Jump to content

Banner.jpg.b89429c566825f6ab32bcafbada449c9.jpg

Fitting a rotary shaft encoder on an eq6 any ideas?

ncjunk

By ncjunk
in DIY Astronomer

Recommended Posts

ncjunk

ncjunk

Posted
Posted

I fancy fitting a rotary shaft encoder on an eq6 and then using something like an arduino to send updates to the mount. With fast enough sampling and correction this would help pec no end.

Any comments welcome...i have no idea on encoder prices so it may shock me this one! (i have no clue on how to fit it either..blutak??)

Software side would be easy...surely...sample...correct...sample...correct. In fact i am sure i can sort out the software side. I am off to search for encoders...

Link to comment
Share on other sites
ncjunk

ncjunk

Posted
  • Author
Posted
You'd need a very high resolution encoder. This would be very expensive and not a viable option IMO. Cheaper to add another scope and camera as a guide system.

Ahhh cant win dont try! Where's the fun in that? (i have 2 guide scopes and a oag bored of that)

I know they are likely to be expensive but how much? and is there a resolution/price point at which it would still give you a good resolution? Eq6 is 902000 micro steps per revolution...

I'm putting on my investigating hat...

...and George:-

1. Damn you i wish i'd just been to the pub.

2. Of course im going to get an eq8....if i find work...this is just for laughs and giggles! along with the astronomy/physics degree i want to do in spanish (1/3 the price here in spain but still looks..interesting)

Link to comment
Share on other sites
Gina

Gina

Posted
Posted
Ahhh cant win dont try! Where's the fun in that? (i have 2 guide scopes and a oag bored of that)

I know they are likely to be expensive but how much? and is there a resolution/price point at which it would still give you a good resolution? Eq6 is 902000 micro steps per revolution...

I'm putting on my investigating hat...

Sorry to pour cold water on your idea, didn't mean to discourage you from trying things. It depends on what you want to do with the data. You mentioned PEC and improving the tracking. That implies improving on the resolution/precision over that of the mount itself.

Let me explain the problem... The encoders I know about resolve to about one degree ie. 1 in 360. You may get 10 times that but to approach the resolution of the EQ6 we're talking about 300 times again - one in a million. Now I believe very high res encoders (also called resolvers) are obtainable but you'll be looking at thousands of pounds. Just think of the accuracy required to approach anyway near the precision required - we're talking about micro millimetres.

The cheapest way of getting high resolution is to use a gear train and count impulses (with digital drives) of the input - as used in the mount. The downside (as you know I think) is that any errors in the gear production result in angular errors at the output (PEC). Higher precision in the manufacture of the gearing will give smaller errors and I imaging mount manufacturers strive for this while balancing it against cost. Professional mounts used in the big observatories have much better precision but cost many times as much as the ones we amateurs use.

Do your research - see what might be achieved at the various pricing points. I have given you the limited knowledge I have on the subject. Good luck :) I shall be interested in your findings. Please keep us informed.

Again, I apologise for being discouraging - I just gave my opinion based on what I've read etc.

Link to comment
Share on other sites
George

George

Posted
Posted

Now I'm sober.......:)

Don't know anything about encoders but I do know that a lot of the inaccuracies of the EQ6 are down to the transfer wheel that drives the worm. If you dig around in the EQ6 Yahoo group there's quite a few people converted the transfer gear to belt drive and its fairly cheap todo and EQmod supports the new gear ratio.

PEC is vastly improved and it quietens the mount down too.

I did a costing and got a list of suppliers with part numbers but didn't get round to doing the mod myself but its on my list of things todo :(

If I can find the info I accumulated I 'll post on here.

Link to comment
Share on other sites
JamesF

JamesF

Posted
Posted

If you do go ahead you may need to do some research on the sampling side, too. I have vague memories from the Robotics lectures on my Computer Science degree (it *was* more than twenty years ago -- it's allowed to be vague :) that sampling and how you do it isn't quite as straightforward as might first appear.

