My low-cost DIY RA drive for EQ5 mount

[kman42]

I recently completed a little project in which I built a low-cost motorised drive for my EQ5 mount. I thought I'd share my results for other people looking for a solution, as commercial drives tend to start at $200 in Australia.

My drive only turns the RA axis, as it is primarily for astrophotography. Because of this, I don't need any tracking systems - it is simply a motor that spins the RA axis at the correct rate of 1 revolution per 24 hours.

My drive system uses a 4 wire, bipolar stepper motor, driven by an A4988 stepper driver and an Arduino Nano, all powered using a 12V battery. The stepper motor runs in quarter-steps, and drives the pre-existing RA fine-turning gear via a V-belt. This way, I still have full manual control over the mount if I wish. Here is a parts list in Australian dollars inc. postage. Everything was purchased from either Ebay or the local Jaycar:

1. Nema 14 4 wire bipolar stepper motor : $22 

2. A4988 stepper driver: $7.55

3. Arduino Nano V3.0: $6.52

4. 2x T2.5 5mm bore timing pulleys: $20.80

5. T2.5 6mm wide, 145mm long timing belt: $6

6. 4-prong locking plug (male and female): $4

7. Jiffy box: $1.95

8. 2x DPDT switches: $4

9. Locking 2 way connector: $3

10. Alligator clips: $4

For a total cost of $79.80

 

The counter-weights for the EQ5 weigh 11.5kg in total. The maximum distance the weights can be placed on the counterweight rod is 0.33 metres, therefore the maximum torque that a motor would ever need to turn the mount is 11.5x2x0.33 = 7.59 Nm.

The RA fine adjustment knob on the EQ5 mount turns 144 rotations for every full rotation of the telescope's RA axis. Therefore, the maximum torque I would need to apply to the RA fina adjustment gear is 7.59/144 = 0.0527 Nm or 5.27 Ncm.

I chose a Nema 14 Stepper motor with 18 Ncm holding torque and a maximum current draw of 0.8A per phase. This motor takes 200 steps to complete a revolution; therefore, if I want to revolve the telescope every 24 hours, I'm looking at 1 rotation of the RA fine adjustment knob every 10 minutes, or 1 step of the motor every 3 seconds. Using quarter steps to smooth out the motion, thats 1 quarter step every 0.75 seconds.

The arduino is programmed with a ridiculously simple code:

const int stepPin = 3; 

 
void setup() {
  pinMode(stepPin,OUTPUT); 
 
}
void loop() {
    
      digitalWrite(stepPin,HIGH); 
      delay(375);
        digitalWrite(stepPin,LOW);
      delay (375);
}

This sends a pulse to the A4988 'step' input every 0.75 seconds. That's all the programming done.

 

The arduino is powered via 12V, which then outputs 5V to the stepper driver logic controls, while 12V also goes to the stepper driver motor voltage input. The unit is controlled with a power switch stop and start the motor, and a switch to reverse the motor's rotation by swapping the wires to one of the poles. 

 

The RA fine-adjustment shaft needed to be modified to fit the T2.5 pulley. First, I bored the pulley to 6mm to match the shaft. Then, the locking nut of the shaft was removed and the brass nut with the external thread was cut down to be flush with the housing. New notches were cut to faciulitate tightening of the brass nut, and a hole was drilled into the side of the housing to accept a locking screw to replace the locking nut, which no longer fit. The result of this was that the RA shaft now had enough room to fit the T2.5 pulley. After this, the motor was bolted in place using mechano peices, and a hole was drilled in the cover for the 3 prong locking connector. Once the cover was fitted to the mount, the 4-prong connector was the only visible sign of any modifications. 

 

 

Below are some pictures to illustrate the build:

 

The original mount with the RA drive case attached:

 

The original mount with the RA drive case removed to show where the motor will go:

 

The modified shaft with the T2.5 gear:

 

 

 

The stepper motor in place. The triangular plate ensures the motor can never move enough that the belt loses tension: 

 

The 4 prong locking connector - the motor's wires run through this:

 

The completed control unit - the circuit boards were held in place using RTV silicone. The entire circuit was soldered before dropping the whole thing into the box. Made things much easier:

 

 

The completed unit. I put a neo magnet on one of the tripod's leg bolts, and the control unit just clips on to the magnet when in use. The magnet is strong enough that it will never get bumped off.

 

All that is left is to test its accuracy with some photos!

[kens]

Now that you have an Arduino controller it would not be hard to code it to do guiding pulses so you can autoguide.

[kman42]

Dammit kens, I thought I was finished. Now you've got me researching autoguiding...

