Making a 300mm f8 CDK mirror set

Jump to:

• Reference sphere
• Convex secondary
• Ellipsoid primary
•   Correction, attempt 1
•   Correction, attempt 2
• Finalization

Preparations

The material to start with are a Schott 300x25mm Boro-Silicate blank for the primary and a pair of 130x15mm blanks of the same material for the secondary. The back side of the primary has been ground flat, down to #320 grit.



The sequence of work is roughly as follows:

• Grind and rough polish the secondary to a set with matching radii
• Core out the convex part to the required size
• Finalize polishing and figure to perfect spheres
• Hog out the primary, and core partly from the back
• Finish grind, polish and figure to required radius and correction
• Finally, take out the core

Mirror parameters

	Primary	Secondary
Diameter	300mm	106mm
Radius	1800mm	940mm
Correction	-0.69	0
Sagitta	6.25mm	1.50mm

The sagitta of the secondary reference sphere is 2.25mm. The secondary can be fringe tested against this reference sphere.
The primary can in principle be null-tested with the lightsource in the near and the knife edge in the far conjugate. These conjugates are at 983mm and 10630mm respectively, making this test quite impractical. The alternative is to use a nulling lens or interferometry.

Reference sphere

Hogged out with SiC #80 until a 2mm drill easily fit under a straight edge. Then ground the surfaces to spherical with #180, keeping track of the RoC by projecting a ceiling light back onto the ceiling. The #180 is applied unitil all deep pits are gone, lastly in the center of the convex part. This can be checked by looking through the glass to a bright light between grinding sessions and study the progression of the shiny pits. Whenever this is equally distributed and changing all the time it is time to move on.

 
Smoothing starts with #320, where still the RoC needs to be kept close to target. Alternate MoT and ToT spending more time on the side that moves to the desired RoC. All in all the phases up to and including #320 have taken about 2:30 hours.

After about 1:30 of polishing time on CX and CV each, there is still a slight haze. I estimate that at least as much time should be put in still.

 
Left an view on the polishing setup, nothing special. Right the current status under filtered luminiscent light (579nm and 546nm lines). Note the TDE on the CV reference surface (bottom) and TUE on CX target piece (on top).

 
After another 30min on each piece, the curves seem to match better. The Foucault image shows slight oblateness and a low edge, fairly typical for not fully polished mirrors.

 
After about 45min polishing with little overhang, changing between MoT and ToT, the reference sphere is getting better (left). Also, the haze on the surface is nearly gone.
Once again, 35min, the edge is slowly getting better (right).

 
One hour more, gradually widening strokes to average 25% diameter overhang. Although overall the zonal roughness decreases, there is now a rolled down edge. The images are increasingly pinched off, and taken with longer exposure times. This presents better resolution of small surface deviations. It is about time to get some quantitative interferometric analysis too.

 
Half an hour of narrow MoT/ToT strokes, up to 10% or so. Left shows the Foucaultgram, right the IF analysis. The edge has improved a bit, but overall correction has gone oblate. So the next round will apply narrow ToT and wide MoT.

 
After 15 minutes, the correction gets better, k=-0.025 and RMS error is about 4% (left). Especially the wide MoT prove to be effective, the edge however has not changed much.
After another 15 minutes of the same some widish MoT and mostly narrow ToT the rolled edge is improving (right). Now try to finalize with some more widish MoT to suppress the center a bit more, and normal MoT-ToT to stabilize the shape.

 
The edge is improving still, after 15 min of the same (left) but the center bulge has increased again. Concentrate more on wide MoT strokes (~35%) and add some 15% ToT.
Next round has been aborted when a scratch appeared, after about 7 min. The wide MoT have resulted in a lowered central region though, now there's a high 70% zone. The next session will apply normal strokes, appr 25% overhang.

 
After about 15 minutes up to 20mm overhang mostly ToT (left). The central disturbances are caused by a direct reflection, maybe from the mirror back, but the idea is clear. The central area gets flatter, while the rolled edge is narrowing but doesn't decrease. Maybe better to switch to max 15mm overhang MoT strokes.
After 15 min of up to 15mm overhang MoT (right), the rolled edge has narrowed. The central bulge will be treated at a later stage, for now let's do another 15min of around 5mm overhang MoT.

