Mount Periodic Error and PEC Explained

What periodic error is

Periodic error (PE) is the repeating, back-and-forth wander a mount's own drivetrain introduces into right ascension tracking, tied directly to the mount's worm-gear rotation. One full cycle of that error corresponds to one full revolution of the worm shaft — on the worm periods common to backyard equatorial mounts, that's commonly around ten minutes, roughly 600 seconds, start to finish astrojolo; Telescope Live. Because the error repeats on that same clock, all night, every night, on that specific mount, it's fundamentally different from the tracking error a guide graph shows from seeing, wind, or a snagged cable — those are random; PE is not.

Why it's periodic

No worm gear is cut to mathematically perfect precision. Small variations in tooth spacing, and a slight eccentricity in how the worm shaft itself was machined, mean the mesh between worm and wheel isn't perfectly uniform through a full revolution — the mount tracks a little fast through part of the cycle and a little slow through another. Because the geometry that produces that variation is fixed into the hardware, the same fast-then-slow pattern repeats identically every time the worm completes a revolution, which is what makes it "periodic" rather than random noise.

What the PE curve looks like

Plotted against time, uncorrected periodic error traces a smooth, roughly sinusoid-like wander — not a sawtooth and not a series of sudden jumps — because the underlying cause is a gradual mesh-quality variation around the whole wheel, not a discrete defect at one point. That smoothness is exactly what makes PEC training possible in the first place: a smooth, repeating curve is predictable, and a predictable curve can be recorded once and replayed later as a counter-correction.

How PE is measured

Two figures are worth separating before going further, because mixing them up is the most common way a PE spec gets misread. A mount's periodic error is a property of the drivetrain, measured with no guiding running at all. Guided RMS is a different, always smaller number that folds in the guide camera, the guiding software, seeing and mechanical flexure on top of the raw PE curve once an autoguider is actively closing the loop. This article keeps the two consistently distinct: every headline class-level PE figure below — the mount-class table and the ~24″ canon figure — is peak-to-peak, the full swing from a curve's highest point to its lowest within one worm period. The full breakdown of that distinction lives in step 5 of our mount-selection framework, and ZWO's guaranteed PE figure for the AM5N is covered in that head-to-head, which also notes Sky-Watcher publishes no equivalent per-unit guaranteed figure for the EQ6-R Pro.

Measuring the curve itself is normally done with dedicated software — PEMPro and PECPrep are the two most common — that logs a guide star's drift, unguided, across one or more full worm cycles and runs a Fast Fourier Transform on the result to isolate the periodic component from noise. The output is the peak-to-peak figure described above, and it's the convention this article's class-level PE figures follow — with one forum-sourced exception below.

On some geared EQ6-class mounts, that FFT can also turn up a second, faster wobble layered on top of the main worm-period curve — a roughly 180-second cycle one Cloudy Nights measurement thread attributes to the second harmonic of the mount's transfer gear, at around 5″ peak-to-peak (about 1″ RMS) Cloudy Nights, “EQ6 Pro unusual periodic error” — the one figure in this article reported both ways by its own source. Treat that specific number as one forum member's own measurement, not a spec — it's included here as an illustration that a real PE curve isn't always a single clean sine wave, not as a figure to plan around.

Typical PE by mount class

None of the figures below are guided RMS, and none are a specific model's manufacturer spec — those are compared side by side in our payload-and-PE roster, which shows, mount by mount, what each manufacturer does and doesn't publish. This table is coarser: what peak, unguided PE typically looks like at each rung of the worm-gear ladder, useful for sizing expectations before you've picked a specific mount.

Harmonic (strain-wave) mounts aren't on this ladder at all — their PE behaves differently enough that it belongs in a different comparison, covered in our harmonic-vs-GEM breakdown and in the roster linked above. For how the ~24″ canon figure plays out on an actual EQ6-R Pro once PPEC training is applied, our deep review walks through owner-measured before-and-after results.

The PEC/PPEC training workflow

"PEC" and "PPEC" get used loosely to mean two genuinely different things: training a correction curve into the mount itself, and PHD2's own guiding-side prediction feature that happens to share the acronym. Both matter, and they're not interchangeable.

In-mount PEC training

Training a PEC curve into the mount is the traditional route, done either from the hand controller or, on Sky-Watcher mounts, through EQMOD's AutoPEC routine. The workflow is consistent across brands: record the mount's own drift, unguided, for at least one full worm cycle — most guidance recommends averaging five to nine cycles rather than training on just one, since a single cycle still carries guide-star noise the averaging hasn't had a chance to cancel out OPT; McWiki. The controller (or EQMOD) then models that averaged data into a single smooth correction curve and uploads it to the mount's motor board. Once it's on the board, the training is permanent — that's the first “P” in PPEC — and it replays automatically on every subsequent worm revolution without needing PHD2 or any other software running OPT; McWiki. Sky-Watcher's SynScan hand controller, on mounts like the EQ6-R Pro, supports exactly this workflow natively — training and playback built into the controller firmware, no added software required Sky-Watcher.

PHD2 Predictive PEC (PPEC)

PHD2 offers its own Predictive PEC (PPEC) — a guiding algorithm, not an in-mount training feature, and a genuinely different thing despite the shared name. It watches the same repeating pattern in the guide corrections PHD2 is already sending and starts predicting the next cycle's error before it happens, tightening the correction loop. The project's own documentation is explicit about what it isn't: PPEC is “complementary to PE correction in your mount, it does not replace it” openphdguiding.org, PPEC documentation. It's also not confined to worm-period harmonics the way an in-mount PEC curve is — it will predict any sufficiently repeating pattern in the guide signal, which is useful for catching things beyond pure PE, and part of why it isn't a clean substitute for training a real PEC curve on the mount.

PEC + autoguiding together

In practice, most imagers run both. A trained PEC curve removes the repeating part of the error before the guider ever has to react to it, which lightens the corrections autoguiding has to make on every exposure — but it doesn't remove tracking error outright, and the PHD2 manual is direct about the residual: some repetitive error remains even with PEC actively running, on top of everything non-periodic — seeing, wind, mechanical flexure — that PEC was never designed to touch PHD2 manual.

When PEC won't help

Everything above assumes a worm-gear mount, because that's where PEC training actually works well: a rigid worm-and-wheel pair produces the same error, cycle after cycle, which is exactly the kind of repeating signal a PEC curve is built to cancel. A harmonic (strain-wave) mount's periodic error behaves differently — it's more variable and depends on how the mount is loaded, a property of the flexspline-based mechanism itself, not a training gap. Rainbow Astro's own FAQ for the RST-135 states plainly that the “magnitude of periodic error varies depending on the weight of load” Rainbow Astro FAQ — a curve trained under one load doesn't hold up once you swap optical tubes or rebalance. The full mechanism behind why worm-gear and harmonic drives produce such different error signatures lives in our harmonic-vs-GEM mechanism breakdown, and we won't re-derive it here. On a harmonic mount, autoguiding does the entire job PEC and guiding split between them on a worm-gear mount.

PEC vs guiding — which to use

Short version: on a worm-gear mount, they're not competing options, they're sequential ones. Train PEC first — it's free, it's permanent once it's on the motor board, and it shrinks the burden on whatever guides afterward. Guiding still runs on top of it, because guiding is the only thing that also cancels the non-repeating error PEC was never built to touch. If you haven't set up autoguiding yet, our PHD2 setup walkthrough covers the hardware and the calibration workflow end to end. And once you're guiding, the natural next question — what guided RMS number you should actually expect to see, and how to read it against your image scale — has its own dedicated breakdown in our guiding RMS thresholds guide; we don't restate those targets here.