Equatorial Platforms, for free!

[wobblewing]

..well, almost free!

I bought my 'space and rocket mad' 9 y/o son a small 150mm table top style Dobsonian (SW 150p Heritage), but being a techy kind of guy I find myself quite interested in it too.  I'm a electronics design engineer (there are quite a few of us in the Cambridge area of UK), I'm into radio/electronics/communications etc...so am familiar with technical things, CAD, and have been using Fusion360 a little - free for home use..

The cloud cover here has been pretty much non-stop rubbish, so we've only got out with it about 3 times between the odd cloud break - so still essentially zero experience in astronomy and I'm finding my feet with an enthusiastic boy!

I have two 3D printers, one really old one which is essentially junk (I'm using it for parts) and another Prusa MK3 which is a pretty decent machine - this will make good quality parts (well, good enough).  I quickly knocked up a phone holder to do a bit of afocal photography with my phone - got some reasonable pictures of the moon, but other things such as capturing Orion's nebula have been challenging due to the need for long exposures (phone only goes up to 10s) - this not being very practical on a non driven Dob mount of course, just cannot track well enough.

I've already got some reasonably expensive hobbies (private pilots license, RC modelling etc), so I'm mainly interested in keeping this low cost, as I don't think I see myself getting totally absorbed in it.  Certainly I don't want to be spending lots on it like I see with some of the very 'pro' looking outfits on this forum.... but 'astrobiscuit' has demonstrated how you can get good results at very low cost.

This lead me into researching very low (i.e. almost zero) cost ways of sidereal tracking for a bit of 'Covid lockdown fun' with the emphasis on using the stuff that I have in my garage/workshop!  I have loads of bits of wood, a few skateboard bearings, and essentially an old 3d printer that can donor stepper motors.. Also the decent 3D printer that I can make parts with (I have a load of PLA).

I then found Neil's great thread (linked below), where he's done a really good job of it.... So I thought I'd have a go as well.. I noted a few issues with the tracking (although most issues may be down to the Dob itself) and thought I'd have a go at a different trick of driving it to try to improve it (whilst still spending the grand total of £0).    Some discussion started on there, but it is rightly unfair to hijack that thread (although the ship may have sailed on that already!), so thought I'd start a new one here.

I've got a lot of it drawn in CAD now, whether I finish it will be another matter 🙂

As an electronics engineer, the motor drive bit isn't too much of a problem.  I am also involved in teaching STEM to kids outside of school and have a collection of BBC micro:bits where we teach block coding as well as Python to primary and early secondary school kids..  A micro:bit could be idea here as it has a simple 5x5 LED display built in, a couple of buttons and (if needed) a radio.. just connect it to a motor driver board I already have from making robots with the kids 🙂

Description follows.......

 

 

Edited January 15, 2021 by wobblewing

[wobblewing]

.. I started using this webpage as a basis for understanding (http://www.reinervogel.net/index_e.html?/Plattform/Plattform_e.html) and hence went off to measure the CofG of the SW Heritage 150P.. approx 25cms give or take,  It's not a heavy telescope, so balancing exactly on C of G probably doesn't need to be critical, and of course it will be lowered a little by the platform itself.

This geometry seemed to suit the telescope, bearing in mind the style and size of its base (it will sit on top, unmodified).  Being in the UK I'm at 52degrees latitude, thus forming the cone angle as shown below.

Edited January 15, 2021 by wobblewing

[wobblewing]

The platform (11mm plywood) and Vertical North Sectors (projected from the circular sector to get the correct movement).  The VN Sectors will be 3D printed.

Edited January 15, 2021 by wobblewing

[wobblewing]

For the South bearing I thought I'd try a friction bearing, based upon the bearing surfaces being the smooth 'first layers' of the 3D printed surface moving against each other, perhaps with some lubrication (although I'm not sure if this is needed).  If this doesn't work, I could use a couple of skateboard bearings I have.  There is a pin in the middle and the top and bottom sections rotate around this axis.

Edited January 15, 2021 by wobblewing

[wobblewing]

So here it is, without the drive and support bearings.. the basic geometry.

 

Edited January 15, 2021 by wobblewing

[wobblewing]

After this bit, much thought on whether just to drive the support rollers (which can be problematic and slip), or try something different.

In my box of 'available bits' I have a stepper motor with a long worm drive as was used on my old 3Dprinter Z axis...  This could give good resolution of control.

A regular NEMA17 stepper gives native resolution of 1.8degress, although this could be further divided down - by 1/2, 1/4 or even 1/8.  The cone above has a circumference of 2200mm, so 6.1mm along the circumference for 1degree of sidereal movement.

with the 3D printer's long worm gear and using the native 1.8deg stepper resolution, this gives about 5 arcseconds (need to check my sums) of resolution (this can be taken down to 2.5arcseconds or less relatively easy in software) - should be good enough?

I can use the available brass worm drive 'nut' to ensure smooth operation and bolt it to a 3D printed part which couples the drive to the platform.  I have tested out 3d printing the ball joints and it works reasonably well (probably good enough).. ok it's not as 'tight' as a metal one (there is ~0.2mm 'play'), but I had PLA, so these are free 🙂.. The play may be a problem though, since a wobble here would be very noticeable if zoomed in at  mag of x200 or similar..