 
After 20 min 5-10mm overhang MoT strokes (left) the edge has almost straightened out. The central bulge has mostly gone due to some occasional very wide (60mm overhang) strokes. There is just a little more narrow MoT required, and only a touch of wide strokes. Then a more elaborate IF analysis is in place...
After 10 minutes of very narrow (5mm) MoT the edge now is slightly turned up (right). Maybe a few ToT strokes with moderate overhang will remedy this. Otherwise the curve seems to be within λ/20 on the wavefront, about 10nm RMS roughness.

 
Well, one minute of 10-15mm ToT didn't help a lot...(left)
After 10 more minutes fairly narrow ToT and Mot strokes, the shape has become oblate (right). Need to do some wider MoT.

 
Two minutes of wide MoT, and some narrow ToT (left)... Apparently need to concentrate on widish MoT to get the center further down.
Five minutes of wide MoT appear a bit too much and dug a hole (right). Need to smooth this with some moderate overhang alternated ToT and MoT.

 
Five minutes moderate overhang alternating MoT and ToT (left).
Another 10 minutes of moderate overhang, mostly MoT (right). The sphere is now quite good, 1.4% RMS wavefront error and a Strehl of over 99%. Roughness in the DFTFringe analysis is largely due to aircurrents. The edge is also visible under the KE though. The RoC is 925mm.

 
As a check, the images under KE null, with a regular white LED (left) and a bare laser diode (right). As measured with the laser diode, the measurable reversal interval is about 20-30μm. This interval is the position difference between noticing a shadow on the left and on the right side. Ideally it should be about 15μm for this mirror.
You can actually distinguish the slight central bulge and the rim just inside the edge. Considering the small magnitude (say, λ/50 wft error) this gives an approximation of the accuracy of the KE test in a null configuration.

Convex secondary

So now is the time to finalize the convex secondary. As a starting point, the RoC is somewhat longer than the reference sphere, and there is a turned down edge. After revitilizing the tool, the start is medium overhang ToT, some MoT and for about 30 minutes or so.

 
Left shows initial contact fringes, right shows what it looks like after re-establishing a well contacting lap and some ToT medium overhang strokes. The RoC of the convex piece is still larger than the concave reference sphere. Also there is still a fairly pronounced TDE.

 
After 20 minutes fairly wide ToT, some narrow MoT (left).
After 25 more minutes of the same (right), it seems that not a lot has changed...

 
After 20 minutes wide ToT with 1kg weight, some narrow MoT (left).
After 20 more minutes medium ToT with weight and narrow MoT (50/50, right). There is about 1/2λ of TDE and the overall curve still has too long RoC. Maybe I should concentrate more on narrower overhang MoT, to rise edge and center and bring down the 85% zone.

 
After 20 minutes up to 20% ToT with 1kg weight, some narrow MoT (left).
After 10 more minutes narrow 5mm overhang MoT (right). It does decrease the TDE, but takes a lot of time...

 
After 40 minutes narrow and later moderate overhang ToT/MoT (left).
After micro-channeling again, and 20 minutes mainly ToT narrow to moderate overhang strokes (right). The low edge is almost gone now. It seems there is some astigmatism, but this could be caused by the parallax in the setup.

 
After a few more sessions moderate (20-25%)overhang mostly ToT (left). The figure is now close to λ/20 wft as analyzed with DFTFringe.
Some more MoT, since I got the feeling that I got it backwards...(right). Added some wide strokes to equalize the RoC over the surface. Below is a DFTFringe analysis of the surface. The relevant bit is now good to within +/- λ/20.

Coring out

 
Pretty scary moment, putting a diamond cutter on the front surface. Still it cuts like the glass is butter, only one thing went wrong against the end: it is better to switch off the power before lifting the drill... The whole room is now full of glass-loaded water spats, and also the annulus dragged over the cored out mirror resulting in a scratch.
The secondary was beveled and then etched all around its perimeter to relieve stresses introduced in the glass.

