Jump to content

  •  

CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.

Photo

J.T.'s 12.5" F/4.3 Hexapod Dob

  • Please log in to reply
246 replies to this topic

#1 jtsenghas

jtsenghas

    Aurora

  • *****
  • topic starter
  • Posts: 4,735
  • Joined: 14 Sep 2014
  • Loc: The flatlands of Northwest Ohio 41.11N --Bloomdale

Posted 02 January 2019 - 11:38 AM

For a couple of months I've been posting in the "What did you work on today?" thread progress on my latest scope, which was approximately halfway done as of New Years Day.  I haven't yet done a build thread on any of my projects during my more than four years on Cloudy Nights, and for a while I've been promising/threatening to write one on this project, which was started two months ago.  Here goes...

 

I had more than a week of vacation time around the Christmas holiday and made enough progress on this dob to get to this point in the build:

 

Halfway done.jpg

 

Anyone familiar with Jonathan Pogson's (Oberon's) excellent 16 inch scope Merope will see that this project bears a passing resemblance, particularly in the shape of the altitude bearings and the nested "mirror box" which is in fact a thick circular ring that is arguably part of the mirror cell.  This scope won't merely be a 3/4 scale of Merope, though; it will have several features that are common with my 10-inch dob, the Tardiscope (which perhaps someday will have its own build thread--I do have the documentation for it).

 

Like Merope this scope will have hexapod adjustment for all collimation, without separate adjustment for the mirror cell and with minimal adjustment of the secondary mirror.  This methodology was introduced on this forum by Jonathan in this thread, and new ideas for implementing hexapod adjustment are still being developed and shared by several of us on that thread.  The keys to effective hexapods are minimal distances between truss ends and, most importantly, length adjustments that are easily enough made yet provide rigid truss attach points without discernible looseness or flexure.  They must be tight, but not too tight for adjustment.  As of this date I'm still working on the final details of my particular setup, but I have established that I will be using left hand and right hand screw eyes from turnbuckles at my pole ends.  These will be fit into customized tube inserts of my own making.  More on those later...

 

In this thread I intend to explain in detail my key design decisions, with a list of my design criteria spelled out early on.  I will then describe my build up to this point, which has jumped around a fair amount among the components made thus far. Many of the photographs I'll share have already been seen in the thread "What did you work on today?", but there will be additional ones for parts made so far.  

 

I intend to give explanations of the underlying thought processes for producing the major components on this, which will be my third scope made from scratch on the mechanical components, but with purchased optics.  I hope to be a glass pusher someday, but am not as of yet.  In this case the mirror is a Hubble Optics 12.5" f /4.3 sandwich mirror that I purchased seven years ago and has embarrassingly been residing in a drawer ever since.  In fairness, it became available at an extraordinary price immediately AFTER the 10-inch mirror which I used on my current scope, the Tardiscope, which saw first light in June 2014 and was presented at Stellafane in 2015.  I wanted to make and massage the 10-inch dob for a few years before starting on the 12.5-inch.  If this project goes well I'll make the 12-hour drive this year and share this one at Stellafane as well.

 

 


Edited by jtsenghas, 02 January 2019 - 09:26 PM.

  • Jeff Morgan, ad701xx, PrestonE and 23 others like this

#2 Bill Schneider

Bill Schneider

    Viking 1

  • *****
  • Posts: 792
  • Joined: 25 Jun 2015
  • Loc: Athens, OH

Posted 02 January 2019 - 12:14 PM

This project deserved its own construction notes all along. I'm glad that you chose to start a thread about it.

 

It's looking great so far!


  • Oberon, jtsenghas, Augustus and 2 others like this

#3 petert913

petert913

    Soyuz

  • *****
  • Posts: 3,837
  • Joined: 27 May 2013
  • Loc: Portland, OR

Posted 02 January 2019 - 01:20 PM

That is some pretty woodwork !!


  • Augustus likes this

#4 jtsenghas

jtsenghas

    Aurora

  • *****
  • topic starter
  • Posts: 4,735
  • Joined: 14 Sep 2014
  • Loc: The flatlands of Northwest Ohio 41.11N --Bloomdale

Posted 02 January 2019 - 03:02 PM

Thank you, all, for the immediate flurry of "likes" to my initial post and comments so far.  Starting as Jonathan did in his Merope thread....

 

Some Background--

 

This scope is arising from a desire to build a scope from a 12.5" Hubble Optics sandwich mirror I happened upon in eBay that I purchased new at an excellent price in 2012.  I had set up a number of automatic online searches for primary mirrors before building the Tardiscope and had no sooner purchased the 10" mirror that later became the primary in that 10" f/4.7 dob than this mirror popped up in my still existing automatic eBay search.  I noticed it had only a couple of bids and was within a day of ending the auction, so I put in a bid only slightly higher than the current high bid and shrugged.  I didn't really need two more primaries than current telescopes.....but if nobody else ran up the price...........I chose not to look again to resist the impulse to bid again...................and got it!

 

I chose to build and use the Tardiscope first though, although that took a couple of years to tweak the final design and build it with about half the work being done in December 2013.  That's for another thread though...

 

As I used the Tardiscope I realized that I wanted this 12.5" to be a very compact scope both for transport and use. At f/4.3 its focal length was only a few inches more than the 10" f/4.7 Tardiscope, and if I got the mirror a few inches closer to the ground the eyepiece would always be at about the same reachable height from an adjustable observing chair.  I really do like to do deep sky observing seated, relaxed, with hands on thighs and reaching occasionally to tug the scope to my target. 

 

For several years I expected that this 12.5"  would be a string scope and designed it a couple of times that way in CAD (I'm a mechanical engineer) and analyzed the design carefully.  I even bought a roll of Vectran twine for the purpose and weight-tested it to determine its Youngs Modulus. Vectran is the low-stretch component of archery bowstring.  The more I got experienced with my 10" f/4.7 the more I realized how critical good collimation is below f/5, and the more dissatisfied I got with my analyses of my 12.5" f/4.3 string telescope designs.  It didn't appear that weight savings would be significant as compared to a well designed truss, and weight, within reason, wasn't a major criterion anyway.  Although I came up with a few of what I consider improvements on Don Peckham's designs, I wasn't confident in holding primary axial collimation within the few tenths of a millimeters that f/4.3 demands.

