An acetylene generator and gas holder is a paired device that produces acetylene gas (C2H2) by reacting calcium carbide with water, then stores the resulting gas in a sealed bell or bladder until a torch or burner draws it off. A typical industrial unit like the Rexarc DA-9 produces 280 cubic feet per hour at 15 psi. The generator controls the reaction rate so pressure never climbs into the decomposition-risk zone, and the holder buffers demand spikes. You see this combination in pipeline welding rigs, foundry burners, and carbide lamps where a bottled-gas supply isn't practical.
Acetylene Generator and Gas Holder Interactive Calculator
Vary carbide feed, gas yield, efficiency, torch demand, and holder size to see acetylene production and bell buffer behavior.
Equation Used
The calculator applies the article relation V_C2H2 = m_carbide x Y x eta as a rate equation. Production is compared with torch demand; a positive net rate fills the floating bell and a negative net rate empties it. Bell time is the holder volume divided by the production-demand difference.
- Gas volume is at standard conditions.
- Yield represents commercial calcium carbide gas output.
- Efficiency accounts for incomplete reaction, leaks, and purge losses.
- Holder time is based on the absolute difference between production and demand.
How the Acetylene Generator and Gas Holder Actually Works
Drop a lump of calcium carbide (CaC2) into water and you get acetylene gas plus calcium hydroxide slurry — the reaction is CaC2 + 2H2O → C2H2 + Ca(OH)2. The catch is acetylene becomes unstable above roughly 15 psi (1 bar gauge) and self-decomposes violently above 30 psi, so you cannot simply collect it in a pressure tank like propane. That's why the generator and gas holder exist as a paired system. The generator meters carbide and water together at a controlled rate, and the gas holder — usually a wet bell floating in a water seal — captures the output at near-atmospheric pressure.
Two generator architectures dominate. Carbide-to-water generators drop measured carbide pellets into a large water reservoir, which gives fast cooling but produces a thick lime-slurry waste. Water-to-carbide generators drip water onto a carbide bed, which is simpler but runs hotter and can overheat if water flow stalls. Either way, the gas leaves the reaction chamber, passes through a hydraulic back-pressure valve (a water-filled trap that prevents flashback into the generator), and rises into the gas holder bell. As the bell rises, it lifts a counterweight or trips a linkage that shuts off the carbide feed — that's the self-regulating loop.
Get the tolerances wrong and the consequences are not subtle. If the water seal in the back-pressure valve drops below the minimum mark — typically 100 mm of water column on a Rexarc DA — a torch flashback can travel back into the generator and ignite the gas pocket above the slurry. If the bell counterweight is set too heavy, internal pressure climbs past 15 psi and you're operating in the acetylene decomposition window. If carbide pellets are too fine (below 4 mm) the reaction runs away in seconds, boiling the water and overpressuring the holder. Particle size, water level, and bell travel are the three numbers you check before every shift.
Key Components
- Carbide Feed Hopper: Holds graded calcium carbide pellets, typically 50-80 mm 'lump' grade for industrial generators or 4-15 mm 'nut' grade for portable units. A mechanical linkage or float-actuated valve releases pellets into the reaction chamber as gas pressure drops in the holder. Pellet grading matters — mixed sizes give erratic generation rates.
- Reaction Chamber: The vessel where carbide meets water. Holds a large water volume (typically 5-10 litres of water per kilogram of carbide) so the heat of reaction (about 1860 kJ/kg of carbide) doesn't boil the charge. Lime slurry settles to the bottom and drains through a sludge cock.
- Hydraulic Back-Pressure Valve: A water-sealed trap between the generator and the downstream piping. Gas bubbles through the water, but a flashback from the torch hits the water column and extinguishes. Water level must stay between the upper and lower marks — usually a 25 mm window — or the seal fails.
- Gas Holder Bell: An inverted cylinder floating in a water tank (a 'wet gasometer'). The bell rises as gas accumulates and falls as the torch consumes gas. Bell travel — typically 300-600 mm — sets the buffer volume. A counterweight on a pulley sets the delivery pressure, usually 0.07-0.14 bar (1-2 psi) for low-pressure systems.
- Purifier: A canister of ferric oxide or heratol packing that removes phosphine (PH3) and hydrogen sulphide (H2S) — impurities that come from industrial-grade carbide and corrode brass torch tips. Purifier media changes from yellow to brown when spent.
- Pressure Relief / Vent Stack: A spring-loaded or water-seal relief that vents to atmosphere if pressure exceeds about 1 bar gauge. Vent stacks must terminate outdoors above roof level — acetylene is lighter than air only when warm, and pools at floor level when cool.
Real-World Applications of the Acetylene Generator and Gas Holder
Acetylene generator and gas holder systems show up wherever bottled acetylene supply is impractical, expensive, or logistically impossible. The reason is economic and practical — bottled acetylene is dissolved in acetone inside porous-mass cylinders, and shipping those cylinders to a remote site costs more than shipping a sack of carbide and a generator. The decision usually comes down to throughput, location, and whether you have trained operators on site.
- Pipeline Welding: Rexarc DA-9 and DA-12 portable generators on remote oil and gas pipeline welds in northern Canada and Siberia, where bottled acetylene transport is restricted by road weight limits.
- Foundry & Heat Treatment: Stationary generators feeding multi-burner heating manifolds at grey-iron foundries, where continuous low-pressure C2H2 demand exceeds practical cylinder-bank capacity.
- Mining & Caving: Premier-style carbide lamps used by cavers and historic mine restoration crews — a miniature carbide-to-water generator and integrated gas holder in a single brass body.
- Glass Working & Lampworking: Small-shop acetylene generators feeding Carlisle and Bethlehem burners in scientific glassblowing studios where natural gas service isn't available.
- Atomic Absorption Spectroscopy: Lab-scale generators producing high-purity acetylene for AAS flame fuel, used when cylinder acetone contamination would interfere with trace-metal analysis.
- Heritage Lighting: Restored acetylene buoy lamps and lighthouse beacons — Aga and Dalén-pattern systems with integrated carbide generators and dissolved-acetylene storage.
The Formula Behind the Acetylene Generator and Gas Holder
The fundamental sizing question is: how much acetylene does a given mass of carbide produce, and how fast does the holder bell fill or empty during a job? At the low end of the typical operating range — small portable units processing 1-2 kg/hr of carbide — you generate enough gas for a single small torch and the bell cycles every few minutes. At the nominal industrial range of 5-10 kg/hr, you feed a multi-torch shop with a bell that cycles every 10-15 minutes. Push past 20 kg/hr and the heat-of-reaction load demands forced cooling, otherwise the water charge boils and gas quality drops. The sweet spot for portable rigs sits at 5-8 kg/hr.
Variables
| Symbol | Meaning | Unit (SI) | Unit (Imperial) |
|---|---|---|---|
| VC2H2 | Volume of acetylene generated at standard conditions (15°C, 1 atm) | litres | cubic feet |
| mcarbide | Mass of calcium carbide consumed | kg | lb |
| Y | Theoretical yield of pure CaC2 (0.349 m³/kg or 5.59 ft³/lb at standard conditions) | m³/kg | ft³/lb |
| η | Practical efficiency factor accounting for carbide purity and dust losses (typical 0.85-0.95 for industrial carbide) | dimensionless | dimensionless |
Worked Example: Acetylene Generator and Gas Holder in a small foundry preheat manifold
You're sizing a stationary acetylene generator to feed a 4-burner preheat manifold at a small grey-iron foundry. Each burner consumes 30 ft³/hr of C2H2 at 5 psi delivery, so the manifold draws 120 ft³/hr continuous. You want to know how much carbide the operator must charge per 8-hour shift, and whether the gas holder bell on a candidate 60 ft³ wet gasometer will cycle at a reasonable rate.
Given
- Demand = 120 ft³/hr
- Y = 5.59 ft³/lb
- η = 0.90 dimensionless
- Vholder = 60 ft³
- Shift = 8 hours
Solution
Step 1 — at nominal demand, calculate total gas required per shift:
Step 2 — solve for carbide mass at nominal 90% efficiency:
Step 3 — calculate bell cycle time at nominal demand. The bell holds 60 ft³ between fully empty and fully full, but practical working swing is about 50% of that to keep the back-pressure valve covered:
At the low end of the typical operating range — say only 1 burner running at 30 ft³/hr during light maintenance work — bell cycle time stretches to 60 minutes. The carbide feed linkage barely trips, and you risk the bell sitting at the top stop for long periods, which lets gas leak past the bell skirt. At the high end, all 4 burners plus a 50 ft³/hr cutting torch tapped in for fitting work pushes demand to 170 ft³/hr:
10-minute cycles are still fine for the carbide feed mechanism, but you're now generating 1.4 lb of carbide per minute peak — heat-of-reaction load climbs, and the reaction-chamber water temperature can rise above 60°C, the point where lime slurry starts foaming over the back-pressure valve.
Result
Nominal carbide consumption is 191 lb (87 kg) per 8-hour shift, with the holder bell cycling every 15 minutes — a comfortable rate that keeps the feed mechanism active without thrashing. At the low end of demand the bell cycles every hour, which is too slow and risks bell-skirt leakage; at the high end the cycle drops to under 11 minutes and you're approaching thermal limits in the reaction chamber. If you measure actual carbide consumption running 15-20% high — say 230 lb per shift instead of 191 — the most common causes are: (1) carbide pellet contamination with calcium oxide dust, which doesn't react but counts toward charge mass, (2) a leaking bell skirt seal letting gas escape into the holder water tank, or (3) the purifier canister oversaturated with phosphine, causing operators to run torches lean and waste fuel chasing flame quality.
