- This topic has 47 replies, 3 voices, and was last updated 7 years, 6 months ago by
.
-
Posts
-
Ah, Now you've made me double check I noticed that maybe I used too much turpentine. In the article in fact you wrote that you used 8 ml for the entire dough, and I wonder why I interpreted that uses 8 ml each 100 g of dough…In everything I have thrown into well 22 ml V 350 g total dough, then a little more than what is indicated (although at least I put less than linseed oil saw that I was no little left)…Well, however definitely do it again before I try it and see how it works…
Thanks again…ciao
Sub-diametri 1 – Mirco 0
From the title already it seems clear that they are not successful in, but going to order here is what happened:First thing I built 3 sub diameter tools of different sizes (50% 30% 18%), of which the one at 50% poured directly into a mold in a star shape.
As support for the tools I used the plywood from 20 mm thick.
Once the mixture cooled, for I had no free time to devote to processing (due to other commitments) I let the tool to adapt to the shape of the mirror for 4 days almost over 20 Kg.At this point I ran 5 revolutions of the table by applying the type of past described in the previous post.
After that I evaluated the progress of work by the Ronchi tests and horror ...... !!!The surface appeared all wrinkled and in some places there were real ruts ...
So I investigated and discovered that the wooden support of 'tool was not nice plan, but it was slightly curved, so during the adaptation phase the effect of almost 20 Kg weight has meant that the curved face of the raddrizzaste wood and would fit to the plane of the supporting table. Once you removed the weights wood for elasticity is returned to its initial shape even if only slightly deforming the curvature of the pitch surface. The result was that only the outer part in a star shape to the well rested tips, while the tool center did not support at all.This pitch mismatched has therefore generated a very rough surface, in addition to having generated a circular groove near the edge of the mirror and one towards the center, under exactly the areas of contact between tool and mirror.
Obviously they are followed 3 hours of despair ...
The next day I decided to resume full tool diameter in hand and apply races 1/3 COC until the complete disappearance of all signs. Some of them are deeper revealed expected and when they finally disappeared I was so close to the ball that I decided to run another couple of dried so that I got a ball that is perhaps even better than the one with which I left a few weeks ago…
A more accurate assessment still will do it only after legacy to rest the mirror for the usual 24 hour.at least one positive thing is that the temperatures rise has meant that even in the room where the working temperatures are raised to 2 ° C making the mixture of the full diameter of the ideal hardness, so much so that to pass from dish to the sphere was enough half a day (not so much effort)…
Now we have to start a new dish and look for a more rigid support and suitable for the sub-diameter ...
See you soon ... hello to all ...
Well, I think it's impossible to play the game with the parable without conceding goals…
I had understood that the tool to 50% I had already, otherwise I would have advised against a freshly made patina, without a minimum of “break-in” for corrections.
From what I've seen, a new patina although it seems perfectly adapted ( but it is not at the level micrometer ), He fails to work well as long as the friction is not uniform across the surface, that is, until you have not created a layer uniformly consumed by cerium. Until then, the patina of the areas with pitch “impairment” that in their first contact immediately create roughness / also important furrows, precisely because they generate a much higher friction than that of the oiled patina, then to send her son away pain…Well it is true a brand new film is not the best to use for purposes correzzioni, but if the fit is done right the tool begins to work almost immediately discreetly…The big problem was that the fit was not good at all seen the wooden support problem I was talking about…
Now that they have returned to the sphere I have the opportunity to run in the sub-diameter in the first phase of parabolizzazzione and at the same time remedy the roughness that is generated with the full diameter… Finally, Finally, Finally…. Finally the first real satisfaction that I take off with machining ...
Then, as I told you are returned to the sphere with machining tool by hand with a full diameter with racing 1/3 COC. The form was not bad, but I could not make her pretty regular, as a slight circular depression was generated in roughly 1/6 D from the edge of the mirror, ie in correspondence with where there was the edge of the tool at the end of the outward or return stroke ...
