A Hazelton boiler is a vertical water-tube steam generator built around a central firebox surrounded by a ring of inclined or curved water tubes connecting an upper steam drum to a lower water drum. The design is best known from late-19th-century American steam launches and small industrial plants built by the Hazelton Tripod Boiler Company. Its purpose is to raise steam quickly in a compact footprint where a horizontal fire-tube boiler will not fit. A typical 4 ft launch unit raises 100 psi steam in under 20 minutes from cold.
Hazelton Boiler Interactive Calculator
Vary tube heating surface, temperature difference, heat-transfer coefficient, and latent heat to see predicted steam evaporation and boiler load.
Equation Used
The Hazelton boiler evaporation estimate treats the inclined water tubes as the useful heating surface. Heat transferred to the water is estimated from overall coefficient U, tube area A, and mean gas-to-water temperature difference dT; dividing by latent heat hfg gives steam production.
- Heating surface is counted on water-tube outside area only; drum area is ignored.
- Uses steady mean gas-to-water temperature difference.
- Overall heat-transfer coefficient is an estimated operating value, not a code rating.
- Latent heat represents the operating pressure range being approximated.
How the Hazelton Boiler Works
The Hazelton boiler runs on a simple thermosiphon principle. Cold feedwater enters the lower drum, the firebox in the centre heats the inclined water tubes wrapped around it, the water inside those tubes flashes upward as a two-phase mixture, and steam separates in the upper drum. Gravity pulls cooler water back down the outer ring of tubes to feed the cycle again. No pumps are involved in this circulation — the temperature gradient between hot tubes near the fire and cooler tubes at the periphery does the work. That natural-circulation pattern is what makes the design so quick to raise steam compared with a Scotch marine boiler of similar output.
The geometry has tight rules. Tube inclination must sit between roughly 15° and 30° off vertical — too steep and the steam slugs choke return flow, too shallow and the upper drum starves at high firing rates. Tube wall thickness on the original Hazelton patent boilers was typically 0.095 inch for tubes of 1.5 inch outside diameter, working pressures up to 150 psig. If you over-fire a Hazelton with low water in the upper drum, the consequence is immediate and severe: the topmost tubes go dry, glow cherry red, and bag or split within minutes. That is the classic Hazelton failure mode and it is why the original designs included a prominent gauge glass directly on the upper drum at eye level.
Scale matters too. The heating surface in a Hazelton is overwhelmingly the outer surface of the water tubes — the drum surfaces contribute almost nothing. So if you are predicting evaporation rate, you count tube surface, ignore the drums, and apply a mean gas-to-water ΔT figure consistent with the fuel and grate area. A unit with 120 sq ft of tube surface burning bituminous coal will not behave like one with 120 sq ft burning anthracite — the radiant fraction in the firebox is different and the tubes nearest the fire do disproportionate work.
Key Components
- Upper steam drum: Cylindrical pressure vessel sitting above the firebox where steam separates from the water-steam mixture rising up the hot-side tubes. Typical diameter 14 to 24 inches, with the steam outlet, safety valve, and gauge glass mounted on it. Drum thickness on a 100 psig design runs around 5/16 inch on a 16 inch drum.
- Lower water drum (or ring header): Annular header at the base of the tube nest that distributes returning water to the bottom of every water tube. On smaller launch boilers Hazelton used a ring header rather than a full drum to save weight. The lower drum sits below the firebox grate level so circulation never reverses.
- Inclined water tubes: The actual heat-exchange surface — typically 1.25 to 2 inch OD seamless steel tubes inclined 15° to 30° off vertical and arranged in a cylindrical nest around the firebox. The hot inner row generates most of the steam; the cooler outer row carries the downcomer flow. Tube ends are expanded into both drums with a slight bell on the fire side.
- Central firebox: Vertical cylindrical combustion space at the geometric centre of the tube nest, with a grate at the bottom and an uptake at the top routing flue gas across the tubes before exiting the stack. Firebox diameter is usually 40 to 60 percent of the overall boiler diameter.
- Uptake and casing: Sheet-steel outer casing lined with refractory or asbestos-substitute insulation that forces the flue gas to make a single upward pass across the outside of the tube nest. A poorly sealed casing short-circuits gas flow past the upper tubes and you'll see evaporation drop 15 to 20 percent without any obvious external symptom.
- Gauge glass and water column: Mounted directly on the upper drum at the lowest safe water level. On a Hazelton this is non-negotiable — because the tubes above water level have no thermal mass to protect them, low-water events damage the boiler in seconds rather than minutes.
Real-World Applications of the Hazelton Boiler
Hazelton boilers found their niche wherever space was tight, weight mattered, and the operator needed steam quickly. The design dominated American small steam launches in the 1880s and 1890s and showed up in stationary plants where a horizontal boiler would not fit through the door of an existing building. Today you'll find them mostly in heritage settings — restored launches, museum demonstration plants, and a small number of working narrow-gauge industrial sites where original Hazelton units survived because they were too awkward to scrap.
- Heritage steam launches: Restored Hazelton-fired launches on Lake George, New York — several 25 to 30 ft Truscott and Racine hulls run original or replica Hazelton tripod boilers at 100 to 125 psig.
- Museum demonstration plants: The Antique Boat Museum in Clayton, NY operates a Hazelton-equipped steam launch as part of its working collection.
- Small industrial process steam (historical): Late-19th-century print shops and small machine shops in Brooklyn used Hazelton stationary units of 10 to 25 BHP for line-shaft drives where a Lancashire boiler would not fit.
- Heritage narrow-gauge industry: A few surviving stone-quarry hoisting plants in Vermont retained Hazelton verticals into the 1940s for intermittent winch service where rapid steam-raising mattered more than fuel efficiency.
- Steam yacht auxiliaries: Donkey-boiler service on larger steam yachts of the 1890s — a small Hazelton handled deck-machinery steam in port without lighting the main Scotch boilers.
- Educational and replica builds: Modern small-scale Hazelton replicas built by hobbyists and trade schools for live-steam launch projects in the 1 to 5 BHP range, typically copying the Roberts and Hazelton 1885 patent layout.
The Formula Behind the Hazelton Boiler
Predicting how much steam a Hazelton actually makes comes down to heating surface, mean temperature difference between flue gas and water, and an overall heat transfer coefficient that captures the radiant-plus-convective behaviour of the tube nest. At the low end of typical firing — say 60 percent of grate rating — you're heat-transfer-limited and the tubes furthest from the fire barely contribute. At the high end, above about 110 percent of rated firing, you're combustion-limited: the firebox runs short on residence time, CO rises, and adding more coal stops adding more steam. The sweet spot sits between 80 and 100 percent of rated firing where every square foot of tube surface earns its keep. The equivalent evaporation formula below converts the actual heat duty into a standardised pounds-per-hour figure referenced to feedwater at 212°F evaporating into dry saturated steam at the same temperature, which is how every heritage boiler manual specifies output.
Variables
| Symbol | Meaning | Unit (SI) | Unit (Imperial) |
|---|---|---|---|
| We | Equivalent evaporation rate from and at 212°F | kg/h | lb/h |
| U | Overall heat transfer coefficient across the tube wall | W/m²·K | Btu/h·ft²·°F |
| A | Total external water-tube heating surface area | m² | ft² |
| ΔTm | Mean temperature difference between flue gas and boiler water | K | °F |
| 970.3 | Latent heat of evaporation at 212°F (constant in imperial units; use 2257 kJ/kg for SI) | kJ/kg | Btu/lb |
Worked Example: Hazelton Boiler in a heritage steam launch on Lake George
Predicting the equivalent evaporation rate of a recommissioned 1889 Hazelton tripod boiler being returned to demonstration steaming aboard a 28 ft Truscott-pattern launch on Lake George, New York. The boiler has a tube nest of 64 tubes at 1.5 inch OD, 4 ft 6 in effective length, working pressure 110 psig, feedwater entering at 100°F, and a measured mean flue-gas-to-water ΔT of 680°F when fired on bituminous coal at the rated grate loading. Estimated overall heat transfer coefficient U = 11 Btu/h·ft²·°F based on field data from comparable restored Hazeltons.
Given
- Tube count = 64 tubes
- Tube OD = 1.5 in
- Effective tube length = 4.5 ft
- U = 11 Btu/h·ft²·°F
- ΔTm (nominal firing) = 680 °F
- Working pressure = 110 psig
Solution
Step 1 — compute the external heating surface area of the tube nest. Each tube contributes π × OD × L of outer surface:
Step 2 — at nominal firing with ΔTm = 680°F, compute heat duty and equivalent evaporation:
Step 3 — at the low end of typical operation, light cruising firing, ΔTm drops to about 480°F as grate rate falls and stack temperature drops:
That's roughly 70 percent of nominal — exactly what you'd expect at half-throttle on a launch idling between docks. The fire is bright but not white, the safety valve never lifts, and a fireman can stay ahead of the engine without breaking a sweat.
Step 4 — push to forced firing on a heavy headwind run, ΔTm rises to about 800°F as the firebox runs hotter and outlet stack temperature climbs:
On paper that's a 17 percent gain over nominal. In practice on a Hazelton you rarely sustain it — the upper tubes start to short on water at this evaporation rate unless the feed pump is well matched, and CO at the stack climbs sharply because residence time in the small central firebox is no longer enough for complete combustion. Above this firing rate the boiler is making smoke, not steam.
Result
Nominal equivalent evaporation comes out at approximately 872 lb/h from and at 212°F, which on a 28 ft Truscott launch with a 6 inch by 6 inch compound engine is enough headroom to cruise at hull speed and still hold the safety valve just below lift. The range from 615 lb/h at light cruising to 1,025 lb/h at forced firing tells you the sweet spot is right around the nominal point — push higher and you trade efficiency for noise and smoke, drop lower and the fire wants more attention than the steaming earns. If your measured evaporation comes in 15 to 25 percent below 872 lb/h, suspect three things in order: scaled tube exteriors from poor-quality bunker coal leaving sulphate deposits on the gas-side surface (the most common Hazelton degradation), a leaking casing seam letting flue gas bypass the upper tube row, or an undersized feed-pump cam letting drum level sag and exposing the top of the tube nest. Each of these has a distinct symptom — sooted tubes show as cold spots on a thermal scan of the casing, casing leaks show as warm patches on the outer jacket, and feed-pump issues show on the gauge glass swinging more than ¾ inch during firing transients.
Choosing the Hazelton Boiler: Pros and Cons
The Hazelton sits in a specific niche: vertical, fast-steaming, compact, but fussy about water level and not built for sustained heavy duty. Compared with the alternatives a launch builder or small-plant operator might consider, the trade-offs come down to how fast you need steam, how much you care about fuel economy, and how forgiving the boiler is of operator error.
| Property | Hazelton vertical water-tube | Vertical fire-tube (VFT) | Scotch marine fire-tube |
|---|---|---|---|
| Time to raise steam from cold to 100 psi | 15 to 20 minutes | 30 to 45 minutes | 2 to 4 hours |
| Heating surface per cubic foot of envelope | High — 8 to 12 ft²/ft³ | Medium — 4 to 6 ft²/ft³ | Low — 2 to 3 ft²/ft³ |
| Tolerance to low-water events | Very poor — tubes bag in seconds | Moderate — crown sheet at risk in minutes | Good — large water mass buffers |
| Steady-state thermal efficiency on coal | 62 to 68 percent | 55 to 62 percent | 72 to 78 percent |
| Maintenance interval (tube/firebox inspection) | Annual — internal tube cleaning required | Every 2 years | Every 3 to 5 years |
| Typical pressure rating | Up to 150 psig | Up to 100 psig | Up to 250 psig |
| Best application fit | Steam launches, donkey boilers, small intermittent plants | Stationary small plants, hoisting engines | Marine main propulsion, large process steam |
Frequently Asked Questions About Hazelton Boiler
That's almost always a circulation problem, not a firing problem. The Hazelton relies entirely on natural thermosiphon flow, and if even one or two tubes have partial blockage from scale or debris pushed up from the lower drum during refilling, the flow pattern destabilises and the upper drum sees pulses of two-phase mush instead of a clean steam-water separation.
Check the lower drum handhole sediment first. On a boiler that's been laid up dry, the lower drum collects iron oxide flakes that get sucked into tube ends as soon as you fill and fire. Symptoms include a gauge glass that bounces more than ½ inch at steady firing and a pop safety that hunts rather than seating cleanly.
Both are vertical water-tube designs from the same era and both steam quickly, but the Roberts uses curved tubes and a more accessible lower drum, while the Hazelton uses straight inclined tubes and a more compact ring header. If your launch hull has tight beam at the engine bay, the Hazelton wins on diameter — its tube nest packs tighter for a given output. If you plan to do your own annual tube cleaning, the Roberts is easier to work because every tube is reachable from a single handhole.
For outputs under 5 BHP the Hazelton is the lighter unit. Above 15 BHP the Roberts pulls ahead because its larger lower drum buffers transient demand better.
Probably not. The formula assumes a single mean ΔT across the whole tube surface, but a Hazelton has strong radiant heat input on the inner tube row that the simple convective form doesn't capture cleanly. On a well-fired bituminous coal grate, the inner row can pick up an extra 15 to 25 percent above the convective prediction from radiant flux off the firebed.
The rule of thumb on heritage Hazeltons: if your measured evaporation is between 0 and 15 percent above predicted, you're seeing legitimate radiant gain. Above 20 percent overprediction, double-check your feedwater meter — that's more likely the source than a radiation bonus.
You're priming. On a Hazelton, priming usually traces to one of two causes: water level set too high in the upper drum, or dissolved solids climbing as makeup feed concentrates over the run. The upper drum on a typical launch Hazelton has only 4 to 6 inches of steam space above the normal water line, so half an inch too high and any boil-up rolls liquid into the dry pipe.
Drop the gauge glass set point by ¾ inch and see if the slugging stops. If it doesn't, blow down 10 percent of the drum volume and refill with fresh feed — TDS above about 2,500 ppm starts to foam aggressively in this design.
No, and an inspector will not certify it. The original tube wall thickness and drum thickness were specified for the original working pressure with a specific factor of safety, and the steel of that era is not the same material as modern SA-178 tubing. Even if the tubes pass a hydro at 1.5× the new pressure, fatigue life at the higher pressure is unknown and the drum welds (or rivets, on truly early units) were never designed for the cycle stresses of 150 psig service.
If you want 150 psig from a Hazelton-style boiler, build a new one to modern code using the original geometry. Several heritage launch builders have done exactly this and certified the result under ASME Section I.
Two reasons, both geometric. First, the firebox sits directly under the upper drum with no buffer water mass between the flame and the upper tube ends — when level drops, the topmost tube ends are exposed to flame within seconds. Second, the tube inclination means returning downcomer water has to travel across a hot lower drum before reaching the cool side, so any local hot spot in the lower drum can vapour-lock a downcomer and stall circulation in that sector.
The fix on heritage Hazeltons is operational, not mechanical: never let the gauge glass drop below the lower visible mark, never fire hard with the feed pump off, and inspect the upper-drum tube ends every season for incipient bagging. Skipping any of those three is what gets these boilers into the news.
References & Further Reading
- Wikipedia contributors. Water-tube boiler. Wikipedia
Building or designing a mechanism like this?
Explore the precision-engineered motion control hardware used by mechanical engineers, makers, and product designers.
