Learning to foucault

[skybadger]

Cant agree more. I actually measure from out to in and then offset in a spreadsheet. I have a digital vernier to remove user error , I just need to work out a way to fit it. 

 

[skybadger]

I made a pair of 50mm laps at the weekend. Cut the circles, drilled the machine foot hole, varnished them and then pitched them up.

I normally use a sink mat with a rectangular grid to make my surface grid on the lap. I used that this time too. 

However in use it was skipping a lot in spite of a warm press and not readily revolving on the machine ball foot. 

Thinking this was due to the uneven distribution of the large grid on the small lap, I re-pressed them by taking them back to plastic in a bath of hot water and re forming using a mould of a punched rubber mat resulting in the spots pattern you can see in the picture. 

Again it skipped and grabbed.

So I repressed cold and then scarified using a wire brush and that improves things so it moved well but ... The Foucault view is now quite scary. The central spherical zone is enlarged and the edge raised zone is reduced but the surface quality is awful, cos siting of surface scars in a concentric pattern. 

I did 4 wets at 80 to blend the zones, centered on zone 6 followed by 4 wets gradually moving the small lap out from zone 4 to 7

Another couple of wets to smooth and blend using the 80mm. 

I'll post the scarigrams later once I have them. So far I've just eyeballed them. 

I also did a wet by hand using the circular motion with just the slow platter rotating, I'll give that another go, was quite grabby. 

 

 

[skybadger]

Here's the lates set of foucault images after the zonal polishing with the new lap. Essentially a mixture of chordal strokes at greater radii from the centre and then small circular loops over the outermost zone and a bit of larger lap smoothing. 

The centre 50% is close to spherical again but the rest shows paraboloid in the Ronchi. 

The aim now I think is to continue to reduce the outer zones and then smooth out the surface by stepping up to using the largest lap again. 

 

[skybadger]

Been quiet of late due to pressures of work. 

Been working on smoothing the mirror back out using a larger lap which had the effect of taking it back to spherical. 

Since it was now almost all spherical, I started parabolising from the centre outwards and , voila, the tde seemed to be gone. 

Continued a bit more and there it was again but parabolising is in progress. 

Using sessions of 10mins with CoC and inspecting after 4 wets out to the edge. 

Ill post some pics to show what I mean in detail. It'll be a week though, off away again..

Key question for me is, the early  Foucault measurements were high in range and reducing over time. For spherical they are essentially zero. The target readings for my 6" f/5 mirror are 0 to 1.8mm but I get 0 to 8mm. Does this indicate strongly overcorrected or strongly under corrected.? I'm working on the theory that it's under corrected since I need to move knife edge much more to get to the desired zone position on the caustic and if it was over, I'd get there very quickly. 

Thoughts ?

[davidc135]

It's very over-corrected if the Foucault images are similar to the previous post with the knife edge coming from the right. ie k.e travel away from the mirror as the nulled zones move from the centre to edge.

David

[skybadger]

The ke does indeed move away from the mirror as the in focus zone moves from centre to edge. That takes about 8mm movement to do that. 

I shall run a smoothing spin polish on it again to return to sphere and very gently correct it this time. 

This is what I mean by getting  experience. I don't recall seeing this sort of advice in books.

[skybadger]

Got some images from the 5th May .

Inside focus to outside. Lines are straight until they reach the edge indicating returned to spherical but still with a bit of TDE 'hook' at the end. 

The hill at the 40% zone and the one in the centre are both reduced. Can see the kink at 40%  in the right hand images outside ROC. 

This concentric bulls eye pattern has gone  in subsequent working but no images of that yet. 

 

[skybadger]

Found the pics I was looking for.

These were the result of parabolising from spherical using CoC strokes through the centre, starting with large eccentric of 35mm, resulting in 10mm overhang, and reducing through 4 wets down to 10mm or so. 

The 70% zone looks raised and could do with a little scrubbing.

