3D Printed 6" Newtonian

[Stub Mandrel]

This is a big project!

I was generously given a 6" mirror and secondary of some age. I had the main mirror recoated but the secondary is fine.

I'm making them into a newtonian, using as much 3D printing as possible, although I have a vintage r&p focuser I want to use.

The tube is an 8" diameter PVC ventilation duct that I split and reduced to 7" diameter with a reinforcing strap along the joint. Prototype reinforcing rings on the ends added lots of rigidity, but I realised that I coudl incorporate these into the mirror cell and secondary support.

Here are my designs for the mirror cell and secondary supports. Annoyingly my 3D rpinter hasn't got enough Y-travel to allow me to print the full diameter, although the bed is big enough. I've ordered the ground rod, allthread and timing belt to let me add an extra ~50mm of travel. This will give me a bed large enough that I can print tube rings which will complete the stiffening of the tube. My one concern is that the 1.2mm thick spider may not be wide enough, although I printed a 0.8mm thick lens hood for a webcam which suggsts it will be strong enough, the question is will it be stiff enough? I can always make it thicker or straight.

Note nut pocket for M5 collimation screw.

The cover for the end tube includes a light tight baffled vent, ideal for 3D printing!

The mirror cell clips into the mirror without obscuring any of the silvered surface, however I will be using adhesive as well, just in case.

My printer has enough area to make the cell, a lovely job in Monoprice PLA Plus, totally unretouched straight off the printer 🙂

[Gina]

Nice one At first I thought you were 3D printing the tube but you would need lots of height or several sections bolted together.

[johninderby]

Interesting little project. 👍🏻 Will be interested in how it progresses.

Orio Optics UK carried out some tests of thick vs thin vanes a few years ago and no one could tell the difference. Some scopes have thicker vanes and doesn’t seem to hurt performance.

 

Edited April 24, 2020 by johninderby

[Stub Mandrel]

58 minutes ago, johninderby said:

Some scopes have thicker vanes and doesn’t seem to hurt performance.

That's useful to know, doubling the thickness will make them four times stiffer and I'd like to go with curved ones.

[johninderby]

Think they were 3mm thick on the old TAL200K Good scope but a very heavy lump! Pity the AZ100 wasn’t around then. 🙂

Edited April 24, 2020 by johninderby

[Stub Mandrel]

2 minutes ago, johninderby said:

Think they were 3mm thick on the old TAL200K Good scope but a very heavy lump! Pity the AZ100 wasn’t around then. 🙂

I've gone up to 2.4mm, that will give me six wall layers at 0.4mm, which is effectively solid PLA, should be plenty at 25mm depth. Now I have to do all the fillets again!

Now I must wait for my bits to come, eBay is givi9ng a 28 day window for second class delivery, hopefully it will be faster than that.

[Chriske]

6 hours ago, Stub Mandrel said:

That's useful to know, doubling the thickness will make them four times stiffer and I'd like to go with curved ones.

Good choice. Remember the sum of the angles of the curves must be 180° to 'rule out' diffraction.
'Rule out' is not correct I know, I should rather say the diffraction normally seen with a regular spider is evenly and completely 'smeared out' all over the field of view using curved spiders.

[Stub Mandrel]

On 24/04/2020 at 19:04, Chriske said:

Good choice. Remember the sum of the angles of the curves must be 180° to 'rule out' diffraction.

I'm going to try somewhere between a full circle and the Orion scope in John's picture above, and see how it goes.

[johninderby]

12 minutes ago, Stub Mandrel said:

I'm going to try somewhere between a full circle and the Orion scope in John's picture above, and see how it goes.

The scope in the pic is actually my old TAL200K not that that really matters. 🙂

Edited April 25, 2020 by johninderby

[Chriske]

This is how it should be build to 'remove' all traces of diffraction.
To be clear : these two circles are the primary and secondary mirror.
I think the spider's scope above is very close, if not perfect.

 

[Stub Mandrel]

7 hours ago, Chriske said:

This is how it should be build to 'remove' all traces of diffraction.
To be clear : these two circles are the primary and secondary mirror.
I think the spider's scope above is very close, if not perfect.

 

Ah yes, I understand now, as each vane makes two spikes they will cover 360 degrees. I've revised my design.

Blimey, I've just noticed the STL is 131 MB!

 

[johninderby]

12 minutes ago, Stub Mandrel said:

Ah yes, I understand now, as each vane makes two spikes they will cover 360 degrees. I've revised my design.

