Birdview reed sensor kit

[bryan.lynch]

I recently acquired a solid Birdview with the H-H mount and the huge pole that it mounts on. The motor has the old potentiometer sensor that does not work with anything other than the Birdview motor controller. I contacted Skyvision to see about purchasing the upgrade kit, but I was informed that the gentleman that makes them is in poor health and they don't know if he will be making any more.

I'd like to rig something up, I have some elementary electronics knowledge, and I found a place that sells some magnetic target wheels. I could rig up a hall effect sensor to read the wheel and send a pulse to the receiver. There is a company by the name of Phoenix America that makes the wheels and various sensors.

What I was wondering is if someone could post some close up pictures of the setup that was available at Skyvision, so I can use it to see if I can come up with something that will work. I will certainly share what I find with the forum and hopefully I can come up with something that will let me use this dish and motor.

Thanks,
Bryan

 

[Anole]

sounds attractive

If you'll check back through this section of the forum, our member Linuxman installed an upgrade kit like that.
He posted some pictures at the time.
Just search on posts by him, and also the parts in question.
Otherwise, he's written extensively about birdviews, so you will get lots of hits.

I have wanted to see something like you purpose, but haven't gotten around to it.
Hall effect sensors could be dug out of 12v fan motors, as a for-instance.

 

[linuxman]

I don't think I ever actually posted a picture of the Skyvision Kit. I do have a picture that came with it, but don't have a scanner to scan it in.

It was only a solid magnet wheel that fit on the main shaft between the motor and the first support arm for the worm-gear. It actually slides on after you take the motor loose and pull it out of the main shaft where the pin is stuck through the U shaped joint.

The other part was just a bracket that bolted onto one of the motor bolts and has an arm that extends to the magnet wheel. There are 8 invisible magnets embedded into the wheel.

I said I wouldn't buy another one because it would be so easy to make one.

The next one I use will be mounted on the other end of the worm gear where the Pot was. It will be really simple to make a piece of plastic with slots in it and attach it to the end gear and use a light sensing switch.

If you take the Pot off, and just leave the little gear, that gear turns at the same speed as the main shaft. You would have the same setup as the kit does and using the same rotation speed.

See the pic:



It should be real easy to make it work.

 

[bryan.lynch]

Linuxman, thanks for your reply. I do think that whatever I use will be put on the end of the worm gear where the old pinion is for the potentiometer based setup. I like the idea of the slots in the large gear, but I'm not sure it would turn fast enough (since it's geared down) to give enough pulses to be accurate. However, the advantage of putting the sensor on that side would eliminate the problem of slack between the motor and worm gear.

I think I just may order one of the magnet wheels and hall-effect sensors from Phoenix America and see what I can come up with. I'd have to find a source of +5v in the VBox or GBox to power the sensor though. Might be more difficult, but I think the hall-effect sensor would be more reliable than the reed switch in the long run.

--EDIT--

I realize after re-reading your post that you are in fact talking about the small gear, not the large one, so disregard my above about the geared-down effect. The company I found makes a 2" diameter wheel with a set screw and a 1/4" diameter center hole. I think it might fit over the small pinion gear after removing the sprocket part by filing or cutting it down.

The other cool thing is they make versions of this wheel with values from 2 up to 120 poles. I'm thinking the 30 pole might work the best.

 

[linuxman]

Hey There again!

I found my picture and took pictures of the pic and the instructions that came with my kit.

I was talking about taking the pot and big gear off and attaching something to the little gear. It turns at the same speed as the shaft.

Here's the pics and you will see:



As you look at the magnet wheel, it is about 2" across. I will have to get accurate measurements, but it rides on the main shaft and turns at that speed. If put on the other end of the shaft where the little plastic gear is now, it would do the same thing.

Hope it helps.

PS BTW, Mr. Anole, you can start making mine before I get the two Birdviews from Nashville.

 

[bryan.lynch]

I've attached a couple of pictures of the magnet wheel that I am talking about purchasing. It's silly how similar it looks to the one that was in your kit. I noticed in your instructions it said that it was a 24 pole magnet, so I think the 30 pole will in fact work. Thanks for sending the instructions, they really help!!

 

[Anole]

I think I found your 30-pole part
Wow, $28 each -
That's the lowest rated magnets.
edit: actually 8-pole (like Linuxman has), 10-pole, 30-pole, and 60-pole all cost the same.

Well, sometimes it's just best to buy a solution and move on . . .
I just didn't know it'd cost so much.