In the end I decided to hold off on making a string scope, at least until I had a good slower mirror that had much more generous collimation tolerances and could perhaps be my "airline travel scope".

 

Putting down my design requirements and desirables in approximate priority order as Jonathan did in his thread I came up with the following.

 

REQUIREMENTS:

 

1) Must be able to be broken down sufficiently to fit into the trunk of either my Toyota Camry or my wife's Kia Spectra. (18.5" maximum height, 20" distance from latch to the top centerline of the opening)

 

2) Must be very stiff and hold collimation well (as seen with my homemade autocollimator and other homemade collimation tools)

 

3) Must balance perfectly with my midweight eyepiece, an ES82 18 mm and Paracorr 2 with the center of gravity right on the altitude axis with those in place and must balance adequately with lighter or heavier eyepieces (If a small weight adjustment is needed with the ES 82 30 mm that is acceptable).

 

4)  Must look good with form following function--smooth lines, good finish, no bits sticking out crudely or weights strapped on.

 

5)  Must have the eyepiece on the right side (as seen from behind the scope) to favor my preference of left eye observing and pulling rather than pushing the scope for tracking when observing objects to the south.

 

6)  Must be short enough for comfortable observing while seated, or for children at outreach events to use from a very short step stool.

 

7)  Must be fun to assemble and use.

 

DESIRABLES:

 

a)  Should have a removable Tardiscope-style azimuth setting circle with an adjustable but stationary circle and moving pointer so that the current bearing is always visible with a downward glance from the eyepiece. (I've always favored the clockwise scales and moving pointers).

 

b)  Should have a practical spot for an LCD non-illuminated display inclinometer gauge such as the newer Wixey I have on the Tardiscope. (These have batteries that last almost forever due to the unlit display and can be easily read by dim red flashlight.)

 

c) Should have trusses that collapse as a unit for a minimum of fumbling and decent collimation at setup.

 

d) Should decently baffle ambient light pollution due to presence of some streetlights near two of my typical observing locations.  This may require a longer UTA  than Merope has and additional baffles including an optional shroud.

 

e) Should angle the eyepiece upwards 30-45 degrees when the scope is angled at the horizon to allow for comfortable viewing at lowest height of observing chair.  (Tardiscope has eyepiece angled at 60 degrees due to being on face of a hexagon and for similar range of observing heights is angled at least 15 degrees more than i would consider ideal)

 

f)  Should be able to be collimated from the top side at least, and better yet from the eyepiece. (Boy, is it a lot easier to collimate the Tardiscope with a helper!)

 

g)  Should have a well-designed spider, likely a wire spider, with minimal diffraction and quick damping times. 

 

CONSTRAINTS:

 

--No welding on assembly as I have no such equipment

 

--Must be able to be constructed with woodworking tools, both unpowered and powered, including a Shopsmith Mark V (A limitation of equipment many would like to have...I bought mine during that brief period in the 1980's when I was both an employed engineer and a bachelor) 


Edited by jtsenghas, 02 January 2019 - 06:50 PM.

  • Augustus, Earthbound1 and ForeignObserver like this

#5 jtsenghas

jtsenghas

    Aurora

  • *****
  • topic starter
  • Posts: 4,735
  • Joined: 14 Sep 2014
  • Loc: The flatlands of Northwest Ohio 41.11N --Bloomdale

Posted 02 January 2019 - 04:33 PM

A bit more about my work space, since the "M" in ATM stands for "Maker" or "Making":

 

I can't emphasize enough the importance of a good work space for effective, efficient and enjoyable projects.  A couple of years ago on this thread I shared some photos of my workbench in progress, which I chose to make entirely with unpowered hand tools in the manner of the Welshman Paul Sellers (who is about 10 years my senior). Later I added retractable casters so that I can move my bench around and either clear the space in front of my dartboard or set up at the overhead garage door for a sunny day.

 

My work space is a 24 foot by 24 foot detached "no car garage".  That is, vehicles are in it only when they are being worked on.  Otherwise, it's a combination garden shed and woodworking shop. This structure was built at least six years before my house in 1956 at the latest and originally was part of the property of the then Methodist church next door.  Within two years after I purchased the property in 2003 I built a block chimney and installed a wood stove to heat it:

 

stove.jpg

 

The man I bought that stove from had purchased it previously as a DECORATION in his living room and hadn't used it.  He only sold it because he found a PRETTIER one! Go figure.

 

You might notice that the heat resistant handle on the stove door is quite similar to the design of the gravity-stabilized damper control which allows me to control the heat:

 

damper.jpg

 

I can't tell you how much I enjoy puttering out in that shop with hand tools, and how much I delight in a sharp well-balanced hand saw, a "wicked-pissa' sharp" chisel that pares softwood like soap, or a hand plane that scritches paper-thin shavings almost effortlessly.  On cold, dark winter days it is a delight to roll my workbench a reasonable distance from the wood stove, retract the wheels, and work on a project.

 

I can't emphasize enough that a workbench is not just a work surface.  An effective workbench is a sturdy fixture to hold workpieces by their faces, edges, or ends.

That may seem obvious on the face of it, and workbench designs certainly vary, but I can't believe how much having a proper workbench that is stable and allows me to place a foot under it for a good stance really makes a difference:

 

routering base.jpg

 

That device holding the face of my telescope base down is an ancient design called a holdfast.  It requires a thick bench and a slip-fit hole but it can be easily clamped and unclamped with a whack of a mallet.  It's so much easier using hand held power tools when there is a good workbench to fixture the workpiece, as can be seen above.  I've done enough crawling around on my hands and knees in the past....

