Guiding RMS Thresholds: What Is Good Guiding?

What counts as good guiding, in one line

Ask the question on any astrophotography forum and the same figure comes back: a total RMS error of less than one arcsecond is acceptable for decent guiding The Ek’s Files, “How Well Does My Mount Track and Guide”, with under 0.5″ generally regarded as excellent stellarnomads, “Autoguiding”.

That number is a reasonable starting point and a genuinely useful gut-check. It is also, on its own, incomplete. Total RMS is a statistic about how far your guide star wandered during an exposure — it says nothing about how big your pixels are, how long your focal length is, or how forgiving your setup is of a given amount of wander. A guided RMS of 0.8″ is outstanding on a wide-field rig sampling the sky at 3″/pixel and could be the visible cause of bloated stars on a long-focal-length rig sampling at 0.5″/pixel. Before comparing your number to anyone else’s, or to the one above, you need three things labeled correctly — that’s the next section — and you need your own image scale, which is the section after that.

Label everything before you compare numbers

Guiding RMS gets misread more often from mislabeling than from bad guiding. Three distinctions matter every time a number gets quoted, compared, or posted for feedback.

Total vs. per-axis (RA vs. DEC) RMS

PHD2’s guiding graph reports a Total RMS figure alongside separate RA and DEC RMS values, and the total is not a simple average of the two — it combines both axes, so a mount that’s excellent in DEC and mediocre in RA can still post a total number that looks unremarkable, and a total figure alone can hide which axis is actually causing your elongated stars. When you’re troubleshooting rather than just checking a headline number, look at RA and DEC separately before you look at the total.

Arcseconds vs. pixels

PHD2 can display guide error in pixels or in arcseconds, and it only gets the conversion right if you’ve entered your guide scope’s focal length and your guide camera’s pixel size correctly in the equipment profile. Get either number wrong and every RMS figure downstream is wrong too, even though the actual guiding hasn’t changed at all — which is a common, invisible source of the “my RMS suddenly got worse” forum post. The underlying math that converts focal length and pixel size into an arcsec-per-pixel figure is covered in full in Pixel scale and sampling explained; treat your image scale as an input you bring into this article, not something to work out here.

RMS is a statistic, not a peak

RMS (root-mean-square) is a statistical summary of a whole exposure’s worth of guide corrections, not the single worst excursion. A Cloudy Nights thread on the topic characterizes PHD2’s RMS as roughly the radius of the circle that encloses about two-thirds of your guide star’s recorded positions, with peak-to-peak excursions running around three times the RMS value Cloudy Nights, “Guiding RMS error below your imaging resolution”. Treat that as a way of reasoning about the number, not a manufacturer spec. It’s also the reasoning behind why a coarser rule of thumb keeps showing up: if you want most of a star’s light concentrated in a single pixel rather than smeared across its neighbors, that two-thirds circle needs to fit inside roughly half a pixel — which is one way of explaining, structurally, why the half-image-scale rule in the next section lands where it does.

Relating RMS to image scale — where the number actually means something

This site’s mount-sizing framework already states the rule this article builds on: guided tracking error should land at roughly half your image scale or better in arcsec/pixel, sourced to Astronomy Now’s 2023 buyers’ guide to equatorial mounts Astronomy Now, “Buyers’ guide: The best equatorial mounts,” 2023 — see Step 5 of that framework for the full context. What that rule doesn’t do on its own is tell you what to make of the flat “under 1″” figure from the section above — and the honest answer is that the two don’t always agree.

Bring your own image scale — if you don’t already know it, the pixel scale and sampling explainer and its calculator work it out from your guide setup’s focal length and pixel size — and find the nearest row.

