they had some reason for choosing that figure.

Is that continuous, or peak, and over single conductors, or parallel pins ? expected lifetime ? Ease of replacing the cooked parts, consequence of a melt-down ?

Does this include a CAT5 connector, or do you mean just the cable itself ?

2A continuous would be a stretch over a single conductor/connector, but probably tolerable over cable, where your main worry is bulk heating ( so low current on other pairs, allows a bit more push on a couple...)

-jg

Have you taken a look at the Power Over Ethernet (POE) literature?

Robert

The conductor cross section area for Cat5 is usually about 0.2 mm² with a resistance about 0,1 ohm/m. In your case the total resistance would be 2 x 1.5 m x 0.1 ohm/m for each pair, in your case 0.3 ohms.

For the 12 V circuit, the current density would be 10 A/mm², the voltage drop 0,6 V and the losses 1.2 W. For the 5 V pair, the current density 5 A/mm², voltage drop 0.3 V and power dissipation 0.3 W. The total power dissipation would be 1.5 W or 1 W/m.

Often current densities around 3 A/mm² are specified for small cables, while individual wires in the free air can take 2-3 times that current density.

It also depends on the length of the bulkhead and the thermal conductivity of that bulkhead how much you can dissipate without melting the conductor insulation. Even if the insulation does not melt, too much heat will deform the pairs and hence distort the impedance. With a bulkhead 100 mm long 100 mW dissipation might be too much, if the bulkhead is made of some material with low thermal conductivity.

Can you run a higher voltage through the bulkhead and generate the 24 and 5 V supplies locally ? This would reduce the current density and hence power dissipation.

Paul

The PoE is quite similar to phantom feeding microphones. The DC extraction points are at the central taps of the Ethernet transformers, thus one pair carries the current to the device and an other pair away from the device. With two 0.2 mm² conductors in each direction, the current density is only 1 A/mm² and the voltage drop only 0.1 V/m.

With the 100 m maximum segment length, the total DC current path would be 200 m, thus the voltage drop would be 20 V, so if the source provided 48 V, only 24 V would arrive at th other end. Apparently the maximum current is limited by the voltage drop in a long segment.

Paul

The connector is not a problem, it is only the actual wire that is the problem. In this application it will be running into screw terminals, not RJ45 :)

Actually it is just two units. The problem is that it was a real pain to seal that bulkhead and test it. I had to borrow a fixture to do the testing. To replace this piece of cable means taking the whole piercing fitting apart and resealing, then retesting to make sure I didn't mess it up.

Thanks for the reply.

I don't think this qualifies as free air, somehow :)

The bulkhead is aluminium.

YES! I didn't think of that. I have a 48VDC bus on the supply side and plenty of room on the other side.

Since my 2A seems a bit bleeding-edge, I'll look at this option. Thanks!

Continuous. Consequence of a meltdown is 70% loss of telemetry. Expected lifetime is 12 months' 24/7 operation, after which time it is free to disintegrate (or, more likely, be scavenged and recycled).

Ease of replacing anything that cooks is not far off the same level of difficulty as going back to the Moon to flip that circuit breaker that popped open on one of the LRVs they were using to videotape the LZ post-LM-takeoff :) This is unattended equipment and no likelihood of access by trained personnel.

_Just_ the cable. It is a spare, not currently connected to anything at all.

It sounds like the 2A is a stretch, so I'm going with Paul's idea of running my bare 48V battery bus over the wire and putting a switcher in the other end. Building and qualifying such a system is less trouble than trying to install a different piece of wire.

Look at the power over ethernet spec. IIRC the current spec is 13W over a single pair of Cat 5. The voltage is in the region of 40 + something volts. The newer spec currently being worked on, hopes to push this up to 30W or more, although over more pairs of wires. The distance of course for the 13W is much longer than 1.5m. Linear Technology and others manufacture components to handle the power of the ethernet cable. Maybe you can get more specs from these datasheets.

Regards Anton Erasmus

In fact the aluminium or copper bulkhead might have a lower thermal resistance from the conductor to the ambient, compared to conductor to free air, especially if you have water on one side of the bulkhead.

I initially assumed that some elastic material with low thermal conductivity was used as the bulkhead.

Running at a higher voltage and lower current should be a good idea, since the forces acting on the bulkhead might deform the conductors even when cold, not to mention if the DC current heats up the insulation material and the bulkhead forces deform the twisted pair, causing impedance mismatches.

Paul