 

I measured the mirror this afternoon and these were the results. 

So my aim is to return to more spherical again - this is is overcorrected - checkout the knife movement range. Doing so will prove that it is overcorrected and let me start comparing the ronchigrams with RonWin. 

I'll also see if I can lower that outer ridge without lowering the edge. 

Edited May 13, 2023 by skybadger
spurious image

[Astrobits]

I am having trouble relating your Ke values to the images. The Ke values are clearly way over corrected but the images look to be near spherical?

Nigel

[skybadger]

The answer to that is that I spun polished back to near spherical and then set to parabolising again with CoC strokes and ended up over corrected. 

The spin polish intent was meant to find out if that was an effective way to smooth out the surface I had which had becomes very rough and since I was getting nowhere with a 50mm lap. The 50mm lap is just too vulnerable to skipping, even after warm and cold pressing. 

So the images are from the return to spherical and the KE settings are from the result after parabolising again. I used some additional weight on the parabolising strokes so I'll remove that next time around.

The intent now is to spin polish back to spherical ( or to reduce the over correction significantly ) and carefully parabolise again. I have some useful ronchi gratings at last and the foucault readings are now repeatable. 

Cheers

Mike

 

[skybadger]

Where I started the session 15th May 

Session details: 

Wet	time (mins)	lap (mm diam)	eccentric (mm)	offset (mm)	overhang (mm)	weight
1	8	80	30	50	10	4kg
2	8	80	30	50	10	4kg

ie I was using a large swing to provide a chordal stroke across the outer zones with a little overhang

#Result of session: 

Inside focus , ROC and outside focus. 

As an outcome, I moved back towards spherical a bit but need to do more on the outer zones. there is a small hole in the centre and the large central zone is still not quite spherical 

 

[skybadger]

So I did a bit more. 

Session details: 

Wet	time	lap	eccentric	offset	overhang	weight
3	8	80	8	55	20	6
4	8	80	15	40	10	6
5	8	80	30	30	-2	4
			had to take the weight down by 2kg from 6 to 4kg because mirror was tipping on extremes

 

So I think Im quite close to a sphere, maybe with a hint of turned up edge now. 

Next steps - 1/3 centre over centre smoothing to spherical using the larger lap. Once I get there, I'll do a gentle parabolising cos I'm itching to start on the 12" mirrors I have accumulated. 

 

 

[Astrobits]

It strikes me that you are using far too much weight on the lap. I never use any extra weight, just the weight of the lap/mirror and any connections needed to move it. With my mirrors up to 10" that was just my hands and I didn't push down on the lap/mirror any more than needed to move it. It might work more slowly but you seem to be going back and forth without getting where you want to be. A 6" mirror doesn't need a lap smaller than 6" until you need to attack very specific zones during the final figuring process.

I was a member of an astro soc back it the 80's who had a stall at an annual event. The members would start and finish a 6" mirror within the one day of the event. Starting with two slabs of glass in the morning they would be able to observe, and show the visitors, the moon that evening with the ( unsilvered) mirror.

It was said to me when I was learning to make mirrors that engineers do not make good mirrors because they tended to do repetitive operations as consistently as possible, without deviation. Using machines tends to do the same thing. This , more often than not, led to a very zonal surface. To make mirrors  one needs to randomise every stroke which machines cannot do by themselves. Having nice loose joints in the machine helps to smooth out the regularities somewhat and horrifies some engineers when they first encounter such equipment.

Nigel

[skybadger]

I think that's a fair shout re the weight once I get back to spherical. It's clear I overestimated the amount of figuring required to get from spherical to parabola. So I will remove the weights and shorten the wets and test frequently at that point. 

You've also seen the vid of the machine, there's a bit of slop there due to hinges having play. I will be replacing the arms with a bigger mech as I move to the 20" mirrors that should also have far less play but that's what the offset settings are for, to blend and avoid zoning as far as I can see. 

Key for me is to get to spherical again.