
Most handbooks give a single formula for pump-down time: t = (V / S) × ln(P0 / P1). Applied to a 1 m³ chamber evacuated from atmosphere to 1 mbar with a DPL-V0100, it returns 4.2 minutes. On site the same chamber usually takes six to eight. The missing minutes are rarely the pump; they come from three things the formula leaves out: the lowest pressure the pump can reach, the loss of nameplate speed in the piping, and outgassing from the chamber itself. Add the three corrections and the estimate lands within about ten percent of measurement.
1. Before using the formula, fix the three assumptions
The textbook form assumes an ideal pump that can reach absolute vacuum, and that all gas load comes from the chamber volume. A safer engineering form subtracts the pump’s ultimate pressure:
t = (V / S) × ln[(P0 − Pu) / (P1 − Pu)]
with P0 = 1013 mbar, Pu the ultimate pressure of the model, V in m³ and S in m³/h; the result is in hours.
This correction explains something seen on site repeatedly: the closer the target pressure sits to the ultimate pressure, the longer the time, and as the denominator approaches zero the time tends to infinity. Asking DPL-V0025A (ultimate 0.5 mbar) for 1 mbar gives 22.8 minutes in theory, and with real outgassing it may never get there. Leaving at least a factor of two between target and ultimate pressure is standard practice.
2. Effective pumping speed: apply the discount
Between pump inlet and chamber there are piping, valves and filters, each with a conductance. In series, 1/S_eff = 1/S_pump + 1/C_piping. From atmosphere down to about 10 mbar the flow is viscous and conductance is high, so short, large-bore piping loses only ten to twenty percent. Approaching 1 mbar the flow enters the transition regime and conductance of the same bore falls with pressure, so losses can exceed thirty percent. For a 1 mbar end point, taking 0.6 to 0.8 of nameplate speed is realistic; long runs with many elbows should be taken at 0.5 or below.
A flat factor of 0.7 keeps the estimate within about ten percent.
3. Example one: 1 m³ chamber to 1 mbar
Duty: net chamber volume 1 m³, start 1013 mbar, target 1 mbar, short piping so S_eff = 0.7 × nameplate.
| Model | Nominal speed (50 Hz) | Ultimate, ballast closed | Ideal time | At 0.7 |
|---|---|---|---|---|
| DPL-V0025A | 20 m³/h | 0.5 mbar | 22.8 min | 32.6 min |
| DPL-V0040A | 40 m³/h | 0.1 mbar | 10.5 min | 15.1 min |
| DPL-2V80 | 74 m³/h | 4×10⁻⁴ mbar | 5.6 min | 8.0 min |
| DPL-V0100 | 100 m³/h | 0.1 mbar | 4.2 min | 6.0 min |
| DPL-V0302 | 300 m³/h | 0.1 mbar | 1.4 min | 2.0 min |
With the end point fixed at 1 mbar, the log term is a constant near 7.0, so time scales as V/S: double the speed and halve the time. There is no trick in roughing beyond raising effective speed — a larger pump or a larger bore.
Note where DPL-2V80 sits: about seventy percent of the V0100 speed, but three orders of magnitude lower ultimate pressure. If the process stops at 1 mbar, the two-stage premium buys no time. If a high-vacuum pump follows and needs a foreline below 10⁻² mbar, that is exactly where a two-stage pump belongs.
4. Example two: reverse sizing for 6 m³ in 8 minutes
Reverse sizing is what production engineers actually use: fix the takt time, then solve for speed.
S = (V / t) × ln[(P0 − Pu) / (P1 − Pu)] = (6 ÷ 0.133 h) × 7.03 ≈ 316 m³/h ideal, or about 452 m³/h of nameplate after the 0.7 discount.
- DPL-V300B: 280 m³/h, 5.5 kW, ≤8×10⁻² mbar. Effective 196 m³/h gives t ≈ 12.9 minutes — short of the eight-minute takt.
- DPL-V630B: 640 m³/h actual at 50 Hz, 15 kW, ≤8×10⁻² mbar. Effective 448 m³/h gives t ≈ 5.6 minutes, leaving 2.4 minutes for outgassing and piping deviation.
One frame size apart: 12.9 minutes against 5.6. Reverse sizing makes the question answerable — V300B is not wrong, it is five minutes short of the takt.
5. After 1 mbar: when a Roots stage earns its place
For a 1 mbar end point, a single-stage oil-sealed rotary vane pump is enough and a Roots blower is wasted. A Roots pump cannot exhaust to atmosphere; its place is the deep-vacuum stage below 10 mbar, engaged once the backing pump has brought the chamber to roughly 10 to 100 mbar.
Size the backing pump at 10 to 20 percent of the Roots speed. LC250 at a nominal 250 m³/h needs 25 to 50 m³/h of backing, so DPL-V0040A (40 m³/h) or DPL-V0063 (63 m³/h) fits; hanging LC250 behind a V0025A leaves the backing pump unable to hold the differential, and both temperature and load rise.
Time each stage separately. Effective speed in the backing stage and in the Roots stage differ by an order of magnitude, and summing two short calculations is more accurate than one average coefficient applied from 1013 mbar down.
6. Check the result: rate of rise beats the formula
Every calculation above assumes a clean, empty chamber. Real chambers outgas from walls, fixtures and load, and leak slightly; the gas load Q lifts the end pressure according to P1 = Q / S_eff + Pu.
Measure it with a rate-of-rise test: pump down to stability, close the valve, and record the pressure rise over one hour; multiply by volume to get the gas load. If the rise is high, fix outgassing and leaks before blaming the pump. ISO 21360-1:2020 lists exactly these influences when volume flow rate is derived by the pump-down method: leak rate, desorption, increasing flow resistance in the connection line at low pressure, and cooling by nearly isentropic expansion during the interval. The formula gives the theory; the measurement gives the sizing.
Use the two together: the calculation tells you whether you are one frame size short or three, and the measurement tells you whether the gap sits in the pump, the piping, or the outgassing.