Vellmark Foodlab — Sustained Energy Group
SATIA-30 SUSTAINED-RELEASE UNIT
One unit holds a single supply line steady for thirty days and releases it on a curve that stays flat. One dose, one morning, one cycle: taken on the first day, finished on the thirtieth.
The thirty-day profile is plotted at Fig. 1, the composition matrix is printed in full, and nine consecutive batch records sit under Fig. 7.
We have a good vocabulary for the time a meal takes and almost none for the time it costs. An hour blocked at noon is legible on a calendar. The forty minutes before it — deciding, queueing, walking, changing your mind about the queue — go unbooked, and so do the twenty after, when whatever you were doing before lunch has to be found again from the beginning.
Across a year, those unbudgeted edges outweigh the meals themselves. Our domestic rhythm survey put the median at one hour fifty-two minutes a day, procurement and preparation and recovery counted together. Asked about it, most respondents described it the way people describe weather.
The second cost is harder to name. Food arrives in pulses. Blood sugar answers in pulses. Attention tracks that rhythm all day. Most people have built their working lives so completely around the three o’clock collapse that they now read it as a personality trait rather than a supply problem.
Then there are the places where attention is the least of it. A resupply window that opens once every twenty-six days. A pressurised hull where every litre of water carries a line item. A winter station where lighting a stove is a decision with paperwork attached. Eating there is a logistics chain rather than a habit, and logistics chains fail from the far end first.
We designed SATIA-30 for the far end, then worked backwards until it fitted a desk.
Vellmark Foodlab domestic rhythm survey, n = 3,880, self-reported across fourteen consecutive days.
Respondents logged elapsed time only. Every question on the form was about clocks.
A broken stretch of attention has to be rebuilt from the start. In the survey that rebuild ran to a median of twenty-two minutes.
Fasting is the absence of supply. This is the opposite: supply — lower, slower and steadier than a meal delivers, and continuous for thirty days.
Three phases share one body: a shell that dissolves inside the hour, a polysaccharide matrix that gives way at a fixed rate, and an inert core that holds the micronutrient depot back until late. One unit is one cycle, taken once on the first morning of it.
The house code for this geometry. It indexes the press file and the release specification together, and it moves whenever either of them does.
Three days to build a straight line. Twenty-one to hold it. Six to take it down on a slope we state in advance. The plateau is the thing we make; both ramps exist to protect it.
Vertical axis is the Vellmark release index, normalised so the Phase II plateau sits at 98. Horizontal axis is days since the unit was taken. The dashed reference is an ordinary three-meal day measured on the same index — same scale, same instrument, different shape.
The outer coat goes into solution inside the first hour and carries the index from rest up to the plateau band. The flavour capsules open here too — all fourteen hundred and forty of them, in the only hour of the cycle that has anything to taste.
The matrix hands over its load on zero-order kinetics. Rate follows the surface it exposes, not the amount left inside, so the last week releases at the speed of the first. Three weeks, one number, ±1.8 either side of it.
Phase II is why the unit exists. The shell in front of it and the taper behind it are both there to defend it.
The inert scaffold comes progressively into view and the index descends at a rate fixed before the unit was pressed. Six days on a straight slope to the end of them, which is the one property of Phase III anybody has ever asked us about.
A thirty-day cycle is a different problem from a large meal spread thin: what the body burns first, what it holds back, and what it leaves alone entirely. The answer is a geometry question before it is an ingredient question, and every ingredient here was chosen to serve the geometry.
Section A–A. Read it from the outside in and you are reading the calendar: one hour of shell, twenty-one days of matrix, six days of core.
The formulation is built around macronutrient partitioning: what the body burns, what it stores, and what it leaves alone. Carbohydrate arrives slowly enough that the body meets it as a trickle rather than as a spike, which is what puts glycogen sparing within reach — the reserve behaves as a buffer instead of as the fuel. Underneath it, beta-oxidation runs as the default, ketone bodies cover the tissues that would otherwise ask for glucose, and gluconeogenesis sits idle rather than taking protein apart to make sugar. The order matters more than any single number in it.
A sustained-release matrix that empties by diffusion runs fast at the start and starves at the end. Thirty days needs the opposite. This one is surface-limited: the polysaccharide erodes from the outside at a rate set by the area it exposes, and the geometry holds that area flat right through Phase II. Take the same powder and press it into a sphere and the curve collapses inside a week.