The seeing floor — where more guiding stops helping

Guiding corrects tracking error; it cannot correct atmospheric seeing, and at some point seeing becomes the limit no amount of mount tuning will move. Jerry Lodriguss put rough numbers on that floor, relayed via a Cloudy Nights post and cited by AstroBackyard:

Notice that the top of that range — 1.0″ RMS under bad seeing — lands right on the flat “acceptable” figure from the sections above. That’s not a coincidence so much as a reminder of what the number is actually measuring: on a bad night, 1″ of guided wander may be close to the best any mount can do, seeing included, not a sign of an undersized or poorly tuned rig. The community rule of thumb that follows from this — guiding is “seeing-limited” once your RMS settles to roughly a third of the night’s seeing Cloudy Nights (forum context) — is exactly that, a rule of thumb, not a Lodriguss figure, so treat it as a way to sanity-check a session rather than a target to chase past the point of diminishing returns.

Reading RMS by image scale, not by target name

It’s tempting to ask “what RMS do I need for galaxies” versus “what RMS do I need for wide-field nebulae,” but target class isn’t actually the variable that matters — image scale is, and target class is just a proxy for it. Long-focal-length galaxy and planetary-nebula work tends to demand a tighter RMS not because galaxies are special, but because that kind of imaging usually runs at a finer arcsec-per-pixel scale, which is exactly where the table above shows the half-image-scale rule pulling ahead of the flat figure. A wide-field nebula shot at a coarse image scale can tolerate the same RMS that would visibly bloat stars on a long-FL galaxy rig. Use the table by your actual arcsec/pixel number, not by what you’re pointing at.

What real rigs actually show

Named, attributed real-world numbers are useful as context for what’s typical — they are not specs, and they are rig-, seeing-, and technique-dependent every time. A Cloudy Nights review of the iOptron GEM45 reports guided totals in the 1–2″ RMS range in PHD2 Cloudy Nights, GEM45 review; an IceInSpace forum report of an HEQ5 fitted with a Rowan belt mod on a permanent pier put guided performance at roughly 0.6–1.0″ IceInSpace. Both are plausible, both are one owner’s rig on one or more nights, and neither is a number you should expect to reproduce exactly with the same mount model.

It’s also worth being clear about what doesn’t exist: neither ZWO nor Sky-Watcher publishes a guided-RMS figure for the AM5N or the EQ6-R Pro. Any AM5N or EQ6-R Pro number you see quoted is a specific owner’s specific rig, not a manufacturer spec — our AM5N vs. EQ6-R Pro head-to-head covers ZWO’s guaranteed ±10″ periodic-error spec for the AM5N, and why that figure isn’t the same thing as guided RMS. Sky-Watcher publishes no equivalent per-unit guaranteed figure for the EQ6-R Pro at all; the closest thing to a reference point is a canon-promoted community measurement — roughly 24″ peak-to-peak, via PEMPro FFT, typical of NEQ6-R-class mounts — covered in our periodic error and PEC explainer and applied specifically to this mount in our EQ6-R Pro deep review, neither of which is a guided-RMS figure either.

None of these examples isolate the guide camera as the variable, and this article isn’t claiming a camera upgrade fixes an RMS number on its own — mount, seeing, and technique all matter more. But a guide camera’s read noise and quantum efficiency do affect how cleanly PHD2 can centroid a faint guide star at short exposures, which is one input among several, not a guaranteed fix.

When to stop tweaking

Round stars in your subs are the real test — not a specific RMS number on a screen. Once your guiding graph and your stars agree that things are fine, chasing a smaller RMS figure for its own sake runs into the seeing floor from above: a Cloudy Nights consensus among posters on the topic is that below that floor, tighter guiding stops producing visibly rounder stars, because seeing rather than the mount has become the limit Cloudy Nights (forum context). That’s community characterization, not a lab result, but it lines up with the Lodriguss numbers above.

The one guiding-graph pattern worth acting on rather than tolerating is a persistent, one-direction drift in DEC rather than the usual back-and-forth waver — that shape is very often a polar-alignment symptom, not a guiding-technique problem, and no amount of PHD2 settings tweaking fixes a mount that isn’t pointed at the pole closely enough. See how much polar alignment your setup actually needs before troubleshooting further.