Sensor FWHM readings versus actual sub FWHM values

andrea tasselli · Jun 7, 2026 at 12:46 PM

Not quite clear what you mean by “sensor” WHFM vs., instrumental (i.e., per sub) FWHM. In my experience (drawn from using remote setups for few years) Atacama, Namibia and southern Spain have an instrumental seeing between 1.5” and 2” on long integrations (300s to 600s) and a shade less for 30s to 60s (down to 1.1”-1.2” for narrowband on very good nights). That is on good nights, even Chile’s Atacama experience bouts of (relative) bad seeing but never registered worse than 2.5” across the board for several hundreds hours’ worth of integrations.

Hi Andrea,

I mean the discrepancy between advertised FWHM values through the night in HCRO/El sauce site (I believe they have specialized sensors for that, that calculate turbulence in arc sec) vs calculated FWHM through your subs.

Interesting to know that your highest FWHM value in El sauce was 2.5, can you please let me know your resolution?

Thanks

I have noted that FWHM tends to be higher on images than on seeing monitors, something I have noted both at home and at my remote site. They are proportional, however. FWHM on images improved when the monitor does. So not the same numbers but the same trend.

I think that seeing monitors should be mandatory at any remote site and am suspicious when they do not have that data available.

Real scope data is great but has the disadvantage of not creating a long term record measured every night in the same consistent way that one can use to evaluate a site. All “amateur” commercial monitors basically operate in a similar fashion and yield a theoretical seeing number based on “staring” at Polaris. Set up properly they should yield numbers that allow one to compare sites even if those numbers are not the same as one will see in real images.

There are three monitors I am aware of. The old SBIG unit no longer made (I have one at home), the Santa Barbara Scientific made by Alan Holmes (formerly of SBIG) and the somewhat expensive ALCOR. I think the most common these days is the SBS.

Just as an example, here is the last night readout from the monitor at my remote site. Note the seeing last night at about midnight was in the area on 1.2 arcsec (SBS brand monitor). A 600 sec Luminance image with my CDK 14 taken about then and at 52 degrees altitude showed a FWHM of 1.93 arcsec….

📷 SAROS.png📷 sfs.png

Rabeea Alkuwari · Jun 7, 2026 at 01:06 PM

andrea tasselli · Jun 7, 2026 at 12:46 PM

Not quite clear what you mean by “sensor” WHFM vs., instrumental (i.e., per sub) FWHM. In my experience (drawn from using remote setups for few years) Atacama, Namibia and southern Spain have an instrumental seeing between 1.5” and 2” on long integrations (300s to 600s) and a shade less for 30s to 60s (down to 1.1”-1.2” for narrowband on very good nights). That is on good nights, even Chile’s Atacama experience bouts of (relative) bad seeing but never registered worse than 2.5” across the board for several hundreds hours’ worth of integrations.

Hi Andrea,

I mean the discrepancy between advertised FWHM values through the night in HCRO/El sauce site (I believe they have specialized sensors for that, that calculate turbulence in arc sec) vs calculated FWHM through your subs.

Seeing monitors (which I suppose are what you are referring to as “sensor FWHM”) and what the FWHM of stars in the images you acquire are, functionally, two different measurements.

Seeing monitors make hundreds of short, 10-50ms exposures of a target star (in the case of SIMM type seeing monitors, that star is the relevant pole star) and track the movement of the star’s centroid over that body of several hundred or thousand images. This gives a somewhat point in time measurement of atmospheric turbulence. In the case of SIMM seeing monitors, the calculated Fried parameter, or atmospheric coherency length, of the pole star is modified with an accepted atmospheric scaling factor to convey what the FWHM would be if it were measured at zenith at 1 airmass. It represents the smallest possible motion FWHM that is possible at that moment.

This is different from your images for several reasons, in two primary ways:

1. Your images are probably acquired over longer timescales. Minutes instead of milliseconds. The light from the star(s) are smeared across more pixels during this longer time and is not a single snapshot of a star’s scintillation but a recording of it over the course of the duration of the exposure.