Fortunately, I've still spend £0 😄

Edited January 15, 2021 by wobblewing

[Spile]

I like your style but it got a bit high-tech towards the end 🙂

This was how I "made" my first equatorial...

Edited January 15, 2021 by Spile

[wobblewing]

Nice.. Is that rubber band powered?  I can't zoom in 🙂

[AlexK]

Following.

FIrst of, I would recommend rethinking your designs from scratch instead of repeating wood/steel EQP designs. For example, the shape of sectors you have copy-pasted is actually quite suboptimal. Folks making them from wood have to use straight angles and surfaces as that just simpler, while you can design a perfect elliptical shapes everywhere with super-optimal load bearing arching supports instead of trivial L-shape. The latter is even dangerous for a hollow 3DP part, as any "crease" with sharp angle (90 deg must be considered a sharp angle here) is a potential snap point weakening the construction.

Consider the same for the South support too, don't make it like that Pisan-tower cylinder you pictured, use wide cones on each side, as all you actually need for the tracking is the contact surface only, the rest must be beefed up to accept strong screws and bolts into platforms not breaking the plastic at high tension.

For the poncet drive, the classic version had a variant where on the rod you had a special cam "nugget" shaped in a special way to push the round (or also special counter-cam shaped) pin on the rocking platform so it's rotating with the constant speed with the motor moving the cam linearly along the threaded rod with the constant speed as well. Can't google-out it right away, but if you simply draw on the paper/in the CAD several phases of a point on the rod shifting and the point on the sector rotating just a bit ahead of it all the time (the dot on the rod should move at the X-axis speed equaling the speed the dot on the sector is moving along the same X axis at the very end of its movement range) you will see some spiral sector forming when you link corresponding dots. Just figure what math curve can approximate that (something like a cycloid: x= r(a-sin(t); y = r(1-cos(t))) just don't forget to make a correction for the rotation of the radius of the pin it is pushing. Though I recall that cam had some limitations, as it will be prone to strong bending forces at higher angles, it still might be worth investigating (see here: https://kisi.deu.edu.tr/saide.sarigul/Cam_Mech.pdf how to design cams in general, the curve we need to cam-follow is the cycloid).

Edited January 15, 2021 by AlexK

[wobblewing]

Well, i started off thinking of doing a drive system on the VN sectors, so that's why they are vertical. 😃  Didn't see much point in changing it as i can map the stepper motor electronically to get the required consistent rate of movement.  

No screw mounts done yet,  that's the easy bit! Also ribbing etc i usually add at the end.

 

I'm trying to picture the cam idea in my mind,  where are you suggesting mounting it?   I'm not familiar with original poncet mechanisms (new to telescopes), i just drew this out according to the required geometries 🤪

Edited January 15, 2021 by wobblewing

[AlexK]

Still, I would highly recommend using a real worm gear (printed cog, and metal rod worm), as you will be able to disconnect the worm any moment and reset the platform to any position manually in zero time. With the classic poncet, like you have pictured, you will have to "fast rewind" for quite a long time every N minutes (though not a biggie, but I could see myself swearing at it's sluggishness ).

 

[wobblewing]

I have a plan for that 😃  a clip on / off worm 'nut' to replace the brass one.

[wobblewing]

The problem with 3dp gears i find is the play, although I will probably end up with similar play in the 3dp ball joints.

[wobblewing]

Also,  metal worm gear isn't free, unless i use the one i have from the printer 😅

Up-cycling!

Edited January 15, 2021 by wobblewing

[AlexK]

53 minutes ago, wobblewing said:

Well, i started off thinking of doing a drive system on the VN sectors, so that's why they are vertical. 😃  Didn't see much point in changing it as i can map the stepper motor electronically to get the required consistent rate of movement.  

No screw mounts done yet,  that's the easy bit! Also ribbing etc i usually add at the end.

I'm trying to picture the cam idea in my mind,  where are you suggesting mounting it?   I'm not familiar with original poncet mechanisms (new to telescopes), i just drew this out according to the required geometries 🤪

Gotcha, so that's just a researching sketch so far.

The cam piece is simply sliding along the rod (X-axis, red in your CAD, I guess) from right to left as on your picture but the point on the sector is moving down first, then up projected on the Z-axis. And its speed is accelerating first, then deccelerating as it's passing the lowest point, if you look at its motion projection to the X axis. Right? Now imagine that depending on Z (height from the rod zero level to the point where the cam is touching a pin on the sector) you are compensating for that speed change (which is dx per dt) by splitting the dx shift vector into smaller dx'/dt shift and some dz'/dt shift dynamically over that cam's surface. It's harder to imagine it on the full circle as it wouldn't work at too steep Zs, I would draw it for a 7 deg sector (7+7 = 14 degrees which is enough to drive the platform for an hour) .

Edited January 15, 2021 by AlexK

[wobblewing]

Ok I see what you mean.

I figured i would do all of that correction in software, maybe that's 'cos I'm an electronics guy 😆

Edited January 15, 2021 by wobblewing

[AlexK]

26 minutes ago, wobblewing said:

I have a plan for that 😃  a clip on / off worm 'nut' to replace the brass one.

