Cat5 and Cat5e are both rated to 100 MHz. That is the part that surprises people, because the letter at the end usually signals more bandwidth. Here it does not.
Figure 1: Twisted pair conductor pairs and outer jacket insulation.The change from Cat5 to Cat5e was about crosstalk, not speed rating. The "e" stands for enhanced, and what was enhanced was the specification for how much signal leaks between the pairs at the top of the same 100 MHz band. That change is what made Gigabit Ethernet a reliable proposition over copper at building distances.
Quick answer
| Cat5 | Cat5e |
|---|
Bandwidth | 100 MHz | 100 MHz |
|---|
What changed | Baseline specification | Tightened crosstalk and return loss limits, tighter pair geometry |
|---|
100BASE-TX | Yes | Yes |
|---|
1000BASE-T (1 Gbps) | Marginal, not a designed target | Yes, to 100 m |
|---|
2.5GBASE-T | Not a designed target | Yes, to 100 m |
|---|
Current status | Superseded by Cat5e | The practical minimum in new work |
|---|
The short version: if you are buying new cable, there is no reason to choose Cat5 — Cat5e costs effectively the same and carries Gigabit Ethernet properly. If you are working on cable that is already in the wall, the question is not which is better but whether the installed link passes.
Why the same bandwidth can behave differently
Every twisted pair works by cancellation. The two conductors in a pair carry the same signal with opposite polarity, and the twist makes external noise couple into both conductors almost equally, so it cancels at the receiver. The better the geometry, the more complete the cancellation.
At 100 MHz, the limits that matter are the ones describing signal that leaks from one pair to another inside the same cable. Cat5’s limits were set when 100BASE-TX was the target — two pairs transmitting, two pairs receiving. Gigabit Ethernet changed the load: all four pairs transmit and receive simultaneously, in both directions, using echo cancellation to separate the traffic. That turns every pair into a noise source for the other three.
Cat5e tightened the numbers that govern exactly that situation. Not the bandwidth, and not the pair count. The margins.
The practical difference in a building
Situation | Cat5 | Cat5e |
|---|
100 Mbps to the desk | Works | Works |
|---|
1 Gbps to the desk | Sometimes works; not a designed distance | Designed for 100 m |
|---|
2.5 Gbps | Not specified | Specified to 100 m |
|---|
Bundles of 40+ cables | Higher risk of failing crosstalk | Marginally compliant |
|---|
Certification against Cat5e limits | Will not pass | Passes |
|---|
The pattern in that table explains the real-world experience people report. Installed Cat5 often does run Gigabit on short runs in a lightly loaded bundle. That is not the same as it being specified for Gigabit, and it is why one office has 1G everywhere on Cat5 and the office next door has links that drop to 100 Mbps.
How to tell which you have in the wall
Read the legend printed on the cable jacket. It will name the category, and it is the only reliable identification: color, brand and installation year do not determine the category.
If the legend is unreadable, the honest answer is a certification test. A permanent-link test against the Cat5e limit table produces a pass or fail with the failing parameter named. That tells you what the link can carry far more reliably than the print on a jacket that has been in a ceiling for fifteen years.
What you find | What it means |
|---|
Jacket says CAT5 | Expect 100 Mbps reliably; Gigabit is not designed for |
|---|
Jacket says CAT5E | Gigabit and 2.5G to 100 m, as designed |
|---|
Legend unreadable | Test a sample of links against the Cat5e limit table |
|---|
Passes Cat5e limits | Treat as Cat5e for planning purposes |
|---|
Fails on NEXT or return loss | The link is the constraint; a retermination may recover it |
|---|
One caveat on that last row: a failing link is not always a cable problem. Poor terminations, damaged pairs at the outlet and untwisted conductor length at the connector account for a large share of failures on cable that is perfectly good. Reterminating a sample is cheaper than replacing a cable run.
What the "e" did not fix
Cat5e was a tightening of limits rather than a new capability, and it is worth being clear about what it left alone.
It did not raise the bandwidth. Both categories are 100 MHz. A Cat5e link and a Cat5 link occupy the same frequency range, which is why neither is a 10GBASE-T cable at building distances.
It did not change the connector. Same RJ45 interface, same wiring assignment, same eight conductors.
It did not make the cable immune to bad terminations. Cat5e’s tighter crosstalk figures assume the pairs stay twisted to the termination. A Cat5e cable terminated carelessly performs no better than a Cat5 cable terminated the same way, and can perform worse than a Cat5 cable terminated well.
It did not change the distance. 100 meters is the channel limit for both, and the split of 90 meters of permanent link against 10 meters of cordage applies to both.
If a network built on Cat5 is failing, the category is one of the last things to suspect. Termination quality, bundle size and pathway temperature all cause more damage than the difference between the two grades.
Scenarios
Buying cable for a new office fit-out. Buy Cat5e or Cat6. Cat5 is not worth specifying; the price difference against Cat5e is negligible and the performance difference is not.
A building with Cat5 installed in 2003, now needing 2.5G access points. Test a representative sample against the Cat5e limit table. Where links pass, 2.5GBASE-T is a supported option. Where they fail, address the termination first.
A tender that lists "Cat5" as an acceptable category. Ask for it to be restated as Cat5e, with 10GBASE-T or 2.5GBASE-T distance requirements named. Quoting Cat5 against a Gigabit requirement commits you to a category that was never designed for it.
A warehouse running Cat5 to PoE cameras. Power delivery is a DC resistance problem, and thin Cat5 conductors dissipate more heat for the same current. Check the conductor gauge before assuming the existing cable carries the PoE load.
FAQ
Is Cat5e faster than Cat5?
Not in bandwidth — both are 100 MHz. Cat5e carries Gigabit Ethernet over the full 100 meters as a designed capability, where Cat5 does not.
Can I use Cat5 and Cat5e in the same installation?
You can, but the link performs to the lower category. If one run in a channel is Cat5, that channel is a Cat5 channel.
Does Cat5 support Gigabit at all?
It often works on short runs. It is not specified for it. Planning on "it works in practice" is what produces a building where half the links negotiate at 1G and half at 100 Mbps.
Is Cat5e going obsolete?
No. It is the practical floor for new work and supports 1G and 2.5G to 100 meters. It is not a 10GBASE-T cable at building distances, but neither is Cat6 beyond 55 meters.
How do I confirm what is installed without replacing it?
Certify a sample of links against the Cat5e limit table. The report names the failing parameter, which usually points at the termination rather than the cable.
Related reading on this site: the next step up the ladder is covered in Cat5e vs Cat6: When the Cheaper Cable Is Still the Right Answer; why twisted pairs work in the first place is explained in How Do Ethernet Cables Work?; what a certification report can and cannot prove is covered in How much should you trust a cable test report?; and the Cat5e product range is on the Cat5e cable page.