 

Several of us, including Jonathan and Mark Cowan, have mentioned that a router is a very VERY useful tool for this hobby of ATMing.  This hobby is built around the circles those tools can generate and the grooves, dadoes, and roundover of a router can make all the difference in a project.   I'll get to what the router was doing in the above picture in time.  It is part of this project.

 

As mentioned in the second post, I also have a Shopsmith Mark V Model 500.  I bought it during that brief period of time in 1985 when I was both an employed engineer (then with GM) and a bachelor.  I bought the jointer planer, bandsaw, and belt sander attachments with no regrets.  It was significantly discounted when the Model 505 came out and I think it actually is a preferable setup despite having slightly smaller tables:

 

shopsmith.jpg

 

The Shopsmith has some limitations, and I'd love to have a great tablesaw, but it has served me well.  After 30 years of hard use I let the smoke out of the motor and had to replace it, but got both a replacement motor and a complete set of replacement bearings for all bearing locations for a very reasonable price in 2015.  It runs like new.  As a table saw it is a bit high and needs strategically placed roller stands, and as a drill press or lathe the slowest speeds aren't quite as slow as I'd like.  All-in-all, though, it has been a very useful workhorse.  I have learned how to do a lot with it that is typically done with metalworking tools through the use of homemade jigs and fixtures.  (I was a tooling engineer for much of my career). One thing I really like is that the table and miter gauge can be left at the same compound angle for sawing, disk sanding, and drilling.  On many occasions that transference of setups that has been very useful.

 

Anyway, before commencing the discussion of this build, I wanted to introduce myself a bit further to all of you and give you and idea of what I have to work with.  I hope I have succeeded.

 

 

 

 

 


  • Jeff Morgan, PrestonE, Pierre Lemay and 8 others like this

#6 ckh

ckh

    Surveyor 1

  • -----
  • Posts: 1,938
  • Joined: 21 Feb 2015
  • Loc: Arizona

Posted 02 January 2019 - 05:39 PM

Nice woodworking.  Looks very robust, stiff and cool.  I've been looking forward to this thread. What's the radius of the rockers?



#7 Lukes1040

Lukes1040

    Ranger 4

  • *****
  • Posts: 320
  • Joined: 26 Dec 2016
  • Loc: Ohio, USA

Posted 02 January 2019 - 06:00 PM

I also have been interested in Jonathan’s build, and will follow your progress as well. You have a nice workspace, thanks for sharing.

I am in the beginning stages of converting my solid tube 12” into one of these hexapod structures. I’m planning on using this as a trial run to see how I like it and how I want to proceed with my future build with a larger sandwich mirror. I will be interested in your opinion of the sandwich mirror inside this structure once complete.
  • jtsenghas likes this

#8 GShaffer

GShaffer

    Knight of Ni

  • *****
  • Moderators
  • Posts: 6,686
  • Joined: 28 Feb 2009
  • Loc: Bogart, Ga USA

Posted 02 January 2019 - 06:11 PM

Yep.....nice so far and I will be following this one :)


  • jtsenghas likes this

#9 jtsenghas

jtsenghas

    Aurora

  • *****
  • topic starter
  • Posts: 4,735
  • Joined: 14 Sep 2014
  • Loc: The flatlands of Northwest Ohio 41.11N --Bloomdale

Posted 02 January 2019 - 06:22 PM

....What's the radius of the rockers?

The outside radius is 16", the inside radius is 14" with a 2" offset of the centers so that the section goes from nearly 2" deep at the ends to 4" deep in the middle where bending moments will be greatest. I'll get into the details soon. 



#10 jtsenghas

jtsenghas

    Aurora

  • *****
  • topic starter
  • Posts: 4,735
  • Joined: 14 Sep 2014
  • Loc: The flatlands of Northwest Ohio 41.11N --Bloomdale

Posted 04 January 2019 - 10:04 AM

Choices of Materials:

 

As mentioned in my opening post, the overall structure of this scope is based on Jonathan's Merope. This design requires a lot of plywood due to all of the circular rings and circular arcs that are laminated several layers thick.  I determined, by shuffling the components around in CAD (more on that later) that even an approximately 3/4 scale version of Merope would require about two full sheets of 3/4" plywood. High quality Baltic Birch plywood is quite expensive and not really needed for most of this project. Lesser quality paneling suffices provided it has no large voids and sufficient number of plies for dimensional stability. 

 

I checked out what the two closest big box stores and a lumberyard had to offer. Both Lowes and the lumberyard had a decent "blond wood" plywood that I've used on the Tardiscope folding base. This has a fairly thin  but handsome outer ply that looks nice but is easily damaged. It also has fewer plies than I'd like and one has to be careful to pick out a fairly flat sheet if this is used.

 

Menard's had a 3/4" (actually 0.710") product I can't find on their website that they call "Utility Panel" priced at less than $40 a 4' x 8' sheet. It is similar to a Lauan plywood in that it has a VERY thin mahogany veneer layer (about 0.010" thick) but instead of a cheap core with thick plies it has eleven layers that are about 1/16" thick each of what is probably birch. The sheets are really sound feeling and extremely flat. I opted to go with this material for the majority of the project with some small pieces of Baltic Birch in a couple of areas I'll describe later.

 

I realized that the mahogany veneer wouldn't match the edges or Baltic Birch surfaces, but could be easily sanded away on external surfaces with 120 grit paper using a random orbit sander. The flatness and smoothness of this product really appealed to me.

 

An experimental sanding of part of one of the altitude bearings confirmed what I had hoped as shown here:

 

20181217_120834_resized_1.jpg

 

As can be seen, the adhesive layer is very red in color, but that came off quickly too. Because the next layer down is a full 1/16" thick I don't anticipate significant chipping.  There are some tiny voids in this material which I'll fill with a stainable blond wood filler and some specks of red adhesive remain in a few pits. I plan on finishing with some red-dyed clearcoat to hide these imperfections to my satisfaction. Such compromises are well worth the material cost savings in my mind. I agree with Jonathan that such work can be done handsomely enough with utility grade plywood.