So I decided to slightly change the shape of the patina by creating a shape of a "star" (though small radial extensions) to see if the reduced abrasive edge capabilities allow me to correct the interference described before.
in fact this has mitigated the negative effect, but reduced the abrasive capacity of the tool edge (COC races in the edge zone of the tool always passes over the corresponding edge zone of the mirror) He always generated an edge which tended to be high.It remained so to remake the mixture to restore the correct tool shape and return to work. But before doing so I wanted to try again (stubbornness) with machining.
The tool used is 310 mm in diameter with a modified shape of a "star".
After several sut-up and testing ...
And little reassuring results… I finally realized where they were and where mistakes and wrong once again re-set the machine we saw right away that she could be the right time to get something decent with this process. And indeed after 6-7 dried by 15 min I got this:
Which I think is one of the most beautiful beads that received lately.
I would say pretty well right from the center to the edge, and with a very good surface quality, no appreciable appearance of roughness (Perhaps the picture is not that great)… I then found that set-up that can be considered as a personal flotation l’ 1/3 COC in machining with the full diameter, and that makes me very happy…
In addition to this base I can easily, increasing or decreasing the stroke of the tool, go to act more on the edge or the center according to the requirements…
If desired, the ball could be even slightly improved, I do not think it's worth it and I think it is time to proceed further…Soon the Parable…
Great Mirco, beautiful sphere and excellent surface “smooth” ! As you hinted, if unable to intervene with the same precision on those two to three millimeters in leading edge ( to me it has never been entirely successful), It will be PERFECT ! These results are very encouraging for your studio with machining which is notoriously more complex than manual, so I understand your satisfaction and I'm glad . Look forward to the next report.
Hi everyone, these days I'm heading in parabolizzazzione very slowly (deliberately) to gradually get closer to the dish often by checking the situation.
After an initial phase in which they arrived at a constant taper of about -0,70 I started doing sessions 3 rounds of table (1-2 with the full diameter and 1-2 with sub 50%) then let go almost 48 hours before making measurements. Tests carried out by initially Foucault test, then replaced by 'Hartmann with which can not analyze the mirror with a mask to 9 areas on the radius against 6 Foucault.
As you can see from the pictures below in the last 3 Sessions are passed by a conical chops -0,7 to a -0,75 up to that from -0,84, with all the PV values good and next to 7-8. Account then be able to proceed as well gradually to approach a lot to the final dish maintaining a good PV and then go to file the details in the final.
Hello to all…
I ran short of parabolizzazzione another session with the full diameter (2 rounds of table) and I ran tests. Here are the results:In the picture are shown 2 different error profiles referred to the same series of measurements. The two profiles differ only because they are related to slightly different reference parabolas. I did it to show more clearly how the mirror's periphery can be considered quite good and should focus (at the moment) all 'deepening of the single central part.
I therefore decided to proceed with the W sub-rides with the diameter at the 50% working within a circle of 140 mm radius.
All for a fairly short session to verify that the road taken is correct.Here are the new results, compared with the old.
Considerations:
indeed embarked on the road seems to be the right one, the central part of the mirror begins to enter the trumpet and the lambda is passed from 2 until 5.
The extension of past adopted, however, seems to be too big because I went to pull down even the eighth area, I would not go to that instead cooked on foods, and which thus led to accentuate the horse hill area 10.
Now I'm realizing 2 smaller tools, one 28% and one on 19% to devote at first to continue with the deepening of the central zone.See you soon…
I find that you are doing well.
And reading your post I see also good from the point of view “instructive” for the benefit of some readers neophyte.
MI referring to the example of the different affrontabilità parabolizzazione, date by reading and comparison of the first two charts blue background!It is illuminated in a different light critical advice “standard” of parabolizzare “from 70%” to limit to a minimum the glass aspostazione, which it is usually found in holy books.
In fact parabolizzando from 70%, from the graph on the left, you must lower the “hump dromedary” central, and also its “head” peripheral, per 140 nanometers from each part, having the problem to be careful not to incur an edge retorted in the critical peripheral areas;
Instead take the brilliantly good for the peripheral part (Massimo docet), avoid the big risk of edge retorted, against the greater job of having to raise 200 glass nanometers from a diameter 125 to reach the center mirror.The alluring advantage is that the latter option turns the worries of mind of the parabolizzazione 70%, in a quieter hike, gradually going to work tolerance zones less restrictive, enjoying a greater constancy of labor improvements represented by lambda which grows slowly, but without the ugly or beautiful unexpected surprises that would result from the parabolizzazione 70%.