A resistant starch fraction passes the small intestine intact and ferments downstream, which keeps the gut microbiome fed through a month with no other substrate arriving. Minerals travel as chelates rather than simple salts — a micronutrient bioavailability decision that turned out to be a taste decision as well, since chelates carry far less of the metallic edge a sulphate does.
Satiety signalling — leptin, ghrelin, GLP-1 — is the regulation lane at Fig. 4, drawn dashed because it acts on the gut rather than carrying substrate through it.
Three routes out of the small intestine and one regulation lane beneath them. Follow the solid lines for substrate and the dashed one for signal.
We report the four layers separately because they behave separately: different onsets, different durations, different ways of going wrong. The other three are slower and quieter, and one of them is still accruing on day thirty.
Four lanes, four start times, all of them panel medians. Layer three runs from hour zero; layer four is still accruing on day thirty.
The mechanism is the one drawn two screens up. What matters here is how it reads from the inside: nothing arrives, and nothing stops arriving.
In four years of panels the word that comes back most often for Phase II is flat, and it is meant approvingly. Two panellists in three put the start day wrong by a week or more; nearly all of them place the taper to within a day of it.
The dip people schedule their afternoons around is a delivery gap rather than a mood. Remove the pulse and substrate arrives as a trickle instead of in waves. In panel work the dip is the first thing people stop mentioning, by day three.
It does not make anyone sharper; it removes a recurring hole.
Load tolerance is the least forgiving thing to design for. A body under real work finds any gap in a supply curve inside forty minutes. So the matrix lets demand pull harder on release than time does: harder work, faster local handover, up to a ceiling the formulation will not cross.
Below that ceiling the curve holds. Above it, headroom runs out and the person notices. The physiologists at Vennhold spent most of a year finding where the line sits. It sits at 248 W sustained, and we print it rather than round it.
Coverage is the slow layer. It fails quietly — a drift, and then six weeks later somebody’s fingernails mention it.
Thirty days is long enough for small things to matter. Fat-soluble vitamins sit in the core and come out late, because the tail of the cycle is where a shortfall would show first. The resistant-starch fraction passes through undigested and feeds the gut microbiome — the one part of this formulation doing work we did not design.
We audit this layer on the retained sample of every batch. It is the slowest assay we run: fourteen days from sample to figure.
Reading the four layers. Each is a panel median with its own onset window, and layer two carries the widest spread between one person and the next.
The question was whether the plateau holds outside a laboratory. Two figures follow, then the composition matrix. These are the parts of the file that fit on a page. The rest goes in the deck, unedited.
Four cohorts, thirty days of daily observation, one endpoint fixed before the first unit went out: retained physical output against each participant’s own day-zero baseline. Micronutrient status and batch release were tracked alongside it. Every enrolled participant is in the figures, including the nineteen who stopped before day thirty.
Vertical axis is retained output as a percentage of each cohort’s own day-zero baseline. Day six is simply where the two curves stop agreeing; the protocol ran unchanged straight through it.
Each mark is one production batch, plotted as its Phase II plateau mean. The shaded band is the release tolerance. These are nine consecutive presses in run order, plotted from the assay that precedes the release decision rather than from the batches that survived it.
| Component | Class | Per cycle | Unit | Released in | Coverage % of Vellmark cycle target |
|---|---|---|---|---|---|
| Sustained carbohydrate | macro | 8.20 | g | II | 104% |
| Composite protein | macro | 4.30 | g | I–III | 97% |
| Structural lipids | macro | 2.40 | g | II–III | 101% |
| Dietary fibre | macro | 1.55 | g | I–III | 118% |
| — of which resistant starch | of which | 1.10 | g | I–III | 105% |
| Medium-chain triglycerides | macro | 0.75 | g | I–II | 93% |
| Iron, chelated | micro | 505 | mg | II | 100% |
| Zinc, chelated | micro | 348 | mg | II | 106% |
| Magnesium | micro | 1.26 | g | I–III | 99% |
| Selenium | micro | 1.65 | mg | III | 91% |
| Potassium | micro | 1.05 | g | I–III | 102% |
| Vitamin A | vitamin | 8.1 | mg | III | 96% |
| Vitamin B12 | vitamin | 72 | µg | II | 109% |
| Vitamin C | vitamin | 408 | mg | I–III | 113% |
| Vitamin D3 | vitamin | 600 | µg | III | 94% |
| Vitamin E | vitamin | 135 | mg | III | 100% |
| Vitamin K2 | vitamin | 3.6 | mg | II–III | 87% |
| Electrolyte buffer salts | functional | 0.62 | g | I–III | 98% |
| Inert core and scaffold | structural | 2.46 | g | — | — |
| Flavour capsules | count | 1,440 | count | I–III | — |
| Unit mass, total | 24.0 | g | — | — |
How to read this table. Every figure is a cycle total held to the tolerance in Fig. 7, and every one is verified against the retained sample rather than the batch record — the record says what we intended, the sample says what we made. The row marked of which is already counted in the row above it. The inert core is weighed and not credited, and the capsule row is a count rather than a mass; the capsules themselves are weighed inside the shell. The per-batch assay sheets carry the same twenty entries at four decimal places.