Linuxman's optical idea is beginning to sound more interesting.
Of course the magnetic solution works through rain, snow, mud, bugs, spider webs, and WW3 . . . - :up


edit: Oh, and one other thing to consider...
If I recall from reading Linuxman's previous posts, those Birdview gearboxes come in two ratios.
One was something like 36:1, and the other around 72:1 (don't quote me)
You might want to read the label on your gearbox and think about how the number of poles relate, before diving off the deep end.

ooh, one more edit: the magnetic disc, without the metal center hub and setscrew, are about $11
Wonder what you could "bond" that to (glue it) ?

 

[bryan.lynch]

My gearbox is the 36:1 ratio, but I never could figure out why the original kit specified that it would only work with the 70:1....

The large worm gear is where the magnet wheel is, and that turns the same number of times from horizon to horizon regardless of how many turns the motor and associated gears make. I'm just gonna run my mount from one end to the other and count how many times the worm gear shaft rotates. I'm thinking that I will end up using a 30 or 60 pole magnet.

I'm definitely going to think about using one of the cheaper ones, along with some J.B. Weld or other type of epoxy. I don't think the set screw is worth the difference in price.

I did order the reed sensor, from Digi-Key (I decided against the hall-effect, too expensive). It's made by a company called Hamlin, and the part number is 59065. I've attached a picture of it. It's only $3 and change. The stupid shipping costs more than the sensor does.

 

[bryan.lynch]

Alright, I ran the mount from one end to the other (about 160 degrees) and determined that the main shaft makes 52 revolutions. If I use the 60 pole magnet that will give me 3120 pulses from one end to the other. I think that will give me enough resolution to position accurately. That's about 20 pulses per degree, so what do you guys think?

 

[Anole]

random notes

I haven't run the tests, but just seat-of-the-pants engineering suggests that you can't get a reed switch to respond to a 30 pole magnet.
By all means, do prove me wrong. -

The hall-effect sensor will handle 30 magnets, but may be touchy on how close it has to run to the disc.
I haven't found sensors for sale on your web site yet, but did look them up on the Honeywell / MicroSwitch site a few months back.
Here is a 126 page Honeywell PDF that'll tell you more than you'll ever want to know on the subject.

What I'd been wanting to work on with Linux man, was something based on just a few magnets, but a multi-phase hall sensor to generate a lot of pulses per revolution.
If I'd known at the time, that 8, 10, and 30-pole magnet discs were readily available, I wouldn't have bothered.!

I mentioned earlier where I was going to get the sensor. . .
Every dead PC fan motor has one of the little parts in it.
They are about 1/8th inch square, 1/16th thick, and have four leads.
I think Allegro Micro makes them, but I don't have my notes in front of me.
Oh, maybe something like this, but I have not confirmed.

Not having a Birdview, I haven't had the incentive to follow up 100% on some of these ideas.
(hint) - :up


edit: regarding your 60-pole magnet, let me suggest you read this thread by Linuxman, for more info on what a reasonable number of counts would be.
Without data supporting that 60-pole magnet, I might favor the 30.
But getting the sensor to work with either will be the challenge.

 

[linuxman]

But getting the sensor to work with either will be the challenge.

Anole is correct about getting the magnets to work properly.

The one on my Birdview has a business card clearance only.

That's not a lot of room for error.

 

[Anole]

curious question

I think we need a resolution to this question.

Linuxman has stated his magnet wheel has 8 poles.
The paperwork that comes with it says 24 poles.

There is a trick with magnets where you take three or four to make one very strong pole with all the flux to one side (direction).
Someone really needs to verify how many pulses the existing Birdview sensor gives per revolution of the magnet wheel.

The same needs to be done with the 30-pole magnet disc.
If it just gives 10 pulses, we need to know.


edit: SECOND QUESTION
To increase the operating range, would one of the stronger magnetic discs do the job better for a few dollars more?

 

[linuxman]

I don't remember where I got the number 8 from. May have been another thread etc.

But bryan.lynch's numbers look right to me. Remember in the thread about count accuracy, I said my Birdview had 1300 something. Well 24 x 52 = 1248 and by using the GBox, I get twice that number.

So Bryans extimation of a 30 pole magnet is not far off. I think something with 30 to 40 counts per revolution should do the job pretty well.

Just my opinion.