 


  • brave_ulysses, Oberon and Lukes1040 like this

#11 jtsenghas

jtsenghas

    Aurora

  • *****
  • topic starter
  • Posts: 4,735
  • Joined: 14 Sep 2014
  • Loc: The flatlands of Northwest Ohio 41.11N --Bloomdale

Posted 04 January 2019 - 11:39 AM

Material Choices Continued:

The material I chose for the mirror cell is cold rolled steel plate and welded 1/8" x 1" x 1" construction tubing with (internal) rockers made of 1/2" square tubing:

20181222_193304_resized_1.jpg

On past scopes I've used aluminum both for weight savings and to reduce thermal mass near the optical path. The heat radiated by steel during cooling, in comparison, can add to seeing and cool down times. On this scope I desired a significantly heavier lower end (more on that later on a balance discussion). In this case I decided the extra counterbalance weight of steel was worth the optical risks, particularly if I kept this scope in my garage or car trunk most of the time and if I added active cooling to the primary and mirror cell. To that end I've purchased cooling components at extremely low prices online and I'll get to that in due time.

The 1/4" x 3" x 4" corner plates provide counterweighting and will also serve as anchors for the trusses through the thick circular ring "mirror box". (Stiffness, stiffness, stiffness). I put "mirror box" in quotes because this ring is little more than an extension of the mirror cell. I realize that I may go through more than one iteration of hexapod adjustment methods and hardware, and such large plates will allow for future changes in tube anchors with a large footprint in which to drill and tap holes.

The trusses will be made of 1" aluminum tubing with a 0.063" wall thickness. I happened to get these in 8' lengths for only $4 each (75% off) at my local Lowes five years ago when this store decided not to continue to stock these, then their only round tubing in stock.

I'm aware that thinner wall thickness is generally preferred in discussions on this forum for trusses and 1 1/4" or 1 1/2' outside diameters are often favored for their dramatically increased stiffness in bending. The consensus seems to be that 0.035" thickness is too prone to denting, 0.049" thickness is preferred for inch dimension tubes, and 0.063" tubes are too heavy. For those where metric sizes are available, 1.0-1.2 mm is generally preferred.

In my case, this is what I have because I got it years ago really cheap. You'll see in time that I tend to hoard components well in advance when they are available cheap. I don't have money to squander and I don't consider my time making components while "playing at building" to be very valuable.

Further analysis has me fairly confident in using this tubing in my particular case, too. My design is a fast enough focal ratio and will have a long enough UTA (for light baffling) that my truss poles will be less than 40" long. This means that bending stiffness is not a major concern and, since my scope will have really large altitude bearings, the center of gravity of the optical tube assembly will only be a few inches below the midpoints of the tubes. As a result, the difference between 0.049" and 0.063" tubing is trivial in terms of balance. I'm not trying for an ultralight here, just a sturdy fast scope whose components for setup are all reasonably light for my sunset years.

The secondary mirror spider will be wire. I'm currently planning on making this in a design of Mark Cowan's that he was kind enough to share, with the wires terminating in a manner recently shared by CN member totvos. This particular spider design requires a fairly long UTA for the vertical offsets required but that's what I want for ambient light baffling.

By the way, I have decent skies and live in a "green zone" in terms of light pollution. I do, however have a fair amount of ambient light at my two usual observing zones. I may employ a light baffle in the UTA that would be my version of the "Turco Dark Bucket" that I'm whimsically calling a "baffling hat". I expect many will find its appearance quite baffling.

I once made a mockup of Mark's crossed wire spider that he shared on this forum in 2006. I found it to have excellent fast damping times from a rap on the UTA, even compared to well designed offset spiders. His design that I plan to use here is what he considers to be his "current SOTA" (which he had to tell me stands for "state of the art"). It's a crossed design with no duplication of angles that has very, very little cantilevering of the mirror from the lines of force. I was impressed by the video he sent me by PM of a secondary in his mockup UTA damping immediately after being whacked hard. That mockup was with dental floss. Let's see how 28 gauge steel wire or thinner goes.

The UTA will be made of wood, a combination of hollow panel plywood rings, wooden dowels, softwood boards, and very thin plywood for its walls. The wire spider will be anchored to the dowels which are being strategically placed for both the spider and the light baffle.

The secondary holder might be made of some carbon fiber sheet molded components I have left over from an R&D project I did at work a few years ago (more hoarding). I have the good fortune of getting to play with water jets as part of my primary work functions. Thin wood would suffice both for weight and a similar coefficient of thermal expansion to the mirror but I might use some of this stuff "because I can ".

Finally, as far as materials go, a lot of small components are being made of a really tough red plastic I still have plenty of scraps of from various past water jet projects. This is a Cass Polymers product they used to call "Red Plank" and now sell as "Red Stuff". It's a urethane tooling board that is extremely tough, dimensionally stable, epoxies very well, and drills taps, saws and turns very well with woodworking tools. Many of you have seen this material used on my various past products. It's not extremely expensive, but is available only in large pieces 12" x 60" by various thicknesses. It isn't available in round stock or small blocks. I'm using it in places where others typically use either nylon or Delrin. It's material properties are slightly better for these things than nylon or Delrin except for weight. It's density is quite a bit less than that of aluminum.

Well, that's my description of the major materials being used in a (rather large) nutshell. I know my descriptions tend to be very long, but I hope in them many of you find gems to use or are prompted to offer caveats if you recognize risks. I'm glad I had a few years of playing with a homemade 10" dob to help me with weighing the design compromises on this project!


Edited by jtsenghas, 04 January 2019 - 08:42 PM.

  • mark cowan, ed_turco, Oberon and 2 others like this

#12 ckh

ckh

    Surveyor 1

  • -----
  • Posts: 1,938
  • Joined: 21 Feb 2015
  • Loc: Arizona

Posted 04 January 2019 - 12:10 PM

Delrin is actually slightly heavier and just a little bit harder. PP-1052 Red Stuff Tooling Board is about 71 lbs/ft3, delrin is about 88 lbs/ft3 and aluminum is 160 lbs/ft3.  Delrin machines like a dream, how about this Red Stuff?  Can you get a good finish on it?