I see, yes, I saw such solutions as well, instead of the brass nut you can print one opening in half or just use a rack piece on the top. Still, quite a cumbersome mechanism, especially for tweaking in the dark, I think.

The worm gear's beauty is that it's asymmetrical, which is creating an uneven wear forces which you can leverage to your advantage. If you print the cog a bit tighter than necessary (just reduce the fit tolerances in half), then the metal threaded rod will be able to shape it into the perfect fit in just a dozen of dry runs (don't run it to fast as plastic will melt from the friction energe release into the temperature buildup). Then add a lithium grease to slow the wear, and don't forget an easy cog replacement mechanism (e.g. bolts attaching the cog sector in place instead of wood screws).

Finally, even metal gears have some slack, to amend that you can simply preload the west side of the platform (just shift your dob's legs a bit West) to create the always positive one side contact between rod thread and gear teeth.

Edited January 15, 2021 by AlexK

[AlexK]

12 minutes ago, wobblewing said:

Ok I see what you mean.

I figured i would do all of that correction in software, maybe that's 'cos I'm an electronics guy 😆

I'm MS in EE and MS in CS, that's why the worm gear and cam's advantage is so clear to me  They are doing the perfect job just passively thus cannot mishap as any EE/CS solution could eventually.

But OK. Let's brainstorm it a bit too.
How do you plan to tell the controller what speed to use at any given moment, especially if you can move the slider back or forward manually any moment?

[AlexK]

By the way, just checked my printer rods, they seem to be OK for the wormgear, you just need to find or create the proper cog profile for their thread. I think, it's called the pipe or tubing thread.

Correction:
I's called ACME thread (though I see different threads might be used in 3D printers, so check your's to be sure), which is very close to the worm thread. The latter is a bit deeper (0.69P vs 0.5P).

If not, a 1/4"x20 steel rod with the standard Sharp V (which is even taller than Worm)  is like $1 per foot at any hardware store.

Edited January 15, 2021 by AlexK

[wobblewing]

35 minutes ago, AlexK said:

I'm MS in EE and MS in CS, that's why the worm gear and cam's advantage is so clear to me  They are doing the perfect job just passively thus cannot mishap as any EE/CS solution could eventually.

But OK. Let's brainstorm it a bit too.
How do you plan to tell the controller what speed to use at any given moment, especially if you can move the slider back or forward manually any moment?

I was going to have a simple look up table from the start position. Since it's a stepper motor the position is predicable. It can count along the length.  The LUT has pre calculated speeds according to the count.  Reset the count when resetting the platform.

Didn't see the need to write a PhD thesis 😉  last time i did one of those it was too time consuming 🤣🤣

Edited January 16, 2021 by wobblewing

[wobblewing]

17 minutes ago, AlexK said:

By the way, just checked my printer rods, they seem to be OK for the wormgear, you just need to find or create the proper cog profile for their thread. I think, it's called the pipe or tubing thread.

Correction:i
I's called ACME thread (though I see different threads might be used in 3D printers, so check your's to be sure), which is very close to the worm thread. The latter is a bit deeper.

Ah OK,  ACME ..I think that's available in fusion 360.  I have seen that there,  but never used it. 

Edited January 15, 2021 by wobblewing

[wobblewing]

I think the pitch on the worm rod is 2mm, so that should probably work.  I've only ever managed to go down to an M10 thread with 3dp using 0.4mm nozzle, but that was with some small tooth profile changes. It did work,  but not fantastic.

Edited January 16, 2021 by wobblewing

[AlexK]

23 minutes ago, wobblewing said:

I was going to have a simple look up table from the start position. Since it's a stepper motor the position is predicable. It can count along the length.  The LUT has pre calculated speeds according to the count.  Reset the count when resetting the platform.

That's another manual control prone to errors...
But OK. Perhaps, use the end switch then and after moving the slider manually, command the controller to rewind to the "zero" (same as on printers). That would give it a more reproducible starting position.

[wobblewing]

4 minutes ago, AlexK said:

That's another manual control prone to errors...
But OK. Perhaps, use the end switch then and after moving the slider manually, command the controller to rewind to the "zero" (same as on printers). That would give it a more reproducible starting position.

Re-position the platform and press reset.  The motor stops when it gets to the end (max count),  so you soon find out when to reset it.  That's easier than accurate polar alignment.

It's only an evening toy used once in a while, i'd be amazed enough if i can get some good images.  My phone doesn't do long exposure!

Edited January 16, 2021 by wobblewing

[AlexK]

11 minutes ago, wobblewing said:

I think the pitch on the worm rod is 2mm, so that should probably work.  I've only ever managed to go down to an M10 thread with 3dp using 0.4mm nozzle, but that was with some small tooth profile changes. It did work,  but not fantastic.

The rod should be metal, you want to print the rack (worm cog sector), which is much easier as you can print it on the side (most recommended for the threads but often needs supports if your parts fan is weak. Definitely a stronger filament will be needed, I would use nylon.

Edited January 16, 2021 by AlexK