USALS for BUDs:

[Sadoun]

Wow, Anole, you've got a great thread going on here.

I like the discussion about trying to use USALS on a BUD and how it would or wouldn't work. It ma not work on an actuator motor, but it may work on a HH Motor, I am not sure. Certainly your input here is most beneficial for any future project.

I may add a few ideas that can be incorporated into a future GBox:

My ultimate GBox will have, in addition to the current features:

• Polarator control
• USALS
• USB connection to connect to a PC/Mac to allow the user to edit/store satellite data (name, location, etc) from/to the GBox.
• Switchable Low/Medium/High power (amps) to accommodate different size motors/actuators/dishes.
• High/Low speed auto setting. This will cut on travel time.

I'm sure you will think up more features, and comment on how to make all the ideas work well together.

 

[Anole]

I'll have to admit to having more fun and spending more time helping newbies and discussing problems and projects here,...
... than actually -watching- anything on TV.
But I digress.

It's worth pointing out that a position sensor directly coupled to the aiming angle of the dish, is far more valuable than one upstream in the gears.
"Slop" is a common term for backlash in a mechanical system.
So, it's possible that even with a sloppy gear train, you could bring the dish to the desired angle.
Keeping it there might be a problem, because the traditional controller philosophy doesn't take into consideration wiggle or wind-blown slop that might displace the dish.
We are not in a fully closed-loop system.
We're open loop, and have to make the best of our situation.

It has been wisely pointed out, that there might be enough play in the system to locate the dish differently when east of true south, as opposed when it's west of true south.
First thought might be to allow a compensation if this is the only source of positioning error.
However, winds and other factors might play a part in the actual aim of the dish, and having a very mechanically tight (not just stiff) gear train is essential.

When slop is more than one click , then you have a pretty poor system.
But, as you reduce slop, and increase the clicks or counts per degree, the system improves.

Wow, Anole, you've got a great thread going on here.

Yes, the interest and technical feedback are first rate!
Hope we come up with something that you might want to bring to market! - :up

My ultimate GBox will have, in addition to the current features:

• Polarator control
• USALS
• USB connection to connect to a PC/Mac to allow the user to edit/store satellite data (name, location, etc) from/to the GBox.
• Switchable Low/Medium/High power (amps) to accommodate different size motors/actuators/dishes.
• High/Low speed auto setting. This will cut on travel time.

I'm sure you will think up more features, and comment on how to make all the ideas work well together.

Thanks so much for your input.
The idea of integrating all these ideas into a Gbox would certainly be better than a single-function design!

Polarotor control is being well discussed in other threads, and I don't want to rehash that subject.
If we come up with any requirements or features which would improve it, I'm sure we'll get the word out!

USB to up/download the diseqc 1.2 positions would be a real time saver, I suspect.
For backup if nothing else!
And if the format the file was simple text-like, then users might edit it with Notepad or some other simple tool.
I can ever see advancements where users program spread sheets to output data to be fed to their positioners!
So, great idea, you have there. - :up

Changing motor speed is something I've been thinking about.
One way to make the motor run slower without losing torque, is to pulse width modulate the 36 volts.
I ran an experiment on a buddy's model train some years ago.
We got his engine to slow down to such a crawl, that it was all but impossible to see it move.
We had to look at a dust mote on one wheel, under high magnification, to be able to recognize movement.
Now, as to how a slow-move might be applied to a dish/motor which had an advanced controller, I do not immediately see.
Could you start the move at a slow speed, then accelerate for mid-move, then slow as you came to final position?
Come to think of it, Stogie once commented that his 12' Paraclipse seemed to shake or oscillate a bit when coming to a stop.
A slow-speed-start, slow-speed-stop feature would certainly put an end to that problem.
So, would that be useful? If so, it's easy enough to do.

 

[linuxman]

I have done my best to avoid commenting in this thread, but after reading through a couple of times, I thought I would offer some thoughts.

I am not the engineering type that some of you are, and don't have a great grasp of the math involved, nor the electronic circuitry, but do have some practical experience with USALS on Ku dishes and BUDs with H-H mounts and also Linear Actuators.

First of all, USALS is a wonderful mechanism for moving a dish. Just set your longitude and latitude and you are done. The problem arises if your dish is not perfectly aligned on the arc.

People here on the forum wonder why it sometimes takes me days to tune in a dish. The reason is that I won't settle for a dish that moves from say 104W to 123W and have good signal on 123W, only to move to 129W, and back to 123W and have poor signal.