  • jtsenghas likes this

#13 Lukes1040

Lukes1040

    Ranger 4

  • *****
  • Posts: 320
  • Joined: 26 Dec 2016
  • Loc: Ohio, USA

Posted 04 January 2019 - 12:43 PM

I like the idea of anchoring the truss tubes into the steel cell. After cutting out my "mirror box" and gluing it up, I was concerned with flexure and longevity of the wood connection. I may have to use this one....

 

Edit: I prefer the long explanations as opposed to just saying here are some pics.  waytogo.gif


Edited by Lukes1040, 04 January 2019 - 12:45 PM.

  • jtsenghas likes this

#14 jtsenghas

jtsenghas

    Aurora

  • *****
  • topic starter
  • Posts: 4,735
  • Joined: 14 Sep 2014
  • Loc: The flatlands of Northwest Ohio 41.11N --Bloomdale

Posted 04 January 2019 - 01:14 PM

Delrin is actually slightly heavier and just a little bit harder. PP-1052 Red Stuff Tooling Board is about 71 lbs/ft3, delrin is about 88 lbs/ft3 and aluminum is 160 lbs/ft3.  Delrin machines like a dream, how about this Red Stuff?  Can you get a good finish on it?

I didn't expect it's density to be less than half that of aluminum.  Thanks for looking that up, Carl.

 

It can generate as good a finish as you want to polish.  I generally sand it up to 400 grit for a flat finish and the components you'll see on this thread will be representative of that. I don't need extra reflections in a scope, though. I have, however,  polished it to a high gloss with cloth and rouge on the lathe so that I can see my face in it.

 

If a 1400 pound water jet nest made of of it is dropped from fork truck blades onto a concrete floor it shatters in to glossy red pieces that look similar to lava glass rock with clam shell chips on the numerous pieces....

 

...ask me how I know.....blush.gif



#15 jtsenghas

jtsenghas

    Aurora

  • *****
  • topic starter
  • Posts: 4,735
  • Joined: 14 Sep 2014
  • Loc: The flatlands of Northwest Ohio 41.11N --Bloomdale

Posted 04 January 2019 - 01:16 PM

I like the idea of anchoring the truss tubes into the steel cell. After cutting out my "mirror box" and gluing it up, I was concerned with flexure and longevity of the wood connection....

Here's an area where truss collimation works to an advantage.  The mirror back supports and edge supports can both be integral to the mirror cell AND the bottoms of the trusses can be anchored to the corners of the mirror cell.


  • Lukes1040 likes this

#16 mark cowan

mark cowan

    Vendor (Veritas Optics)

  • *****
  • Vendors
  • Posts: 9,543
  • Joined: 03 Jun 2005
  • Loc: salem, OR

Posted 04 January 2019 - 04:57 PM

I like the idea of anchoring the truss tubes into the steel cell. After cutting out my "mirror box" and gluing it up, I was concerned with flexure and longevity of the wood connection. I may have to use this one....

 

Edit: I prefer the long explanations as opposed to just saying here are some pics.  waytogo.gif

It does cut down on any sources of flex, unlikely as they might be here.


  • Lukes1040 likes this

#17 GShaffer

GShaffer

    Knight of Ni

  • *****
  • Moderators
  • Posts: 6,686
  • Joined: 28 Feb 2009
  • Loc: Bogart, Ga USA

Posted 04 January 2019 - 05:08 PM

JT that is an interesting mirror cell design.....I am assuming the rockers inside the tubing are metal.....how are the tooling board posts attached to them and how did you get the assembly inside the tubing? Threaded posts on the rockers and then screwed the tool board posts on to them thru the holes in the 1" tubing?



#18 jtsenghas

jtsenghas

    Aurora

  • *****
  • topic starter
  • Posts: 4,735
  • Joined: 14 Sep 2014
  • Loc: The flatlands of Northwest Ohio 41.11N --Bloomdale

Posted 04 January 2019 - 06:41 PM

...and how did you get the assembly inside the tubing? Threaded posts on the rockers and then screwed the tool board posts on to them thru the holes in the 1" tubing?

Close enough. I used set screws threaded into both the rockers and the plastic knobs.  Greg, you helped  me to decide on what I'll discuss next..  This weekend I plan to do a detailed writeup on the mirror cell.


Edited by jtsenghas, 04 January 2019 - 07:06 PM.

  • GShaffer and Lukes1040 like this

#19 Lukes1040

Lukes1040

    Ranger 4

  • *****
  • Posts: 320
  • Joined: 26 Dec 2016
  • Loc: Ohio, USA

Posted 04 January 2019 - 08:16 PM

Maybe you’re planning on getting to this, hope I’m not jumping the gun. But, did you use a different method than Jonathan to cut the slots for the mirror box in the bearings and the slots in the ground board for the bearings to ride in?

Edited by Lukes1040, 04 January 2019 - 08:17 PM.


#20 jtsenghas

jtsenghas

    Aurora

  • *****
  • topic starter
  • Posts: 4,735
  • Joined: 14 Sep 2014
  • Loc: The flatlands of Northwest Ohio 41.11N --Bloomdale

Posted 04 January 2019 - 10:31 PM

Maybe you’re planning on getting to this, hope I’m not jumping the gun. But, did you use a different method than Jonathan to cut the slots for the mirror box in the bearings and the slots in the ground board for the bearings to ride in?

Actually, I've not gotten to one of those items in the build yet, but I'm doing both of those quite differently than he did.  The following may be clear only to those readers very familiar with Jonathan's build in particular. For the rest of you, I assure you I'll show in detail my methods in due time. 

 

Jonathan and I didn't cut notches in the altitude bearings the same way, although we both did make cylindrical notches in the altitude bearings for the mirror box to rest into. He devised a method to hold the bearing fixed and rotate a small circular saw to remove material. I chose to make a jig to rotate the altitude bearing past a fixed router bit in the Shopsmith configured in the drill press position. I'll show what I did in detail as soon as I complete the mirror cell discussion, possibly within the next few days.

