What's the main difference between VPM-B and Bühlmann?

Bühlmann (ZHL-16C) is a dissolved-gas model that limits tissue supersaturation; VPM-B is a bubble model that limits the growth of pre-existing micro-nuclei. VPM-B prescribes deeper first stops, while Bühlmann with moderate gradient factors concentrates decompression time in the shallow stops.

Which one should I use?

Both are widely used and respected. The deep-stops evidence (notably the NEDU study) has shifted mainstream practice toward shallow-weighted profiles — i.e. Bühlmann with moderate gradient factors — but VPM-B remains popular and defensible. Choose with training, and compare the two on the same dive.

Where does VPM-B come from?

From David Yount and colleagues' Varying Permeability Model at the University of Hawaii, built on bubble-nucleation experiments in gelatin published in the 1980s. The B revision adds Boyle's-law compensation for bubble expansion during ascent, with Erik C. Baker central to the diving implementation.

Is VPM-B more conservative than Bühlmann?

Neither is uniformly more conservative — it depends on the dive and the settings. VPM-B typically demands more time deep; a low GF-high Bühlmann plan demands more time shallow. On long, deep exposures VPM-B can produce shorter total runtimes than conservative gradient-factor settings, which is part of the deep-stops debate.

{ GUIDE }

VPM-B vs Bühlmann — what's the difference?

These are the two decompression models technical divers argue about most. They start from different physics, carry different histories, and produce differently-shaped ascents. Here's the whole story — where each comes from, what each assumes, and what the evidence now says.

The shared problem

At depth your tissues absorb inert gas in proportion to its partial pressure; ascend and that gas must leave. If tissue gas tension exceeds ambient pressure by too much, gas comes out of solution as bubbles, and bubbles in the wrong place are decompression sickness. Every algorithm is a hypothesis about where the danger line sits — the models below draw it in two different places.

Bühlmann ZHL-16C — a dissolved-gas model

The Haldanean lineage: model the body as 16 parallel compartments with nitrogen half-times from ~4–5 to 635 minutes (helium in parallel), track dissolved gas exactly, and limit each compartment's supersaturation against its M-value — a pressure ceiling defined by the published a/b coefficients in Albert Bühlmann's Tauchmedizin. Because Bühlmann published everything, ZHL-16 became the open reference model of diving, and it is what most dive computers run.

Conservatism is tuned with gradient factors (GF-low/GF-high — explained in depth here), which cap the fraction of each M-value you may occupy. The shape this produces: with moderate gradient factors, Bühlmann lets you ascend well off the bottom before the first stop bites, then front-loads the decompression into the shallow stops, where the pressure gradient driving off-gassing is steepest. The model's bet: dissolved gas is harmless until supersaturation crosses the line, so get shallow efficiently and off-gas hard there.

VPM-B — a bubble model

The Varying Permeability Model came out of the University of Hawaii, where David Yount and colleagues studied bubble formation in supersaturated gelatin and published the resulting model in the 1980s. The experimental surprise: bubbles form far more easily than pure physics of nucleation says they should — unless microscopic bubble seeds already exist. VPM models these micro-nuclei as tiny spheres with surfactant skins whose permeability varies with pressure (hence the name): under normal pressures gas diffuses through the skin; crush them hard enough and the skin goes impermeable.

The model's bet follows: since seeds always exist, the danger isn't supersaturation itself, it's letting the seeds grow. VPM therefore caps the supersaturation gradient so that only seeds above a critical radius can grow, and its critical volume algorithm then relaxes that cap just enough to admit a small, tolerable total volume of free gas. The B revision adds Boyle's-law compensation — bubbles expand as ambient pressure falls during ascent, so the allowable gradients tighten on the way up — with Erik C. Baker central to the diving implementation used in planners today. Because bubble growth is most explosive when a saturated diver first leaves the bottom, VPM-B prescribes deeper first stops than a moderate gradient-factor plan; conservatism is set with levels +0 (least) to +5 (most), which scale the critical radii rather than reshaping the profile the way gradient factors do.

How the profiles actually differ

A concrete reference: for a 120 m, 20-minute trimix 18/50 dive at conservatism +1 — the published VPM-B reference profile DecoLog validates against (see how we validate) — VPM-B calls the first stop at 63 m, barely half-way up the water column, with ~141 minutes of total runtime. A Bühlmann plan with moderate gradient factors starts noticeably shallower and spends correspondingly longer at 9, 6 and 3 m.

First stop: VPM-B starts deeper — often much deeper on big dives.
Shape: VPM-B distributes time across the deep and middle stops; Bühlmann+GF concentrates it shallow.
Total runtime: often similar on moderate dives; on long, deep exposures the models diverge, and which is longer depends on the conservatism settings chosen for each.

What the evidence now says

For years the bubble-model intuition — protect the fast tissues early, stop deep — was winning by argument. Then it was tested. The US Navy Experimental Diving Unit ran a controlled trial (Doolette and colleagues, published 2011) comparing shallow-weighted and deep-weighted air decompression schedules of equal total time: the deep-stops schedule produced more decompression sickness. Time spent deep protects fast tissues that tolerate supersaturation better than assumed, while the slow tissues keep on-gassing. Simon Mitchell's widely watched lectures on the study carried the result into technical-diving practice, and mainstream guidance has since drifted toward shallow-weighted profiles — in gradient-factor terms, moderate rather than very low GF-low. VPM-B remains popular, defensible and internally consistent; but "deeper stops are inherently safer" is no longer a claim the evidence supports.

Which should you use?

Both are well-established; neither is "the safe one." The honest answers: dive the model you were trained on, at conservatism you understand; be suspicious of very deep first stops on long exposures, per the NEDU result; and above all compare the two on the same dive — runtime, first-stop depth, gas required — so the difference is something you've seen rather than something you've heard.

See them side by side

DecoLog runs both: ZHL-16C with any gradient-factor pair, and a VPM-B engine ported from Subsurface's reference implementation with conservatism +0…+5. Switch algorithms on the same plan and watch the schedule reshape. Free, offline, no signup.

Try it now — free, in your browser. • OPEN THE PLANNER

⚠ Educational only — not a substitute for training or a dive computer. Sources: A. A. Bühlmann, Tauchmedizin; D. E. Yount & D. C. Hoffman, Varying Permeability Model papers (1980s); Erik C. Baker's VPM-B and gradient-factor papers; Doolette et al., NEDU deep-stops study (2011); Simon Mitchell's deep-stops lectures. Decompression diving requires certification; verify every plan independently.