When to Use a Acetylene Generator and Gas Holder and When Not To
The choice between an on-site generator and bottled acetylene is rarely about gas chemistry — both deliver the same C2H2 to the torch tip. It comes down to throughput economics, site logistics, and operator skill. Here's how the realistic alternatives compare on the dimensions that actually drive the decision.
| Property | Acetylene Generator + Gas Holder | Bottled Dissolved Acetylene | Propane / MAPP Substitute |
|---|---|---|---|
| Delivery pressure | 0.07-1.0 bar (1-15 psi) | Up to 15 psi regulated from 250 psi cylinder | Up to 5 bar from cylinder |
| Throughput capacity | 1-50 kg carbide/hr (3-170 m³/hr gas) | Limited by manifold size, typically 6-20 m³/hr per cylinder | Effectively unlimited from bulk tank |
| Capital cost (industrial unit) | $8,000-$25,000 (Rexarc DA series) | $0 — cylinders rented, $200-500 per refill | $500-2,000 for tank + regulator |
| Flame temperature | 3,160°C with oxygen | 3,160°C with oxygen | 2,820°C with oxygen — won't weld steel |
| Operator certification | Generator-operator licence required in most jurisdictions | Standard cylinder handling training | Standard fuel-gas training |
| Practical run time before refuel | 8-12 hr per carbide charge | 4-8 hr per cylinder at typical welding rate | Days to weeks from bulk tank |
| Best application fit | Remote sites, continuous high-volume demand | Fabrication shops with cylinder delivery | Cutting, brazing, heating where weld-grade flame isn't needed |
Frequently Asked Questions About Acetylene Generator and Gas Holder
You have a leak somewhere between the bell skirt and the discharge piping. The most common culprit is the bell-skirt water seal — if the water tank level dropped below the skirt during the day's operation, gas escapes into the atmosphere through the skirt gap rather than out the discharge line. Top up the seal water to the high mark and the bell should hold position.
Second most common cause is a slow leak at the purifier canister gasket. Brush soapy water on every flanged joint downstream of the back-pressure valve. If the bell drops more than about 50 mm overnight, you have a leak large enough to be a fire risk and the system shouldn't be relit until you find it.
No. Low-pressure generators are physically designed so the gas holder cannot exceed about 1-2 psi delivery pressure — the bell counterweight sets the ceiling. Forcing higher pressure means modifying the bell weight, which puts the gas in the unstable decomposition window above 15 psi. That's how acetylene generators explode.
If you need medium-pressure delivery you need a medium-pressure generator (the carbide-to-water type with a sealed reaction vessel), and even then the design limit is 15 psi at the generator outlet. Anything beyond that has to come from dissolved-acetylene cylinders where acetone stabilises the gas.
Rule of thumb: holder working volume should equal at least 5-10 minutes of peak demand. Less than that and the generator can't keep up with surge draw — bell drops to the bottom stop and the torch starves. More than 15 minutes of buffer and you're paying for tank steel that never gets used.
For a single 30 ft³/hr welding torch, a 5-7 ft³ working volume bell is right. For a 4-burner manifold at 120 ft³/hr, you want 15-20 ft³ working volume — which usually means a 30-40 ft³ nominal bell since you only use about half the swing.
Phosphine (PH3) breakthrough at the purifier. Industrial-grade carbide carries 0.05-0.1% phosphorus impurities that come out as PH3 gas. The purifier's ferric oxide or heratol media absorbs it, but media has finite capacity — typically 4-6 hours of run time on a portable unit before you see breakthrough.
The yellow flame is partially burnt phosphine. Replace the purifier charge. If the new charge browns within an hour, the carbide grade is too contaminated for your application and you need to switch to a higher-purity grade or accept more frequent media changes.
You're seeing real-world efficiency. The theoretical 5.59 ft³/lb assumes 100% pure CaC2 reacting completely. Industrial carbide is typically 80-85% CaC2 by mass with the balance being calcium oxide, silica, and free carbon. That alone caps yield around 4.7-5.3 ft³/lb depending on grade.
If you're getting significantly less — say below 4.5 ft³/lb — check three things: carbide age (old carbide absorbs atmospheric moisture and pre-reacts in the drum), water temperature (cold water below 5°C slows the reaction and traps unreacted carbide in the slurry), and pellet size uniformity. Mixed fines and lumps cause channelling in the reaction chamber so chunks pass through unreacted.
Carbide-to-water for almost every portable job. The reaction is in a large water bath, so heat dissipates fast and the gas comes off cool — you can lift the lid 5 minutes after shutdown. Slurry handling is messier but predictable. Rexarc, Ozark, and most Chinese-built portables use this design for that reason.
Water-to-carbide generators run hotter and drier, which gives slightly better gas purity and less slurry waste, but they're prone to local overheating if water flow stalls — and on a cold pipeline site where water lines can freeze, that's a real risk. Choose water-to-carbide only for stationary indoor installations with reliable water supply.
References & Further Reading
- Wikipedia contributors. Acetylene. Wikipedia
Building or designing a mechanism like this?
Explore the precision-engineered motion control hardware used by mechanical engineers, makers, and product designers.
