Overtrace is a browser-based industrial telemetry overlay tool. You load datalogs from multiple modules — compressors, gen skids, engine ECMs — and overlay every channel on one shared time axis so you can see what happened, and when, across the whole system at once.

Which datalog formats does Overtrace support?

Overtrace auto-detects Emit (Dynamic Controls), Caterpillar, Waukesha and Excel datalogs, with more formats coming. Just drop the files in and the channels are parsed and plotted automatically.

Is my telemetry data uploaded to a server?

No. All parsing happens entirely in your browser using local JavaScript in a Web worker. There is literally no upload endpoint — your datalogs never leave your machine. That privacy-by-architecture is a core part of the design.

Do I need to install anything?

No install required. Overtrace runs in any modern browser on desktop, laptop or field tablet, and works even on a spotty plant network because everything is client-side.

Can I try Overtrace before buying?

Yes. The full-featured demo runs on built-in sample telemetry, so you can try measuring, overlays and alignment for free. A license unlocks uploading your own logs, and it is tied to your account rather than a single device.

Unit profile — the compressor's own numbers

A datalog tells you what the machine did. It does not tell you what the machine IS. Bore, stroke, rod diameter, cylinder clearance, the polytropic exponent the OEM used, the frame's rod-load limits, and which throw feeds which stage — none of that is in a panel export, and every derived number beyond the log itself needs some of it.

So the profile is saved, not asked for repeatedly. It changes only when somebody restages the machine or moves a pocket. Open it from Tools → Unit profile.

Nothing leaves your device. The PDF is parsed in the browser by a bundled copy of pdf.js — no upload, no service, and no network needed.

Three ways in

Type it. Around three minutes from a run sheet. Every field is optional — site elevation and the throw→stage map alone already improve what the other views can say.
Paste the sheet's text. Select all in any PDF viewer, paste, and the same deterministic parser fills the form.
Pick the Ariel PDF. Identical parser; the PDF route just saves the copy and paste. On the reference sheet it reads all three stages, all six throws, and matches every throw to its stage.

A scanned or photographed sheet has no text layer, and that case falls back to typing rather than to OCR. A wrong digit in a clearance percentage is silent and changes every number downstream, which is worse than three minutes of typing.

The throw→stage map

This is the field that matters most, and it is not printed on the sheet as a map. It is inferred: each throw's internal suction pressure sits a few psi below its stage's suction flange pressure, and the stages are far apart — 38.6, 177.9 and 618.2 psig on the reference unit. When two stages are similarly close to a throw the tool refuses the match and asks you to set it, because a stage label given with false confidence is worse than a blank.

With the map set, channel names get a plain-language second line: “TC-19 DISCHG 5A TEMP” becomes “throw 5 · stage 1 · 20-1/8″ · head end”. The raw channel name always stays first — it is what the panel in front of you says.

What the profile unlocks

Inferred suction temperature per stage, on units that log no suction thermocouple. On the reference unit: 88 °F at stage 1 against a design 80, 128 °F at stage 2 against 130, and 111 °F at stage 3 against 130. Trending it tracks interstage cooler fouling with no sensor installed.
The case's predicted discharge temperature at today's measured ratio. This is the same equation solved the other way, and it is presented as such rather than as a second, agreeing tool.
Rod-load headroom per stage against the frame's limits from the sheet, beside the case's own percentages.
Capacity as a percentage of the run case. The efficiency model is run twice — once at the case's conditions, where the sheet already states the flow, and once at today's — and the ratio between model and sheet at the case point is kept as a per-stage calibration. Whatever the model gets wrong about the machine's geometry cancels. Never an absolute rate: there is no meter.
A warning when today's measured ratios have walked away from the case, because the case describes how the machine was set and pockets move.

These carry an uncertainty band and are labelled as needing the run sheet, both in the section and on the verdict card. They are gauges built on the OEM's exponent and a flange pressure, not measurements.