2. Whereas a physically-stationary SIMM seeing monitor can subtract the minute amounts of error caused by the sidereal movement of the pole star across the sample stack, and a DIMM seeing monitor’s dual view of the a star permits it to subtract out tracking and other mechanical effects of the mount it is on, you do not have the same capabilities on your imaging system. Tracking error and lagging guide corrections will further bloat the stars. The FWHM of your sub may also be an average of all stars across the image whereas a seeing monitor is measuring a single star near-ish to or at the center of the frame, so bloated stars out towards your image corners can further increase the a single averaged FWHM assessment of your image.

So the distinction to be made is that seeing monitors provide assessments of sky turbulence based on a measuring a single star. Your image’s FWHM assessment may be an average of all the stars in the image across a possibly imperfect field, on mount where mechanical imperfections or external disturbances such as slight breezes can further influence that calculation. So, yes, atmospheric turbulence will influence the FWHM in your subs, but it is just a component in addition to several other factors that also influences that value; some of which are in your control and some are not so much.

John Hayes · Jun 11, 2026 at 07:21 PM

Rabeea Alkuwari · Jun 7, 2026, 09:46 AM

I have a question about FWHM measured values through sensors Vs actual FWHM values measured in subs

I’m currently evaluating sites & options regarding establishing a remote observatory - the route I decided on pursuing was to look into the market for established setup to buy to potentially save on used equipment & shipping costs…this approach has the advantage of testing said skies and setup prior to any commitments.

I’m coming to this discussion a little late but I have a few things to add.

First, you are wise to consider buying a system in place at a remote facility. You are limited to what you can find for sale but that approach can save some money and a LOT of hassle and it’s a good way to start out with a remote imaging experience.

As for FWHM values, that’s a more complicated issue. There are a lot of different methods to measure seeing and some are definitely better than others. Most monitors are a good way to gauge relative performance. In most cases, a monitor that reads 3” on one night and 1” on the next tells you that the conditions should definitely be better on the second night! The harder question is how to tell what you should get from your scope when the monitor is reading 1” and here is what I’ve seen. IF the telescope is really well tuned for optimum guide performance, for achieving “perfect” focus, there is no wind, the optics are well aligned and you are looking near the zenith, with a moderately large telescope (<10”), the FWHM that you get out of your scope is the correct “long-term” seeing value. This assumes an exposure in the range of 1-5 minutes. IF the local seeing monitor is well tuned and operating correctly, you should see “similar” values for the seeing. Understand that most seeing monitors do not look at the zenith. They normally look near the pole. So, there will almost never be perfect agreement. It’s not uncommon to see a disagreement in the range of 0.2” - 0.5” but if you see 3” out of your system when the monitor is measuring 1”, something is probably wrong with one of those two measurements.

The seeing monitor at Obstech is a DIMM (differential image motion monitor) that was designed and built by Vincent Suc. It is similar to what most large observatories use. Here’s a picture of the DIMM seeing monitor used at Rubin (Vincent is on the right).

📷 image.pngIn my experience at Obstech, the FWHM values that I get from my 600 mm telescope using 300s exposures generally correlate fairly well with the local DIMM monitor. It’s not always perfect but the monitor can indicate problems if there’s a large disagreement.

In my experience years ago at DSW, the agreement was a lot less robust. So, while seeing monitors are useful, in my view, the results that come from the telescope are the last word. Just be careful. If the telescope always reports say 3” FWHM in a location where the typical seeing from the monitor is showing 1.5”, there might be something wrong with the scope.

John

Thanks for confirming my doubts John! I was about to send you this thread actually to get your opinion since in most of your image descriptions you mention something about FWHM measurements…so I knew you had the answer

Do you by chance know if FWHM data collected by the instrument you have shown in chile is publicly available? I have searched for El Sauce FWHM values and found a site that seems to be out of date with no graphs/data

A note about terminology here, that being what’s a DIMM and what isn’t when it comes to the category of “seeing monitors”.

