PARAMETRIC SOLAR FILTER ADAPTER - THREADED DEW SHIELD TO SLIP-ON FILTER CELL

PARAMETRIC SOLAR FILTER ADAPTER - THREADED DEW SHIELD TO SLIP-ON FILTER CELL

This is a template, not a single part. It fits any refractor whose dew shield has a female thread at the front, to any felt-lined solar filter cell whose bore is larger than that thread. That is the common case: you own, or find secondhand, a filter cell made for a slightly bigger telescope, and nothing bridges the gap.

You change eight measured numbers at the top of the OpenSCAD file and you get your adapter. The thread is parametric, the cell interface is parametric, and the model refuses to compile if your measurements cannot produce a sound part.

The ready-made STL are two worked examples: William Optics Megrez 72 (M92x1) with a Thousand Oaks Type 2+ cell sized for a Meade ETX-90 / Canon 70/200 f2.8 with custom Baader AstroSolar cell . Print them directly only if that is exactly your combination. Otherwise edit the SCAD file.

The custom Baader Astrosolar cell was made with the excellent model from ThomasQuinot  and I have left a STL of this design which fits well with my Canon 70/200 filter adapter.

The filter is held by friction on the cell's own bonded felt, so it comes off with one straight axial pull. That matters during a total eclipse: at second contact you have seconds to remove the filter in near darkness, and unscrewing a 90 mm ring takes several turns, two hands, and puts torque straight into the tube - enough to shift your pointing on a lightly loaded mount.

There is no flat flange on this part. A single continuous taper runs from the friction diameter out to just past the cell bore, and the cell seats on that taper. That is not styling - see WHY A TAPER AND NOT A SHOULDER below.

Made for the eclipse of 12 August 2026. Useful for any white-light session.

-------------------------------------------------------------------------

WHY A SCREW-ON ADAPTER RATHER THAN A FILTER THAT JUST SLIPS OVER THE TUBE

-------------------------------------------------------------------------

A filter cell that slips straight onto the dew shield is simpler, and for visual observing it is the right answer. This adapter exists for a different job: an automated, unattended, two-camera eclipse run.

I shoot the eclipse with two chains running in parallel - AstronoMolly's N.I.N.A. eclipse sequence on the telescope, and SETnC driving a Canon EOS 6D alongside it. Both are on the clock, both are triggered on absolute time, and neither waits for me. The only thing I do by hand during the whole event is pull both filters off a few seconds before second contact and put them back at third contact.

That is 100 seconds apart on a lightly loaded alt-az mount carrying two instruments. Anything that transmits torque or a sideways shove to the tube between those two moments costs pointing - and pointing is what the sequences assume has not changed.

So the interface is split in two:

• a threaded ring that goes on before the session and never moves again,

• a filter that seats on it by friction and comes off with one straight axial pull.

Screwing a 90 mm ring on and off takes several turns and two hands, and puts a couple straight into the tube. Slipping a cell over the dew shield itself is better, but the dew shield is often the part that is retractable, or has a seam, or is not round enough for a repeatable fit. The ring gives you a surface machined to your own tolerance, in the right place, that you can rehearse against twenty times before the day.

Which is the real point: the geometry here is chosen so that the removal is a single reflex movement you can perform in near darkness without looking, while two computers keep shooting. That is worth a printed ring.

-------------------------------------------------------------------------

WHAT IS IN THE DOWNLOAD

-------------------------------------------------------------------------

• thread_gauge.scad Parametric test ring. PRINT THIS FIRST.

• thread_gauge_M92x1.stl Ready-made gauge, M92x1

• thread_gauge_M93x1.stl Ready-made gauge, M93x1

• solar_filter_adapter.scad The adapter. Edit the top block.

• solar_filter_adapter_M92x1.stl Ready-made, Megrez 72 + Thousand Oaks Type 2+

• solar_filter_adapter_M77x0.75.stl Ready-made, Canon 70_200 f2.8 + Custom Baader AstroSolar cell

• solar-filter-nubs-Canon.stl , Ready-made custom solar filter for Baader AstroSolar made with the ThomasQuinot design.

The SCAD files need BOSL2, unpacked into your OpenSCAD "libraries" folder:

https://github.com/BelfrySCAD/BOSL2

The STL files need nothing.

-------------------------------------------------------------------------

BEFORE YOU PRINT: VERIFY YOUR THREAD

-------------------------------------------------------------------------

Manufacturers rarely publish dew shield front threads, and they are not the same as the rear tube thread quoted for focuser upgrades. Print the gauge, not the whole adapter - fifteen minutes instead of a wasted 30 mm part.

Measure the male part of whatever screws into your dew shield now:

1. Major diameter over the crests, three positions, take the largest.

2. Pitch across TEN crests, divided by ten. Not five - the error doubles. Telescope tube threads are nearly always 1.00 mm. 0.75 turns up on photographic fittings.

3. Cross-check: the female minor diameter should be about nominal minus 1.0825 times pitch.

Then print the gauge at your best guess and at the next size up.

PASS = starts by hand with no tool, four full turns with no hard spot.

A gauge that forces for one turn then jams is the WRONG size. That is crest interference, not engagement.

If neither size starts, raise COMP to 0.35 before you give up on the size - FDM over-extrudes on thread crests and that is the usual cause.

-------------------------------------------------------------------------

WHY A TAPER AND NOT A SHOULDER

-------------------------------------------------------------------------

The obvious design is a flat flange the cell bottoms out against. It is also unprintable without support, and not for the reason you would guess.

A flat seat necessarily sits at a LARGER diameter than whatever is under it.

Printed thread up, its first layer is an isolated ring hanging in free air around its whole circumference - about 340 mm of it here. It curls, and it can detach. No cooling or speed setting fixes that: there is nothing underneath to anchor to.

A conical seat has no such layer. Material grows outward from the register at 30 degrees, every layer resting on the one below. The cell seats on its inner rim edge instead: self-centering, and it cannot wedge, because a cone only self-locks below roughly 7 degrees from vertical.

SEAT_FLAT will give you a flat seat back if you insist. The file refuses to compile until T_WEB is also raised enough to keep the slope, and it echoes the resulting flat width so you can see what you are committing to.

One consequence worth noting: because the cell lands where the taper reaches its own bore diameter, the widest part of the adapter is the cell BORE plus 1 mm, not the cell outside diameter. Anything wider touches nothing and just stands proud of the dew shield.

-------------------------------------------------------------------------

PRINT SETTINGS

-------------------------------------------------------------------------

ORIENTATION: THREAD UP. This is not a preference.

Every overhang in the part is chamfered at 45 degrees or tapered, so it prints with NO SUPPORT in that orientation. Printed thread-down the slicer fills the bore with supports and the friction surface comes out too rough to use.

The gauge prints axis-vertical for the same reason. Printed flat, the thread is built up layer by layer across its profile and is unusable.

MATERIAL: PETG or ASA. NOT PLA.

PLA's glass transition is around 60 C. A black ring in full August sun reaches that easily and deforms - at the worst possible moment.

• Layer height 0.15 mm

• Perimeters 3 to 4

• Support None

• Brim Yes. The first layer is a narrow ring.

• Infill 15 to 20 percent. The walls are thick enough on their own.

STRINGING: the usual culprit is retraction, not cooling. High-flow PETG on a high-flow nozzle holds more melt and needs MORE retraction than standard, not less. Profiles shipping 0.4 mm are far too short; 1.2 mm is a better starting point. Also enable retract-on-layer-change, drop the nozzle to the low end of the recommended range, and set the overhang fan to full with a low engagement threshold. Put the seam on Aligned or Rear: a seam blob on the taper is the one surface defect with a mechanical consequence.

-------------------------------------------------------------------------

ADAPTING IT TO YOUR HARDWARE

-------------------------------------------------------------------------

Eight measurements with a caliper. Take them from male surfaces where you have the choice - outside surfaces measure more reliably than bores.

• D_NOM, PITCH Dew shield thread. See the section above.

• D_DEW Dew shield outside diameter.

• D_CELL_OD Filter cell outside diameter.

• D_CELL_ID Cell bore, measured in the BARE zone near the mouth, not over the felt.

• T_FELT Felt thickness per side: (bare bore - felt bore) / 2.

• DEPTH_CELL Cell inside depth, mouth to whatever stops you.

• L_FELT Width of the felt strip.

Everything else derives from those. The spigot length follows your felt width automatically. The centering register follows the unfelted part of the bore, and disappears cleanly if your cell is felted right to the mouth.

WORKING ENVELOPE, enforced by assertions in the file:

• cell bore greater than dew shield thread plus about 8 mm

• cell OD at least dew shield OD

• cell depth greater than felt width

Outside that the error message names what is actually wrong, rather than leaving you with a part that looks right on screen and does not fit.

D_APERTURE Clear aperture of your objective. Not a dimension of the hardware you are adapting - it is the number the model uses to check that it will not vignette you. Set it honestly. For a camera lens use the entrance pupil at your working f-stop, plus the off-axis allowance: a 200 mm lens at f/4 on full frame needs about 62 mm, not 50.

The bores are derived automatically from the thread and the spigot, so they follow your measurements. Two assertions refuse to compile if either bore would vignette D_APERTURE.

Note the physical limit that exposes. Bridging a filter cell to a dew shield needs roughly 8 mm of wall between thread and bore, so a small thread cannot pass a large aperture. If the vignetting assertion fires, no parameter will fix it: that combination of cell, thread and objective is impossible.

Do not hollow the bores out by hand to save filament either. Opening them by 3 mm saves about 15 g and takes the wall under the thread down to 2 mm, which is not enough under screwing torque in PETG.

Three more parameters you may want:

• ANG_OVH Overhang angle, degrees from vertical. 30 by default. Drop to 20 if your filament droops; the angle actually achieved is echoed at compile time, since the available length can force something steeper.

• SEAT_FLAT Leave at 0. See the taper section above.

• LAND_REG 0 merges the two slopes into a single taper. Any non-zero value inserts a cylindrical band between them and the part visibly grows a second flange for no functional gain.

-------------------------------------------------------------------------

FITTING

-------------------------------------------------------------------------

One parameter sets the friction on the felt:

CLR = -1.0 NEGATIVE = smaller diameter = LOOSER

The default is deliberately on the loose side. Bonded felt has local high spots, and a fit that grips in three places releases with a jerk that drags the telescope off target. Go to -1.5 if it still grabs, -0.6 if it slips.

Target behaviour: holds when the assembly is inverted, releases with one firm axial pull. Aim for REPEATABLE rather than tight. You will make this movement twenty times in rehearsal.

-------------------------------------------------------------------------

CHECKS AFTER PRINTING

-------------------------------------------------------------------------

1. THREAD. The web must seat flat on the nose of the dew shield. Rock the ring by hand. If it moves, the relief groove is too shallow and you are binding on the last thread.

2. TAPER. The cell must slide on and come to a firm stop on the taper, square, with no rock. The taper aligns; the felt does the gripping.

3. FRICTION. See the section above.

4. LIGHT LEAK. In full sun, filter fitted, look through the empty focuser. No stray light may reach the tube around the joint.

5. REHEARSE. Twenty removals and refits, until the movement is automatic.

-------------------------------------------------------------------------

SAFETY

-------------------------------------------------------------------------

A solar filter fits on the FRONT of the telescope, never at the eyepiece end.

Before every session, hold the filter up to a bare lamp in a dark room and look for pinholes, scratches through the coating, or any lifting of the metallic layer at the edge of the cell. If in doubt, do not use it.

This adapter is a mechanical mount and nothing more. It does not make an unsafe filter safe.

-------------------------------------------------------------------------

Dependency: BOSL2 - https://github.com/BelfrySCAD/BOSL2