That is not the fault of the USALS system, nor slop in the motor. It is the fault of an imperfectly aligned dish. Keep in mind this is with very flat .6 F/D ratio dishes that can tolerate being off the satellite by a half degree and still have good signal.

I have been through the scenario above and thought there was something wrong with the motor etc, only to solve the problem by fine tuning the dish on the arc. When I finish tuning a motorized Ku dish on the arc, I get the same signal on every satellite, every time I move it from 43W to 148W, back and forth, starting and stopping every time.

Apply that principle to BUDS.

I have two Birdviews using diseqc 1.2 with H-H mounts. I did put a bolt through the shaft to eliminate the only slop in the mechanism. Once these are aligned, they act very much the same way as the USALS motor does. I can move from one end of the arc to the other starting and stopping in between with the very same signal on every satellite know matter from which direction I arrived at the satellite. It doesn't matter whether Ku signal or C-Band.

I will soon know if the AJAK H-H mount is as accurate. I have had one of these for a couple of years, but seldom play with it because it operates my family's setup.

The dishes with Linear Actuators is another story. These type mounts have more mechanical parts and are subject to more "slop". Plus by their very nature, the counts between degrees changes dramatically once you move past the center of the arc. I am sure you engineers are working on that problem with your movement detectors at the pivot point on the dish rather than pulses taken from the motor.

Even with the problems stated above concerning linear actuators, with a properly aligned dish, one can expect to move about from satellite to satellite from one end of the arc to the other and have the same signal quality no matter where you land, no matter what the speed and no matter where you came from.

The point I am making is this. For USALS to work, it is imperative that the dish be perfectly aligned on the arc to begin with, and to have as much of a "slop" free mechanism as possible driving the dish, and finally have a highly accurate movement detector for letting the GBox or whatever box know where you are and where you are going.

EDIT: If you find that you have to "bump" your dish to get the best Ku signal on a satellite on a BUD, there are only two reasons possible.

1. Too much slop in the mechanism.
2. An improperly aligned dish.

Just my opinion!

 

[Cadsulfide]

All modern motion control systems use some sort of trapizoidal velocity profile when making a point to point move. IE: smootlhy ramp motor speed up from a dead stop to full speed and ramp it down as it approaches the setpoint. Much gentler on the hardware than the Bang-Bang control curently used with linear actuators on sat dishes.

The point may be moot, as the circuit complexity goes up, the price point of the product shifts up also and sales drop.

 

[B.J.]

I have done my best to avoid commenting in this thread, but after reading through a couple of times, I thought I would offer some thoughts.

I am not the engineering type that some of you are, and don't have a great grasp of the math involved, nor the electronic circuitry, but do have some practical experience with USALS on Ku dishes and BUDs with H-H mounts and also Linear Actuators.

First of all, USALS is a wonderful mechanism for moving a dish. Just set your longitude and latitude and you are done. The problem arises if your dish is not perfectly aligned on the arc.

People here on the forum wonder why it sometimes takes me days to tune in a dish. The reason is that I won't settle for a dish that moves from say 104W to 123W and have good signal on 123W, only to move to 129W, and back to 123W and have poor signal.

That is not the fault of the USALS system, nor slop in the motor. It is the fault of an imperfectly aligned dish. Keep in mind this is with very flat .6 F/D ratio dishes that can tolerate being off the satellite by a half degree and still have good signal.

I have been through the scenario above and thought there was something wrong with the motor etc, only to solve the problem by fine tuning the dish on the arc. When I finish tuning a motorized Ku dish on the arc, I get the same signal on every satellite, every time I move it from 43W to 148W, back and forth, starting and stopping every time.

Apply that principle to BUDS.

I have two Birdviews using diseqc 1.2 with H-H mounts. I did put a bolt through the shaft to eliminate the only slop in the mechanism. Once these are aligned, they act very much the same way as the USALS motor does. I can move from one end of the arc to the other starting and stopping in between with the very same signal on every satellite know matter from which direction I arrived at the satellite. It doesn't matter whether Ku signal or C-Band.

I will soon know if the AJAK H-H mount is as accurate. I have had one of these for a couple of years, but seldom play with it because it operates my family's setup.

The dishes with Linear Actuators is another story. These type mounts have more mechanical parts and are subject to more "slop". Plus by their very nature, the counts between degrees changes dramatically once you move past the center of the arc. I am sure you engineers are working on that problem with your movement detectors at the pivot point on the dish rather than pulses taken from the motor.

