#soda-fountain #carbonator #chilling #diy # Overview Chilling is required, but a commercial remote chiller is not. Coca-Cola's fountain standard is a finished drink of **40°F or colder** at the valve, measured without ice, because warmer drinks foam and lose carbonation. Diet Coke's surfactant sweeteners make the foaming worse. Each 12 oz pour needs about 26 BTU removed in roughly four seconds. No small compressor can do that in real time, so every fountain relies on a reservoir of cold liquid or ice in contact with the heat exchanger. The earlier build most likely failed because the cold plate sat in fridge air, and probably also because the syrup and tower were warm. A bigger chiller was not the missing piece. The cheapest fix is your follow-up idea. Submerge the existing aluminum cold plate in a lidded liquid bath inside the mini fridge, run both the soda water and the Diet Coke syrup through it, and stir the bath. An Inkbird controller with its probe in the bath then holds about 33–34°F. That costs roughly **\$70–\$140 on top of the gear you own**, or about \$125–\$195 with a jug of glycol (estimates; prices captured 2026-10-11). Fill the bath with **30% inhibited food-grade propylene glycol** in distilled water and change it about once a year. The alternative is plain distilled water, changed when it turns cloudy. Skip the bleach-dosed, periodically cycled bath, because every dose leaves behind chloride that pits aluminum. If you need more than about 3–5 drinks in a row, the next step is a \$208–\$239 chest freezer with a plain-water ice bank. Only after that do the costlier options make sense: an \$800–\$1,400 homebrew glycol chiller, a \$1,011 Lindr dry cooler whose pressure rating is unverified, a \$1,599.90 Lancer 500 line chiller, or a \$2,025 remanufactured single-flavor system with a remote chiller. Keep the carbonator outside the fridge on its own CO2 regulator at about 100–105 psi, put the Flojet on a separate regulator at about 65 psi, and keep copper out of both the carbonated-water path and the bath. # Warm Diet Coke foams past Coca-Cola's 40°F line, so chilling is mandatory Coca-Cola's own operator standard answers the first question. Its fountain training tells crews to pour a test drink without ice and stir it for 15 seconds. It should read **40°F or colder**. If the valve has been idle for 30 minutes, they first pour and discard a 20 oz drink. Drinks above 40°F "tend to foam" and lose carbonation ([Coca-Cola Fountain Customer Training](https://crewconnect.coca-cola.com/wp-content/uploads/2022/02/14345-Fountain-Customer-Training-CA-v1.pdf)). Water Technology adds that at 50°F "nearly a quarter of the carbonation will have been lost" ([Water Technology](https://www.watertechonline.com/waters-impact-on-fountain-beverages-and-beverage-systems-part-two)). None of the commercial sources reviewed dispense uncooled soda water. Standalone ambient carbonators like the Big Mac run at 105 psi and feed ice-cooled cold plates. Cold carbonators inside refrigerated units run at 75 psi ([Coca-Cola Fountain Customer Training](https://crewconnect.coca-cola.com/wp-content/uploads/2022/02/14345-Fountain-Customer-Training-CA-v1.pdf); [Lancer Drop-in 1522 manual](https://lancerbeverage.com/uploads/tools/68000027-056-Dropin-1522-Dual-Bargun-T-Block-Manual-270416.pdf)). The physics turns on what happens at the nozzle, not in the carbonator. University of Florida solubility data show that at atmospheric pressure, water holds **1.60 volumes of CO2 at 32°F but 1.01 at 59°F**, while colas target 3.5–4 volumes ([UF/IFAS FS379](https://ask.ifas.ufl.edu/publication/FS379)). The research team extrapolated that table. By their estimate, a 3.7-volume drink leaving the nozzle is about 2.4 times supersaturated at 35°F and about 4.4 times at 70°F. That is why warm soda throws off bubbles so much faster, even though a 100 psi carbonator could in principle exceed 4 volumes at either temperature. Diet Coke makes this worse. Aspartame and potassium benzoate are surfactants that lower surface tension so bubbles form more easily, and diet soda's slightly higher viscosity makes the resulting foam last longer ([McGill Office for Science and Society](https://www.mcgill.ca/oss/article/did-you-know/why-diet-coke-so-fizzy)). The syrup sets the real temperature target. At the standard 5:1 water-to-syrup ratio ([Water Technology](https://www.watertechonline.com/waters-impact-on-fountain-beverages-and-beverage-systems-part-two)), the drink temperature is roughly (5 × T_water + T_syrup) ÷ 6. No Diet Coke-specific ratio was found, so confirm it on the BIB. With 70°F syrup, the soda water has to reach the valve at about **34°F** for the drink to hit 40°F. Water at 36°F already gives about 41.7°F (team estimates). The syrup therefore needs chilling too, and Coca-Cola rules out the obvious shortcut. Its guidelines say to store BIBs at 40–77°F and not in refrigerated coolers, and they warn that diet products are "particularly vulnerable to higher temperatures" ([Coca-Cola Syrup Quality Guidelines](https://www.cokesolutions.com/content/dam/cokesolutions/us/documents/foodservice-quality/foodservice-quality-syrup-quality.pdf)). Instead, run the syrup through its own cold-plate circuit, as commercial cold plates and remote chillers do. With both streams chilled, the practical target is **34–38°F soda water at the valve**. That matches the 34–40°F range an industry cold-carbonation patent gives for soda water ([US7021077B2](https://patents.google.com/patent/US7021077)). # A 26 BTU pour needs stored cold, not a remote chiller A home fountain does not need a commercial remote chiller, because its problem is peak rate, not total energy. The team worked this out from standard constants ([Tempco Engineering Data](https://www.tempco.com/Tempco/Resources/Individual-Catalog-Pages/16-2.pdf)). Cooling a 12 oz drink from 70°F to 36°F removes about **26 BTU**. At a pour rate of 1.5–3 oz/s, that is an instantaneous load of about **12,000–24,000 BTU/hr**. For comparison, a 1/10 HP appliance-class compressor is rated at 380 BTU/h ([HomElectrical, Embraco EMI40HNR1](https://homelectrical.com/refrigeration-compressors/embraco-emb-emi40hnr1.1.html)). Five drinks a day add up to only about 135 BTU (estimate). No compressor at any price chills a pour as it flows. Every working design stores cold in liquid or ice and lets a small compressor recharge it between pours. A commercial remote chiller is mostly this storage plus a stirrer. Lancer's 604 Delta is a 24 lb ice bank with an agitator ([HowdyBrewer](https://www.howdybrewer.com/products/lancer-604-delta-line-chiller-4-circuits)). Soda Dispenser Depot's compact home unit is a 1/4 HP water bath with a 15 W agitator motor ([SDD L3100](https://www.sodadispenserdepot.com/pics/L3100.html)). The stored cold has to be liquid in contact with the plate, and that is where the old setup fell short. By textbook ranges, heat moves from a surface into still air an order of magnitude or more slower than into water (estimate). Assuming the plate weighs 15–20 lb, it stores only about 3–4 BTU per °F (estimate). One drink therefore warms an air-cooled plate by several degrees, and the plate then takes hours to recover. Coca-Cola says a cold plate works best "COMPLETELY covered with ice" ([Coca-Cola Fountain Customer Training](https://crewconnect.coca-cola.com/wp-content/uploads/2022/02/14345-Fountain-Customer-Training-CA-v1.pdf)). HomeBrewTalk posters told a would-be builder that air alone won't do it and "you really need an ice bank" ([HomeBrewTalk, "Cold plate in a refrigerator"](https://www.homebrewtalk.com/threads/cold-plate-in-a-refrigerator.287745/latest)). That same thread has the closest match to your build. A user running a **McCann's Big Mac** with a 100 psi regulator packed the cold plate in water-filled zip-lock bags inside a marine/RV fridge-freezer set to 30°F. The result was **38°F seltzer**, for about \$700–\$800 total and an estimated \$26 a year in electricity ([HomeBrewTalk](https://www.homebrewtalk.com/threads/cold-plate-in-a-refrigerator.287745/latest)). That was plain seltzer, though. With warm Diet Coke syrup, the same 38°F water would mix to roughly 43°F (estimate), which is why the syrup circuit matters. The corny-keg attempt probably failed for reasons beyond the plate. The keg already sat in the fridge, so the water arrived cold and the air-cooled plate added little. The likely culprits are warm syrup, an unchilled tower, and under-carbonation. By team estimates from the UF data, holding 3.7 volumes at 38°F takes about 23 psi, and 4.0 volumes takes about 26 psi. A keg at a typical beer pressure of 10–15 psi delivers only about 2.4–3.2 volumes. Low keg pressure may also have upset flow and ratio at a valve designed for a 75–105 psi supply. These causes are probable, not confirmed. The Big Mac fixes the carbonation and pressure side, which leaves chilling as the remaining gap. # A cold-plate bath in the mini fridge works if it behaves like an ice bank Your follow-up idea is the right one: **keep the existing plate, submerge it in a lidded liquid bath inside the mini fridge, and run both soda water and syrup through it.** This puts the plate back in the job it was designed for, a heat exchanger sitting in a large cold reservoir, and it is the cheapest configuration the evidence supports. The details all target the weak link, which is moving heat from the fridge's evaporator into the bath. Use an opaque HDPE or PP tub just big enough for the plate plus about an inch of fluid around it. Stand the plate on plastic feet so fluid reaches its underside, and press the tub against the freezer box or the cold back wall. Add a small 12 V fan blowing across the evaporator, and a small all-plastic submersible pump aimed across the plate, to break up the stagnant layers of air and fluid. HomeBrewTalk posters made the same point: a plate in a bucket of water in the fridge chills far better than one in air ([HomeBrewTalk](https://www.homebrewtalk.com/threads/cold-plate-in-a-refrigerator.287745/latest)). Plug the fridge into the cooling outlet of an **Inkbird ITC-308**, which listed at \$36 and sold for \$27.36 on October 6, 2026. Put its probe in the bath at plate mid-height, and set a 1–2°F differential and a compressor delay of 5 minutes or more ([HomebrewFinds](https://www.homebrewfinds.com/inkbird-itc-308/)). The KegLand 110 V submersible pump (\$39.99) is sold for circulating ice water through coils ([Williams Brewing](https://www.williamsbrewing.com/Home-Brewing-Equipment/Wort-Making-Equipment/Wort-and-Water-Pumps/Copy-of-Mark-II-Wort-Pump-Head-O-Ring)). Whichever pump you use, its wetted parts must contain no copper (see the next section). Keep every brass or copper fitting above the fluid line. Notch the lid for the tubes and the probe lead, and seal the notches with closed-cell foam so condensation does not drip in and dilute the bath. Expectations should be modest. These are team estimates, because no measured data exist for a mini-fridge bath: - **Storage:** a 3 gal bath allowed to rise 5°F holds about **125 BTU, roughly 4–5 back-to-back drinks**. - **Recovery through fridge air:** only about 20–90 BTU/hr, or 1–3 drinks per hour. - **Recovery with a cabinet fan:** roughly 70–190 BTU/hr. - **After six quick pours:** recovery could take 2–8 hours. Direct contact between the tub and the evaporator is the biggest single improvement. It is not known whether your mini fridge has an accessible freezer-box evaporator. A 50 ft × 3/8" 304 stainless jockey-box coil (\$98.99–\$99.95 at [MoreBeer](https://morebeer.com/products/stainless-draft-coil) and [KegOutlet](https://kegoutlet.com/50-c-stainless-steel-draft-coil-for-jockey-boxes.html)) can replace or supplement the plate on the soda-water side. Retail coils publish no pressure rating. Swagelok rates 3/8" × 0.035" 304 tubing at **3,300 psig** ([Swagelok tubing data](https://dieselduck.ca/machine/08%20programs/2017-Tubing%20Information.pdf)), so at 100 psi the fittings, not the tube, are the weak point. The last foot of line from fridge to valve matters as much as the bath. One HomeBrewTalk builder had a dedicated ice-making undercounter chiller and still saw soda water at about 42°F at best ([HomeBrewTalk, "Chiller and carbonator questions"](https://www.homebrewtalk.com/threads/chiller-and-carbinator-questions.362267/latest)). Keep the run short and insulate it; Lancer specifies at least 19 mm of insulation on its chilled lines ([Lancer R6000S manual](https://lancerbeverage.com/uploads/tools/008-R6000S-Superchiller-Manual-051110-1.pdf)). If the first pour still foams, two fixes are available. KegLand's tower fan kit (\$53.99) blows fridge air up the tower ([Howdy Brewer](https://www.howdybrewer.com/products/draft-tower-fan-kit)). BYO describes a passive copper heat-sink tower that conducts cold up from the evaporator and never touches product ([BYO](https://byo.com/project/kegerator-tower-cooling-an-alternative-chilling-system)). A standard-flow valve setting also helps compared with fast-flow, because calmer flow keeps more CO2 in solution ([US7021077B2](https://patents.google.com/patent/US7021077)). Slower flow also gives the plate more contact time (inference). # Inhibited glycol, not a bleach-cycled bath, protects the aluminum plate The bath never touches the drink. The product circuits are sealed stainless tubes cast into the aluminum, so bath "gunk" is a problem of corrosion, odor and heat transfer rather than food safety, unless a circuit ever leaks. Aluminum chemistry therefore drives the fluid choice. Aluminum stays passive between roughly **pH 4 and 8.5**. Within that range, its main failure is pitting "mainly caused by chlorides." It is also very sensitive to copper: just **20 ppb of dissolved Cu²⁺** can plate out onto it and drive deposition corrosion ([SUEZ Water Handbook](https://www.suezwaterhandbook.com/Water-and-generalities/Corrosion-in-metal-and-concrete/Corrosion-in-metallic-materials/aluminium)). In AMPP lab tests, a copper-bearing solution perforated 1.25 mm aluminum strips within weeks ([AMPP](https://www.ampp.org/blogs/webmasternaceorg/2024/05/29/examining-galvanic-and-deposition-corrosion)). Stainless cast into aluminum is how every cold plate is built, and those plates last for years in melting ice. The real hazard is anything you add to the bath: brass or copper fittings, copper pump parts, copper coils, or copper-based pool or aquarium algaecides. Fill with distilled water, which avoids tap-water chloride and any copper picked up from household plumbing. That chemistry rules out the bleach-cycled bath. Cornelius does sanitize cold plates with household bleach, but only as a short step that "must not exceed 200 PPM of chlorine," followed by cleaning and rinsing ([IMI Cornelius ice-cooled dispenser manual](https://www.cpsohio.com/media/import/manuals/COR/COR002.pdf)). Bleach breaks down into ordinary salt. The team estimates that each 200 ppm dose eventually leaves about **200 ppm of chloride** in the bath. That is roughly eight times the 25 ppm chloride limit Dow sets for glycol dilution water ([Dow DOWFROST Engineering Guide](https://www.blichmannengineering.com/media/wysiwyg/pdf/Dowfrost_Engineering_Guide.pdf)), and it builds up with every dose. No chiller maker reviewed doses bleach into a bath. Lancer allows only an algae and microbe "water conditioner" and specifies an **annual drain and flush** ([Lancer R6000S manual](https://lancerbeverage.com/uploads/tools/008-R6000S-Superchiller-Manual-051110-1.pdf)). An automatic turnover rig would also add pumps, valves and a flooding failure mode inside the fridge. "Cycle it out periodically" is sound only as a manual full replacement: drain, wipe the tub and plate with mild detergent, optionally sanitize at 200 ppm or less and rinse thoroughly, then refill. Inhibited food-grade propylene glycol works as both coolant and anti-gunk additive. Dow states the following ([Dow](https://www.blichmannengineering.com/media/wysiwyg/pdf/Dowfrost_Engineering_Guide.pdf)): - Glycol below **25%** "may be at risk for bacterial contamination." - Its inhibited fluids protect aluminum below 150°F. - In its corrosion table, uninhibited PG removed 1.80 mils/yr from aluminum, while inhibited DOWFROST coupons showed no loss. - Mix with distilled water and hold pH at 8–10. Replace the fluid once pH drops below 7, or if it darkens or turns to sludge. Commercial makers back this up. Lancer's Everest and R6000S chillers both call for a **30% glycol/water mix**, checked monthly ([Lancer Everest manual](https://lancerbeverage.com/uploads/tools/007-Everest-Glycol-System-Chiller-Manual-051110-1.pdf); [Lancer R6000S manual](https://lancerbeverage.com/uploads/tools/008-R6000S-Superchiller-Manual-051110-1.pdf)). An instructor at the Micro Matic Dispense Institute recommends changing glycol yearly and warns that automotive testers misread PG ([Micro Matic forum](https://www.micromatic.com/forum/us-en/forum/bar-pub-restaurant-dispensing-commercial/system-components-beer-side/1841-how-often-do-you-change-the-glycol-in-your-systems)), so use a PG-scale refractometer. Never put automotive ethylene glycol near the system: about 4 fl oz can kill an adult ([MedlinePlus](https://medlineplus.gov/ency/article/000774.htm)). Pink RV antifreeze costs from \$4.27 a gallon, but its composition is undisclosed ([Walmart](https://www.walmart.com/c/kp/winterize-rv)), and ethanol blends are flammable and dry out seals ([RVtravel](https://www.rvtravel.com/rv-antifreeze-myths-busted-mistakes-avoid-1234/)). Use it only as a test fill while debugging. | Bath fluid | Freeze floor | Aluminum | Biofilm control | Change interval | Cost (captured 2026-10-11) | |---|---|---|---|---|---| | **30% inhibited food-grade PG in distilled water** (default for the mini fridge) | None; the bath itself freezes near 9°F ([Dow](https://www.blichmannengineering.com/media/wysiwyg/pdf/Dowfrost_Engineering_Guide.pdf)), so the controller must protect the product circuits | Protected by inhibitor | Resists bacteria at ≥25% | Check with PG refractometer; replace about yearly | CellarScience 1 gal \$52.99 ([MoreBeer](https://morebeer.com/products/food-grade-propylene-glycol-100-gallon-cellarscience)); DOWFROST HD 1 gal \$66.56 ([Walmart](https://www.walmart.com/ip/5434721925)); about 0.6–0.9 gal per 2–3 gal bath (estimate) | | **Plain distilled water**, lidded and dark, optional 3% H₂O₂ | Cannot drop below 32°F while liquid; allows a partial ice bank | Fine if free of chloride and copper | Weak; relies on cold, darkness and lid | When cloudy, roughly monthly at first (team suggestion); Lancer drains ice banks yearly | Pennies | | Bleach-dosed or auto-cycled water | Same as water | Chloride builds up, so pitting risk | Good while dosed | Continuous | Not recommended | | PG-only RV antifreeze | Low | Inhibitor not rated for continuous chiller duty | Concentration undisclosed | About 6 months, test use only (team suggestion) | From \$4.27/gal | | Automotive ethylene glycol | — | — | — | Never | Toxic | If you choose plain water, hydrogen peroxide is the aluminum-friendly additive. Solvay names high-purity aluminum as the most compatible material for storing H₂O₂, and peroxide leaves no chloride behind ([Solvay](https://www.solvay.com/sites/g/files/srpend221/files/2019-10/H2O2%20Materials%20of%20Construction%20for%20the%20Storage%20of%20H2O2.pdf)). A dose of 1–2 tablespoons of 3% peroxide per gallon after each refill is the team's suggestion, not a sourced figure. Do not add peroxide or bleach to a glycol bath, because oxidizers use up the inhibitor (inference). Glycol's catch is freeze protection. A plain-water bath cannot drop below 32°F while it is liquid, which is why ice banks never freeze soda circuits. A 30% glycol bath has no such floor. A stuck relay or a probe that falls out could let the fridge pull it toward 10°F. The team estimates that carbonated water at about 4 volumes freezes near 31.4°F. Diet syrup contains far fewer dissolved solids than sugared syrup ([University of Illinois Physics Van](https://van.physics.illinois.edu/qa/listing.php?id=16983)), so it probably freezes only slightly below 32°F (unverified). Do not copy beer-chiller settings such as the Lancer Everest's −2°C default setpoint. The Everest itself alarms 4° below setpoint ([Lancer Everest manual](https://lancerbeverage.com/uploads/tools/007-Everest-Glycol-System-Chiller-Manual-051110-1.pdf)). For a glycol bath: - Run it at **33–34°F**. - Set a low-temperature alarm near 31°F. - Keep the probe next to the plate, not near the evaporator. - Set the fridge's own dial to the warmest position that still reaches setpoint, so it acts as a crude backup limit. If that discipline is unappealing, or if the build moves to a chest freezer with an ice bank, plain distilled water is the better fluid. # Step up to a chest freezer long before a \$1,600 chiller Ranked cheapest first, on top of the gear you already own. Burst figures are team estimates from bath mass and ice storage. | Rank | Setup | Price evidence (captured 2026-10-11) | Burst at ≤40°F (estimate) | Choose when | |---|---|---|---|---| | 1 | Existing plate in a 2–3 gal bath in the mini fridge, ITC-308, pump, fan | ~\$70–\$140, plus ~\$53 if using glycol (estimate) | ~3–5 drinks; 2–8 h recovery | A few drinks a day | | 2 | 5–7 cu ft chest freezer, 5–8 gal plain-water bath with partial ice bank, plate and/or SS coil | Vissani 5 cu ft \$207.90 ([Home Depot](https://homedepot.com/p/VISSANI-5-cu-ft-Manual-Defrost-Chest-Freezer-in-White-MDCF5WH/314186968)); Magic Chef 7 cu ft \$239.00 ([Home Depot](https://homedepot.com/p/Magic-Chef-7-0-cu-ft-Chest-Freezer-in-Black-HMCF7B4/311895693)); ~\$280–\$420 all in | ~8–12 drinks from water alone, plus ~27 per 5 lb of ice | Back-to-back pours, guests | | 3 | Homebrew glycol chiller feeding a coil bath | BrewBuilt IceMaster Max 2 \$799.99; Blichmann 1/2 HP \$1,399 ([MoreBeer](https://morebeer.com/category/glycol-chillers.html)); Ss Brewtech 1/5 HP \$999 ([MoreBeer](https://morebeer.com/products/glycol-chiller-15hp-ss-brewtech)) | Ample | Wants a packaged unit or a long tower run | | 4 | Lindr Pygmy 25/K 115 V dry cooler (soda line only) | \$1,011.15 new, 38 lb ([TapYourKeg](https://tapyourkeg.com/products/pygmy-25-k-115v-60hz-green-line-2)) | 25–35 L/h continuous | Only after getting a written pressure rating | | 5 | Lancer 500 Series 2-circuit line chiller | \$1,599.90 new ([HowdyBrewer](https://www.howdybrewer.com/collections/line-chillers)) | Not stated | Cheapest new standalone remote chiller found | | 6 | Soda Dispenser Depot reman 1-flavor tower system with 1/4 HP remote chiller | \$2,025 + freight ([SDD t00201R](https://www.sodadispenserdepot.com/products/systems/remanufactured/tower/t00201R.html)); \$3,560 new ([SDD L3100](https://www.sodadispenserdepot.com/pics/L3100.html)) | Purpose-built | Turnkey, but duplicates your tower, pump and carbonator | | 7 | Lancer 604 Delta (24 lb ice bank) or Lancer 1500 1/3 HP | \$2,094.18 ([HowdyBrewer](https://www.howdybrewer.com/products/lancer-604-delta-line-chiller-4-circuits)); \$2,316.66–\$2,352.84 ([HowdyBrewer](https://www.howdybrewer.com/products/lancer-chiller-1500-1-3hp-115v-60-6-circuits)) | Purpose-built | Plans for more valves | Use Coca-Cola's no-ice thermometer test as the trigger. If step 1 fails it after three or four back-to-back pours, the next rung is a chest freezer, not a commercial chiller. A chest freezer's walls are its evaporator, so a bucket pressed against them should recover much faster. The team estimates a few hundred BTU/hr, but this has not been measured. A plain-water bath also lets you grow a partial ice bank, and each pound of ice stores 144 BTU, enough for about five drinks ([Tempco](https://www.tempco.com/Tempco/Resources/Individual-Catalog-Pages/16-2.pdf)). Freeze risk is also highest in a freezer: HomeBrewTalk posters warn that liquid can freeze inside a plate kept there ([HomeBrewTalk](https://www.homebrewtalk.com/threads/cold-plate-in-a-refrigerator.287745/latest)). The team suggests an untested DIY control that mimics commercial ice-thickness sensing. Mount the controller probe ½–1" from the coldest wall, so ice grows out to the probe and then stops. Never let the open water around the plate fall below 32°F. Glycol chillers earn their \$800–\$1,400 only if you want quiet packaged operation or need a long python run. Even then, a python only keeps cold product cold; it "won't cool" warm product ([AHA forum](https://forum.homebrewersassociation.org/t/home-glycol-system/6880)). The purpose-built options carry caveats: - **Lindr Pygmy 25/K:** the cheapest new single-line electric chiller found. Neither the listing nor Lindr's manual gives a maximum coil working pressure, and the manual says to set the thermostat no colder than "5" for non-alcoholic drinks to avoid freezing ([Lindr manual](https://www.lindr.cz/storage/helios/sheet/1aba876a6e2a4a4016e3e010bac7413e.pdf)). Get a written rating before placing it after a 100 psi carbonator. It chills one line only, so the syrup would stay warm. - **Lancer line chillers at HowdyBrewer:** sold mainly into the draft-beer market. Whether their coils fully chill soda water from room temperature, rather than just holding it cold, is unconfirmed. - **Soda Dispenser Depot compact chiller:** 1/4 HP, 7 A, 10 × 20 × 23.5 in, 108 lb shipped and 145 lb filled. It is the class of unit the vendor markets for home fountains, but it was found only bundled into complete systems ([SDD t00201R](https://www.sodadispenserdepot.com/products/systems/remanufactured/tower/t00201R.html)). Hidden costs: - **Freight:** budget \$152.71–\$499 for heavy units, as seen on eBay ([eBay](https://www.ebay.com/shop/lancer-soda-machine?_nkw=lancer+soda+machine)). - **Python:** \$11.55–\$14.65/ft for a 2-line trunk if the run is long ([WholesaleBeerParts](https://wholesalebeerparts.com/collections/2-products)). No used standalone chiller prices turned up. The one used comparison was a complete home single-valve fountain with a remote refrigerated chiller, listed at \$1,500 OBO on Craigslist ([Craigslist](https://www.craigslist.org/view/d/state-park-complete-home-soda-fountain/7K5ZrjAYDDPb591mS1zbAU)). Local Craigslist and Facebook Marketplace finds remain the wild card that could undercut new prices. Two units commonly called "electric" are actually ice-cooled. - **Lancer IBD 4500** is an ice/beverage dispenser with 180 lb of ice storage. It draws 3 A at 115 V, ships at 255 lb, and takes its carbonated water from an outside source ([Lancer IBD4500-22 spec](https://lancerbeverage.com/uploads/specsheets/IBD4500_22_SS.pdf)). - **Cornelius Enduro 150**'s spec sheet states that "Ice used to cool cold plate is separate from ice dispensed into cups." It draws 3 A and ships at 240 lb ([Cornelius Enduro 150 spec](https://cdn.shopify.com/s/files/1/2336/4729/files/ED150_unit.pdf)). A 3 A draw runs agitation and valves, not a compressor. An eBay title like "Lancer 4500 12 Head Electric Ice Beverage Dispenser" (\$2,575.99 plus \$350 shipping, [eBay](https://www.ebay.com/shop/lancer-soda-machine?_nkw=lancer+soda+machine)) does not mean refrigerated, and these units still need daily ice. Neither is a remote chiller. Truly ice-free countertop units do exist. Soda Dispenser Depot's CX2000 2-flavor unit costs \$3,845 new ([SDD CX2000](https://secure.ultracart.com/catalog/SODA/soda-fountain-systems/new/CX2000.html)), and its reman 6-flavor units at \$2,150–\$2,535 are out of stock ([SDD CE00600](https://secure.ultracart.com/catalog/SODA/soda-fountain-systems/remanufactured/CE00600.html)). Lancer's small S2H Superchiller is a 240 V/50 Hz Australian unit with its own carbonator built in ([Lancer S2H](https://lancerbeverage.com/product/s2h-superchiller/)). Several other options fail on specs or cost: - **Aquarium chillers** fail on setpoint: the Active Aqua 1/10 HP (\$358.19) bottoms out at 39°F ([Hydrobuilder](https://hydrobuilder.com/products/active-aqua-water-chiller-with-power-boost-1-10-hp)). - **Drinking-fountain chillers** are preset to 50°F ±5°F and rated for 40–105 psi, which leaves no margin for carbonator output. Examples are the Elkay ER101Y at \$2,627.48 ([BottleFillingStations](https://bottlefillingstations.com/products/elkay-er101y-remote-chiller)) and the Elkay ECH8 at \$637.86–\$1,463.88 new on eBay ([eBay](https://www.ebay.com/p/24034142098)). - **A new 5,000 BTU window AC** for a DIY glycol build now costs \$314.32, or \$264.32 with a Home Depot card ([Home Depot](https://homedepot.com/p/Midea-5-000-BTU-115-Volt-Mechanical-Cool-Only-Window-Air-Conditioner-in-White-MWA05CM71/305311272)). That is more than a chest freezer, and the build risks damaging refrigerant lines ([HomeBrewTalk window AC build](https://www.homebrewtalk.com/threads/the-ole-window-ac-glycol-chiller-build.650039/latest)). - **Elkay's DSFBF180K carbonation chiller** (115 V, 3.2 A, 20 GPH, 37–50°F, internal carbonator; [KaTom](https://www.katom.com/189-DSFBF180K.html)) could replace the McCann's. It has no syrup circuit, though, so drinks would land near 42–43°F with warm syrup (estimate). Its price was not captured. # Hook up the McCann's with two regulators and no copper past the check valve | Item | Setting | Source | |---|---|---| | Carbonator CO2 | ~100–105 psi; tank relief is 180 psi | Coca-Cola standalone 105 psi ([Coca-Cola training](https://crewconnect.coca-cola.com/wp-content/uploads/2022/02/14345-Fountain-Customer-Training-CA-v1.pdf)); Lancer 90–110 psi ([Lancer 1523 manual](https://manualmachine.com/lancer/1523/5982037-user-manual/)); relief ([McCann's spec](https://res.katom.com/products/055/055-E400397/055-E400397.pdf)) | | Flojet BIB pump CO2 | ~65 psi on a **separate** regulator; 20–80 psi range, never above 80 | [Flojet N5000 datasheet](https://wholesale.yodify.com/Documentation/Document/nwgWn/297097-N5000-Series-Technical-Datasheet.pdf); [Coca-Cola training](https://crewconnect.coca-cola.com/wp-content/uploads/2022/02/14345-Fountain-Customer-Training-CA-v1.pdf) | | Water supply | Regulate to ~40–50 psi, at least 10–20 psi below CO2, with 20–25 psi minimum while flowing; above 70 psi the water may not carbonate | [Cornelius VA13](https://www.cpsohio.com/media/import/manuals/COR/COR0051.pdf); [Lancer Superchiller manual](https://lancerbeverage.com/uploads/tools/68000002-036-5-S4E-V3-Hi-Carb-Superchiller-Manual-120320.pdf); [Water Technology](https://www.watertechonline.com/waters-impact-on-fountain-beverages-and-beverage-systems-part-two) | | Pump protection | McCann's Pump Saver 16-0856: shuts off below 15 psi, restarts at 20 psi or more | [McCann's spec](https://res.katom.com/products/055/055-E400397/055-E400397.pdf) | | Filtration | Carbon/scale cartridge, e.g. Everpure 4CB5-S at \$112.99, rated 6,000 gal or 12 months; Coca-Cola says usually every 6 months | [Howdy Brewer](https://www.howdybrewer.com/products/everpure-qc-cartridge-4cb5-s-09d02122); [Coca-Cola VoG](https://www.cokesolutions.com/content/dam/cokesolutions/us/documents/foodservice-quality/foodservice-quality-ValvesOfGoldBrochure.pdf) | | Backflow | ASSE 1022 vented dual check with ≥100-mesh screen, upstream of the carbonator and downstream of any copper; never plug the vent | [Virginia 2VAC5-585-2270](https://law.lis.virginia.gov/admincode/title2/agency5/chapter585/section2270/); [Minnesota Dept. of Health](https://www.health.mn.gov/communities/environment/water/docs/factsheet/bevdisp.pdf) | | Ports | 3/8" male flare water in, 1/4" male flare CO2 in, two 3/8" male flare soda out | [McCann's spec](https://res.katom.com/products/055/055-E400397/055-E400397.pdf) | The team's suggested order on the water side is shutoff, filter, regulator, Pump Saver, ASSE 1022, then carbonator. Food Code as adopted in Minnesota bars copper between the backflow device and the carbonator ([Minnesota Rule 4626.0465](https://www.revisor.mn.gov/rules/4626.0465/version/2019-01-02T09:58:45-06:00)). Cornelius forbids copper tubing and brass fittings anywhere CO2 can reach them ([Cornelius VA13](https://www.cpsohio.com/media/import/manuals/COR/COR0051.pdf)). The reason: if a check valve fails, carbonated water backs into copper lines and dissolves the copper. Poisoning symptoms appear within 5–30 minutes, and syrup and cold mask the taste ([Contractor](https://www.contractormag.com/plumbing/article/20874422/post-mix-poisoning-must-be-addressed)). The used unit's backflow device is of unknown age. Inspect the vent, and replace the device if it leaks ([Minnesota Dept. of Health](https://www.health.mn.gov/communities/environment/water/docs/factsheet/bevdisp.pdf)). The same no-copper rule protects the aluminum plate in the bath, so one rule covers both. Keep the carbonator outside the fridge. A HomeBrewTalk thread warns that many carbonators have open motors and unsealed controls that are unsafe in a humid fridge ([HomeBrewTalk](https://www.homebrewtalk.com/threads/carbonator-co2-chiller-etc-for-home-seltzer-tap.250351/latest)). McCann's says the Big Mac's electronics are encapsulated against moisture ([KegOutlet](https://www.kegoutlet.com/sc-bm2-mccanns-big-mac-soda-carbonator-115v.html)). Even so, its 1/3 HP motor draws 6.5 A ([McCann's spec](https://res.katom.com/products/055/055-E400397/055-E400397.pdf)). Inside a small cabinet it would dump heat, and its fittings would sweat in a humid Connecticut summer (inference). Ambient carbonation at about 105 psi is what it was designed for. Electrical: the carbonator plus a 7 A purpose-built chiller would load a 15 A circuit close to its limit. Steps 4–7 therefore warrant a dedicated circuit (estimate). CO2 safety: - CO2 is about 1.5 times heavier than air and pools in basements. Use CO2 detectors, not oxygen monitors ([Missouri Division of Fire Safety](https://dfs.dps.mo.gov/safetytips/carbondioxide.php)). - OSHA's exposure limit is 5,000 ppm, and NIOSH's immediately-dangerous level is 40,000 ppm ([NIOSH](https://cdc.gov/niosh/idlh/124389.html)). - Flojet calls for good ventilation where its gas-driven pumps exhaust CO2 ([Flojet N5000 datasheet](https://wholesale.yodify.com/Documentation/Document/nwgWn/297097-N5000-Series-Technical-Datasheet.pdf)). - To leak-check, close the cylinder and confirm the gauge holds steady ([Coca-Cola VoG](https://www.cokesolutions.com/content/dam/cokesolutions/us/documents/foodservice-quality/foodservice-quality-ValvesOfGoldBrochure.pdf)). # Conclusion A remote chiller is three things the first attempt lacked: a cold liquid reservoir, agitation, and a thermostat that reads the reservoir. Your follow-up idea rebuilds all three from parts you mostly already own. Compressor capacity barely matters at home volumes. What decides the result is whether liquid touches the plate, whether the syrup is chilled, and how much heat the last foot of line and tower picks up. All three plausibly contributed to the earlier disappointment. Replacing a commercial chiller also means taking over its built-in freeze protection. A glycol bath needs an alarm and a backup temperature limit. A water bath with an ice bank needs some control over how thick the ice grows. The sensible sequence is to spend roughly \$100–\$200 on the step-1 bath first. Then run Coca-Cola's no-ice, 15-second-stir thermometer test after three or four back-to-back pours, and let that reading decide whether a chest freezer is worth adding. The unknowns that most affect the outcome are the mini fridge's evaporator layout, the cold plate's circuit count and internal volume, and, before buying a Lindr, a written coil pressure rating. # TODO - ~~Write a cleaning & sanitizing note for the final setup~~ (done 2026-10-11): [[Home Soda Fountain Cleaning and Sanitizing]] - ==Confirm the Diet Coke water:syrup ratio on the BIB label== - ==Check the mini fridge's evaporator layout and the cold plate's circuit count== - ==Inspect the used carbonator's ASSE 1022 backflow device; replace if it leaks== - ==After building the bath, run the no-ice 40°F test after 3–4 back-to-back pours== # Changelog ## 2026-10-11 - Note created from a deep-research report: chilling requirements, DIY chilling options, commercial chiller market, cold plate bath fluid - Prices captured 2026-10-11; sources linked inline - Resolved the cleaning-note TODO; linked [[Home Soda Fountain Cleaning and Sanitizing]]