×

INDI Library v2.0.6 is Released (02 Feb 2024)

Bi-monthly release with minor bug fixes and improvements

DIY Flatfield Dustcover with RPi

  • Posts: 271
  • Thank you received: 72

Hey.
I uploaded the files to thingiverse :)
I updated the original thread I had on the 'Dark Cap' so It won't hijack @pauledd's thread

indilib.org/forum/ekos/1452-feature-dust....html?start=12#16659
Proud owner of Observatory 17b - A remote Linux observatory.
Website: Observatory 17b
Build thread @ SGL: Starting summers observatory project
6 years 10 months ago #16660

Please Log in or Create an account to join the conversation.

  • Posts: 2247
  • Thank you received: 223
Hey Magnus, this cover looks awesome, nice work.
The following user(s) said Thank You: Magnus
6 years 10 months ago #16662

Please Log in or Create an account to join the conversation.

  • Posts: 314
  • Thank you received: 34
I made "some" progress with my dustcover. I decided to make parts of it made of wood because that was the easiest. In the course of this I realized that I was maybe wrong with the assumption that my littly tiny servo would have enough power.

I left the EL-Flatfield Foil away to just concentrate on the dustcover. The cover itself as baseplate for the foil was made of acrylic glas.




Its weight is about 150g. Some tests showed I would need about 800g force to lift the arm. That approx. matched my calculations I did before:



However this does not work. I can unpark the cap but the servo never reaches the end position and stays making noise.
I tried to increase the servo voltage up to 6.7 Volts but it still doesn't reach the end position. I dont know whats wrong here because with that 3.8cm servo arm it should be able to handle up to 1.2Kg on that arm.

Before I consider to buy another stronger servo I'll try to replace the dustcover with an aluminium one, 0.5mm thickness, that should give a weight of about 20cmx20cmx0.05cmx2.6g= 52g and that should give a load on the servo of about 52x4.73= 245g, one third of my current weight.

Here's just a short video of my dustcover in action, looks good but doesn't reach end position and continuously makes noise.

--= human, without Windows™ =--
pls excuse my bad english! :)
Last edit: 6 years 9 months ago by pauledd.
6 years 9 months ago #17387

Please Log in or Create an account to join the conversation.

  • Posts: 271
  • Thank you received: 72
Great work, and nice to see you are making progress :)

Something you can try to be 100% sure that it is a weight issue is to see that your servo in fact can move further with the lid disconnected (just remove the bolt that is connecting the brass / anodized servo horn to the wood linkage.

The reason for this is that from watching the video I'm not entirely sure if the lid stops when it has past the point where it's the heaviest, or just after.
(When I think about it, you don't have to unscrew anything. Just point the scope at the zenith. If it fails to move at all or if it opens to the expected position, it's the weight for sure).

However it might be servo centering. When you hear the bussing at the most open position, and it stops to move then it is binding.
There is three ways a servo can bind.
1. The object that is being actuated reaches a mecanical endpoint, and can move no further.
2. The object being moved is to heavy and the servo motor is skipping.
3. Internal binding.

If the weight test above do not proove it's the weight (i.e servo motor skipping) then it's allmost certanly internal binding.
Servos work at a PWM pulse from 0.5 - 2.5, at least some of them! There is no standard that ceeps a servo from actually woriking in PWM range 1.0 - 2 as an example.
If you tell a servo to go to PWM 0.5, and it has internal limits at 1, then the servo will appear to bind, even if it is not binding at an endpoint or skipping from overload.

If you have not done so allready, you can disconnect the servo horn completely, and try to Park / Unpark while encreasing / decreasing the (Max travel limits [set open travel (ms), set close travel (ms)]) in the Calibrate tab. You should aim for the limits closest to 0 and 3 that you can acchieve while the servo makes no sound at the endpoints. You should do this while the (Preffered travel limits [set open limit (%), set close limit (%)]) is set to 100, 100

After that set your close limit to 90 - 95%, park the cap and attach your servo horn and linkage while the lid is closed against the telescope. (Then you have 10 - 5 % to work on when the lid is connected to the servo, to ensure that it can close fully). Leave your preferred open limit at 100%, as it does not look like your setup will bind.

Hopes this helps to determen if you need new servo / lighter lid, or if you can calibrate it out, (not sure if I explaned thiswell enough on the github readme, so tought I'll go into some detail :) )
Proud owner of Observatory 17b - A remote Linux observatory.
Website: Observatory 17b
Build thread @ SGL: Starting summers observatory project
6 years 9 months ago #17397

Please Log in or Create an account to join the conversation.

  • Posts: 314
  • Thank you received: 34
Thank you Magnus_e for your detailed ideas!

I am nearly 98% sure its a weight issue. The cap moves fully if the scope is slightly off the home-position so that the cap is not in its "heaviest position". I currently discuss my torque/weight calculations on a physic-board and its a little bit more complex than I thought.
Someone presented a formula to get the caps torque when its opened at 90° degree (the most heavy position):

Where
Mk: is the needed cap torque
m: is mass of the cap
g: Gravitational acceleration (9.81)
l: length of the cap (diameter)

With that formula he got about 13Ncm force needed to move the cap.
My motor has 46Ncm divided by 3.8cm of the servo arm that makes ~12Ncm.
I have no confirmation yet that its the right way to calculate it but it seems my motor
specs missed the caps torque very close. So I am quite shure replacing the acrylic cap
with 0.5 or 1 aluminium will do the trick. I will order two sheets of it, its quite cheap.
--= human, without Windows™ =--
pls excuse my bad english! :)
Last edit: 6 years 9 months ago by pauledd.
6 years 9 months ago #17398

Please Log in or Create an account to join the conversation.

