Equatorial platform type 'BM'

[Chriske]

For most of you it could be completely new, this way of tracking stars.
In the days when I was 'a bit' younger I made my first 'Boxmount' (as I called it then) to track the stars with my first big telescope, a selfmade 10" newt with Coulter optics. I even wrote an article about that big mount in S&T. Sinds then I made several of this equatorial mounts. During the following years I always tried to make them less higher every time again, compared to that very first 'big box' Ive built in the early '80.
The one I'm building right now is intended to be used in different ways.
I can install a big (bino-)scope on it, or I can put one or more camera's on the top of that platform. The platform has a diameter of 500mm and in 'neutral' position the entire device is  only 66mm high.
These early tracking platforms I made in the early days  could track for more then 1 hour. The one I'm building now I reduced the tracking time to 25minutes max. During observing I notice I never observe an object longer then (about) 10 to 15 minutes, 20m, maybe more, when someone else joins me at the eyepiece. So why mechanically bother to make them track for one hour or even longer.

Purpose now is to install a motor and ad some electronics that automatically will go to its starting position(at a higher speed of course) when it is asked to do so, say at the end of a observing session. The mount will then wait there until I give the order to start tracking again. I need some time to find me a new object...
To eliminate tracking errors at the most lower or higher point of the run I might install a curved threaded rod. But because of the short running time(and the distance between axis and rod) there's maybe no need for that curved rod. Need to test it. Anyway I will cut the threaded rod myself on a lathe because the commercially available rods aren’t very accurate. Most of these threaded rods are 'rolled' and not cut. They're fine to be used in constructions all around but not good enough for this task.
In the drawing the motor and threaded rod are not 'installed' yet.

Platform seen from the east, in neutral position.
Polar axis is in these two white blocks.

Seen from the North in high and lowest position

 

Seen from the South-East

 

Seen from the West, looking at the (printed)gliding blocks(still wood in the drawing). Between these blocks I always glue sheets of teflon.
The threaded rod hold it all in position(not yet in the drawing).

 

Start of tracking

 

End of tracking session

[starman345]

This is interesting.  I'm not sure I understand completely, maybe a picture of the unit with the threaded rod installed will be more clear. I hope you will show your build progress here.

[Chriske]

Made a small animation. seen from the west. Threaded rod not yet 'installed' in the drawing.
The central part  (is in the back - east side of the mount) is in fact the polaraxis.
The gliding blocks are all printed and are located at the Eastside of the mount, so in fact in the front of the drawing.
The first motion -upper part closing toward the lower part- that is the tracking direction. Speed is way to high of course. I only show that motion partly, it can in fact close in a bit more, anway...
The second motion is when the upperpart  needs to be rewind for a next session.

I hope it's clear..?

 

 

[A P B]

This is really interesting. I hope you post more pictures as you go along. Would love to understand how it works.

[Chriske]

Together with a Nema17 stepper I will use this planetary gear system to drive my platform.

To make it more printable I changed a few things.
On the left is the original unit, and right is my modified PGU.

[Chriske]

My friend Henri made me a PCB to drive a Nema17 stepper (Nano + stepper driver) and it worked just fine.

A few months ago I could vary the speed of that stepper over the complete range of the potentiometer.
Yesterday I started the stepper again to see if all was still ok and it didn't work at all. Turning the pot it did it's job but turning even further to speed things up the stepper stopped and started to make noises....😟
Problem is I do not know what is wrong with that PCB

So I need to change the main body of that planetary reduction unit to install another motor(dc most probably)

 

[Chriske]

The planetary reduction unit works just fine.
At low speed there's almost no noise. But what I did not expect was that this first step is completely free of backlash.
Two additional levels need to be installed. It will end up with a reduction of 1:520(about).

 

[Chriske]

Change of concept....
Because the mount has no constant speed over it's racking length I'll try another approach. The stepper will rotate very accurately at the same speed, but because the mount has  a straight threaded driving rod it will not track the stars at the correct speed along the length of it's treaded rod. As a matter of fact it has the same tracking problem as the well known barn-door trackers.
A friend of mine(electronics guru) said he could solve that problem. After some calculation he sent me a sketch to upload in my arduino in which he programmed the stepper's speed variations along the 20 minutes of tracking.

I'm very happy with this solution.
Needless to say this solution could be adapted for any 'barn-door-isch' mount.

[markse68]

Hi Chriske, your gearbox looks great! have you tried/thought about 3d printing a harmonic gearbox?

Mark

[Chriske]

Not yet, but I'll try it alongside with a cycloidal gearbox, that too looks promising.

[RobinH]

Hi Chriske,

do  you have any more info on your friend's  varying speed strategy that you can share? I am building a straight threaded rod barn door tracker but  have only a constant  speed on my stepper. I think that I will need to vary the speed but have not yet thought of a simple strategy to do this.

Robin

[Chriske]

I'll ask Sven whether it is do-able for a real barndoor capable of tracking one hour or even more.
I mean, my Boxmount only tracks for about 20 minutes. For my goal it is more then enough as I explained higher up in this thread. So the differences in speed between centre position and highest/lowest point are very small.

I suppose it will work for longer tracking, it's only a matter of extrapolating I guess.

[RobinH]

Thank you,

I have designed my barn door for a max of one hour movement. After doing a spreadsheet with lots of rows it does seem that there is only a very small speed difference at one hour. 

Robin

[Sven_r]

First some theory:
The speed of the motor needs to change over time; this can be done with a regular dc motor or with a stepper motor.
The regular dc motor requires very precise tuning or a feedback loop with an encoder. To simplify things I opted for the stepper motor (with reduction) these motors will turn against any required speed.

To change the speed over time the time between each step the motor should be different, but since the resolution of microcontroller timers is limited the time between two steps will not always differ if required. And calculating the next step tick (or timer interrupt) requires cpu power which is rather limited if you opt for an Arduino based system.
To overcome this one can calculate the number of steps for each time; For each period of time (lets say each second) one needs to take the required angle of the tracker platform and calculate the number of rotations to reach that angle. The number of rotations translate into number of steps using both the number of steps of the stepper and any reduction that would be in place.
An example with fictive numbers to keep things simple: 0sec:0steps, after 3 seconds 10 steps (taken), after 6 sec a total of 22 steps (taken), 9 sec 36 steps, ...
This means that the speed will differ for each period of time: 0->3sec: 10steps/sec, 3->6sec: 12steps/sec, 6->9sec 14steps/sec.

To minimize the errors it's better to calculate the total number of steps for each period of time (this number always increases) this way rounding errors are taken into account because the error will not increase for each entry in the table.

These pre-calculated values can be loaded into a table that can be stored in flash memory. It's a long list of the number of steps for each period in time. The microcontroller reads the first entry in the table calculates the required step frequency and starts the movement for the number of steps listed in the table. When the movement is ended the next entry in the table is taken, the number of taken steps is subtracted and with this results the next step frequency can be calculated. A good thing to do is to calculated the passed time since for the given movement, this again should lower the error.

Diving deeper: Typically one should use an interrupt based timer to get the best accuracy: Everyone working with an Arduino and timer interrupts knows that calculating the best value for the OCR register involves a loop and a table of the prescaler values to get the best match for that OCR register.
Since you microcontroller will not change this can also be done "offline" so the table can include a list of OCR values and the number of steps for each period.

I hope this explains the theory used.

I did some additional calculations on the maximum error; using full step mode (highest accuracy) and a reduction on the stepper should limit the error -0° 0' 0,02" in the configuration use by Chriske with a straight treaded rod with an new speed every 30 seconds. Without the correction the error will be -0° 0' 45,86".

How long you're able to track depends on how many table entries can be stored in the flash of the microcontroller.

[Chriske]

Thanks Sven for joining in...!!!

[RobinH]

Thanks Sven,

I will read  your very detailed explanation and see how I can use your ideas .

Robin

[Chriske]

The most important part in a tracker this accurate is the quality of the threaded rod.
A standard threaded rod out of a 'normal' hardware shop is a no-go. Even with a integrated PEC system it would not be feasible imo. These threaded rods are good enough to track the stars if you don't mind the objects floating up and down constantly in the FOV.
As my little lathe can't handle thread cutting I will use the threaded rods very often use in 3D printers for the Z-axis. They're available with different pitch. From 2 to 8mm if I'm not mistaken. The 2mm pitch would be the best choice.
If I'd have a lathe capable of cutting threads I'd go for that option. These would be the perfect choice.

My point : using 'everyday' threads in combination with Sven's software-solution would be a waste of time.

Lead screw with T8 brass nut should do the job nicely.

[dark star]

I wanted to ask if this equatorial platform could be modified to make it a dual axis platform? I would like to make one of these so that I could use my 12 inch dobsonian for EAA and maybe  some astrophotography. There is a commercial dual axis equatorial platform that can even have a guide camera added to it. But it is expensive and also very heavy.

David

 

[Chriske]

A dual axis platform...? Explain please...

[dark star]

My understanding is that a dual axis platform is also driven on the North/South declination axis as well as right ascension. Commercial versions are sold by Tom Osypowski in America and TEPP in Hungary.

I have successfully made a 14 inch Dobsonian. But since it was not square enough to use with digital setting circles, I will most likely attempt a simpler EQ platform first. As I think making even a single axis one with sufficient accuracy will be a challenge for me.

The platform you are making does not look extremely complicated, and I have lots of experience of cutting circles out of wood with a router.

[Chriske]

Nope there's no Decl axis involved. Just RA.
But now that you mention it, maybe  a good idea in case the polar axis is not correctly aligned and small corrections are needed.
A second axis would be overkill imo. Just a polar-axis correction should suffice.

Anyway a second axis would not correct Decl errors. Yes it would but only when observing at culmination point. But observing more East or South there's no way you can compensate for Decl errors.
A device could be added at the rocker of the Dob itself...just maybe.

Now that I come to think of it, how do regular-equatorial-platform users correct Decl errors while taking pictures. For years now I used this Boxmount, never had any issues with it because I exclusively used to observe visually . But now, planning to use it for astrophotography, it's a complete different problem.
Sven's solution is absolutely perfect to track in RA, but what if (small)corrections are needed in Decl....?

Think with us...🤔

[Chriske]

2 hours ago, dark star said:

The platform you are making does not look extremely complicated, and I have lots of experience of cutting circles out of wood with a router.

Nope not at all, the 'polaraxis' is tucked away nicely between these two sheets of plywood. You need two printed parts(that's what I do) and a few angled blocks. The polaraxis is set at 51° and the (gliding) blocks at the opposite site are cut at 39°(in my case) But you living in or near London it would not differ that much from mine I suppose.
Look at the drawings higher up and you'll notice a few holes in both sheets of plywood. These holes are there to allow passage to both blocks at the upper and lower sheet of plywood. This is not necessary of course. The only reason I made it this way is the challenge involved to make this EQ platform as low as possible, 65mm in my case. It is indeed a ridiculous low platform, I know, but I like a challenge from time to time..:-)

If I were you I'd make it like the Boxmount I made for my Kutter Bino a few years back. Far more easy to construct.
This page explains it all...

[RobinH]

thanks for the suggestion of 3d printer rod and brass nut.  I have used a m6 threaded rod (construction type) on my barn door tracker but have been experiencing problems sometimes with the stepper motor stalling.  The 3d printer drive thread is not expensive like I had imagined so I will be changing.

Robin

 

[Chriske]

7 hours ago, RobinH said:

thanks for the suggestion of 3d printer rod and brass nut.  I have used a m6 threaded rod (construction type) on my barn door tracker but have been experiencing problems sometimes with the stepper motor stalling.  The 3d printer drive thread is not expensive like I had imagined so I will be changing.

Robin

 

A stalling stepper when using a barndoor tracker..? You do have serious problem. There are in fact no forces involved at all in a regular barndoor tracker.
You have a mechanical problem or the current on the stepper is far to low. Or you need a bigger stepper.

Edited May 23, 2022 by Chriske

[Sven_r]

16 hours ago, RobinH said:

thanks for the suggestion of 3d printer rod and brass nut.  I have used a m6 threaded rod (construction type) on my barn door tracker but have been experiencing problems sometimes with the stepper motor stalling.  The 3d printer drive thread is not expensive like I had imagined so I will be changing.

Robin

 

You might check the power supply, if it cannot deliver enough current it will also cause the stepper motor to stall. The Arduino might survive because the onboard voltage converter and/or capacitor(s).