 
On the left is an igram showing the surface before etching, on the right the same after etching.

 
After 5 minutes of polishing on the original lap the surface gets rougher, the sharp turned edge doesn't change. Now I am tempted to build a new lap in order to improve the surface.
On the right the refurbishedlap is shown.

 
After 15 minutes of polishing with moderate overhang the edge is fairly good and the overall correction is well within diffraction limit. Also the surface has become a bit smoother overall.
On the right the DFTFringe analysis, showing all to be wihin a +/- λ/20 WFT band. Analysis of the wavefront sugests a conic of 0.005, a Strehl of 99% and an RMS deviation of better than λ/60. I call this mirror done...

Primary

 
The show begins, dig out a 6.25mm deep curve in the 300mm blank, quite a job. To the right, sagitta at about 4mm, ground on the back of my old 250mm mirror. Interesting to note that the edge is hardly touched, this is due to the extreme overhang used in hogging out the blank.

Grinding

 
When the target sagitta is almost reached (6mm) the curve needs to be spherized. After roughing in the hole the center is usually deep, and the curve a strong hyperboloid. Therfore the edge zones need to be worked by rducing the overhang and also working ToT. This slowly reduces the flat edge as shown in the left image. Adding extra weight (15kg) helps a lot in speeding up the grinding. When the curve is as deep as it is now, almost F/3, progress seems to slow down otherwise.

 
The right sagitta is reached, and the smoothing with #180 and #240 started: the RoC is now spot on, 180cm (+/- 5mm). So it is time to partly core the mirror from the back. To do this, a puck is drilled out of 8mm float glass, with the 70mm holesaw also used for coring. This puck is positioned exactly in the center of the blank, and attached with a dab of silicone glue. This is used to guide the holesaw and obtain a core centered to better than 1mm.

 
It takes about an hour to get to a 15mm depth. The process required addition of some #80 grit, since the holesaw quality appeared not as good as expected. The groove is filled with plaster to mechanically stabilize the core a bit more.
Next grit is #320, which takes about 2 hours of alternating MoT and ToT, with moderate overhang strokes (i.e. 0.2 - 0.3 D). Then slowly down through #600, #800 and finally #1200. Right shows the mirror and pitchlap to be. It is made of a relatively hard casting plaster, which (after giving it sufficient time to harden and dry) is coated with a generous treat of Richards' Secret Sealer Sauce (i.e. pitch dissolved in turpentine).

Polishing

 
The pitch hardness tester I made according to Texereau Annex J sinks 3.5mm in 5min, at a temperature of about 21°C. The pitch is 26° as obtained from Stathis. To the right the lap as it is after an hour of polishing. Channels have just been redone.

 
Polishing progresses well, but the core gives a problem. As can be seen it is definitely not caused by stress introduced while cutting it out. Possibly it is the slight concaveness of the back, that allows the core to move a little under pressure.

 
The back has been re-ground, with #80, #240 and #320, to get rid of the concave center. Also polishing is now done on a firmer substrate, consisting of a single layer wet chamois. The fabric type has consistent thickness, and the underlying stone slab is quite flat as well. The result of 30min narrow strokes, most ToT, is shown on the left. Not much has changed regarding the area around the core.
To the right, the result after 20 minutes of 25% average overhang MoT strokes. Not a lot of change, center seems to be deepened more. Maybe best to first concentrate on the edge and correct the center last.

 
After 40 min moderate ToT strokes. The tool did not make good contact fow a while, and the air was quite restless. Anyway, the edge does not seem to progress a lot. On the other hand, the images may not really be comparable very well, due to differing conditions. It would be better now to obtain absolute values with the interferometer.
To the right the mirror after 40 minutes of edge correction with very narrow (10%) mostly ToT, followed by 30 minutes of moderate (25% average) strokes alternating MoT and ToT every 5 minutes. The shape look better now.

