Mirror making log for my 300 mm f/5; 30 mm thickness pregenerated pyrex mirror

This is the story of the making of my first mirror: a 300 mm f/5. The grinding and polishing went rather straightforward, but parabolizing was another thing! I parabolized using a full sized pitch lap, as advised by Mike Lockwood at his homepage.
The log is rather long. The links below can be used to jump to the important topics in the log.

Rough grinding
Fine grinding
Inspecting and repairing the pitch lap
A new pitch lap
High edge problem
Primary ripple
A good mirror at last!!

Rough grinding


1/15, 22, 29/2005

Started with  carborundum 80 MOT. After about 6 hours, mirror is reasonably spherical according to the micrometer readings  and has about 4% longer F then f/5. Micrometer reading = 0.56. Must be 0.54 for f/5. Proceeded with carborundum 120.


2/5 120 carborundum

MOT. Ground too long without testing. Reading 0.53 now. Proceed with 120 grit, TOT.

After about 2 hours arrived at 0.545.


2/12 180 grit

Started grinding at home to make more speed (can only go to the observatory once a week). Ground about 75 minutes with 180 - 320 grit. Checked for pits every 75 minutes. No pits. Started sharpietest at 220 grit. Not enough contact at the edge. Did 1 hour  TOT, which I shouldn't have done. Ground 3 hours MOT with 220 grit to get better contact. Did about 2 hours with 320 grit. Sharpie test OK now.



Fine grinding


2/13 - 2/18  400, 600, 800 grit

No problems. About 75 minutes per grit, about 2/3 MOT, 1/3 TOT. Sharpietest every time before proceeding to finer grit. Micrometer reading = 0.54 now.


2/18 1200 grit

Worked 1.5 hours, about 50/50 MOT and TOT. Last 20 minutes without changing grit. No pits. Stopped fine grinding.


2/19 Check for pits with lager magnification loupe.

Two tiny pits, which I overlooked with my 10x loupe, but saw with the (30 or 40x) loupe at the observatory. About half an hour 1200 grit. Didn't check again.




19/2 Plaster tool.

Made a plaster tool.


26/2 Pitch lap.

Made a pitchlap on the plaster tool. Pitch is Gugolz64, using a silicone pitchlap mold as sold by Kevin McCarthy. Works perfectly! After pressing for an hour I polished for about an hour: about 1/4 diameter W-strokes.



Contact between tool and mirror not good enough on two spots at the edge. Cold pressed with 10 kilo extra weight. Contact OK now. Polished for several hours, cold pressing every 20 minutes.



To early to do a Foucault test, because the mirror isn't polished out yet. But I couldn't wait and wanted to see how the mirror looked. First I used a Ronchi screen, which showed reasonable straight lines (first time I did this). Foucault test shows a high edge.



Bit of a 'jerky'  movement of the tool after cold pressing for 15 minutes with 10 kg of extra weight. Pressed again for 20 minutes. Polished half an hour, cold pressed again for 15 minutes. No problems and good movement.



Contact after cold pressing not good enough. Warm pressed for 10 minutes. Let it rest for the day.



Started cold pressing for 20 minutes. Movement easy in one direction, heavy in the opposite direction. Cold pressed for 15 minutes. No cure for the above mentioned problem. Cold pressed for half an hour. Movement much better but not good enough.


Cold pressed for 40 minutes using 10 kilo's extra weight. Trimmed the edge of the  pitch lap


After trimming the edge, pressed again for 20 minutes, with 10 kilo's extra weight. After that, polished for some time. Reasonably smooth movement, but can be better.


Cold press for 40 minutes, 10 kilo's extra weight. Polishing again not smooth enough. After ten minutes, pressed again for 15 minutes. Polishing OK now. Good contact mirror and tool. Added some dish washer detergent to make polishing a bit easier.



Polished for two hours. Good contact now between mirror and tool. Wonder why I never had to re-open the gaps between the pitch squares, about which I have read. But I wouldn't know how to do that without making a mess, so I stopped wondering. Foucault tested again, just to practice testing. Test shows a high centre and a high edge. Piet Hans Bruinsma, my instructor, says I should be pleased, because this is a better starting position to parabolize then a turned down edge and or an hole in the center.



Polished for 1.5 hours. Good contact. Foucault shows the same.



Polished for about an hour. Laser test shows mirror is polished out well.



Test shows that the mirror is not perfectly spherical, but still has a bit of a high center and a high edge.





It took a long time to get good contact: much pressing, warm and cold.  When contact was good, parabolized for 20 minutes: W stroke, between 1/3 and 1/2 mirror diameter.


Cold press after 20 minutes. Test shows very weak parabola. No sign of TDE



Two times 20 minutes normal parabolization stroke, with pressing in between. Test shows the parabola much better now. Made a foucault mask. (5 zones). Experimented a bit, but don't know well enough how to take measurements. And not sure I read the center zone accurately.



Made a three zone mask (advise Piet Hans). Results of measurement below (blue line = ideal readings).






Parabolized for about 45 minutes, with normal parabolization stroke. Pressed three times in between. Results going better, but the edge stays high. Called Piet Hans. His advise: a narrower W stroke to work more on the edge.



Parabolized with narrow W stroke, after a 30 minute cold press. After 15 minutes, another cold press for another 30 minutes. Movement still not smooth enough. Cold press again for 10 minutes.  Movement OK now. Measurements not consistent. Will go to the observatory to let someone else advise me in testing.



Tested with 5-zone mask. Results for three measurements below (stragiht line is ideal reading).  About 1/3 lambda PV wavefront, according to FXP. A good edge zone would bring the mirror to 1/10 lambda, taking the best reading as the good one. But the readings are not consistent enough to justify this conclusion.



27/3  About 10 minutes TOT with accented edge pressure. Edge a little bit down, but not enough.



Repeated yesterdays 'edge pressure' technique. Not much difference. Practiced testing for some time. The results get more and more consistent. Again two rounds accented pressure. Asked Piet Hans to test the mirror as well. Got the same results as I did, shown below.


Deformation is 1/6.32 lambda PV, Strehl ratio is 0.94, according to FXP. The edge is still too high. The results are not bad, but I decide to proceed, because my own standard of  'better then 1/10 wave PV on the wavefront'  is not met. But on vacation to France first.

Back from rainy France.

There's more then one way to proceed and I get different kinds of advise (fullsize lap, subdiameter lap, Zulu lap).

I decide to go for correcting wit a fullsize lap, inspired by Mike Lockwoods webpages. From here, I follow Mikes advise. After each correction session I send the FXP readings to Mike, who is so kind as to advise me in how to proceed.


8/4 and 9/4

After a long cold press, 10 minutes TOT, very short stroke to undo the overcorrection in zone 3 and especially zone 4, with little succes. After that worked the edge a bit, which came down a little. Again some edge work and after that 10 minutes TOT, very short stroke, to try to decorrect zone 3.


Result below: zone 4 decorrected , zone 3 overcorrected, center and edge too high.






Corrected for 16 minutes, TOT 1/4 diameter with small sideways movement. Extra weight of 5 kilo's on the tool. Pressed twice during this session. The results show (see below): evenly (under)corrected zones 1-4, zone 5 much too high.
Followed Mikes advise to make a 7-zone mask.



Tested with 7-zone mask.

The results are in line with the 5-zone test, but give more detailed information. See below. .





Followed Mikes advise to work on the edge with 1/4 diameter MOT stroke. Did'nt work out well. See below.




Inspecting and repairing the pitch lap

Followed Mikes advise to inspect my pitch lap. About half of the edge squares don't work well and have no or only a little contact with the mirror. Not good if one wants to work the edge. Maybe this was the problem, without me being aware of it. Tried to make them have contact by a long warm press. Good result, except for 6 squares.


The long warm press had closed the channels and I decided to open them up with a soldering iron. Very messy job and resulted in high edges on all squares, which I had to remove with a knife.


Repaired damaged squares by dripping pith upon them. All this reapir bussiness made it necessary to warm press again, which closed the (3 mm wide) channels again. Opened them up with a wider (6 mm) soldering iron. Messy again, but it worked. Warm press again.


Repaired some more squares and warm pressed again. Pitch lap looks good now.


Correcting again.

1. MOT, COC 1/4 diameter 10 minutes. Zone 7 came down a bit. Zone 6 didn't.

2. MOT, W-1/8 - 1/7 overhang. Zone 6 down now.Result below. Correction of zones closer to each other, less total corrrection, zone 3 extremely low.




Edge parabolization stroke (narrow W), with much COC action and little overhang to work the edge zones.

Result below.



Somewhat complicated stroke: a combination of much edge action, much center action and only little action in between, for two times 5 minutes. Result (see below): much too high zone 3 and the edge correction has gone away. Not clear how this could happen.



Followed Mikes advise to micro-facet the lap with a knife. After that, a half hour cold press, with 14 kilo's extra weight. After that 2 times 6 minutes narrow W stroke (MOT) to work the edge again. Between the 2 6 minutes sessions a half hour cold press with 14 kilo's extra weight.

Result (below) not the expected effect on the edge and overcorrection of (especially) zone 3 and zone 4.



A new pitch lap

Mikes conclusion is, that something must be wrong with the pitch lap. Advise: make another one.
At the observatory new pitch has arrived. The pitch I used was very old and very hard. That must have been the reason that the channels did not close before I warm pressed for a long time and that I had to press very often, with much extra weight to let the pitch make good contact with the mirror. 



Made a new pitch lap. Pressed until contact was good. No further activities today.



Started with a warm press for good contact. After that, half an hout MOT 1/3 diameter COC. Test shows decorrection clearly. Zone 3 less overcorrected. Edge stays too high.  See FXP image below.



After a cold press, 2x10 minutes TOT (with 15 minutes cold press in between) 1/3 diameter., mostly COC but also some time with 1/6 diameter overgang on both sides.


Result: a bit more decorrection, but the edge stays high. Two times 10 minutes (pressing in between) COC 1/3 diameter, for more decorrection at the center. Result below. Edge will not go down.




High edge problem

I begin to suspect that the edge is still higher then I think, because I can see, without a mask, a graying of the edge somewhere before zone 6 should about come to focus.

1/5 - 4/5

Two times 10 minutes MOT, COC, 1/3 diameter. After that two minutes edge parabolization stroke: MOT, long strokes, narrow W. Practiced testing the edge with and without mask. The tests clearly show now, that the edge-zone (7) is high and focuses before zone 6. Results below, showing the very high edge clearly.




Two sessions:

1. A bit less then 10 minutes TOT, 2-4 cm COC strokes (after cold
pressing). The results are not what we wanted. Decorrection is too far to the
edge, on zones 6 and 7. The middle zones didn't change in the right direction. See below.


2. The edge-correcting version of the parabolizing stroke (MOT), to get correction in zone 7.

The edge came down a bit (but stays 'negative'), but the strokes had also a parabolizing effect on zone 3 -6.

Results below.




I think, seeing the already high corrected middle zones,  the best thing I
can do is go back to a sphere, a bit further than I did some days ago and
take the edge from there.


7 /5

Microfacetted again with a knife. After that, 15 minutes TOT - COC, to work the edge. Not without a little succes, as can be seen below. But it's still much too high and still focuses before zone 6.



After that, with 4 kilo's extra weight, 10 minutes MOT, COC 1/3 diameter to try to decorrect the center.


Result: center and edge stay high, see below.



Now 15 minutes MOT, 1/3 diameter to decorrect the middle zones.

Result is on the center zone (decorrection) and middle zones are highly corrected now. See below.



To work the middle and edge zones, the same stroke as above, but TOT instead of MOT. After three minutes, a 15 minute cold press. Had effect on the middle zones and a bit less on the edge zones, but not enough as can be seen below.




For 15 minutes TOT, 1/3 overhang, 1/3 diameter stroke.

Result: all zones below 50% correction, except for zone 1,2 and 7. So this stroke worked on predominantly on zones 4 and 5. See below.



Tried to lower the edge wit  MOT, COC 1/2-2/3 diameter. Didn't work out. Instead it came up a bit. Decorrection of the other zones increased (from about 50% - 29-35%), except for the center zones. The result here is more correction.

See below.



Worked the edge with an accented pressure technique: push the mirror away while pushing the edge down, for 5 minutes:  MOT, COC, 1/3 diameter. Pressure only while pushing, not by pulling back. Not the expected result.  


Then  TOT with accented pressure and 1/4 overhang. Result: zone 5 and 6 are lower, but not zone 7. See below. Begin to think zone 7 has something against me.





About 18 minutes TOT 1/3 diameter, 1/3 overhang. Correction mainly in zone 5 and 6. Edge zone a bit. See below.




About 12 minutes MOT COC 1/2 diameter. Tested with 5 zone mask. Edge came down a bit, other 4 zones a bit more corrected. See below.





1. Four times 5 minutes COC MOT parabolizing. Effected mainly zone 4. As good as no effect on zone 7.

2. Two times three rounds accented edge pressure. Edge stays high.

3. Back to sphere: 20 minutes MOT and 20 minutes TOT normal stroke. Result: closer to sphere and edge came down. Shows much correction now!



Two times 10 minutes (pressing in between) narrow W stroke to parabolize the edge. Contact between mirror and tool not good enough. Result: little difference.


Channels need to be opened. Didn't use a solering iron this time, but a small chisel. Worked much nicer then the soldering iron and made only one chip. After that, 10 minutes 3/4 diameter overhang, MOT - COC to correct zone 7.


Result below. Doesn't look bad, but zone 7 high again and relatively much correction in zone 5.





Two times 5 minutes MOT COC 1/3 diameter. Result below. Zones a bit closer together, but 7 still high.




Again two times 5 minutes MOT COC 1/3 overhang.

Result: zones a bit more equally corrected. See below.


Primary ripple

For the first time I saw (severe) primary ripple, maybe caused by temperature difference of mirror and tool during pressing. But thy might also be caused by the tool 'biting' in the mirror at the turning points of a stroke (I read this somewhere).  Looks as if the squares are pressed into the glass. Worked 14 minutes MOT COC to get rid of the ripple. Didn't go away.


I let the mirror rest now for some days, because I get too focused.




Started again. For 15 minutes MOT COC 1/3 diameter,very light 8 pattern. Ripple almost gone Result below. Undercorrection mainly at the center of the mirror.




Parabolized with white powder I purchased in Essen (Germany). Made some '8' movements to smoothen the movement. There is some kind of a bump visible now at the center of the mirror. Maybe caused by the '8' movement??


Proceed with a more straight stroke, hope the bump goes away and the ripple doesn't return. Following Mikes advice, I will therefor use a very diluted mixture and make the movements as smooth as possible, with no extra pressure then the weight of the glass. Did some more parabolization, with accent on the edge. Most important result (not shown): edge zone a bit more corrected then the other zones. Not too much difference in correction between the zones.


A good mirror at last!!

Normal parabolizing stroke for two times 5 minutes.
Result better then expected and a certainly not expected by me: correction error ony 1/11 lambda, Strehl ratio 0.98. I think I'll stop now.



Maybe I can make it still better. First decorrect a bit. Result below. The 1/5 lambda hits my soul hard.





Normal parabolization stroke for about 5 minutes. 1/8 Lambda: we're going up again.



Again about 5 minutes normal parabolization stroke. Result: 1/13 lamda, according to FXP, Strehl ratio about 0.98. See below. This is the time to stop.





And thanks Piet Hans Bruinsma to guide me through the grinding and polishing proces and the first parabolization steps and thanks Mike Lockwood, who was my 'long distance guide' to a very good first mirror.


Can't wait to hit a second mirror.

Go to: main menu

Go to: A 20 inch f/3.6 computerized Dobsonian Go to: Building a trilateral computerized 20 inch f/5 Dobsonian
Go to: Project: five 12 inch lightweight Dobsonians Go to: Motorizing a 12 inch lightweight Dobsonian
Go to: 20 inch telescope Go to: Equatorial platform Go to: Scotch mount Go to: Binocular mount
Go to: Dobsonian tips Go to: Bending aluminium Go to: Collimating Go to: Making a Krupa collimator
Go to: Dotting the primary Go to: A ballhead type telrad/finder mount Go to: Mirror making log of 300 mm mirror Go to: Astronomy Go to: Using digital finder charts at the eyepiece Go to: Astronomy links
Go to: Building a spherometer Go to: Building a Bath interferometer Go to: A Foucault-Ronchi-Lyot tester
Go to: Building a mirror making machine Go: Home

email: Jan van Gastel