Cat 5e vs Cat 6 vs Cat 7

100mhz and 250mhz??? where are you getting these numbers from?

Standard rates in terms of mhz is 5mhz-33mhz... pci-e certainly would be higher. 133-200+ but even still i think most ethernet devices of the 100Mbit range only produce clock speeds of 33mhz at the wire. So an xbox or ps3 using a pci interface would be here...and a cat5e would be more then suitable?? Particular with the stock hardware not being able to be set higher then 10mbits??

I currently don't have a solid understanding of how an ethernet device can produce 100-250mhz signal in a wire...there are several article citeing capabilities of the wires. But don't mention how an ethernet device produces a signal outside of it's clock speeds...Or how this effects bandwidth or data transmission. It doesn't sound like they are descibeing the speed of the data, but more the resistance and it's ability to produce EMI. And the wire's ability to handle this without problems???

http://www.belden.com/pdfs/techppr [...] _6_Q&A.pdf

This article goes more in depth, the physical differences and capabilities of each wire.

I currently use a cat 6 24 awg. Because i intend to upgrade to a 1gig network. I believe using these on a 100mbit network leaves one or 2 wires un-used...but is still compatable with previous wire sets. The router and ethernet devices also detect what wire is used, and sets it's bandwidth appropriately. Those unused wire in theory help prevent crosstalk and lost information due to attenuation? Or emI produced by the signals sent and recieved in the wires.

cat7 may not be compatable??

Also some routers network balanceing, will downgrade bandwidth when lesser wire sets are used on the same network. You can check this in your routers status when the other devices are in use. You may see connections that you set for 100 at the pc, but when you check the router it's set them for 10. This may not be true for all routers. also cat5e is 100Mbit compatable where cat5 is not.

So, i don't think you would see any gains past cat 6....???

ewwww adaptive link speeds even at the WAN, where you connect your service line...

http://www.routers-review.com/netg [...] fications/

So even with cat-6, it may throttle back, and throttle up whenever it likes, this would create pauses and disconnects. Lag in fps...low framerates in videos and media...ugh. It would never reach full capability for sustained periods of time.

hawkeye:

half-duplex uses 2 wires, full duplex uses four...?? i thought.

This would suggest that. but in reality you'd only be seeing the data as fast as the system can present it to you. So it is a factor.

So how does the signal get high enough for this to be needed? We are discussing the speed of the electical pulses that carry data on a copper line...How do we figure out what that is? So we can decide what we need? And what we desire.

Does modifying this rate of transfer make the signal un usable at the other end?

I would think you could attain almost any speed of transfer rate for a period of time...what limits this?

Hi there guys. You seem to be getting all confused about frequencies, cabling bandwidth and system clock speeds.

1. The fundamental frequency of the data transmission down your cable is not the same as the bandwidth. This is just the main frequency of the signal being sent out down the cable.

2. The fundamental frequency of the transmission has nothing to do with the spec of your pc bus frequencies such as PCI or USB frequencies. The networking card will generate whatever frequencies it needs, either by its own oscillators or clock multipliers.

3. The data rate is a function of the data link layer protocol that is in use. All of these protocols you've been talking about have a symbol rate of 125MHz. Where they vary is the number of bits per symbol, or essentially how many possible levels there are on each cable pair multiplied by the number of pairs. On 100baseTX you have one pair used in each direction (only two out of the four pairs are actually connected), and only two data levels per symbol (0 or 1) which means you've got 125Mbit/sec. Lastly the protocol uses 4B5B encoding which actually sends 5 symbols for every 4 data useful data symbols. Part of this is to make sure that the system can't ever send a long sequence of zeros which would have no level transitions for the receiver to synchronise to. End result: symbols are being sent at 125Mbit/sec, but the useable data rate is 100Mbit/sec, and the fundamental frequency (i.e. the highest spike you'd get on a spectrum analyser is 31.25Mbit/sec.

When you go to 1000baseT, you have 5 possible levels, and now you use all 4 pairs in both directions simultaneously. Overall, you get 1000Mbit/sec, but the fundamental frequency of the data is the same as 100baseTX.

4. Twisted pair cables are always used in pairs for high frequencies like ethernet. It is a differential signal. None of this 'one signal per conductor' nonsense.

6. The bandwidth of your cable is the maximum frequency that can get down the cable, and yes, this is a function of the physical construction of the cable. There will be capacitance and inductance in the cable itself which all end up acting like a low pass filter, which means that any components of a signal above a certain frequency start to get 'attenuated' (i.e. reduced), down the wire. If your cable has a bandwidth figure given, this will probably be the 3bB frequency, i.e. when the signals have dropped to half their maximum level. Frequencies above this will only be worse.

7. So far I've just been talking about the data link layer. By the time you add the other layers of the protocol (session, transport, network), plus the wait time that the data link layer has after sending each frame, plus any error retransmissions you start eating into your 100Mbit/sec, meaning the final data throughput you can get drops down to well below 100Mbit/sec.

6. The Cat5/6/7 of the cable is a specification of the cable construction. I don't know the detais, but this covers things like inductance, capacitance, characteristic impedance, resistance, twist rates, bend radius and shielding. These affect the bandwidth and also the noise rejection and crosstalk between conductors. The more noise gets onto your signal, the more likely you are to have a bit error, and therefore require a retransmission, slowing things down. Clearly a system with 5 possible signal levels, and 4 pairs all transmitting at the same time (1000baseT) is much more vulnerable to crosstalk and noise. Ultimately this means that the higher Cat cables (e.g. cat 7 with individual screens) have better signal to noise ratios, and work the best in high noise environments such as offices where you have possibly thousands of cables running in close proximity. You will get the lowest error rates with these cables, but perhaps more importantly, you have the future potential to use even faster network protocols that may exist in the future.

7. The voltage / watts / resistance thing has nothing to do with data frequencies, you can't do a conversion. Fan speed is not necessarily related to voltage btw, but you are right that wattage is like the horsepower.

8. Wifi will almost certainly always be slower than 100baseTX. You have more overheads, and more errors, plus multiple devices using the same channel on wifi.

9. I always see slower data rates on wifi, something like 10-20% of the speed I get on 100baseTX ethernet.

10. The centre frequency of the wifi standard (2.GHz), has nothing to do with the data rate. Actually this is governed by how many Hz either side of of the channel frequency that the data is allowed to use, this actually *is* a legitamate use of the phrase 'band' 'width'. You can get an idea of the bandwidth by looking at the wifi channel frequency spacings. 802.11g wifi uses 54Mbit/sec.

11. Ping times are a function of network propogation delay (i.e. the speed of light time that it takes your data to get down the wire), number of nodes between you and the destination (hint try typing 'tracert google.com' at a command prompt, to get an idea of how many hops there are), and the computing time required at each node to process the data. Forget overclocking your network card, all you will do is break it.

12. Minor correction. If you mean Mega Hertz, please use MHz, not mHz or mhz. Milli Herts is much much slower!

More info:
http://en.wikipedia.org/wiki/Category_7_cable
http://en.wikipedia.org/wiki/100BASE-TX#100BASE-TX

In answer to the first post, cat 5e is fine for you now on your small network. You may want to go for something higher if you are thinking much longer term (e.g. 10-20 years), when we might be all streaming 4k 3D TV at 120fps! But it's a gamble not a safe bet!