 
Moving the knife edge inward a few mm shows above series with various lateral knife positions. I guess this implies that the shadow contrast in the previous foucaultgram is pretty high, and exaggerates the surface deviations. Indeed the light is almost pinched off and the shadows are hardly visible with the naked eye, an exposure time of 0.5 seconds is required.
To the right the Bath interferogram analysis, 4 igrams at each of 4 orientations. Needs some work stil...

 
After 5min wide MoT, 5 min narrow ToT and 5min wide MoT (left). Not much change...
After 20 min of very wide MoT, up to and sometimes beyond the groove (right). Central bulge is mostly gone now, mirror in principle is diffraction limited. Need to address the roughness though, by somewhat more moderate strokes, MoT and ToT. Also next time: measure more accurately in 8 orientations.

 
After 5min D/6-D/3 MoT, 5 min D/6 ToT and 5min D/6-D/3 MoT (left). The shape has reversed, checked with the thumbtest.
After 15 min Mot/Tot up to D/6, MoT a few 45% overhang to get the center down (right). Crater has returned, so probably the recipe is to do a few wide MoT, but not beyond the groove, and some moderate ToT. After that I might as well just go for the target -0.69 ellipsoid. Hopefully this can be done with the current pitch, maybe I need to pour a softer and smaller lap.

Correction attempt 1

The correction towards a -0.69 ellipsoid is done by means of a 2/3 subdiameter tool covered with softer pitch (5.5mm/5min at 21°). The pitch used here is a bit softer, the 23° from Stathis.

 

Since circular hand strokes are used, the relative speed of tool over mirror is homogeneous. Hence the amount of wear is assumed to be proportional to the percentage of contact per stroke, for a given radius and using a certain radial displacement. According to this graph, using a stroke that has barely overhang will deepen the center and correct the outer zones. This stroke will be used to start with.

 

After 15min with 0.35 offset (left), not a lot has happened.
After 30min with avg 0.3 offset 1kg added weight (right), the center seems to have deepened a bit, but too marginal. Also the effect of the core cut is more pronounced, time to grind the back flatter.

 

After 15min with 0.30 offset and 5kg added weight (left). Before this the scary process of grinding the back with #80 has been carried out. It is a Q&D Bath anaysis, but you can see that things have reversed now, the center is actually sinking.
After 30min with slightly larger, average 0.40 offset + 5kg (right). No real change, the system seems to have reached an equilibrium. Maybe it is better to lower the center first with narrow strokes, 0.20 offset.

 

After 20min 0.10 - 0.30 offset + 5kg (left), the center actually seems to be sinking but progress is too slow. Correction is now -0.02 of the target -0.69. Maybe it is an idea to try MoT on a subdia tool, or use a smaller polisher.
The right image shows the surface after bringing it inot contact with the 10cm tool described below. Note that the central depression required for correction develops much faster now.

So it appears a smaller polisher is required to get better control over the development of the correction. For this purpose a 10cm (D/3) plaster tool is cast and covered with the medium pitch that was also used for polishing out (left under).

 

The expected effect of this polisher is recalculated and is shown on the right. Note that the polishing effect is now much more localized, where obviously the relative contact on the inside the center trajectory is larger than that on the outer side. It mat be a good idea to either soften the polisher edge a bit by cutting away parts of the pitch. We'll see whether this is actually neccessary based on further interferometry.
Also the mirror back has been ground once again, to get it truly flat, down to #320. Even now there remains a small dimple in the center, but it is smaller than the core diameter. A lot of damage can be done with just a few #80 chordal strokes...

 

After 15 minutes of trial strokes with the 1/3 tool, no weight and varying over zones, the figure hasn't changed a lot. So I guess that the weight will add some action as well a concentrating on the center a bit more. Left image shows as compared to a sphere, right image as compared to the target conic. There is still a lot of glass to be removed...

 

After 30 minutes with the 1/3 tool and 1kg weight, 0.3-0.5 tool offset, analysis defocused -0.3λ (left). The center is deepening now compared to the relatively untouched edge, but progress is slow... After 30 min 1/3 tool + 1kg, 0.1-0.3 offset, defocused -0.8λ (right). Correction is now about -0.1.

