Views: 0 Author: Site Editor Publish Time: 2026-11-11 Origin: Site
Almost every sigma knead recipe carries a time, ninety minutes or two hours, and somebody ran a trial years ago before the number got copied forward ever since. The operators are new, the suppliers are new, and one reformulation nobody wrote down.
Time is not the goal; it is a proxy. What the recipe actually wants is a material state — filler wetted through, no dry pockets, viscosity on the plate, temperature inside its band, colour even. Time tracks that state under one set of conditions and no others. Change the conditions and the correlation dies quietly. So read the end point off the batch, not off the clock.
Three properties you can measure, one you can see. Viscosity stops falling and turns upward. Temperature sits at the jacket set point instead of drifting. Filler wets out completely, and a sample tells you so. Then a spatula from the front of the trough matches one from the back.
Anything before that is a guess. Anything after it is waste.
Four variables drift under a fixed recipe.
Incoming material comes first: filler off a humid warehouse needs longer to wet out than last week's delivery. Resin viscosity shifts lot to lot. Neither one shows up on the batch sheet.
Fill ratio matters next. A trough at 60 percent does not knead like one at 80 percent at the same blade speed, because the blade area is the same but the mass is larger, which means more heat and slower surface renewal. Ninety minutes then produces a different material.
Jacket temperature plays its part as well. Cold walls raise the effective viscosity near the wall, so the batch kneads unevenly, hot in the middle and stiff at the edges. The gauge average looks perfect, but the compound is not.
Blade wear is the last variable. A blade that has lost a few millimetres off its working edge stops wiping the one opposite. The high-shear zone at the saddle fades. Nobody notices until the dispersion complaints arrive eighteen months later.
Every knead draws a power curve. That curve knows more about the material than any timer does.
Loading spikes first. Dry filler hits moving blades and torque peaks before anything is wet. Current drops as binder coats the particles and the mass starts to flow. Then it climbs as viscosity builds and the filler network forms. Then it flattens. That plateau is what fully wetted, evenly kneaded looks like in electrical terms.
Knead past it and the curve rises a second time. That rise isn't better mixing. It's solvent loss, cooling, or partial curing. The batch is getting harder to discharge and harder to use.
Curve segment | What the batch is doing | What it means for the end point |
Loading spike | Dry filler hits moving blades; torque peaks before anything is wet | Expected; size the drive for it, don't judge the batch by it |
Current drop | Binder coats the particles and the mass begins to flow | Normal progress |
Second climb | Viscosity builds and the filler network forms | Keep kneading |
Plateau | Fully wetted, uniformly kneaded | Candidate end point — valid only while temperature stays stable |
Second rise | Solvent loss, cooling or partial curing | Stop; the batch is getting harder to discharge and to use |
Three things let you read that plateau: a variable frequency drive, so blade speed is really held constant; a current signal the control system can log; and a stable temperature. Lose any one of them and the shape blurs. Karvil's 3000L explosion-proof sigma mixer is built around exactly that: Siemens drives, a Siemens PLC and HMI, and an explosion-proof cabinet. Process data goes to the control layer instead of a needle on a panel. Variable speed is worth a look here: guide to VFD control on sigma kneaders.
Power draw and temperature travel together. A batch that heats past its set point thins out, draws less current, and gives you a plateau that looks like completion but isn't. A cold jacket does the reverse. Current stays high while the material at the blades is already overworked.
The rule is blunt. No current plateau counts unless temperature held stable across the same interval. When the two disagree, fix the jacket first.
Stage one costs nothing. Operators log amps on paper at fixed intervals. The shift supervisor reads the logs against dispersion results. Patterns show up inside a month, usually along the lines of "the afternoon batches run hotter and finish sooner than they should".
Stage two adds instruments. A drive with an analog output feeding a PLC gives you a continuous trend, exportable per batch.
Stage three closes the loop. The recipe holds a target current window plus a minimum dwell inside it, and the batch ends when both are met. Alarms fire on the exception, not on the clock.
This is where batch records start to earn their keep. A customer audit asks how you know a batch was fully mixed, and you hand over a data file instead of a shrug. Still comparing machines? A 5L laboratory vacuum kneader is the cheapest place to build that curve, before production scale makes the mistakes expensive.
Every plant has that one operator who finishes the batch early. Usually they're right. Usually nobody has asked why.
Ask why and the answers stay concrete. Edges looked dry, so in went oil. Material climbed the blade instead of rolling, so speed came down for ten minutes. The last delivery of filler felt coarser in the hand. Eyes and a spatula, reading the same signals the instruments read.
Leaving those corrections invisible is the mistake. An operator who shortens a cycle should log the reason beside the number, for example 2 kg of oil at 60 minutes, half speed and a stiff discharge. Do that for a few months and the log explains half the dispersion complaints a plant gets.
Blade and seal wear speed up once the material stiffens: shaft seals run hotter, packing goes faster, and vacuum loss follows. On a vacuum kneader that's the start of a long, expensive sequence.
Quality moves as well, because colour shifts and fines rise, since agglomerates — the particle clusters the knead was meant to break down to a useful size — get crushed past usefulness. Solvent loss changes the rheology the formula was built around, and the next batch inherits a machine that's a little further out of adjustment.
Energy is the smallest of the three costs, and the only one that shows up on a utility bill. It's rarely the reason to fix this.
A clock says when a batch used to be done. The curve says when this one is. Log current and temperature together. Set a plateau window per formula, at the fill ratio you actually run. Write operator overrides into the batch record beside the number they replaced. The ninety-minute figure can stay as a sanity limit. The plateau decides, not the minute hand. And where curve and operator disagree for more than a few batches, that's evidence on its own: an instrument drifting, or a material that changed. It is worth finding out which.
Karvil supplies sigma kneaders with VFD drives and PLC/HMI control that log current, temperature and speed per batch, from lab units up to production vessels. Send over what you run now — time, fill ratio, jacket temperature, dispersion complaints — to info@karvil.com. Or start from the mixer range, and the team will help turn it into a curve you can hold.
Yes. On a given drive, torque is proportional to current, and plenty of VFDs report estimated torque directly. The practical difference is instrumentation. Current is available on almost every drive. Torque usually needs a parameter enabled and a scaling check against a known load.
No. Dispersion improves until the agglomerates are broken and wetted, then it plateaus. Past that, extra time buys heat, wear, and — in shear-sensitive compounds — a filler network rebuilding in the wrong shape.
Start with a panel ammeter and a manual log. The curve shape is visible either way. You lose resolution, not the signal. Retrofit a drive with an analog output when the budget allows.
Run three trials at the same fill ratio and jacket temperature. Log current against time. Sample dispersion at intervals. The plateau where dispersion stops improving defines the window. Write the fill ratio beside it, because the number means nothing without it.
It changes the curve. Pulling vacuum removes air and solvent, and both raise apparent viscosity, so current climbs earlier and the plateau sits higher. Set the window with vacuum on, at the level you actually run in production.
