Stage performance — ratios, work balance and sensor checks
A compression ratio is a ratio of ABSOLUTE pressures, and almost every field number gets this wrong. On the reference unit — 40.8 psig suction, 180.5 psig first-stage discharge, 2000 ft elevation — dividing the gauge readings gives 4.43. The truth is 3.57. That is not a rounding difference; it is the difference between a stage you think is overloaded and one that is fine.
What it computes
- Absolute ratio per stage — (discharge + barometric) over (suction + barometric), where barometric comes from the site elevation you enter. At 2000 ft that is 13.66 psia against the reference unit's run sheet value of 13.651.
- The panel's own ratio beside it, when the log carries one. On the reference log the two agree to within 0.05, and the panel's channel is rounded to 0.1 — so this is a confirmation, not an argument. When they disagree by more than rounding, the difference is shown so you can find out why.
- Share of the total work per stage, and how far each stage sits from an even split. An even split is the same ratio in every stage, which is the geometric mean of the overall ratio.
- Overall ratio across the whole train, from first-stage suction to last-stage discharge.
What it deliberately does not compute
Cylinder-internal ratios and predicted discharge temperatures are absent, and that is a decision rather than an omission. Both need valve losses and the polytropic exponent, which live on the unit's run sheet, and a plausible-looking guessed number sitting next to a measured one is worse than no number at all. They arrive with the unit profile.
Two assumptions are stated on screen next to the numbers they affect. When a panel has no interstage suction taps — most do not — each upper stage is fed the stage below's discharge, which ignores the interstage cooler's pressure drop and makes upper-stage ratios slightly optimistic. And the work split assumes the same suction temperature at every stage; a fouled interstage cooler breaks that, and the fix is a suction thermocouple per stage rather than a cleverer formula.
Sensor checks
The same surface runs four checks over every channel in the log, because a beautiful trend of a dead sensor is worse than no trend.
- Stage pressures out of sequence — a stage discharging below the pressure feeding it. Reported with how long it lasted and how soon after a start it began, because an unloaded stage during startup produces this legitimately: on the reference log every one of the ten windows followed a start, so they are marked suspect rather than impossible.
- A dead member in a live family — “Lube Flow 3 never changes while 2 siblings do”. An unconfigured slot or a failed sensor, and either way not data.
- A transmitter frozen while its family moved — the member that sat at exactly one value while its siblings shifted several times this channel's own resolution. Judged against the siblings on purpose: a thermocouple logged to the nearest degree legitimately sits on one integer for hours at steady load, and a naive “held one value for N minutes” rule produced sixteen false findings on the reference log.
- Physically impossible values — the −9999 of an open thermocouple, a negative pressure, a temperature no gas in this machine reaches.
Every finding carries a “Show me” that plots the channels involved and moves the chart to the window in question, so a check can be confirmed rather than believed.
- Open the View menu and turn on Stage performance.
- Enter the site elevation. This is the one number the view cannot read from the log, and every ratio depends on it.
- Leave the basis on psig unless your panel reports absolute pressure already — field panels report gauge.
- Read the ratios, then the balance line: a stage more than about 20% off an even split is doing work the others are not, which shows up as heat and rod load on those throws.
- Work down the sensor checks and use “Show me” on anything you have not seen before.
- “Copy summary” puts the table, the assumptions and every check on the clipboard for a report.
Privacy
All of it is computed in your browser. No log, no pressure and no unit identifier is uploaded to a server.
FAQ
How do I calculate compression ratio from a datalog?
Add barometric pressure to both the suction and the discharge reading, then divide. Barometric comes from site elevation — about 13.66 psia at 2000 ft, against 14.696 at sea level. Dividing raw gauge readings is the most common mistake in the field and it overstates first-stage ratio badly: 4.43 instead of 3.57 on the reference unit.
Why does Overtrace's ratio differ from my panel's?
Usually it does not. On the reference log the two agree within 0.05, and the panel channel is rounded to 0.1 while Overtrace shows two decimals. Both are displayed side by side, so a real disagreement is visible rather than hidden — and a real one normally means the panel is using a fixed 14.7 psia or a different suction tap.
What does the work split tell me?
Which stage is doing the compressing. Stages within about 10% of an even split are balanced; beyond 20% one stage is carrying work the others are not, and that shows up as discharge temperature and rod load on those throws. The split assumes equal suction temperature at each stage, which is stated on screen.
Why are cylinder-internal ratios not shown?
They need valve losses and the polytropic exponent from the unit's run sheet. Guessing them would put an invented number next to measured ones, which is how a tool loses trust. They arrive with the unit profile.
Why is my stage-3 discharge flagged as reading below stage 2?
Because it did, and a stage cannot discharge below the pressure feeding it. Check when it happened: if it began within 45 minutes of a start it is very likely an unloaded stage during startup, which Overtrace says explicitly. If it did not follow a start, look at the transmitter, its range setting, or a blocked sensing line.