[kens]

I have some code that enabled auto-guiding on my Super Polaris mount but that was using unipolar steppers. I'm happy to share that FWIW.

Auto-guiding will help to overcome periodic error in the RA drive train. Periodic error exists to some degree in every mount due to slight imperfections. It causes tracking to speed up and slow down slightly - enough to cause problems with astrophotography. Autoguiding uses a second scope and camera to track a guide tar and send corrections to the RA motor that keep the star stationary.

When you get your camera, aim at a star and take a long exposure (5 minutes or more). That will show you graphically the effects of periodic error and polar alignment.

[kman42]

That was the plan - take some pictures and see how much drift there is. How big a scope do you need for the guide scope?

[kens]

That depends on your imaging scope. Autoguiding software like PHD2 http://openphdguiding.org/ can resolve down to around 0.1 pixels. As a simpe rule of thumb, a guidescope should have a focal length in the order of 1/4 the focal length of your imaging scope. Many people adapt a finderscope for the purpose.

[kman42]

I'm using a Skywatcher 200p, 1000mm focal length. Is aperture important at all for the guide scope?

[kens]

You need to be able to find a decent guide star near the object you ar eimaging. If you are imaging small galaxies, a bright star can be hard to find and more aperture can help. Typical aperatures for a guidescope range from 50 to 80mm. You can also use an off axis guider (OAG) which is an attachment on the focuser that lets you attach both your imaging camera and guide camera so you don't need a guide scope. The OAG uses a small prism to pick off some of the light coming through the focuser and send it to the the guide camera.

[John78]

So just swap the nano for a mega and load AstroEQ on it then you have a full goto EQMod setup once you've built a Dec axis... 

https://www.astroeq.co.uk/tutorials.php?link=/doku/doku.php?id=buildown

 

[kman42]

A Dec axis motor is definitely on the list. But for now, I have some example photos from a weekend of astrophotography. 

Carina Nebula: 4 minute exposure, 800 ISO, f/2, 58mm Helios lens on Pentax K7. It looks like streaking but that is just coma.

 

Jupiter: 5 second exposure, 100 ISO, f/5, 1000mm through Skywatcher 200p on Pentax K7.

Acturus: 15 second exposure, 500 ISO, f/5, 1000mm through Skywatcher 200p on Pentax K7. The stars are just starting to streak.

MIlky Way: 7 1/2 minute exposure, 1000 ISO, f/4, Pentax 17-70 f/4 @ 17mm on Pentax K7.

 

The verdict: the mount can track some brighter DSO's - I'll have to find some more impressive objects to see what results I can get from it. However, it excels at shorter focal lengths.

Edited May 1, 2017 by kman42

[jase1973]

On 22/04/2017 at 04:42, kman42 said:

I recently completed a little project in which I built a low-cost motorised drive for my EQ5 mount. I thought I'd share my results for other people looking for a solution, as commercial drives tend to start at $200 in Australia.

My drive only turns the RA axis, as it is primarily for astrophotography. Because of this, I don't need any tracking systems - it is simply a motor that spins the RA axis at the correct rate of 1 revolution per 24 hours.

My drive system uses a 4 wire, bipolar stepper motor, driven by an A4988 stepper driver and an Arduino Nano, all powered using a 12V battery. The stepper motor runs in quarter-steps, and drives the pre-existing RA fine-turning gear via a V-belt. This way, I still have full manual control over the mount if I wish. Here is a parts list in Australian dollars inc. postage. Everything was purchased from either Ebay or the local Jaycar:

1. Nema 14 4 wire bipolar stepper motor : $22 

2. A4988 stepper driver: $7.55

3. Arduino Nano V3.0: $6.52

4. 2x T2.5 5mm bore timing pulleys: $20.80

5. T2.5 6mm wide, 145mm long timing belt: $6

6. 4-prong locking plug (male and female): $4

7. Jiffy box: $1.95

8. 2x DPDT switches: $4

9. Locking 2 way connector: $3

10. Alligator clips: $4

For a total cost of $79.80

 

The counter-weights for the EQ5 weigh 11.5kg in total. The maximum distance the weights can be placed on the counterweight rod is 0.33 metres, therefore the maximum torque that a motor would ever need to turn the mount is 11.5x2x0.33 = 7.59 Nm.

The RA fine adjustment knob on the EQ5 mount turns 144 rotations for every full rotation of the telescope's RA axis. Therefore, the maximum torque I would need to apply to the RA fina adjustment gear is 7.59/144 = 0.0527 Nm or 5.27 Ncm.

I chose a Nema 14 Stepper motor with 18 Ncm holding torque and a maximum current draw of 0.8A per phase. This motor takes 200 steps to complete a revolution; therefore, if I want to revolve the telescope every 24 hours, I'm looking at 1 rotation of the RA fine adjustment knob every 10 minutes, or 1 step of the motor every 3 seconds. Using quarter steps to smooth out the motion, thats 1 quarter step every 0.75 seconds.

The arduino is programmed with a ridiculously simple code:

const int stepPin = 3; 

 
void setup() {
  pinMode(stepPin,OUTPUT); 
 
}
void loop() {
    
      digitalWrite(stepPin,HIGH); 
      delay(375);
        digitalWrite(stepPin,LOW);
      delay (375);
}

This sends a pulse to the A4988 'step' input every 0.75 seconds. That's all the programming done.

 

The arduino is powered via 12V, which then outputs 5V to the stepper driver logic controls, while 12V also goes to the stepper driver motor voltage input. The unit is controlled with a power switch stop and start the motor, and a switch to reverse the motor's rotation by swapping the wires to one of the poles. 

 

The RA fine-adjustment shaft needed to be modified to fit the T2.5 pulley. First, I bored the pulley to 6mm to match the shaft. Then, the locking nut of the shaft was removed and the brass nut with the external thread was cut down to be flush with the housing. New notches were cut to faciulitate tightening of the brass nut, and a hole was drilled into the side of the housing to accept a locking screw to replace the locking nut, which no longer fit. The result of this was that the RA shaft now had enough room to fit the T2.5 pulley. After this, the motor was bolted in place using mechano peices, and a hole was drilled in the cover for the 3 prong locking connector. Once the cover was fitted to the mount, the 4-prong connector was the only visible sign of any modifications. 

 

 

Below are some pictures to illustrate the build:

 

The original mount with the RA drive case attached:

 

 

The original mount with the RA drive case removed to show where the motor will go:

 

 

The modified shaft with the T2.5 gear:

 

 

 

 

 

 

The stepper motor in place. The triangular plate ensures the motor can never move enough that the belt loses tension: 

 

 

 

The 4 prong locking connector - the motor's wires run through this:

 

 

The completed control unit - the circuit boards were held in place using RTV silicone. The entire circuit was soldered before dropping the whole thing into the box. Made things much easier:

 

 

The completed unit. I put a neo magnet on one of the tripod's leg bolts, and the control unit just clips on to the magnet when in use. The magnet is strong enough that it will never get bumped off.

 

 

All that is left is to test its accuracy with some photos!

Nice work, I can't tell from the pictures, but did you tie ms2 high to enable quarter stepping?

[kman42]

I can't remember sorry. It was a long time ago. I think instructions on quarter stepping motors are pretty well documented though.

[jase1973]

On 03/11/2018 at 11:16, kman42 said:

I can't remember sorry. It was a long time ago. I think instructions on quarter stepping motors are pretty well documented though.

no problem. I've gotten past that now but I've killed 2 drivers by being silly.

Are you still using it or have you moved on. I'm just trying to get something going while I build an rduinoscope controller.

[Dr_Ju_ju]

A very good controller is Tom Carpenters', AstroEq (https://www.astroeq.co.uk/tutorials.php) which is Arduino based, and can be controlled from a PC etc....

[jase1973]

I've revised the code to suit my application and my irresistible urge to change things.

I've used an 8255 driver as I had them to hand and I've wired the microstepping up to the arduino so I can see what difference a change in stepsize makes and it's easy to reverse in code the motor direction after soldering everything.

this particular code is using 1/16th steps and a 16t and 40t pulleys, so needs to step 2.5x faster

 

#define STEP_PIN 2
#define DIR_PIN 3
#define Mzero 10
#define Mone  11
#define Mtwo  12
#define Denable 5

 

 
void setup() {
  digitalWrite(DIR_PIN, HIGH);
  digitalWrite(STEP_PIN, LOW);
  pinMode(DIR_PIN, OUTPUT);
  pinMode(STEP_PIN, OUTPUT);
  digitalWrite(Mzero, LOW);
  digitalWrite(Mone, LOW);
  digitalWrite(Mtwo, HIGH);
  digitalWrite(Denable, LOW);
 
}
void loop() {
    
      digitalWrite(STEP_PIN,HIGH); 
      delay(37);
        digitalWrite(STEP_PIN,LOW);
      delay (38);
     
}

[jase1973]

10 minutes ago, Dr_Ju_ju said:

A very good controller is Tom Carpenters', AstroEq (https://www.astroeq.co.uk/tutorials.php) which is Arduino based, and can be controlled from a PC etc....