James

Link to comment
Share on other sites
JamesF

JamesF

Posted
Posted
Don't know anything about encoders but I do know that a lot of the inaccuracies of the EQ6 are down to the transfer wheel that drives the worm. If you dig around in the EQ6 Yahoo group there's quite a few people converted the transfer gear to belt drive and its fairly cheap todo and EQmod supports the new gear ratio.

I wasn't aware of that. Now you've mentioned it I shall go and look it up.

James

Link to comment
Share on other sites
ncjunk

ncjunk

Posted
  • Author
Posted

No appology needed! You are correct in what you say but there are more accurate encoders than 1 in 360 which, as you say, is not good enough.

You are not discouraging me you are just pointing out, effectively, a list of items that need to be taken into accout with the project. (but thats the whole fun of it! No point doing it if its easy...i would buy the commercial one if i wanted that)

There is a commercially available one for 1,500 so the encoder is not thousands of pounds. There are 500 ppr encoders for 120 so i would expect to find one for around 200 to 500 pounds but i would be after 10,000 ppr.

The method is to correct for all those mechanical errors on the fly and this will reduce the pe seen at the telescope. I need to look at the sidereal velocity and the minimum/maximum possible error change and determine what resolution of encoder is required to stay within 1arc second say.

The eq6 has a peak to peak error of around 30 arcseconds over the whole cycle...but i only need to keep nudging corrections to keep that down.

Therefore the encoder doesnt have to match the same resolution as the eq6....although in a perfect world it would be nice to measure at twice the resolution of the drive system. (overkill but think of the accuracy!)

Then there will be the question of the software and how to apply a correction if your last measurement was x out. Any change will then affect the next reading so it will have to be a rolling average so that an error and correction over ten samples, say, will even out to give you current sidereal rate plus any residual error.

Of course the effectiveness will be limited to the step in the cycle with maximum mecanical error multiplied by the response of my corrections to the error but that will still be significantly better than an uncorrected system.

Anyhow this is all after finishing the cloud sensor or George will kill me!

Link to comment
Share on other sites
ncjunk

ncjunk

Posted
  • Author
Posted

Should have updated my browser before replying...

Thanks for the responces..

Peter...i sat down then couldn't find the price so stood up again. I'll have a look from the deskop later but knowing you i'll probably be better having a few shots to ease the shock as well!

James- sample rate is twice

that of encoder pulse rate to ensure you dont miss a pulse. If i grab an old plc from the next job i work on at work then these sometimes have pulse counters on the inputs that are independant of the plc scan cycle as they need to work at higher sample rates. You can also get digital inputs that run of interrupts but you have to be carefull with those...dont want it continually interupting. These can also provide a current frequency value in some cases.

Link to comment
Share on other sites
George

George

Posted
Posted

Anyhow this is all after finishing the cloud sensor or George will kill me!

Talking of which I ordered another sensor yesterday from somewhere else rather than wait 2 weeks for RS to get them back in....Hopefully should be here today, had to bite the bullet on price tho :(

Might have a play with the LDR while I'm waiting :)

Link to comment
Share on other sites
ncjunk

ncjunk

Posted
  • Author
Posted
Hi

These encoders have a physical resolution of 5000 tics but somehow turn that into a resolution of over 10 million :)

Harry

Now thats interesting...i could see how you could have slots and detect a value for light level through the slot to give an analogue indication of position rather than digit...or sets of slots? (but then i'm thinking large slots here! Not 5000 per 360 deg ;-). )

Link to comment
Share on other sites
ncjunk

ncjunk

Posted
  • Author
Posted
Talking of which I ordered another sensor yesterday from somewhere else rather than wait 2 weeks for RS to get them back in....Hopefully should be here today, had to bite the bullet on price tho :(

Might have a play with the LDR while I'm waiting :)

Get a shift on George! Which input? What value range on the input? What resisters are you using? Do you want a graph with that? What are the lottery numbers for tonight?

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.

Go to topic listing

  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...

Important Information

We have placed cookies on your device to help make this website better. You can adjust your cookie settings, otherwise we'll assume you're okay to continue. By using this site, you agree to our Terms of Use.