 

After 30 minutes with the 1/3 tool and 1kg weight, 0.3 avg tool offset (left). The correction is now -0.13, but the graph shows it as compared to a reference of -0.2. The center requires some more attention, so narrow the strokes. Same time, some edge work might pay off as well. The roughness is a consequence of the small tool, and should at some point be smoothed with the larger tool.
After 30 minutes narrow strokes with 1/3 tool and 1kg weight (right). The correction is -0.23 and the graphs shows against -0.3 reference curve. The strokes should be widened to also deepen the zones up to 70%. First try to decrease the roughness with the 2/3 tool.

 

After 15 minutes with the 2/3 tool and 5kg weight, narrow strokes (left). The bump at 80mm zone is less acute now, but the center as a whole has risen. Next continue with the smaller tool.
After 30 minutes 1/3 tool and weight, mostly center (0.25R offset) to keep dead center in 100% contact, again -0.3 reference (right). Some excursions to the edge zone were made as well. Correction is now at -0.27, so I need to continue this stroke.

 

After 60 minutes, 1/3 tool and 1kg weight, 0.25R offset, compared to k=-0.3 reference surface (left) and k=-0.5 (right). Correction is now -0.35, center half looks more like -0.4, and where the deviation seems to increasingly come from the edge zone. Continue the stroke for another hour perhaps increase the average width somewhat to propagate the correction outward, and don't forget the very edge every now and then.

 

After 60 minutes, 1/3 tool and 1kg weight, 0.2R~0.4R offset, compared to k=-0.5 reference surface (left). Correction is now -0.44 still with the high edge zone. It might pay-off to get some more edge work done, apart from further deepening the center.
The new DFTFringe 3.0 output (right) does not show any differences compared to the version 1.12 analysis, so this is what will be used furtheron.

 

Same treat for an hour, compared to k=-0.5 reference surface (left) and to k=-0.6 (right). Correction is now -0.51.
The increased edge treatment has sorted some effect, comparing left analysis with the previous session. More attention needs to be paid to lowering the 85mm zone, by spending more time with 0.5R offset. In any case, the center needs to be taken down even more, but progress is still on track with about -0.1 per hour.
For now: one more session with the same stroke, then spread the correction with the 2/3D tool.

 

Same treat for an hour, compared to k=-0.6 reference surface (left) and to target k=-0.69 (right). Correction is now -0.57.
Next session is for smoothing, use 2/3D tool with the edge approximately 10mm overhanging.

 

After 20 minutes with 2/3D tool, edge just overhanging. Curves as compared to k=-0.6 reference surface (left) and to the target k=-0.69 (right). Correction is now -0.54, this is less than after previous session. Ths seems not to be the proper way forward, although it seems that the only things changed are indeed the decreased bump at 85mm, but also the risen center. Back to 1/3D tool, maybe overshoot target correction first and then revert with the 2/3D tool.

 

After 1 hour with 1/3D tool plus weight, mainly working on center, some over 85mm and some over the edge (left). The curve can from now on be compared to k=-0.69 target reference surface. The overall correction is now -0.64. In general the whole portion inside 125mm needs to worked on, but the 85mm radius needs more work, as well as the very center.
After 20 minutes 1/3D concentrating on center and also 85mm zone (right). (Meanwhile DFTFringe switched to 3.1)

 

After 15 minutes with 1/3D tool plus weight, mainly working on 85mm zone (left). After 20 more minutes with 1/3D tool plus weight, mainly working on 85mm zone (right).
Seems that another 30 minutes is needed, then focus on edge zone.

 

After 30 minutes with 1/3D tool plus weight, mainly 85mm zone (left). Time for some smoothing with 2/3D tool and about 10-20mm offset.
After 15 min with 2/3D plus weight, mostly 20mm offset (right). Now is probably a good time to start flattening the edge, and do some center work (25% offset) on the fly.