It's an interim solution while I am putting an rduinoscope together. http://rduinoscope.byethost24.com/rDUINO_step1.html

[kman42]

5 hours ago, jase1973 said:

Are you still using it or have you moved on. I'm just trying to get something going while I build an rduinoscope controller.

I don't do astrophorography often, but when I do I still use it. If I can get the polar alignment right, it's very accurate and consistent. 

[jase1973]

On 04/11/2018 at 20:09, kman42 said:

I don't do astrophorography often, but when I do I still use it. If I can get the polar alignment right, it's very accurate and consistent. 

just tested mine on the breadboard, inbetween the clouds and seems to be staying on track so far. need to solder it up and make a box.

I just found visual frustrating due to the speed things move and not always being able to reach the slowmo control to keep things centered.

Startests are suddenly easier

Thanks for the handy project.

[jase1973]

On 04/11/2018 at 20:09, kman42 said:

I don't do astrophorography often, but when I do I still use it. If I can get the polar alignment right, it's very accurate and consistent. 

Question, do you have a polarscope. I'm looking for one but havn't found one to fit yet.

 

[kman42]

6 hours ago, jase1973 said:

Question, do you have a polarscope. I'm looking for one but havn't found one to fit yet.

 

I'm in Australia so a polar scope would be noe use to me unfortunately... I have to resort to other means!

[jase1973]

11 hours ago, kman42 said:

I'm in Australia so a polar scope would be noe use to me unfortunately... I have to resort to other means!

they cater for the southern polar axis too...

 

[kens]

10 hours ago, jase1973 said:

they cater for the southern polar axis too...

 

There's no known polarscope that can make sigma Oct bright enough to find easily ?

I thought most polarscopes were pretty standard. My three mounts (Vixen SP, EQ6 and Avalon) all take the same scope. So I have three unused bits of kit. 

If you want it for AP then there are better, sometimes simpler, methods for polar alignment using imaging and/or guide camera.

[jase1973]

20 hours ago, kens said:

There's no known polarscope that can make sigma Oct bright enough to find easily ?

I thought most polarscopes were pretty standard. My three mounts (Vixen SP, EQ6 and Avalon) all take the same scope. So I have three unused bits of kit. 

If you want it for AP then there are better, sometimes simpler, methods for polar alignment using imaging and/or guide camera.

http://www.avalon-instruments.com/support/13-troubleshooting/104-polar-alignment-using-the-new-skywatcher-polar-scope

how about the 4 stars that are listed here for polar allignment?

[theropod]

On 04/11/2018 at 10:53, jase1973 said:

I've revised the code to suit my application and my irresistible urge to change things.

I've used an 8255 driver as I had them to hand and I've wired the microstepping up to the arduino so I can see what difference a change in stepsize makes and it's easy to reverse in code the motor direction after soldering everything.

this particular code is using 1/16th steps and a 16t and 40t pulleys, so needs to step 2.5x faster

 

#define STEP_PIN 2
#define DIR_PIN 3
#define Mzero 10
#define Mone  11
#define Mtwo  12
#define Denable 5

 

 
void setup() {
  digitalWrite(DIR_PIN, HIGH);
  digitalWrite(STEP_PIN, LOW);
  pinMode(DIR_PIN, OUTPUT);
  pinMode(STEP_PIN, OUTPUT);
  digitalWrite(Mzero, LOW);
  digitalWrite(Mone, LOW);
  digitalWrite(Mtwo, HIGH);
  digitalWrite(Denable, LOW);
 
}
void loop() {
    
      digitalWrite(STEP_PIN,HIGH); 
      delay(37);
        digitalWrite(STEP_PIN,LOW);
      delay (38);
     
}

I think you should look into using “unsigned long millis” to control/eliminate your sktech halting delay statements. I am currently rewriting my code for the DIY alt/az I built employing not only this feature but also employing proportional motor speed based on joystick mapping. Since I am not a C expert by any means this is a steep learning curve for me, but the Arduino is capable of great precision if the code is written tightly. Even you simply change your current delay statements to delayMicroseconds you can achieve far greater timing control.

Yeah, the same desire to mess with things is part of my makeup too.

[kman42]

I just did some tests with the tracking mount.  After getting perfect polar alignment using the drift technique, I achieved a tracking accuracy ratio of 9000/FL (compared to the general rule of 500/FL for untracked astrophotography). In other words I was able to shoot 45 second exposures at 200mm and 30 second exposures at 300mm with no streaking. This was pointing at Sirius (almost at the zenith)  where tracking times will be the worst.