What a High Split Upgrade Requires

A high split is often described as a spectrum change. In the plant it is a replacement programme, a documentation programme and — the part that surprises people — a compliance programme.

Moving upstream to 204 MHz does three things at once. It reclaims spectrum for the return path. It puts the upstream band on top of frequencies that legacy equipment is still using. And it moves the strongest upstream signals into a band the FCC watches, at the point in the network where the plant is oldest and least documented.

This is a field-level reference on what the upgrade actually asks of the plant.

The split ladder

North American HFC was built sub-split: a narrow return path, because subscribers were assumed to consume far more than they sent. Upstream demand broke that assumption, and operators have been climbing the ladder ever since.

Boundaries per CableLabs. See Band Splits 101.

What has to be replaced

Every active in the signal path has to pass the new split, and everything that does not has to be found, documented and swapped. Amplifiers and other legacy equipment come out; diplex filters change; the node has to support the new split in both directions.

Less obvious, and more expensive to discover late: the RF port level design changes. Loading the upstream band harder means operators need a lower RF port level design requirement to prevent gain stage distortion and analog laser clipping. That is not a like-for-like swap. It is a design revision applied across the plant, which means the design has to be built on an accurate picture of what is out there.

The practical consequence is that the constraint on an upgrade programme is rarely the equipment. It is knowing where the equipment is, what model it is, and what it feeds.

The legacy signalling problem at 70–130 MHz

This is the one that catches upgrade plans built from a spectrum diagram rather than from the plant.

Legacy set-top boxes are managed over an out-of-band control channel, and both of the protocols still in service put their forward carrier in the 70–130 MHz range:

• SCTE 55-1 — forward 70–130 MHz at 2.048 Mbps; return 5–42 MHz.
• SCTE 55-2 — forward 70–130 MHz at 1.544 Mbps (Grade A) or 3.088 Mbps (Grade B); return 8–26.5 MHz.

A high split moves the start of the downstream band to 258 MHz. The OOB forward carrier at 70–130 MHz is now inside the upstream band, and the legacy boxes that depend on it stop being reachable. Mid split does not escape it either — downstream starting at 108 MHz cuts straight through the middle of that range.

So the upgrade has a dependency that is not an RF problem at all: which nodes still have legacy set-tops on them, and how many. That is a records question, and records are exactly what tends to be stale. An operator that cannot answer it per node cannot sequence the upgrade, because the answer decides whether a given node can be cut over or has to wait for a box swap.

Leakage inverts

Under a sub-split plant, leakage monitoring has a comfortable shape. Downstream carries the strong signals, the aeronautical band sits in the downstream, and the strongest signals are on the trunk side. Detection was built around that.

A high split turns it upside down. Upstream now extends to 204 MHz, which means it contains the 108–137 MHz VHF aeronautical band — the band used for aircraft navigation and communication. And the transmitters in the upstream are the subscribers' modems, running at high power. The result is the strongest aeronautical-band signals at the home, and the weakest at the node or headend — the opposite of the arrangement every legacy leakage method assumed.

The FCC rules that apply are not new, but where they bite is:

• 47 C.F.R. § 76.611 sets the basic leakage criteria: a cumulative leakage index where 10 log I∞ ≤ 64, aeronautical bands at 108–137 MHz and 225–400 MHz, airspace field strength not exceeding 10 µV/m RMS at 450 m altitude, and individual leaks in new plant held to 20 µV/m at 3 m (analog) or 17.4 µV/m at 3 m (digital). Compliance is demonstrated by sampling at least 75% of cable strand annually, or by annual airspace measurement.
• § 76.610 brings additional obligations once signals exceed 75.85 µW over 25 kHz in the aeronautical bands, and § 76.1804 applies at 10 µW over 30 kHz. Even modest OFDMA configurations can cross those thresholds.

What changes operationally is that leakage detection has to work on an upstream signal that is bursty and subscriber-generated rather than a continuous downstream carrier. The industry approaches are CW time-division bursts, CW frequency-division tones in an excluded band, and OUDP test bursts — the last of which is generally preferred because DOCSIS 3.1 already specifies it and it needs no firmware beyond what a high split deployment already requires.

The field consequence is blunt: drop, connector and plant integrity stop being a maintenance backlog and become a compliance dependency. Shielding faults that were tolerable when the aeronautical band carried a downstream signal are a different matter when a modem is transmitting into them.

Reference: Leakage in a High Split World, SCTE Expo · 47 C.F.R. § 76.611

High transmit, and why plant condition shows up as a service problem

Bonding more upstream carriers pushes modems toward their transmit ceiling. The OFDMA maximum is 44.25 dBmV in a 1.6 MHz reference bandwidth, or 50.27 dBmV in a 6.4 MHz reference bandwidth. A modem already near that has nothing left for another channel.

The extended band also produces significant upstream tilt. Modems can accommodate tilt as high as 12 dB, but compensating for it through pre-equalisation costs transmit power — from the same budget.

When the budget runs out the failure is quiet rather than obvious. The modem stays online. What happens instead is transmit channel set reduction, IUC reduction, or a channel marked impaired — capacity silently missing rather than a subscriber calling. Every dB of unnecessary loss in the plant and the drop makes it more likely.

Reference: High-Transmit in a High-Split Architecture, Broadband Library

What all of this asks of the documentation

Read the three sections above together and the same requirement falls out of each one.

The replacement programme needs an accurate active and passive inventory. The OOB dependency needs legacy set-top presence known per node. The leakage exposure needs plant and drop condition known before cutover, not after. The transmit budget needs loss understood span by span.

Every one of those is a field engineering and documentation problem before it is a construction problem — and it is where upgrade programmes slip. A discrepancy between the records and the plant that is found during design costs a redraw. The same discrepancy found during construction costs a truck roll, a change order and a schedule day, and it is found at the worst possible moment.

Where TESINC fits

TESINC does the field engineering and documentation side of HFC upgrade work across the Southeast:

• Plant mapping and verification — walking the plant and recording what is actually there, not what the records claim
• Upgrade documentation — identifying the actives, passives and plant elements that will not pass the new split
• Node split support — field data and documentation for segmentation programmes
• Fiber deep and DAA migration support
• As-built surveys — closing the loop so the records match the plant when the programme ends
• Make ready and pole loading — where upgrade work changes the aerial path, including make-ready engineering and joint-use coordination

Field data is GPS-synced and comes back in Katapult Pro, KMZ or custom GIS formats, with CAD in AutoCAD and MicroStation.

We are a field engineering and documentation firm, not a headend or RF design house. The value we add to an upgrade programme is that the design and the construction crews are working from a picture of the plant that is actually true.

HFC and high split upgrade engineering services · Talk to us about a programme

Frequently Asked Questions

What frequencies does a high split use?

Upstream runs 5–204 MHz and downstream begins at 258 MHz. That compares with sub-split at 5–42 MHz upstream and 54 MHz downstream, and mid split at 5–85 MHz upstream and 108 MHz downstream. Ultra-high split extends upstream as far as 684 MHz.

Why do legacy set-top boxes break in a high split?

Both legacy out-of-band control protocols put their forward carrier in the 70–130 MHz range — SCTE 55-1 and SCTE 55-2 both. A high split moves the downstream band start to 258 MHz, so that carrier now sits inside the upstream band and the boxes lose their control channel. Mid split has the same problem, because downstream starting at 108 MHz cuts through the same range.

Does a high split change signal leakage requirements?

The FCC rules do not change, but the exposure does. Upstream to 204 MHz now includes the 108–137 MHz aeronautical band, and the transmitters are subscriber modems — so the strongest aeronautical-band signals end up at the home rather than at the node, which is the reverse of what legacy leakage detection assumed. 47 C.F.R. § 76.611 still requires a cumulative leakage index of 10 log I∞ ≤ 64 and annual sampling of at least 75% of strand.

What has to be physically replaced in the plant?

Amplifiers and other legacy actives that will not pass the new split, along with diplex filters and node equipment. Operators also apply a lower RF port level design requirement to avoid gain stage distortion and analog laser clipping, so it is a design revision rather than a like-for-like swap.

What is upstream tilt and why does it matter?

The extended upstream band produces tilt across its width. Modems can accommodate as much as 12 dB and compensate through pre-equalisation, but that compensation costs transmit power from a budget already stretched by carrier bonding — which is how plant loss turns into missing capacity.

How does a modem running out of transmit power show up?

Usually not as an outage. The modem stays online and the symptoms are transmit channel set reduction, IUC reduction, or a channel flagged impaired. Capacity is quietly absent rather than a subscriber complaint, which is why it is often found late.