Even with the problems stated above concerning linear actuators, with a properly aligned dish, one can expect to move about from satellite to satellite from one end of the arc to the other and have the same signal quality no matter where you land, no matter what the speed and no matter where you came from.

The point I am making is this. For USALS to work, it is imperative that the dish be perfectly aligned on the arc to begin with, and to have as much of a "slop" free mechanism as possible driving the dish, and finally have a highly accurate movement detector for letting the GBox or whatever box know where you are and where you are going.

EDIT: If you find that you have to "bump" your dish to get the best Ku signal on a satellite on a BUD, there are only two reasons possible.

1. Too much slop in the mechanism.
2. An improperly aligned dish.

Just my opinion!

I agree with most of what you said above, however I don't agree with the last EDIT part, in that I think you should have included a couple other possibilities. For one, lost counts should be one of the most common reasons for needing to bump the dish one way or another. You can usually identify lost or extra counts by all sats being off, and it's fixed by re-syncing.
Another issue that perhaps could be included in the improperly aligned part though, is not having the focal length adjusted properly, because if it is off, the ku signal will appear to be a donut shape, and show up at two places off center, and weaker at the center point where you see C-band. This could also play havoc with USALS since it's assuming the center point.

I'm curious about your comment about the bolt eliminating the slop in your system. Where did you put this bolt? I've been wondering if there was a way to snug up the worm gear to the big semi-circle gear in my H-H, and was wondering if maybe you were taking the slop out by pushing gears closer together, or if you were just talking about some other part of the mount?

But slop wise, none of the mounts I've had were free of slop. All of them have given me different landing points if I come into the sat from the west vs coming from the east, particularly if I cross the south point.

My repeatability has been getting worse and worse, so that I'm not sure what of the many possibilities is the cause. I think I'm going to put a scope on the counter signal to see if it looks like a dirty signal, as it's possible that my Drake is starting to lose it's power supply. I've noticed that the LNB voltage on Ku is less than it should be. I may end up switching to my spare IRD, which is a Monty 50/55, just to see if it works better.

Also, when the snow goes away, I think I'm going to see if I can somehow take that mount apart, without taking the whole dish down, but I don't think that's possible. But at least maybe I can at least lubricate the gears well, and see if there is ANY possible way of snugging them up tighter to take out the slop. But I'm thinking that I really need to replace bushings that will require me to take the whole dish down, and I really don't want to have to do that.

 

[Sadoun]

Could you start the move at a slow speed, then accelerate for mid-move, then slow as you came to final position?

Yes, that is exactly what I had in mind. The speed of the motor will have a bell curve.

 

[dunnsept]

hmmm.. I'm not an engineer (never played one on TV either).
I wonder if instead of some sort of encoder, you could use something like a compass module?
put the dish to true south sat, "mark it" and then from that point maybe a microcontroller could calc sat positions based on USALS.
maybe I'm way off here,, no clue.. I've only dabbled in micros.. robots, coffee roasters etc. and I've just started with GPS modules and the like.

but one thing this thread has inspired me to do is hook my o-scope up to the 4dtv and watch the 'counts' come back in ;-)

 

[Anole]

example of slop:

Some time back before I had a Birdview, I asked PhlatWound for some pictures of the motor, deckplate, and magnet wheel.
This is from his excellent collection.

To the right, the silver thing is the motor and first gearbox.
The shaft comes in front of the ruler, right to left to the coupler, and is pinned in place (see the small hole?).
The coupler has a C-shaped notch that engages a big pin on the worm gear shaft at the left of the picture.
All that slop at the C-to-big-pin junction is unacceptable.
(presumably, there is no slop at the small roll-pin)

 

[linuxman]

I agree with most of what you said above, however I don't agree with the last EDIT part, in that I think you should have included a couple other possibilities. For one, lost counts should be one of the most common reasons for needing to bump the dish one way or another. You can usually identify lost or extra counts by all sats being off, and it's fixed by re-syncing.
Another issue that perhaps could be included in the improperly aligned part though, is not having the focal length adjusted properly, because if it is off, the ku signal will appear to be a donut shape, and show up at two places off center, and weaker at the center point where you see C-band. This could also play havoc with USALS since it's assuming the center point.

Perhaps I should have used the term imperfectly "tuned" dish.

Improperly aligned/tuned dish could mean anything from the feed-horn to the elevation bar needs to be "tuned".

As far as lost counts, I have never experienced any lost counts. I have had the actuator move because I didn't have the clamp tight enough which caused me to have to re-sync, but never any actual lost counts.