 

As for the notching of the lower ring for the altitude bearings, I'll be doing that quite a bit differently too. I'm likely to do that with hand tools such as saws, mallets and chisels aided by squares and sliding t-bevels for precision and personal safety. I'm considering having the right hand (nearer to observer) altitude bearing captured between rollers to act as close tolerance thrust bearings, and the left hand bearing to have more clearance on the inboard and outboard faces to avoid any binding from modest variation on the distance between the altitude bearings.  The reason I want one altitude bearing captured fairly tightly is for increased accuracy and repeatability with using the azimuth setting circle, which will be an optional item to install with each setup.



#21 jtsenghas

jtsenghas

    Aurora

  • *****
  • topic starter
  • Posts: 4,735
  • Joined: 14 Sep 2014
  • Loc: The flatlands of Northwest Ohio 41.11N --Bloomdale

Posted 07 January 2019 - 01:21 PM

Designing the mirror cell support point locations:

Before I show how the mirror cell was built, I would like to show how I arrived at the dimensions of the components. Please bear with me on this, because this post may in fact greatly change how many of you approach mirror cell design and the dimensions you choose.

The analysis for determining the number and placement of support points was performed with that excellent free downloadable software package "GUI PLOP". PLOP was developed by David Lewis and Toshimi Taki in the mid-1990s in the programming language FORTRAN and the results of it were published in Sky and Telescope in April 1996. PLOP stands for "PLate OPtimizer" and is a finite element program that models the deformation of parabolic mirrors under gravity loading. The support points are modeled as single points with zero friction on the back of the mirror, and the model has gravity forces normal to the back plate of the mirror, which is the orientation of greatest deformation when the mirror is aimed at zenith. Even at 45 degrees altitude these forces and the resultant deformations don't even drop by 30%. Prior to the finite element analyses from this and similar programs such as NASTRAN (NASA STRuctural ANalysis, which I had the opportunity to use in college in the early 1980's) support points were generally placed further from the center of the mirror than was actually optimal for minimizing deformation. The mistaken assumption, which can be found in many ATM books, even some of those written through the early 90's, was that one wanted to place the supports so that equal areas were supported by each point. Thus, for a three point cell the recommendation was to put the points at 71% of the mirror's radius. The actual optimal radius turns out to be closer to 40% for a three-point cell and about 60% the radius for a six-point cell. An additional brilliant insight by Lewis and Taki that changed results a little more was that minimal deformation that resulted in a slightly different focal length was acceptable since trivial changes of telescope focus were generally considered acceptable. The optimal support placement, therefore, was that in which the resultant shape of the surface had the minimal deviation from a new best fit parabola.

My mirror is one of those Hubble Optics (funny name for a Hong Kong based company in my mind) sandwich mirrors. It has a thin mirror fused to numerous cylindrical spacers, which are in turn fused to a flat rear glass plate. This design doesn't reduce the weight by more than a third as compared to similar thickness monolithic mirrors, but I can see why they have appeal for rapid cooling. Air can circulate freely between two thinner plates. This mirror presents a problem for analysis in PLOP because it is not homogeneous. It must be nearly as stiff as a solid mirror of the same overall thickness since its mass is concentrated towards the faces, but it must deflect slightly less under gravity due to the reduced overall mass. Still, there must be SOME deformation in the pattern of the spacers from gravity loading even if it may be less than a nanometer. I chose to model my mirror in PLOP as if it were a monolith of the same overall dimensions, figuring that would give a conservative result of more than actual error. The edges of the mirror present additional challenges for appropriate edge supports since there is no material at the plane of the center of gravity at the edge. I'll show shortly how I addressed that with vertical whiffletree details.

I like to use as a permissible amount of RMS deviation on the profile a value of 5 x10-6 when using PLOP. Results in PLOP are in millimeters, so 5 x 10-6 mm is 5 x 10-9 m, or 5 nanometers. What's necessary and sufficient for this tolerance is a matter of debate, depending on what percentage of an overall error budget one wants to use up on mirror supports. On this value I defer to the judgement of those more experienced than I. It is safely under 25% of the error of all but the very best mirror profiles.

A three point mirror cell with dimensions optimized by PLOP, as I expected wasn't sufficient:

PLOP 3PT.JPG

In this graphical plot the low areas in red are about 17.5 nanometers low. Even if the RMS (root mean squared averaged) error of 6.4 nanometers is arguably sufficient, I dislike support points so close together for maintaining collimation or for keeping edge supports at precise heights. Either a slight loss of collimation or a flexing of the edge from deviations on edge support heights can have a more significant adverse effect on images than a few nanometers in figure.

A six point cell with support locations optimized by PLOP got things well within tolerance:

PLOP.JPG

As is typical for PLOP analyses I've seen, the error for a 6-point cell was less than 1/4 the error associated with a 3-point cell. This low value for error makes me feel better about my assumptions regarding the mechanical properties about this sandwich mirror construction; even if the true amount of deformation is  more than this model shows, it more than certainly has a RMS error well under 5 nanometers. 

I rounded the value for the calculated radius to the nearest 1/8 of an inch to put it at 3.625", or 3 5/8". I reran PLOP with that value for the radius fixed and verified that the RMS error only got a couple tenths of a nanometer worse. So why did I rely on PLOP if I was going to change the radius to something more convenient anyway? I was more confident I could execute my design more precisely if the distance from my rocker fulcrums to each pivot was rounded to the nearest 1/16 of an inch and PLOP verified to me that such a change resulted in a negligible increase in error. Execution is as important as design, and with round numbers I'm far less likely to make mistakes and have components rubbing or asymmetrical, either of which would have a much more adverse effect on quality than the trivial change in dimensions.

