The Radar Goes Up the Mast

Radar is the one instrument that tells you about things nobody is broadcasting. AIS shows you ships that want to be seen; radar shows you the fishing panga with no lights, the squall line building astern, and the headland you're closing with at two in the morning. For a 29-foot boat going down the Baja coast, that's not a luxury item.

Why we moved it

The Rev A pole was a good piece of engineering, and we're not going to pretend it wasn't — it's why the old drawings are summarised further down rather than deleted. But three things pushed the radar up the mast, and they all point the same way.

• A stiffer thing to hang it from. An eight-foot pole with a dome on top is a cantilever, and the whole Rev A design was about keeping that cantilever short by bracing it off the solar rack. A mast is a spar built to carry the rig. Thirteen pounds of radome on its forward face is not something it will notice.
• More range. Radar range on a small boat is set by antenna height. The pole put the dome about 12 ft above the water; fourteen feet up the mast puts it at about 18.5. That's roughly an extra nautical mile on every target, which at six knots is ten more minutes of warning.
• A quarter of the money. A turnkey radar pole kit we'd trust was coming in around $1,450 before a radome plate; a Scanstrut SC20 mast mount is $395, and it's rated for the HALO20 out of the box.

The last three rows are the honest part. The mast mount isn't free: it is very slightly heavier aloft, it gives us a blind wedge astern, and the install happens fourteen feet up on a mast that already has a mainsail rolled up inside it. All three are dealt with below.

The gear

We bought the radar as a bundle: Defender's Simrad GO9 XSE with HALO20 and a Navico 9-pin high-speed skimmer transducer, part 000-16294-001, at $1,999. The same retailer lists the HALO20 on its own at $1,999 — so the bundle is, in effect, a radar with a free 9-inch plotter, a sounder and a C-MAP Discover chart.

That does change one line from Rev A. We used to say the radar would talk to OpenCPN “without a $2,000 plotter in the way.” Now there's a plotter — it just didn't cost anything extra, and it gives the boat a factory radar display that doesn't depend on a Raspberry Pi staying booted. The Pi still gets the radar too. Both see it at once.

How high? The earth has opinions

Radar is line of sight, bent slightly by the atmosphere. The working rule is that the radar horizon in nautical miles is about 1.22 × √(height in feet), and you detect a target when the two horizons overlap — yours plus the target's.

That gives you a number that has nothing to do with the 24 nm on the box. For surface targets, a small boat's radar range is set by how high you can get the antenna — which is the whole argument for the mast.

Range creeps up with the square root of height, and weight-moment goes up in a straight line with it. That's the trade in one table: the last two feet buy 0.3 nm and cost 40 ft·lb. Fourteen feet is a design assumption, not a measurement — the real number is “as high as fits under the spreaders, below the steaming light, and clear of everything that exits the front of the mast.” We'll know when someone's been up there with a tape.

Where on the mast

The forward face, on the centreline, below the spreaders. That's where everyone puts a mast-mount radome, for good reasons: it's out of the way of the mainsail, which on this boat furls into a slot on the aft face; the forward face is where the bracket's swivel feet can find a fair surface; and the dome looks forward over the bow, where the targets we're closing with are.

The mount is a Scanstrut SC20: marine-grade aluminium, four swivel feet that adapt to the mast's section, and a twelve-point fixing in Monel rivets. Defender lists the HALO20 and HALO20+ as compatible.

The blind wedge astern

Put a radome in front of a mast and the mast is in the beam behind it. Everything dead astern, in a narrow wedge, is in the mast's shadow. The pole never had this problem — it stood clear at the stern with nothing behind it but open water.

Heel, and the 25-degree problem

Moving the radome doesn't change this one. The HALO20's beam is 25° tall — call it ±12.5° either side of horizontal. Heel the boat and the whole beam tilts with it, because the mount doesn't know the boat is leaning.

Up to about 12° of heel you lose nothing. Past that, the horizon slides out of the beam: to leeward the beam is pointing into the water, and to windward it's pointing at the sky. At 20° of heel you are substantially blind abeam.