 

[bryan.lynch]

I think I may have found another solution - incremental rotary encoders. I pulled the small nylon pinion gear from the end of the main shaft on the Birdview motor last night and found that the diameter of the shaft is roughly 1/4 inch.

A company by the name of CUI makes a modular rotary encoder that will adjust to different shaft sizes and mounting arrangements. Better yet, it is configurable by DIP switch from 48 - 2048 pulses per revolution. I think with a +5v line from the Vbox and a transistor and a few other electronic components I could come up with a pulsed output that would give 48 pulses per revolution.

Digi-Key has it for $29.95 in a single quantity, so it's not the cheapest solution. However, it doesn't require any fabrication on my part. I think I'll be ordering one soon.

http://www.amtencoder.com/

 

[linuxman]

That might be as you say a workable solution. It is certainly a lot less expensive than Skyvisions kit. If you can make it work without a lot of electronic accessories.

Let's see what our friend Anole comes up with too.

 

[gabshere]

with this modular rotary encoder you could set the pulse per revolution higher and since i only use 110 degree of the arc could set it to have more counts per revolution and be accurate correct ? this does sound interesting. i have a couple of birdviews lined up to inquire about removal and will be changing things over when i aquire them. but this thread is very interesting

 

[Anole]

not optical

For all the technical info given in the docs on that AMTencoder site, there sure is also a lot missing!
I particularly wanted to know about dust/moisture resistance.
It says it's capacitive sensing, but that the index pulse might be screwed over by magnetic leakage of an adjacent stepper motor.
That just makes no sense at all !
I was curious about power consumption (I have a clever trick if it's low enough).

Here's something not necessarily apparent 'till ya look closely, too:
The A or B outputs are "in quadrature" at 1/4th the rate of the specified PPR (Pulses Per Revolution).
This is actually a good thing 'cause now the too-high pulse rate becomes:

[FONT=Fixedsys]...
rated  actual
-----  ------
 2048 = 512
 1000 = 250
  800 = 200
  384 =  96

 1024 = 256
  500 = 125
  400 = 100
  192 =  48

  512 = 128
  250 =  62.5
  200 =  50
   96 =  24

  256 =  64
  125 =  31.25
  100 =  25
   48 =  12[/FONT]

Obviously, the lower numbers cover a very attractive range:
12, 24, 25, 31, 48, 50, 62, and 64 PPR.

edit: The outputs are square waves, so the Gbox 3000 would count twice (once on the rise and once on the fall) of each pulse.
That's not necessarily a good thing, so maybe you'd want to try this with a Vbox 2, if those are still around.
(or use the four very lowest rates with a Gbox: 12=24, 24=48, 25=50, & 31=62)

What was the cost of the official Skyvision replacement kit?
One reason it might have been specified for the higher gear ratio motors, is cause they would probably move slower.
Linuxman: have you ever had any trouble with your kit?
If I recall, it's hall-effect based, and not a reed switch?

 

[linuxman]

What was the cost of the official Skyvision replacement kit?

I just found the invoice, it was $69.95 and with S/H total was $77.95. Pretty expensive for a switch, bracket and magnet wheel.

One reason it might have been specified for the higher gear ratio motors, is cause they would probably move slower.

That would be my speculation too.

Linuxman: have you ever had any trouble with your kit?

Haven't had to touch it since it was installed. Trouble-free!

If I recall, it's hall-effect based, and not a reed switch?

I think it is a simple reed switch. My electronics neighbor commented on how much they were charging for $20.00 worth of parts.

Edit: I just remembered something else from a comment you made earlier.

In the count accuracy thread, I tried to use a magnet I got from a local craft store for the 2nd magnet in the Von Weise LC actuator and it didn't work because it was too strong.

The ones I got from Von Weise were very low strength magnets. My speculation is that they hold on too long to get the pulses in.

On the other hand, the magnet wheel on the Birdview is very strong. I dropped a washer through a top crack in the mount, and it rolled near the magnet and jumped on it. Had to take the housing back off to get it loose.

Second edit:
I may just have my neighbor make me one of the light sensing switches and give it a try.

 

[bryan.lynch]

Does anyone have a definitive answer on how many counts the GBOX will do from one end to the other? I'd hate to overdo it on the counts and outrun the limits of it.

 

[linuxman]

Does anyone have a definitive answer on how many counts the GBOX will do from one end to the other? I'd hate to overdo it on the counts and outrun the limits of it.

I am pretty sure it will handle 4000 counts total.

It might do more, but I just don't know for certain.