The previous paragraph makes me think of Nils Olof Carlin, who we had the great misfortune to lose in August of 2017. For those of you who never had the privilege to converse with him on these forums, or benefit from his sage advice, it is hard for me to express the huge impact this Swedish physicist had on this hobby. It was Nils Olof (and in his culture first and middle names are used together) who had the inspiration to use a barlowed laser for collimation, with the dramatic improvement on primary axial error that made possible. He wrote up this technique and published it in Sky and Telescope Magazine in 2003. It is the very principle on which the Glatter collimation tools are made (R.I.P., Howie Glatter) and still marketed by Starlight Instruments. Nils Olof also performed the math that computed the size of coma-free sweet spots on paraboloids, and expressed primary collimation tolerances as a function of the f-ratio only. Primary axial error tolerance is now accepted in millimeters as 0.005 x F3, with F equal to the focal ratio of the primary. This means that under f/5 we really need to keep the primary mirror aimed at the focuser within 0.6 mm, and this tolerance gets halved again at f/4. That is why I concentrate more on execution than trivial changes in dimensions from PLOP designs.

In 2015 I was sharing some private messages with Nils Olof about equatorial platforms and he sent me a picture of his 150 mm scope on his home built platform. i was surprised to see that he had a square mirror cell made from aluminum on that scope with four support points at nearly two-thirds the mirror's radius and a single rocker between two of them. I inquired about that design, since it would be easier to construct in a lot of ways as compared to a six point cell. He replied with words to the effect that, as long as the error from support points is sufficiently low as analyzed by PLOP, there is no reason not to diverge significantly from optimized designs.

I would like to demonstrate that last point by making a few more calculations with PLOP on my particular mirror that might be very useful on an ultra-light design (which my scope is not attempting to be). A 4-point PLOP-optimized cell for my mirror would also have acceptable error and would have all four points at a radius of 0.53 the mirror's radius:

PLOP 4PT.JPG

Note that the error is less than HALF that of the 3-point cell. Now, suppose further that to add a rear cooling fan one needed those supports to be at 0.6 the mirror's radius:

PLOP 4PT .6R.JPG

That result is also perfectly acceptable if one uses a maximum RMS error of 5 nanometers (5 x 10-6 mm) as the tolerance. I show this to try to open up possibilities in your minds for possible designs. The addition of cooling fans has been shown to have a very useful effect in this hobby even for these (now) modest apertures. Square holes are very easily adapted for the installation of square computer fans. With a single rocker one can have half the error of a three point cell. That's not useful to me in this case, though, because I want a triangular frame for my mirror cell for a more robust "space frame" construction...


Edited by jtsenghas, 07 January 2019 - 09:38 PM.

  • mark cowan, Pierre Lemay, ForeignObserver and 1 other like this

#22 jtsenghas

jtsenghas

    Aurora

  • *****
  • topic starter
  • Posts: 4,735
  • Joined: 14 Sep 2014
  • Loc: The flatlands of Northwest Ohio 41.11N --Bloomdale

Posted 07 January 2019 - 02:17 PM

Designing the mirror cell itself:

 

With the support points determined at 3 5/8" radius, or 3.625", the distance between points becomes 3 5/8" also on this regular hexagon:

 

DS Mirror cell closeup.JPG

 

I chose to make this mirror cell similar to that of the 10" Tardiscope, which I shared on this forum in 2015.  That cell had been made deliberately lightweight, in an attempt to minimize thermal mass near the optical path, which could slow cooling and add to "seeing", and because that scope originally had a lighter upper end.  In 2014 I had only a cheap rack and pinion focuser, no Paracorr and no heavy wide angle eyepieces.  It took only one visit to the Cherry Springs Star Party for me to get tempted over to the "wide side".  After adding better finders and a Paracorr in 2015, and accumulating slowly (mostly through CN classifieds) a fairly complete set of Explore Scientific 82 degree eyepieces I had to add a significant counterweight to the bottom of the Tardiscope in the form of a 1/4" thick steel plate bolted to one of the door panels.  Another significant weight addition was the Antares low profile focuser, which was the fourth in a series of focusers I've had on that scope, including a homemade helical one that was on the scope when it was presented at Stellafane in 2015.

 

I'm determined to make this 12.5" scope  capable of supporting a similar focuser, Telrad finder, Paracorr, and the ES 82 degree eyepieces.  I haven't done a thorough balance analysis, because I'm not sure exactly how heavy my top end will be by the time I'm done dressing it up with light baffles and a shroud.  With some "back of the envelope" calculations (literally) I decided to put several pounds into the mirror cell and its mounting plates for balance.

 

I've become less concerned about the added thermal mass of this cell near the optical path because I'll probably keep this scope in my garage and expect to add active cooling to it with 12-volt fans. 

 

One neat thing about CAD is that it easily helps one to arrive at such dimensions.  Drawing my circumscribed hexagon with a 3.625" radius, I could offset the lines of that rectangle 0.5 inches either way and and extend those offsetted lines to determine the triangle required.  I cut three pieces about 0.050" long to the 10.371" dimension shown in the above drawing and then sanded them at 60 degrees.  When I screwed them together they matched the full-sized drawing I had printed out and hung on the circular bulletin board with my oversized push-pins:

 

drawing on dartboard.jpg

 

Note that the paper drawing shows the 1/4" x 3" x 4" mounting plates arranged in an array every 120 degrees.  The actual arrangement I used is shown on the electronic drawing above it, with two of the plates rotated 15 degrees downward so that they provide a better mounting surface for the edge supports.  They are each 45 degrees from the lower edge of the mirror rather than 60 degrees.


Edited by jtsenghas, 07 January 2019 - 09:53 PM.

  • Lukes1040, Earthbound1 and ForeignObserver like this

#23 jtsenghas

jtsenghas

    Aurora

  • *****
  • topic starter
  • Posts: 4,735
  • Joined: 14 Sep 2014
  • Loc: The flatlands of Northwest Ohio 41.11N --Bloomdale

Posted 07 January 2019 - 03:47 PM

Mirror cell, continued:

 

This mirror cell, as Greg pointed out last week, is unusual in that the rockers are within the tubes with the contact knobs protruding through clearance holes in the top.  This is definitely a case of design form following function, and perhaps it isn't that unusual among engineering-type builders.  Jonathan Pogson (Oberon), Mark Cowan, TopherTheME, and I have all built previously this way.  Jonathan and I were building this way simultaneously in 2014 and 2015, although Jonathan supported triangles on an 18 point cell with such rockers on Merope, Mark showed with a photograph that he had done this earlier, and Topher, more recently on the scene did this in his build thread in 2016 on his own.  Such a design is not only clean-looking, it also has no twisting forces on the rockers due to support on both sides.