Blimey, I've just noticed the STL is 131 MB!

 

And the print time? 🙀🙀🙀

[Gina]

WOW!!

[Stub Mandrel]

1 hour ago, johninderby said:

And the print time? 🙀🙀🙀

Probably only about 4 hours. I've found that if I run the printer quite fast I actually get better prints, plus most of the complexity is an abundance of fillets. That means lots of 3D curves which an STL has to represent with (literally) millions of little flat triangles.

 

[Gina]

Yes, I've found that.  A big STL file doesn't necessarily mean a long printing time.  And the reverse can be the case too.  Some of my all night prints haven't been enormous STL files.

[johninderby]

Yes some prints are complex. Took about 12 hours total to print this T-rex skull but it was a fair size.🙀

[Chriske]

It's a simple concept, how do you manage to end up making that spider : 130Mb...??

[Chriske]

I did a quick drawing, ok no holes, a spider for a 200 mm scope stl file was 389KB slicing that thing it was just 9.8Mb

Edited April 26, 2020 by Chriske

[Stub Mandrel]

1 hour ago, Chriske said:

I did a quick drawing, ok no holes, a spider for a 200 mm scope stl file was 389KB slicing that thing it was just 9.8Mb

Yes, but that's a prismatic object with constant cross section. I've got fillets all over the place to remove stress raisers, make it easier to assemble and more pleasant to handle. They hugesly increae the complexity of the STL but barely affect print time.

Imagine the difference between a crude cylinder, with say 20 segments, so 20 triangles for each face, top and bottom that's 40. Plus 2 for each of the 20 'sides' of the curved part - total 80 triangles.

Now put a fillet around the top, 90 degrees, so 1/4 circle that's 5 rows of 20 segments to render, Each segment will require two triangles, so that's an extra 5 x 20 x 2 = 200 triangles.

The crude cylinder now has 280 triangles, more than doubled from adding a fillet.

 

If you use a realistic number of segments, say at least least 100, maybe 360, to get smooth curves on large objects, and the number of fillets increases as the square of the number of segments. 

100 segment cylinder has 400 traingles, five times as many.

Add the fillet and its 25 x 100 x 2 = 5,000 triangles, so now it's more than a ten-fold increase in file size.

Of course some triangles will be too small to render and be ignored, but I'm using a 1-degeree segment as I don't want a faceted look so it's easy to see how a 9.8Mb stl can become a >100Mb stl just by adding fillets.

[Stub Mandrel]

Amazing!

The ground rods arrived today, took about an hour to cut to size, swap over and then rebuild/calibrate the printer.

Then a test print.

Then ages bed levelling as 200mm circle requires the bed to be dead level right across!

Then over five hours to heat up and print the secondary support, took it to the workshop and in utter disbelief it was a perfect light push fit onto my reduced diameter tube!

I was really expecting to have to tweak the diameter. Plus it feels really rigid, no worries about the strength/stiffness of the vanes.

Photo to follow tomorrow!

[Gina]

Great stuff Neil

[Stub Mandrel]

A couple of pics:

 

[Stub Mandrel]

First Light tonight! Well the mirror I was very kindly given seems to be excellent.

Hard to collimate with my laser collimator as the dot is almost invisible on the freshly coated mirror!

I tried it with all my eyepieces on the moon. The 2.5mm UWA was usable (I can use it in my 10" scope, but my 6" x 1200mm scope doesn't like it), and it's the first time my cheap 4mm plossl has given a view worth looking at. Every other ep from 5mm  to 40mm gave its best, although a few of the cheaper ones (especially the 8-24 zoom) showed CA or softness at the edges. The 18mm ortho was, as ever, outstanding giving a full disk view, the 5mm UWA was great going along the terminator. The ortho, cheap aspheric 23mm and longer plossls all showed a touch of colour - subtle blues and browns on the maria but surprisingly the 23mm aspheric which cost about £5 showed the colour best!

I'm sure contrast will get even better when painted/flocked inside!

Really very pleased. It perhaps isn't rigid enough to be a serious astrophotography scope but being between my 1200 and 590mm fl imaging newts it's an ideal visual scope at just over 900mm.

Going to see if I can find some galaxies now.

Definitely needs flocking... also temporary collimation screws

 

[johninderby]

Glad to see it has worked out. 👍🏻

Flocking or paint?

[tooth_dr]

That’s awesome!  Quick question - the focuser seems quite far racked out and the eyepiece quite a distance from the tube.  Is this fixed and determined by the mirrors?