The image of Rubin’s DIMM-type seeing monitor in John’s post is a true DIMM. These are generally a 10-12” RC or Cassegrain OTAs with a Hartmann mask fitted to the front. One of the two apertures of the mask has a wedge prism over it. In the referenced image, the tube sticking out the front of the top aperture of the mask is what contains the prism, and the other aperture is clear. The prism-equipped mask projects two spot images of each star in its field of view onto the sensor, the separation of which can be adjusted by rotating the prism.

The software that runs the camera and ingests the images will pick a spot pair of a single star and analyze it. Usually it’s of the brightest star in the field, as the exposure times must be fast (~10-50ms) to freeze the seeing-induced movement while maintaining adequate spot SNR. This is where the “Differential” in “DIMM” comes in. What is being measured is how the dual spots of the same star moves relative to each other. Motion induced by the mechanical motion of the OTA (buffeted by breezes) and the mount (tracking error, periodic error) can be nullified because both spot images will move in the same direction by the same amount. It is the differential movement of the spots relative to each other that will reveal atmospheric turbulence. The differential movement will also reveal some differential flexure in the optical path, but that could be negligible and also calibrated out. The quality of the OTA and mount matters here to some degree.

A DIMM type seeing monitor’s pointing is generally synced with the main observatory optic so that it can measure turbulence in the area of the sky that the main research optic is also looking at, with the requirement that the seeing monitor have sufficiently bright stars to use in that same neighborhood. They may also be operated in a way that tracks bright stars as they cross the meridian at zenith. How they’re operated in this manner depends.

The other type of seeing monitor that is in wide use is what I refer to as the “SIMM” type, where the “S” means “Single”. It’s a single optic and camera, but without a Hartmann mask. This means there is a single image of each star in the field of view. These types of seeing monitors would be the Alcor Cyclope or Santa Barbara Scientific SM series of units. These are designed to stare at the polestar and measure seeing by tracking its wobble across a sample set of images. These types of units are relatively inexpensive and require no mechanical tracking or movement as they generally utilize sensor sizes and lens combinations that permit the polestar to stay in its field of view all night and all year. They are limited to middle and higher latitudes, though, as readings become unreliable 20 degrees or less from the equator.

What many people refer to as a “DIMM” seeing monitor are often really “SIMM” types. There is no differential movement analysis going on; it’s just centroid tracking a single star. There are other seeing monitor designs such as C-DIMM, which is kind of a SIMM hybrid that utilizes two such units, and SH-DIMM, where “SH” means “Shack-Hartmann”, a maskless DIMM design that accomplishes seeing analysis via wavefront inspection.

-Dale

John Hayes · Jun 11, 2026, 07:21 PM

If the telescope always reports say 3” FWHM in a location where the typical seeing from the monitor is showing 1.5”, there might be something wrong with the scope.

Or maybe something is set up wrong with the monitor. I suspect that could be fudged to look better than it should pretty easily.

Bill McLaughlin · Jun 12, 2026 at 03:54 AM

John Hayes · Jun 11, 2026, 07:21 PM

If the telescope always reports say 3” FWHM in a location where the typical seeing from the monitor is showing 1.5”, there might be something wrong with the scope.

Or maybe something is set up wrong with the monitor. I suspect that could be fudged to look better than it should pretty easily.

I hope thats not the case in HCRO because the CDK 17 i experimented with looked tuned to me via ASTAP and defocusing a star to verify collimation. Tracking looked perfect as well with an encoded paramount scope.

Love to hear more from HCRO customers

andrea tasselli · Jun 12, 2026 at 07:38 AM

Tip and tilt errors aren’t the only components of seeing (the one recorder by the typical SIMM or DIMM), higher order do exist, signally defocus, which aren’t and they affect way more larger apertures than the motion components. My ever recorded worst seeing (~6”) was due to such a defocusing term.

Obvious things such as focusing and tracking issues were not present in my experiments…the only variable i was not able to verify 100% was the colimation due to the fact that the camera used in that setup didnt have video. But i took some test shots on a defoucsed star and it seemed like colimation is good enough and not to the extent that it reports x2 FWHM values compared to the monitros. Though I could be missing something hence why I opened this thread