  • Posts: 35
  • Thank you received: 5
you might try a count weight along the plane of the flap.
provided you have clearance along the OTA, then only a very small weight would be required.
Last edit: 6 years 9 months ago by Rob Jones.
6 years 9 months ago #17400

Please Log in or Create an account to join the conversation.

  • Posts: 271
  • Thank you received: 72

This would of cource depend on the orientation of the scope when unparking.
I park my scope at alt +1, so a counterweight would not work when unparking the cap at that position.

However... You could add a spring to the equation. Fasten it so the spring is slightly tensioned when the lid is at it heaviest. This might cause the lid to not stay completely closed when the servo is not powered, so possibly not a god idea? Ebay id for some springs 281291676039
Proud owner of Observatory 17b - A remote Linux observatory.
Website: Observatory 17b
Build thread @ SGL: Starting summers observatory project
6 years 9 months ago #17401

Please Log in or Create an account to join the conversation.

  • Posts: 314
  • Thank you received: 34
That sping is a good idea, nevertheless I had to exchange my payback coupons so I decided for a new servo.
Today I recieved the ds6020mg, normal sized an MONSTER power of 95Ncm@4.8V.

Another point for a new servo was the concern that the cap might not close completely
(in the meaning of an scope thats always outside in my telescope box, dust/pollen/wind).
I maybe want to put the first small servo on the opposite side of the cap to tighten the
cap somehow when the cap is in parked position because I doubt the sealing strength by
just one servo.
--= human, without Windows™ =--
pls excuse my bad english! :)
6 years 8 months ago #17493

Please Log in or Create an account to join the conversation.

  • Posts: 314
  • Thank you received: 34
I make some progress. The cap itself is finished so that the EL-Foil is now attached to it. It opens now fully thanks to the new stronger servo.

I found out the new servo motor sucks about 0.1A maximum additional power from my system when its opened but I get
a little headache on how to get the cap really good (light/dust) sealed when its closed. I will try to use some very soft black
ring cutted foam.


While trying to get the INDI WiringPi GPIO working for the EL-Foil relay I encountered a problem with the
INDI ServoBlaster Cap driver:

@Magnus_e
In the ServoBlaster Cap --> Calibrate Tab --> set flat light device Number I can only set a number from 0 to 9, but I would need WiringPi PIN 11 because I occupied allmost all other pins on my RPi2.
Can I simply change the line 124 in servoblaster_cap.cpp
from:
IUFillNumber(&LightSwitchN[0], "LIGHT_SWITCH", "set flat light device", "%6.2f", 0, 9, 1, 0);
to
IUFillNumber(&LightSwitchN[0], "LIGHT_SWITCH", "set flat light device", "%6.2f", 0, 11, 1, 0);
?
--= human, without Windows™ =--
pls excuse my bad english! :)
6 years 8 months ago #17559

Please Log in or Create an account to join the conversation.

  • Posts: 271
  • Thank you received: 72
Hi.
Nice to see you are making sutch good progress :)

The fix you suggest to access pin 11 should work if I remember correctly. The limit 0-9 was set so I did not nedd any special handeling if USBRelay2 Roof driver was used for lightswitch (it only supports 10 relays)

One more fix you can do while re compiling is to comment out these lines
if (ParkCapS[0].s == ISS_ON && !isMoving)
    {
        DEBUG(INDI::Logger::DBG_ERROR, "Cannot control light while cap is parked.");
        return false;
    }
from the function
bool ServoBlasterCap::EnableLightBox(bool enable)

I was not expecting someone to make such a nice cap with buildt in light, so I disabeled turning on the light when cap is parked. I will push a update in a few days, so you don't have to keep a custom version.

When it comes to keeping closed 'm going to try adding a neodyne magnet to the back of the lid, and mount one on the scope side that I can adjust how close it is. Mount it so that it has the strongest force that the servo opens. You can get these magnets cheap as simple fridge magnets.
Proud owner of Observatory 17b - A remote Linux observatory.
Website: Observatory 17b
Build thread @ SGL: Starting summers observatory project
The following user(s) said Thank You: pauledd
6 years 8 months ago #17568

Please Log in or Create an account to join the conversation.

  • Posts: 314
  • Thank you received: 34
The code change worked, I can now switch the flatfield in the ServoCap Tab on and off :).

Nice idea with that magnets and by sheer coincidence I left some on my fridge, I will try that.
--= human, without Windows™ =--
pls excuse my bad english! :)
6 years 8 months ago #17572

Please Log in or Create an account to join the conversation.

  • Posts: 314
  • Thank you received: 34
Getting closer... There was no way for me to get the cap really good closed without that magnet trick. Because of my sloppy drilling for the caps main hinge there is a lot of slackness. I had to make some kind of "guide clip" for the magnet so that they really meet together at their middle. Those tiny magnets have a lot of force but only if they are really close and centered.

Now the cap closes quite good and the motor has enough force to devide the two magnets :)



After proper rewiring everything I hopefully can do some real tests outside when the scope is back on the mount.
I fear a little bit the flatfield cables might interfere other cables because of its 1kHz@100V power. No idea how this might influence the servo pwm. I also consider a complete rewiring because with the servo and the flatfield I have 5 new cables, one more step down converter and a low/high converter for the flatfields relay (because its active-low). Its getting really messy in my box

I consider to etch a new pcb and put everything possible on it but only if outside test were successful.
--= human, without Windows™ =--
pls excuse my bad english! :)
6 years 8 months ago #17729

Please Log in or Create an account to join the conversation.

Time to create page: 2.449 seconds