 

After 15 minutes with 1/3D tool plus weight, mainly edge and some center (left). Not a lot has happened, the tool may be too wide to address the edge only. Alternative is to deepen the central part instead with the 2/3D tool, that would lower the central part, shorten RoC and virtually lower the edge. As a result I expect a central hill due to reduced correction.
After 30 minutes 2/3D plus 5kg (right). It yields some progress, but is probably still a lot of work. Let's get back to lowering the center and overshoot it.

 

After 30 minutes with 1/3D tool plus weight, center and 85mm zones (left). The bump has moved from 85mm to 100mm radius, so I need to widen those specific strokes. The work on the center part should leave the inner 20mm radius covered, so max 30mm offset. Overall correction is now -0.68.
After 30 minutes 1/3D tool plus weight on 100mm and center zones (right). The spread of the lines is a bit worrisome, I need to check whether this is real or a consequence of the way it is tested (only averaging 0°, 45°, 90° and 135° angles). Maybe a better strategy is to return concentrating on the hill at 95mm. Make 1/3 tool strokes that vary with the edge over 95 - 130mm zones, to reverse the slope.

 

After 30 minutes with 1/3D tool plus 1kg, edge varying between 95 and 130mm radius (left).
After 15 minutes with 2/3D tool plus 5kg, keeping edge within 135mm (right). The intention was to make the slope between 100 and 135 less steep, but all that has happened is that the high edge is lowered. I am tempted to apply some rounds of localized pressure over this stubborn ridge at 100mm.

 

After 2 rounds of accentuated pressure over 80-135mm zones, using 2/3D tool (left). Each round means two full passes over the zone (out-in-out-in-out), for each of 12 orientations around the mirror. It seems not a lot has happened to the zone in question, I guess the mirror responds slowly... The bit in the center probably is caused by insufficient thermal stabilization.
Same recipe, 3 rounds, but slower strokes with more pressure (right). The ridge finally shows some erosion, I think 4 more rounds will take it close to target. Tempertures are quite low, lots of turbulence due to heating and hence noisy igrams. Next one should be taken at all 8 orientations.

 

After 4 rounds of same stroke (left). What I feared seems to have become true, astigmatism and other mayhem (right). The interferometer is quite merciless.
So, what to do next? Maybe start with an attempt to achieve some damage control, small tool, working down the high spots and astigmatism, measure often...

 

First attempt to correct roughness (profile left, 3d view right).
As related to 0° orientation: some passes with accentuated pressure over 120mm zone for bottom half, tool center over edge for top half of mirror. This was using the 1/3D tool.
The edge on the top side seems to have improved a little, the 120mm zone on the bottom not at all. It might be an idea to repeat the previous accentuated pressure stroke for the bottom half only.

 

Second attempt to correct roughness (profile left, 3d view right).

 

I made a new tool, approximately 65mm diameter (1/5D), to attack high zones more specifically. The above images show the result of the first round using this tool, approximately 10 minutes treating south half 120mm zone and north half edge zone. It seems to be slowly moving into the right direction, although the igrams were a bit noisy. Let's have one more of those sessions.

 

One more 10 min session later with 1/5 tool over 120mm zone and center. The curve (left) is rotated to match the 3d picture, left is bottom and right is top of the mirror.

 

Another 10 min session, same treat. Again, the curve (left) is rotated to match the 3d picture. Strangely, the glass seems to be hardly affected...

 

After 10 min on center up to 80mm and 5 min over 120mm zone, using 1/3D tool. This combination has helped to get the trench at 80mm up, but the edge is now sinking. So next 10 minutes use the 1/5 tool and only work on the 120mm zone, keeping the tool edge right on the mirror edge. Maybe also try to doubly work the high part of the ring between 3:00 and 12:00. (The 3-d image is left-right mirrored!)

 

After 10 min over 120mm zone, using 1/5 D tool. This tool works slowly, but that's ok at this stage. A few more sessions on the 120mm zone are required, targeting the high parts especially.

 

After 10 min over 120mm zone, using 1/5 D tool, using a little more pressure. Need to work the north edge, the south 120mm zone and the 35mm zone. Last one reacts fastest, so requires less time. Do it in one go, few minutes each.