Physiology put the layer-three ceiling at 248 W. Polar medicine broke the coating on purpose, then explained how it had broken. Matrix chemistry put a rate on the polysaccharide. Human factors told us the case was wrong, twice, and was right both times.
Every one of them has told us something we would rather not have heard. That is the arrangement.
Ran the load-and-output series, eleven sessions a cohort, and put the layer-three ceiling at 248 W sustained.
Cold-cycled the coating until it failed, then explained precisely how it had failed. The second half is the rarer service.
Held the four thirty-day cohorts and the substrate partitioning behind them. Fig. 4 is theirs, and so is the endpoint it was drawn against.
Took the polysaccharide apart and put a rate on it. The surface-limited curve at Fig. 3 was drawn on their bench before it was drawn on ours.
Payload and waste accounting, to the gram, against a fixed consumables budget.
Storage stability under pressure. The thirty-six-month shelf-life figures are theirs, not ours.
Everything about the unit beyond its chemistry: how it is carried, how it is opened, how it gets lost.
Sixty-one flavour iterations. They rejected fifty-eight.
Terms of collaboration. Partner work is contracted per programme and paid on delivery, never on result. Every one of the eight sits outside the company's cap table, and every one of them keeps the right to publish what it finds — including the parts that go against us.
Wherever the supply line outlasts the appetite.
Five briefs the formulation group works against. The desk is the easy one.
Five notes from the first thirty-day cohorts, printed as they were written — including the one about the case.
“On day twenty-seven I realised I had not washed a pan all month. The observation log for that stretch is the most continuous one I have kept in nine years, and I do not believe those two facts are unrelated.”
“There is no such thing as a mealtime during a polar night. You work when the light is worth working in, which in February was about two hours, and everything else in the day moves to make room for those two hours. That month the light was the only thing setting my schedule.”
“Nobody lost the unit. Eleven people lost the case.”
“Everything in a habitat schedule is accurate to the minute except eating, which is a variable with opinions. This turned it into a constant. It sits on line one of the consumables manifest now — one entry, thirty days — and we have not re-planned that line since.”
“I expected to miss chewing. Day nine was the first time I thought about it at all, and only because somebody asked me what I was having for lunch.”
Most of the process notes in the file began as a failure somebody had to sit with. This is the one the formulation group still tells first.
“The first forty-one units came out of the press cracked. Not hairlines — they split along the seam like a walnut, every single time, and only ever on the afternoon shift.
Five weeks to find it. The room ran two degrees warmer after lunch because an extract fan sat on a timer somebody had set years earlier, and nobody thinks to look at a fan.
The thing I would not give up was the flavour. Everyone else thought the capsules were vanity — thirty days with nothing to chew, why bother with basil. I bothered because a mouth that has stopped expecting anything is a strange thing to carry around for a month, and I was not willing to ship that.”
Thirty-seven origins, compared batch against batch. Only the narrowest particle spread goes forward.
Held below freezing end to end, so nothing heat-sensitive ever meets warm air.
Each phase formed in sequence. Every interface between one layer and the next comes out clean, which is where the forty-one went.
Fluid-bed deposition sets the Phase I to Phase II handover. Then the batch waits — and three units out of it wait for thirty-six months.
The house document the tolerances opposite are written against. It sets what a press run has to demonstrate before its units leave the room, and it is revised only after a batch has taught us something.
Every tolerance we publish is the tightest we can hold nine times running, not the tightest we have ever hit once.
Twenty-one of the thirty days are Phase II, and Phase II is the part with nothing to look at: a plateau held inside ±1.8 index, press after press, while whoever is carrying it thinks about something else entirely. Eight years to make that boring, most of them spent losing arguments to eight laboratories. The forty-one cracked units are still in the retention rack, three shelves under the batch that finally held.