Weight aloft — roughly a wash

This is the number we expected to kill the mast idea, and it didn't. The pole design carried 39 lb — dome, 2 in stainless, base, clamp, plate — at an average of about 8.6 ft above the centre of gravity. The mast mount carries far less weight, but much higher.

About 9% more weight-moment than the pole, for a radome 6.6 ft higher. In stability terms that's about 0.05 in more rise in the centre of gravity — lost in the noise of how much water is in the tank. The solar page's weight-aloft table carries the new figure as Rev N.

What it costs to leave it on

Twenty watts sounds like nothing until you multiply it by a passage — and now there's a 9-inch display to multiply as well.

Left transmitting around the clock, the HALO20 alone eats roughly 37 amp-hours a day — about a third of what the 400 W array is expected to return in decent sun. Leave the GO9 lit next to it and that's another 22 Ah. Together that's most of a day's solar harvest. So:

• Timed transmit is the big lever. The HALO will sleep and wake on a schedule — a sweep every few minutes is most of the collision-avoidance value for a fraction of the amp-hours.
• Guard zones let you run intermittently and still get told when something enters the ring.
• One screen at a time on passage. The Pi is already running for the nav stack. At night, radar on the Pi and the GO9 dimmed or off is about 22 Ah a day back.
• On continuously when it's earning it — fog, shipping lanes, night approaches — and asleep when it isn't.

Simrad doesn't publish a standby figure, so we'll measure the real duty-cycle draw with a shunt and put the numbers here rather than guessing. This is one more argument on the pile for the lithium conversion the energy budget already points at.

The network — and why it stays open source

Here's the part that makes a Simrad radar a reasonable thing to put on a boat running OpenCPN: the HALO talks Ethernet multicast, and it does not care who is listening.

The GO9 XSE has exactly one Ethernet port, so the radar doesn't plug straight into it. It goes into a small switch, and the switch feeds both the GO9 and the Raspberry Pi. The GO9 gets the full factory feature set; the Pi gets the radar overlaid on OpenCPN charts through the radar_pi plugin, which supports the Navico HALO directly.

The cable is the real job

Everything above is a bracket and a few settings. The cable run is where an afternoon becomes a weekend.

• Length. The HALO20 ships with a 10 m cable — 32.8 ft. Fourteen feet up the mast, down to the step, through the deck, and aft to the nav station, with a service loop at each end, could easily be more than that. Simrad sells a 20 m HALO dome cable. Measure the run with a string before deciding; don't splice an Ethernet radar cable.
• Which way it goes down. Internal conduit if there is one that's separate from the furler cavity; otherwise an external track up the forward face. See the furling note.
• Which end goes first. The dome end plugs into the radome. The other end — Ethernet plug and power tails — is the one you feed down the mast and out at the step, so size any exit hole for that end and nothing larger.
• The step. The cable has to get from the base of the mast into the cabin. A proper deck gland beside the mast step, bedded in butyl, with a drip loop on the outside. If the mast is ever unstepped, that's the joint that has to come apart — leave enough slack.

Parts

Load checks

Design case is the HALO20's own 100 kt rating — dynamic pressure 33.9 psf. If the dome survives it, the thing holding the dome should too. The new load the pole never had to think about is roll: fourteen feet up the mast, the dome swings through a much bigger arc than it did eight feet up a pole.

None of these numbers is close to anything. That's the point of a mast mount: the structural engineering that filled half of Rev A — tube sizes, a brace, a backing plate, a margin table — mostly disappears. What's left is workmanship: fair feet, sound metal under them, isolated fasteners, and rivets that stop where they should.

Measure before you order

Most of these need a trip aloft. The first three can each move the design.

1. Spreader height above the coach roof, and how much clear mast face there is below it. The radome plus the SC20 need their height, plus room to swing a rivet gun.
2. What exits the forward face of the mast below the spreaders — the steaming/deck light, any halyard sheaves, a spinnaker or spare halyard exit, a pole track. The radome goes below the steaming light and must not obscure it.
3. What's inside the section at mount height. Where the furling foil sits, whether there's a separate wiring conduit, and the wall thickness — which sets the rivet grip length.
4. The mast's section, fore-and-aft and athwartships, at mount height. Confirms the SC20's feet will sit fair, and it's the athwartships width that sets how wide the blind wedge astern really is.
5. The cable run, with a string — mount, down the mast, out at the step, through the deck, to the nav station, plus a service loop at each end. Under about 30 ft: the 10 m cable. Over it: order the 20 m.
6. The jib on both tacks — leech, clew and sheets against the radome at its height. Fractional rigs usually keep the jib well forward of the mast up there, but check before deciding on a guard.
7. Mast-up or mast-down. If there's a haul-out with the rig coming down on the calendar, this whole job is easier with the mast on trestles. If not, it's a bosun's chair, a second halyard as a safety line, and a calm day.
8. The nav station. Where the switch, the breaker and the GO9 live, and whether the GO9 goes at the helm or below.

Build sequence

1. Commission the radar on deck first, before anything goes up the mast. Dome on a bucket, cable to the GO9, power on. Prove the radar works and the cable reaches before it's fourteen feet in the air.
2. Bring up radar_pi on the Pi against the same dome through the switch, and check both see it at once. Still on deck.
3. Go aloft and take the measurements above. Mark the mount position on the forward face, centred, plumb to the mast — not to the horizon.
4. Check what's behind the wall at each rivet position. Settle the rivet length on that, not on the bag it came in.
5. Offer up the SC20, set the swivel feet fair to the section, mark and drill the twelve holes. Isolating compound in every hole and under every foot.
6. Rivet it off. Check the mount platform is level to the boat at rest on her lines — the swivel feet are there to make this possible.
7. Run the cable: down the conduit or into its external track, out at the step, through a bedded gland, aft to the nav station. Service loop at the dome and at the switch.
8. Hoist the dome on a line rather than carrying it up the chair. Bolt it to the mount, connect, drip loop.
9. Breaker, switch, GO9, Pi. Commission on the GO9 first, because it's the simpler path.
10. Calibrate the heading offset against a known object — a breakwater end or a charted buoy. An uncalibrated radar overlay is worse than none.
11. Map the blind wedge against a fixed target, turning slowly through the stern sector. Publish the real angle here.
12. Put a shunt on the radar circuit and log a week of real consumption before deciding how to run it on passage.

What happened to the pole

Rev A, published 28 September, put the HALO20 on an eight-foot, 2-inch stainless pole on the port quarter, braced off the solar rack's aft corner. The brace was the clever part: it cut the pole's cantilever to under four feet and let 2-inch tube do the job of 3-inch, saving sixteen pounds. It also propped up one corner of the rack.

All of that was sound, and none of it is needed now. With the radar on the mast:

• The rack's R3 brace clamp is deleted, and the solar rack goes back to carrying 100% of its load into the arch — which it was always sized to do, because we never let the rack depend on the pole.
• The radar beam no longer has to clear the solar array, which is what set the pole's minimum height.
• The aft deck keeps its clear centreline and quarters for boarding, the ladder and a stern anchor — no pole base, no backing plate, no new hole.
• The GPS/AIS antenna that was going to ride on the pole needs a new home. The GO9 has its own GPS, so this is now an AIS antenna question only.

Still to sort out

• The trip aloft — spreader height, what's in the way on the forward face, and what's inside the section
• Internal conduit or external track, and whether the 10 m cable reaches
• The SC20's real weight, on a scale
• A heading source for the GO9 — dedicated sensor or Signal K through a gateway
• Mapping the blind wedge astern, measured rather than estimated
• Real duty-cycle power draw for both the radar and the GO9
• A new home for the AIS antenna
• Where the skimmer transducer goes on a sailboat's transom — its own page

All of it goes on YouTube as we work through it — including whoever ends up in the bosun's chair — and the long version lands on Substack.

Radar and display: Simrad HALO20 · GO9 XSE / HALO20 bundle at Defender · HALO 20 m cable. Mount: Scanstrut SC20. Plugin and networking: radar_pi · requirements wiki. Hull data: Sailboatdata Hunter 290.