MI like to dwell on this subject simply because I had not I ever thought, and I have to do to you and Maessimo complimented on strategy! You always learn something new and interesting!
Wow ! It is already a beautiful parable
nice job Mirco ! And, I also believe that further refinements should use the smallest sub ( but not too small to work downtown ), with classic tangential races...Otherwise error, working to lower areas 3, 8 e 9 would you take in the entire surface “"trumpet bell"” tolerance.
P.S: A strictly personal taste on processing graphics, I would prefer a chart with broken lines rather than with the fittings of the curves (Especially when the areas are many, and a fitting can hide a short area with negative trend).
This due to the fact that it is true that in reality there is a continuity of connection between the various zones of the mirror, but the end of the graph (that is what counts)not interpret reality, but highlight the slope denouncing the error, which best as such would see when not interpretable “medicated” by realistic curved fittings.
(I emphasize that it is a very personal taste, and that everyone does well to do as they wish).
It Giulio, I understand what you mean, but the graphics as you see them arise from a different way to rebuild the profile of the mirror with respect to how it does the Excel spreadsheet that you usually use, why they are curvilinear and not of broken.
It is not that is the graphic set: displays with fitting, are precisely the points calculated by the program that generates the curved profile, just to make the error and form most likely to make it more likely the PV value calculated. There are 30 different points on that curve and between broken and curvilinear line is not seen almost no difference.
That's because I used a parabolic polynomial on each trait that allows me to impose dell 'error value in the first node and the gradients in 2 end nodes (for each pair of average values of the windows). This way of doing imposes the slope calculated in the mean radius of the areas (as it should be). The sheet you have yourself instead uses a different strategy, or it imposes the slope calculated in the most inner radius of the window (instead of in the middle range). This method uses a simpler process to implement, but less accurate and gives a PV value of less realistic (although all things considered unimportant, this to me was wrong, and I then modified the spreadsheet as the most liked).
As I mentioned in the previous post I run 3 -4 rounds of sub with tool table diameter at 28% with modified edge-to-star and very light pressure with W rides within a circle of radius 100 mm.
After the usual 24 hours I performed the analysis, and this is what is the average of 5 different measurements:
the best PV has increased from 4,99 a 7,61…
Comments:
To begin with I found that the sub-diameters and dig very quickly especially if they are small.
It 's why I wanted to take a little hand with this tool before going to work for the central area that has a greater margin for error, because if I had also dug a bit more than necessary I would not affect all the work.
In fact, I dug a little too, but the figure remains well within the tolerances.Now let me go and bring down the hill at the turn of the 10th post area, with always tangential rides with the tool to 28% for a maximum 1-2 rounds per work session table.
Then again with tangential races to go 1-2 turns straddling the zones table 4 e 5 to lower a bit also that area, always with slight pressure.Reflection on the results:
In this session I worked within a circumference of 100 mm radius, then the entire outer area has not been changed. What also similar to previous work session.
One sees in fact well from 'above image as the data obtained from the tests have a correspondence almost identical in the non-machined area. This thing seems obvious in theory, It is not so simple to obtain in reality due to the variability of the data that are generally obtained from different measurements with the test foucault. that, Also this confirmation is convincing me more and more that the Hartmann test I'm using is a valid test, and that gives me good results and consistent with theoretical predictions ...A taken for further developments ... hello..
Great Mirco
, I'm really happy ! And now with small corrections in the right places you can achieve a compelling result ! Eye now with tangential races, meditates well what to do next, because if the area on horseback jobs 4 e 5, not lower this intermediate zone (which will remain unchanged in its radius of curvature ), lower the 5-6 and consequently raise the 3-4, in practice it come back a little back to the previous analysis…the same goes for the area 10, Currency carefully !
- You must be logged in to reply to this topic.