Bandwidth and PHY Architecture: Comparing 100BASE-T1 and 1000BASE-T1 Standards

100BASE-T1 (IEEE 802.3bw) and 1000BASE-T1 (IEEE 802.3bp) are both single twisted-pair automotive Ethernet standards. The difference is speed and what it demands from the cable and connectors: 100BASE-T1 carries 100 Mb/s and suits sensors and actuators, while 1000BASE-T1 carries 1 Gb/s and suits backbone links that collect high-bandwidth sensor data.

In short:

• Choose 100BASE-T1 when 100 Mb/s is enough and you want lower PHY cost and power.
• Choose 1000BASE-T1 for zonal gateways, multi-camera aggregation and other links that need 1 Gb/s.
• The faster standard is harder on the whole channel, so cable symmetry, connector impedance and the number of inline connectors matter more.

Quick Comparison

• IEEE amendment — 100BASE-T1: 802.3bw; 1000BASE-T1: 802.3bp
• Data rate — 100BASE-T1: 100 Mb/s full duplex; 1000BASE-T1: 1 Gb/s full duplex
• Symbol rate — 100BASE-T1: 66.67 MBd; 1000BASE-T1: 750 MBd
• Modulation — 100BASE-T1: PAM3; 1000BASE-T1: PAM3
• Nyquist frequency (half the symbol rate) — 100BASE-T1: about 33 MHz; 1000BASE-T1: about 375 MHz
• Forward error correction — 100BASE-T1: None; 1000BASE-T1: RS-FEC, RS(450,406)
• Cable — 100BASE-T1: Single balanced twisted pair, 100 Ω; 1000BASE-T1: Single balanced twisted pair, 100 Ω
• Reach — 100BASE-T1: 15 m link segment; 1000BASE-T1: 15 m (Type A, unshielded); 40 m (Type B)
• Inline connectors — 100BASE-T1: Up to four in the link segment; 1000BASE-T1: Up to four per segment

Reach figures come from the project objectives for each amendment and are minimum reaches. Confirm the final channel limits in the standard and the OPEN Alliance channel specifications before you design a link.

How the Two Standards Differ

Both use a single balanced pair and echo cancellation, so each end transmits and receives at the same time on the same wires. Both use three-level pulse amplitude modulation (PAM3). The difference comes from the symbol rate.

Symbol rate and bandwidth

At 750 MBd, 1000BASE-T1 runs more than eleven times faster than 100BASE-T1 at 66.67 MBd. Half the symbol rate gives the Nyquist frequency, which is about 33 MHz versus about 375 MHz. Channel limits are specified to higher frequencies than this, so check the channel specification for the exact range. The result is that attenuation, return loss and high-frequency noise become the dominant problems at 1 Gb/s.

Error correction

1000BASE-T1 uses Reed-Solomon forward error correction, RS(450,406). IEEE material describes it adding 396 bits of parity to each 4050-bit frame. The 100BASE-T1 PHY does not use RS-FEC. At 1 Gb/s the extra margin helps the link cope with noise, at the cost of some added latency.

Cable and reach

The 1000BASE-T1 project objectives describe two link segment types. Type A is unshielded balanced cabling supporting up to four inline connectors for at least 15 m. Type B is an additional segment supporting up to four inline connectors for at least 40 m. 100BASE-T1 targets a 15 m link segment on unshielded twisted pair. Shielded cable is common in noisy locations, but the standards do not require it for every link.

Connectors for Each Standard

Each inline connector adds insertion loss and a reflection point, so the connector family needs to be rated for the speed you run. These are vendor-published ratings, so confirm them against the current datasheet:

• 100BASE-T1: compact connectors such as Molex Mini50 (designed for 100 Mb/s, with sealed and unsealed versions) and TE MATEnet.
• 1000BASE-T1: connectors with a published 1000BASE-T1 rating, such as TE MATEnet and Rosenberger H-MTD.
• Multigig (2.5, 5, 10 Gb/s) and above: Rosenberger lists H-MTD for 2.5/5/10GBASE-T1.

For a family-by-family comparison of H-MTD, MATEnet and Mini50, plus impedance, mode-conversion and latching rules, read Automotive Ethernet Connectors Explained. For sealing and ingress ratings, see our sealed vs unsealed automotive connectors comparison.

Which Should You Choose?

Choose 100BASE-T1 for:

• Domain sensors and smart actuators
• Entry-level displays
• Body electronics that replace CAN where 100 Mb/s is enough
• Designs where lower PHY cost and power matter most

Choose 1000BASE-T1 for:

• Central zonal gateway backbones
• High-resolution radar and LiDAR processing links
• Multi-camera ADAS aggregation
• Links that must have headroom for future data growth

If you are unsure, count your worst-case bandwidth, then decide how many inline connectors the path needs. A path with several connectors at 1 Gb/s needs a careful channel budget. For the wider family of single-pair standards, see our guide to single-pair Ethernet and automotive Ethernet standards, and for cable construction, why Ethernet wires are twisted and cable shielding types.

Frequently Asked Questions

Can 100BASE-T1 and 1000BASE-T1 share the same cable? They use the same type of single balanced pair, but a cable and connector set that works at 100 Mb/s may not meet the channel limits at 1 Gb/s. Check the channel specification for the faster rate before reusing cabling.

Does 1000BASE-T1 require shielded cable? Not by the standard's objectives: the Type A segment is described as unshielded. Many designs still choose shielded cable for EMC margin in noisy locations.

Why does 1000BASE-T1 use error correction and 100BASE-T1 does not? The faster symbol rate leaves less noise margin, so 1000BASE-T1 adds RS(450,406) forward error correction.

What is the next step up from 1000BASE-T1? IEEE 802.3ch defines 2.5, 5 and 10 Gb/s automotive Ethernet, and 802.3cy (approved June 2023) covers 25 Gb/s.

How many connectors can a link have? The 1000BASE-T1 objectives describe segments supporting up to four inline connectors. Fewer connectors leave more channel margin.