Actuator --> add sensors to up the count - - Part 1

[gabshere]

ok in another thread we talked about adding magnets to make the actuator counts go up

so here i thought since i was going another way and all actuators aren't build the same. thought i would discuss adding sensors to the motor to make the count go up.

first this is from an old SuperJack II+ actuator that is broken so its my parts actuator ( motor is bad but i can use it to get the ideas and placement ) i have the twin actuator out on my 7 1/2 ft unimesh dish .

first the Superjack II+ has one sensor and a magnet wheel with a arrow/ triangle pointing to where its magnetized ( where it activates the sensor).

second i can't remember (but i will double check when i can) but i think this actuator has a 32 count per inch

layout of the sensors ( i first planned to add 3 additional sensors for a total of 4 X 32 counts = 128 but space being limited inside the area around the magnet wheel i probably can only get 3 . I also thought that 2 or 4 sensors would be equally spaced so going with 3 sensors i would probably have to evenly space the sensors so i would get an accurate count both ways.

old motor



magnet wheel



three sensors




thoughts and comments welcome , i don't know how long this project will take but basically walking through it now

 

[FaT Air]

My 2¢ on the subject: Don't think you can arbitrarily add some number of extra sensors. Your aim is to have one of the sensors 'triggered' equidistantly(angle) about the rotation.
Scenario 1 Doubling the count per rev.
On a wheel with 8 magnets spaced 45° around it, you'd want the second sensor spaced 22.5°(or the equivalent) from the original sensor to produce a pulse between the original sensors pulses. This would double the count to 16 counts/rev.
Then, to double that, 2 more sensors would need to be added, 11.25° away from the 2 existing sensors.
Each additional increase of sensors must double what's there, taking into account the angular position of the magnets.
Scenario 2 Triple the count per rev
Same wheel, magnets spaced 45°. Add two sensors (45/3) 15° from the original sensor(or it's equivalent) Here one of the sensors is 'triggered' every 15° of rotation. #1(original) at 0°, #2 at 15°, #3 at 30° and at 45° #1 again.
I can't see where the poles are on that particular magnet so some investigative work would be in order.

 

[gabshere]

yes i can see your point ... and thats what was being discussed in the other thread where magnets were added from 4 to 8 or 10 to a plastic / nylon wheel
this particular actuator has a wheel that the whole thing is a magnet but only has one pole marked with a arrow / triangle i thought this was the area where the magnet was the strongest and activated the sensor.

 

[FaT Air]

It's been a while since I've played with them, but think there are more than one place about those, that will activate the sensor switch.
But if it is only one active region, should be able to place as many positions around the thing as you like. Could confirm it with small lights or LED's. See that only one switch is on at a time as the magnet rotates.

 

[Magic Static]

Not to rain on your parade, but increasing the counts doesn't make the motor able to stop any better than it did before. Can it perform at the increased accuracy rate? I think a better way to improve the performance of dish tracking is to change the geometry. Make the lever longer and put a longer actuator on the dish. Improved power and accuracy.

 

[Pixl]

My dish has about 6 counts per degree of dish rotation in the middle of it's travel. The dish is 2 degrees wide, so a 1 deg pointing accuracy exceeds the specs of the dish. Now you take one of those degrees and chop it up into 6 pieces and you have an almost unnecessary pointing accuracy. Doubling that again with your mod to 12 pieces of one degree will do nothing.
Your efforts would be better spent as Magic suggests improving the mechanics of the assembly. Also try to get the slop out of mount and attachment points. This will give you repeatable positioning no mater what the pulse count is.

 

[Lak7]

second i can't remember (but i will double check when i can) but i think this actuator has a 32 count per inch

About how many counts between 2 Degree Sats are you getting?
If you are only using Dish for C Band, a stock Arm should be good enough. However, being a 7-1/2 foot Dish, it may not be 2 Degree compliant.
On my 10 Foot Dish, I average 35 counts per 2 degrees, plenty for Ku - it's all in the alignment, and a mount that is mechanically sound - no loose or worn parts.

 

[gabshere]

yeah its a tight dish its setup and everything is aligned (mount that is mechanically sound - no loose or worn parts --> almost new ) so no give there
the actuator has no slop its actually a newer ( old actuator that has little use on it) 18 inch SuperJack II+

it provides only 10 counts per 2 degrees ( on the V-box III ) and the V-box III might be the problem , i just don't know
I know it can get KU but currently i have a c band only lnbf on it ( removed the c/ku lnbf) the reason being i could tune it for ku on one sat, but the low count prevented me from fine tuning it enough to gain KU on other sats across the arc. C-band tuned in Fine across the arc..... Except C-band S2

so i was thinking of adding more counts to increase fine tuning to gain the ability to tune Ku or C-band S2

I will have to go back out and play with it a bit more but like I said when i started my time is limited so hence the Part 1

 

[shopman]

I've taken a much different approach to the problem and have had great success. While watching my dish moving one afternoon it occurred to me that it was moving too fast and was overshooting the stopping point. It really wasn't an issue of actuator counts, it was speed and momentum that were my enemy. So, I popped open the Gbox and inserted a variac between the output of the transformer winding that drives the DC supply for the actuator and it's associated rectifier. After lowering the voltage with the variac I've had outstanding accuracy and haven't had to re-align the system in over a year.

Actuators are available with different voltage requirements but as far as I know all the Gbox/Vbox units output 36vdc. Some people may have 24vdc actuators being driven by 36vdc movers(which is what I suspect was going on with my system) and the drop in voltage can make all the difference in the world.

 

[gabshere]

ok i checked i get 32 clicks per inch
i can go from 131W to 70w takes 283 clicks
2 degrees takes about 10 clicks
didn't get enough time to complete the sensor experiment so next time when i have more time i will try it .

 

[Magic Static]

It will interesting to hear of your results. I understand your need for more accuracy. I'm trying to improve on 20 counts for 2 degrees I get with a PMA24 & VBox X on my 10' Laser C/Ku

 

[FaT Air]

If overshooting is a problem, check the mover. It should be placing a short across the motor leads(when not moving the dish). That load will stop a motor quick! If it's 'spinning down' to produce an overshoot, it's likely the short isn't there.