 

I first turned knobs that could be tightened with a 3/8" wrench from my ever-present "Red Stuff":

 

rocker knob.jpg

 

The way I made these was that I first sliced a piece of material into a 3/8" x 1/2" strip.  I then drilled and tapped a series of 1/4"-20 blind holes all carefully to the same depth.  Next, I sliced them into blocks about 9/16" wide and threaded them onto a piece of threaded rod chucked into my lathe.  Finally I turned them with a skew chisel, sanded the spherical tip with 400 grit sandpaper, and removed each with a wrench as shown.

 

I also made Teflon spacers to fit the 1/16" gap on either side of the 1/2" rockers inside the tubes.  These were made just under 1/16" thick x 11/16" wide by a couple inches long.  They were drilled 9/32" to be clearance for a 1/4" bolt.  An exploded view of the assembly is shown here:

 

mirror cell exploded.jpg

 

You can see that each knob is fitted with a 1/4"-20 set screw which is cranked in tight.  

 

Tool tip:  When needing to drill pairs of holes precisely the same distance apart (which I had to do six times here), first drill a jig from a piece of scrap wood as shown with the hole locations very carefully laid out and drilled:

 

drill jig.jpg

 

Then, after the first hole is drilled in your workpiece, pin this jig to that first hole with another drill bit of the same size.  Drill the second hole through the jig for perfect spacing.

 

To assemble everything I first slid the rocker into the middle of the tube. Next, with a thin awl I slid a Teflon spacer on either side of that rocker until the 9/32" holes aligned with the 1/4" tube holes.  Next, I reached into the tube with a pencil and lifted one end of the rocker until all the holes aligned and slid the bolt through.  Finally, I screwed the red knobs with set screws already in them into the previously tapped rocker and tightened them with a 3/8" wrench.

 

After I also attached bottom 1/8" thick gussets with 10-32 flathead screws and top 1/4" thick plates with flathead 1/4"-20 screws the result looked like this:

 

Mirror cell top plates on.jpg

 

Tool tip:  To precisely align all of the holes for flathead screws, first clamp them together and then drill through all layers with the diameter needed by the tap--here #7 or about 0.201".  After that, open up the hole in one piece--here 1/4"--and countersink it and tap the other piece.  Cutting fluid helps with both the tapping and countersinking.  If the work can't be securely clamped for all holes at first, install one screw this way, and then drill the tap diameter for the others through both layers.  Remove the first screw and drill, countersink and tap as required.  The result should be the good fit shown in the above photograph.

 

Design Tip:  If you want a longer smooth diameter on a bolt, such as the axle on these rockers, choose a longer bolt--here I used 2 1/4".  When the bolt is cut down to barely engage in the nut--here a Nyloc--the smooth diameter all but enters the nut as well.  This 2 1/4" bolt had no threads within 1" of the bolt head.  A 1 1/4" bolt would have threads digging into both clearance holes and getting looser over time.


Edited by jtsenghas, 07 January 2019 - 07:11 PM.

  • GShaffer, Lukes1040 and ForeignObserver like this

#24 jtsenghas

jtsenghas

    Aurora

  • *****
  • topic starter
  • Posts: 4,735
  • Joined: 14 Sep 2014
  • Loc: The flatlands of Northwest Ohio 41.11N --Bloomdale

Posted 07 January 2019 - 04:16 PM

Mirror cell, continued:

 

To fit the mirror cell flush with the "mirror box" ring, I scribed the pattern with a sharp knife and used chisels and a mallet and a router plane to make perfect 1/4" deep cavities.  For small jobs with sharp corners like this I prefer to use a router plane to an electric router:

 

routing cell pocket.jpg

 

perfect fit.jpg

 

To resolve the problem of edge support location, I used shower sliding glass door rollers with a pivot between them.  These rollers used to be stocked at my local Lowes but aren't anymore.  Fortunately, I found them at Menards for $5 for a set of four of them.  They are available with either flat or rounded wheels.  I've used both in the past for various projects. For this application I used the flat ones:

 

mirror cell installed.jpg

 

The red "mirror clip" fits between the plates and is there to make sure the mirror never tips forward.  It clears all glass surfaces by about a millimeter when the mirror is in place and against back and edge supports.

 

The top roller is at the center of gravity of the upper concave/plano mirror plate.  The lower roller is at the height of the center of gravity of the lower plate.  the pivot between them is at the overall center of gravity:

 

edge roller details.jpg

 

Frankly, these sandwich mirrors are a bit of a nuisance to set up edge supports for, but I think I've got one of the best solutions.  At 45 degrees on either side of the lower edge of the cell a mirror this size is adequately supported and I like hard stops for maintaining collimation.  The manufacturer suggested I use a seat belt type sling, but I'm not confident I can maintain mirror position within a few tenths of a millimeter that way.

 

Oh, and finally, those clearance notches for both edge supports were made with a Forstner bit and a mallet and chisels--wicked sharp chisels!

 

 

 

Attached Thumbnails

  • chopping edge clearance.jpg

Edited by jtsenghas, 07 January 2019 - 07:18 PM.

  • Diego, GShaffer, Oberon and 4 others like this

#25 GShaffer

GShaffer

    Knight of Ni

  • *****
  • Moderators
  • Posts: 6,686
  • Joined: 28 Feb 2009
  • Loc: Bogart, Ga USA

Posted 07 January 2019 - 06:24 PM

Another variation on the internal rocker theme I noticed right after you had posted the 1st pic of yours......


  • jesustheplumber, jtsenghas and Earthbound1 like this


CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.


Recent Topics






Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics