High-temperature short-time pasteurization heats fluid milk to at least 161 F (72 C), holds it there for at least 15 seconds, and cools it to 40 F or below in a single continuous pass. The heat exchanger is not one exchanger but three or four in one frame: a regeneration section where cold raw milk meets hot pasteurized milk through the plate, a heating section fed by hot water, a holding tube that fixes the residence time, and a cooling section on chilled water or glycol. A flow diversion valve at the end of the holding tube sends any sub-temperature milk back to the balance tank.
Regeneration carries most of the load. Raw milk at 40 F is lifted to roughly 145 to 155 F by the outgoing product, so the hot water section only has to add the last 10 to 20 degrees and the cooler only has to remove what regeneration left behind. Regeneration efficiency of 85 to 95 percent is normal on a gasketed plate unit; each point of efficiency is heat you do not buy twice.
Request a Quote
Raw milk at 40°F picks up heat from the outgoing pasteurized stream until it is within 10 to 20 degrees of setpoint, at 85 to 95 percent efficiency. A 15 second hold at 5 ft/s needs roughly 90 feet of sloped tube, timed on the fastest particle.
The product decides the section type; the utility and the pressure decide the rest. These are the pairings HeatX quotes most often.
| DUTY / CONDITION | RECOMMENDED CONSTRUCTION | WHY |
| Fluid milk, skim to whole, 5,000 to 60,000 lb/h | Gasketed Plate | Highest regeneration per square foot, opens for inspection, cheapest heat transfer area in the plant |
| Cream, ice cream mix, 30 to 40 percent fat | Wide-Gap Plate | Twice the channel gap keeps velocity and pressure drop sane at cream viscosity; plates still regenerate well |
| Milk with pulp, cocoa or fiber; flavored milk | Corrugated Tube | Tube-in-tube passes solids, corrugation keeps the coefficient up on the product side |
| Hot water set for the heating section on plant steam | U-Tube | Steam on the tube side, free expansion, no bellows; keeps live steam off the product exchanger |
| Chilled water or glycol final cooler on a tubular line | Straight Tube | Fully drainable, brushable, no U-bends to trap product; sanitary tri-clamp both ends |
| Regeneration section where a plate leak cannot be tolerated | Double-Wall Plate | Two plates with a vented air gap between raw and pasteurized milk; any leak shows at the vent, not in the product |
Start with the holding tube, because it is the one component the inspector times with a stopwatch. Its length is set by the fastest particle, not the average velocity: for turbulent flow the fastest particle travels about 1.2 times the mean velocity, so a 15 second legal hold at a mean velocity of 5 ft/s needs a tube on the order of 90 feet, sloped upward at least 1/4 inch per foot with no restrictions, and the indicating thermometer and recorder sensor at its outlet. Change the flow rate and the hold time changes with it, which is why the timing pump is a positive displacement pump or a magnetic flow meter loop with a sealed setpoint.
The regeneration section is sized on the approach you are willing to pay for. At 90 percent regeneration the raw milk leaves the regenerator about 12 degrees below the pasteurized temperature and the pasteurized milk leaves it about 12 degrees above the raw inlet. Pushing to 95 percent halves that approach and roughly doubles the plate count in the section; the payback is usually two to four years on a line that runs two shifts. The pasteurized side must run at least 1 psi higher than the raw side through the whole section so a pinhole leaks the safe direction, which is why the timing pump draws raw milk through the regenerator and any booster pump goes on the raw side, interlocked so it stops whenever that margin is lost.
The cooling section removes what regeneration leaves. With 90 percent regeneration the product arrives at the cooler around 52 F and must leave at or below 40 F; on 34 F chilled water that is a routine duty, on tower water it is impossible, and on propylene glycol at 28 F it takes about half the plates. If the product then goes to a filler at 36 F the cooler is sized on glycol and the chilled water section drops out.
The PMO requires that pasteurized product in the regenerator be at higher pressure than raw product, and that the relationship hold during the whole run, during CIP, and when the flow diversion valve is in divert. In practice that means a raw-side booster pump interlocked to the timing pump and the diversion valve, a differential pressure switch across the regenerator, and a design where the raw side drains before the pasteurized side when the system shuts down.
The flow diversion valve is a three-way sanitary valve at the outlet of the holding tube, driven by the safety thermal limit recorder. It has to move to divert in under one second, it has to be leak-detect type so that product can not pass to forward flow during a divert, and the exchanger pressure drop calculation has to include the divert leg so that the timing pump cannot over-speed the holding tube when the forward leg closes.
A fluid milk pasteurizer with a 5 degree hot water approach and 1 psi of gasket margin should run 8 to 12 hours before the pressure drop across the heating section rises 20 percent. Cream lines run shorter, high-solids mixes shorter still. If the plant needs a 20 hour run, the fix is more area in the heating section at a lower wall temperature, not a bigger pump.
Product and fat content, flow rate in lb/h or gpm, raw inlet temperature, pasteurization setpoint and hold, final outlet temperature, hot water and chilled utility available, and whether the unit is plate, tubular or your choice. With those in hand a pasteurizer is a one-conversation quote.
Call and talk it through with an engineer: 1-805-484-2992
Usually two: the product with the highest legal setpoint sizes the heating section and hot water loop, and the thickest product sizes the cooler and the holding tube. Under the PMO, products with 10 percent fat or more, or added sweetener, pasteurize at 166 F for 15 seconds instead of 161 F, and eggnog at 175 F for 25 seconds, the pair state frozen dessert rules also apply to ice cream mix. A plant running whole milk at 40,000 lb/h and cream at 8,000 lb/h sizes the hot water loop for the 166 F duty and the regenerator for the milk flow, then gives the cream its own timed, sealed setting. If the cream goes laminar, its fastest particle runs at 2.0 times mean instead of 1.2, so its hold cannot be scaled from the milk test: at the milk's flow, the tube that held milk 15 seconds would hold laminar cream about 9.
Yes, the PMO allows it, and the heating section and the holding tube both change. The listed pairs climb from 161 F for 15 seconds to 191 F for 1 second, 194 F for 0.5 second and 212 F for 0.01 second. At 191 F the hot water loop runs about 196 to 201 F for the same 5 to 10 degree approach, which puts the wall squarely in the range where whey protein deposits fastest, so plan 20 to 30 percent more heating area than the 161 F design and shorter runs between cleanings. A 1 second hold at a 5 ft/s mean velocity is only about 6 feet of tube on the fastest particle, so valve response and sensor lag become a large share of the hold, and at the 212 F pair the line needs back pressure so the milk does not flash.
The holding time, the diversion valve, the recorder and the regenerator pressure relationship, in that order of attention. Expect a salt conductivity timing of the tube at the sealed pump setting, a check that the valve diverts when the recorder pen crosses the cut-in line, a comparison of the indicating thermometer against the recorder, and proof that the pasteurized side stays above the raw side in forward flow, divert and shutdown. Put the test ports, sample valves and sensor fittings where a person can reach them with a meter and a stopwatch; a sensor well three feet above the platform turns a short inspection into a scaffold job.
Yes; both fix where the regenerator is split and what pressure the sections see. A separator wants milk at 110 to 130 F, so the raw regenerator is built in two sections with a draw-off between them, and the skim and standardizing cream return into the second section at nearly the same temperature. A homogenizer between the regenerator and the heating section has a pulsing suction, so the plates feeding it need a positive head and the regenerator pressure relationship is proven with the homogenizer running, not read off the pump curve. Settle both positions before the plate count is final; moving a draw-off after the frame is built means new connection plates and a new pressure test.
Fouling in the heating section, which carries the hottest wall; the regenerator and the cooler run clean for days. The tell is the hot water valve opening a little wider each hour to hold 161 F, with the heating section pressure drop climbing behind it; by the time the outlet temperature wavers the plates are coated. The lever that works is wall temperature: at 90 percent regeneration the heating section adds only the last 12 degrees, so a 5 degree approach means water at 166 F, while an undersized section run on 180 F water puts the wall 19 degrees above the milk and fouls in half the time. Air is the second cause, since dissolved air released at the hot wall seeds deposit, so keep the balance tank float working and the raw line free of leaks that draw air. To stretch 8 hours to 16, add heating plates and lower the water temperature; a bigger hot water pump does nothing.
Often, if the carrying bar has room and the plate pattern is still made, but everything downstream of the plates has to follow. Going from 40,000 to 50,000 lb/h raises the velocity in the holding tube 25 percent and cuts the hold from 15 seconds to 12, so the tube is lengthened and retimed, the timing pump is resized and resealed, the diversion valve and booster interlock are rechecked, and the chilled water or glycol has to carry 25 percent more load. Where the bar is full, a tubular regenerator or cooler section beside the frame takes the added duty without a new skid. Give us the plate count, the free bar length and the current sealed flow and we will tell you which route costs less.
Indirect sterilizers for 275 to 285 F and ESL heaters for 250 to 260 F, built to run a shift between CIPs.
+ Learn More
Raw milk to 38 F in one pass, cream heated without channeling, whey cooled without scaling.
+ Learn More
Mix heaters, culture-temperature coolers, cheese-milk warmers and whey heaters that respect cultured product.
+ Learn More