 

After 5 min over 120mm zone (south) and 140mm zone (north), and 5 min over 35mm zone, using 1/5 D tool.

 

After 5 min over center, 30mm offset, with 1/3D tool and 5 minutes working high areas in 120mm zone, using 1/5 D tool. Overall correction is now -0.683, but the surface is too rough.

 

After 5 min roughness combat, with 2/3D tool, no weight or pressure. Rapid radial strokes, between 30mm overhang and 30mm inside edge, while slowly walking around the mirror. The 3d plot shows that the roughness tangentially has somewhat reduced, so this may be the way to go. Let's do another 10 minutes or so.

 

After 10 min rapid radial strokes, with 2/3D tool, no weight or pressure. The ridge at 120mm has mostly gone now, and the surface is more regular. Next will focus on center, 1/3D tool with edge up to 90mm zone, max 5 minutes. The edge will seem to flatten at the same time as the center sinks.

 

After 5 min with 1/3D tool and weight, max 40mm offset. Center is lower according to plan, but the pit needs to be treated with the 1/5D tool over 40-50mm zone (i.e. not covering the mirror vertex).

 

After some local treat with 1/5D tool, 50mm offset. This small tool certainly has little effect, but still the few high spots that were treated locally have decreased. Next is taking the center down some more and smooth with the larger tool.

 

After 5min over center with 1/3D tool and weight, 50mm offset, followed by rapid radial strokes with the 2/3D tool, no weight or pressure. Progress is minimal, let's get the center down now with a combination of 1/5D over 40mm and 1/3D with 50mm offset.

 

After 5min 1/5D tool over 40mm and 5min 1/3D tool and weight with 50mm offset. Not sure how much of this is the measurement and how much real, since not only the center was affected. Average correction is spot on at -0.69, but the surface still quite rough. It might be a good idea now to us the full size lap with short strokes, alternating MoT and ToT.

 

After 10min short strokes with full tool, MoT and ToT. Bad idea, not a lot of smoothing, but I do have some scratches now: contaminated lap. Back to intermittent use of 2/3D and small tools, which seemed to give better results... Start with 5min 1/3D edge up to 100mm, then 5 min 2/3D radial strokes up to 30mm overhang.

 

After 5+ min 1/3D + weight 50mm offset, follwoed by 5min rapid radial strokes with 2/3D no weight. Correction slightly overshot (k=-0.7), overall surface seems less rough. Need to get the 70-100% zone down a bit, using 1/3D tool reaching just inside the edge.

 

After less than 5 min 1/3D + weight, edge just inside mirror edge. The 100mm zone has sunk, but obviously the center has risen as a consequence. So now try a short session treating the ridges on 35mm and 120mm with the 1/5D tool, evaluate and probably smooth afterwards. It's a small steps up and down process, but slowly the surface gets less bumpy while the overall correction remains under control. What keeps worrying though are the two persistently low areas (blue in the right image).

 

After some local retouche with the 1/5D tool, on 35mm and 120mm zones. It gets touchy, 35mm zone is now too low instead of too high. Treat with 1/3D over 100mm.

 

After just a couple of rounds with the 1/3D tool + weight on the high area between 50mm and 150mm. Seems like a repetition of moves... Also, it matters enormously how long the mirror and garage are allowed to stabilize. Another point of attention may be the sharp ridge at 35mm, possibly indicating that the mirror also requires more stabilization before the polish starts. Next session let's focus on the center with 1/3D tool ofset only about 20mm or so.

 

After just a couple of rounds with the 1/3D tool + weight on the high central area up to 30mm offset. This time the mirror has thermally stabilized for a few hours before starting the polish. The overall correction is -0.683 now, and the Strehl 86%. It needs a few more licks over the 35mm ridge and some over the 120mm ridge, small steps.

 

After some tweaking with the 1/5D tool on 35mm and 120mm offset. K=-0.685, Strehl=89%.

 

After two session of rapid radial strokes with the 2/3D tool, no pressure. The stroke length is about 60mm, between +/- 30mm overhang, slightly elliptical. During this I walk once around the mirror, and do this for 12 mirror orientations.
Even though the scale is smaller, the roughness seems decreased. Now let's get the center down again.

 

After a session with the 1/3D tool, no pressure, offsets varying between 10mm and 40mm. Center is a bit lower now, continue.

 

After a session with the 1/3D tool plus weight, offsets varying between 10mm and 40mm. Apart from the very center, not a lot has happened, no idea why...

 

After a session with the 1/3D tool plus weight, offsets varying between 10mm and 40mm, 8 directions, appr 5 minutes. Obviously the edge zones seem to turn down now, so now use the 1/5 tool to treat that ridge at 120mm. Then get back to lowering the center.

 

After localized work with the 1/5D tool on the 120mm zone, appr 5 minutes. Not much change...

 

After a session with the 1/3D tool plus weight, offsets varying between 10mm and 40mm, 8 directions, appr 5 minutes. Let's now try to lower the ridges at 35mm and 120mm. They've been there forever...

 

Some ridge riding with the 1/5D tool, over 120mm and 35mm offsets, 4 directions each, appr 5 minutes. Strehl is about 90% again, correction almost on target, surface still looks ugly though. The high ridge at 35mm is exactly over the location of the coring groove. I think it needs some more attention, since not much changed since last session...
But first, lets shave off the high parts with the 2/3D tool.

 

Smoothing with 2/3D tool, 2 rounds of 4 orientations, 30mm overhang. Now back to lowering the ridges again.
I also need to pay more attention to thermal issues during interferometry. The analyses are all over the place, not so sure anymore about the value of the average, maybe transient effects during a series bias certain artefacts. This could be tested by starting a series from another orientation.

 

After several sessions with localized work using the 1/6D tool.
The Strehl is now 91%, correction -0.691. Roughness is still a problem

 

After several more such sessions interleaved with smoothing, the Strehl is now 94%, correction -0.692. Roughness remains, and now also some scratches have appeared. Moreover, the vertex of the ellipsoid appears to be several mm offset from the center of the mirror. This imposes to much disalignment to be corrected in the cell.
So I have decided to grind with #600 and #1200 powders and do the polishing / correction again.

Correction attempt 2

The primary has been ground back to a smooth sphere, with a nice edge zone, so it is ready for a second attempt at correction.
The approach is again to obtain the desired correction with a mix of tools, of 2/3, 1/3 and 1/5 diameter size. The effect of each tool has been evaluated, so a more educated guess can be done regarding effectivity of a treatment. The other variables are added weight and duration of a polishing session. The figure is measured by means of interferometry.



Three examples of polishing sessions are shown above, the curves represent the differential surface analysis before and after the session.
The 2/3 tool (200mm) will primarily be used to correct outer regions in the first stages of figuring, while leaving the center relatively unaffected. Obviously it does wear away, but without any gradient. The gradient is the zone where the tool has contact for less than 100% of the time.
The 1/3 tool (100mm) has a more localized effect, and will be used to bring correction further towards the center, while leaving outer zones relatively unaffected.
The 1/5 tool (60mm) will be used mainly to correct small remaining deviations, mostly towards the end of figuring. The trick with this tool is to take care not to introduce additional ripple while removing other.
The process is trial and error, since this is the first mirror that I figure by exclusively working ToT.

Finishing up

It's been a while since last work on the mirrors. Now, late 2021, it's time to re-evaluate the status and decide how to finalize these optics. So first off, the secondary reference and convex secondary were measured again:



The reference to test the secondary against is a very good sphere, better than λ/50, Strehl ratio larger than 0.99.



The secondary was contact-tested against the reference, with diverging laser and camera placed in the CoC. Due to the refraction in the secondary glass the distance is about 2/3 of the real radius of curvature. The astigmatism due to parallax seems to be absent...
The curve as analyzed does not show gross deviation from the reference glass, so I conclude that the curve is close enough to a sphere.