I'm curious about your comment about the bolt eliminating the slop in your system. Where did you put this bolt? I've been wondering if there was a way to snug up the worm gear to the big semi-circle gear in my H-H, and was wondering if maybe you were taking the slop out by pushing gears closer together, or if you were just talking about some other part of the mount?

In the picture that Anole posted just above, the big pin in the C shaped junction is what I knocked out and drilled a bigger hole and replaced the pin with a bolt.

It eliminated all slop int he Birdview mount.

But slop wise, none of the mounts I've had were free of slop. All of them have given me different landing points if I come into the sat from the west vs coming from the east, particularly if I cross the south point.

My repeatability has been getting worse and worse, so that I'm not sure what of the many possibilities is the cause. I think I'm going to put a scope on the counter signal to see if it looks like a dirty signal, as it's possible that my Drake is starting to lose it's power supply. I've noticed that the LNB voltage on Ku is less than it should be. I may end up switching to my spare IRD, which is a Monty 50/55, just to see if it works better.

Also, when the snow goes away, I think I'm going to see if I can somehow take that mount apart, without taking the whole dish down, but I don't think that's possible. But at least maybe I can at least lubricate the gears well, and see if there is ANY possible way of snugging them up tighter to take out the slop. But I'm thinking that I really need to replace bushings that will require me to take the whole dish down, and I really don't want to have to do that.

I have replaced bolts, bushings, and washers to get my various mounts to be as slop free as possible. The other biggest single slop is the acme nut inside an actuator.

As far as moving a dish with an analog box, scrap the analog and buy a GBox.

I guess I never did address whether USALS would work on a BUD.

I think it would work fine keeping in mind the issues I addressed above.

It is difficult to make USALS work perfectly on a small dish, so magnify those problems for the big dish not taking into account the differing hardware.

Not insurmountable, but it will be difficult.

 

[Anole]

Lost counts are an unacceptable and fixable problem.

I'm somewhat concerned about installing my own magnet wheel and reed sensor, as far as how to find the center of the sweet spot.
The matter IS a subject for this thread, in that such a procedure would be needed for setup for any new V/Gbox or the like.


Back to the matter of USALS on a BUD, I will go out on a limb, here.
Based on the general knowledge base provided by reading the forum for several years, I'd suggest:
- USALS is probably doable on a good H-H mount on C-band.
- USALS seems possible (if harder) on a good H-H mount on Ku-band.

The Ku would be the bigger challenge.
And frankly, even if unsuccessful, I might be happy with USALS on C-band only.
Of course, you shoot for the moon, do your best, and Ku might just surprise ya. -

 

[Lone Cloud]

I just thought of something that is somewhat off the point of the thread as it has developed.

In my post I was proposing a mount that could be adjusted by the remote control for both movements of the dish - across the arc and for elevation.

I am just wondering - if two actuators were installed on a dish, and 2 Gboxes - 1 to each actuator, 1 actuator in the usual place and the other to adjust elevation, let's say when you find the best aim for the dish on that particular satellite, you save the location in each G-box, and name it the same as well.

This might involve covering the remote sensor on one or the other G-box while saving.

Then later, using one remote control, you could just go to that satellite position, with both G-boxes taking the remote command to its own saved location.

Got me thinking.

 

[SatelliteAV]

Jaeger manufactures a dual axis motorized system SMR-1224EL similar to your description. The unit is designed for inclined orbit satellite reception. It is not DiSEqC controlled.
http://www.jaeger.com.tw/tvro/superjack/mount_superjack.htm#1224

I just thought of something that is somewhat off the point of the thread as it has developed.

In my post I was proposing a mount that could be adjusted by the remote control for both movements of the dish - across the arc and for elevation.

I am just wondering - if two actuators were installed on a dish, and 2 Gboxes - 1 to each actuator, 1 actuator in the usual place and the other to adjust elevation, let's say when you find the best aim for the dish on that particular satellite, you save the location in each G-box, and name it the same as well.

This might involve covering the remote sensor on one or the other G-box while saving.

Then later, using one remote control, you could just go to that satellite position, with both G-boxes taking the remote command to its own saved location.

Got me thinking.

 

[Lak7]

LoneCloud-
You could use 2 of the same receivers, 1 for Elevation and 1 for Azimuth.

 

[Anole]

direct dish axis feedback:

I am not particularly anxious to get into linearizing a jack actuator, but for those who might be ....
This sensor was mentioned a couple of years ago here on the forum.
It is capable of at least 2000 pulses per revolution.

When I considered using it on a Birdview, it was with the understanding that one of its lower resolution modes would be more than adequate.
Eventually, I gave up on the sensor, mostly due to the difficulty of interfacing it to existing equipment.


On another subject, the reed/magnet in an AJAK motor are well upstream of the final worm gear.
This gives a lot of total counts with a 5-magnet wheel.
However, the down-stream slop goes unmonitored, so it's not all roses.

For purpose of discussion, we should document how many pulses are generated for 180° of motion of the dish.
Not only for the AJAK with 5 mangets, but also for the Birdview with 8-, 24-, and 30- magnet wheels.

 

[Lone Cloud]

I sort of have a plan for an elevation motor add-on for my Birdview dish.

I have taken the motor off of the Superjack actuator I had surplusing around. One bolt and it slipped out. The place where the screw could sit is about 1/2 inc h in diameter and an inch or so deep.

I'm thinking of putting a rod in there fine threaded. With some strong angle iron pieces bolted onto the mount, a little welding, maybe some bicycle inner tube pieces to keep the threads clear, it just might work.

I'd need another G-box, but if it had the same remote, when I hit the power button, theoretically it would turn one unit on and the other off.

The wiring to the dish looks like the most work.

If I do it, I'll start a thread and try photos.

 

[a33]

... however the counts that come back from an actuator motor are NOT directly proportional to the USALS angle. There is a relationship, and it can be calculated, but you'd need to enter parameters like the distance from the rotation axis to the two connection points of the actuator, and the angles relative to a right angle connection (I think). Plus, all the necessary parameters would have to be fed into this motor controller by some other connection to a computer, since a generic STB isn't going to allow sending these parameters.
Basically, it's pretty obvious why no-one has ever tried to sell a USALS box for a BUD, and that is that basically every brand and size of actuator driven BUD has different parameters.

I have no experience with actuators myself, and I don't know much about pulse-counters and actuator-controling devices and so on, but I thought a bit about calculating the actuator extension length for different USALS angles (or rather: for MODIFIED Hour angles, as the modified motor angles track the arc better than normal USALS commands).

I haven't considered the practical applicability of the calculations (can the needed measurements be done precise enough, are the mounts constructive elements without annoying abberations, and can sensor and control unit technology easily support it?), but the math for the calculation is doable.

The math would 'only' be some trigonometric calculations, and exactly five parameters/measurements of the dish/mount/actuator combination would be needed, I would think.
So the math is extensive, but not too difficult I think.
The five measurements of course always have to be done 'in situ', by yourself, and must be done with proper attention (to get the needed precision of outcomes).
Also the measurements must be done with a dish that already follows the arc correctly; otherwise you get meaningless measurements. And after a readjusment/repositioning of the mount/actuator, you'd have to do the measurements all over again...


Would there be interest for such an actuator-extension-length calculator?

greetz,
A33

 

[a33]

So, nobody is interested. That is a pity.

For all those NOT interested:

For the calculation, I considered the following:
Assuming the actuator is in a plane, exactly perpendicular to the mount's axis, you need 1) the distance axis--actuatorbolt on dish; 2) the distance axis--bolt on mount, holding the actuator-clamp; and 3) with dish outlined in zero-position: the distance between these two bolts.
As the bolt on the mount, holding the actuator-bracket/clamp, is usually not co-located with the middle of the actuator, you also need 4) the distance from bolt to middle of the actuator.
And as the assumption that the actuator is in one exact perpendicular plane to the axis might not allways be correct, you will also need 5) the distance (measured on the axis) between the plane of the actuatorbolt in 1) and the plane of the actuatorbolt in 2).

See my drawing in the attached picture:
1) = A-T
2) = A-BB
3) = T-BB
4) = B-BB
5) = H

With these measures you can form three triangles, needed for the calculation.
From a changeing angle at the axis (following the calculated modified hour angle for a certain satellite position) you can thus calculate the new actuator extension, as compared to its known zero-position extension.
(I hope as a not native english speaker and not-mathematician, I have made myself somewhat clear...)

If these calcuations proof to be correct, with adding 6) the amount of pulses per length unit of the actuator, you would even be able to use the actuator really like USALS.


QED: A USALS-like calculation is doable for actuators.

greetz,
A33

 

[Magic Static]

Trigonometry was not my best subject in school. Taking for fact you have a working formula, how do you implement your formula to a USALS like motor control? Or is that a question for another day?

 

[Titanium]

It isn't that we aren't interested in your project on a resurrected 9 year-old thread... LOL

I am certain that a hobbyist would want to implement if they have the passion to experiment. Yes, we have previously discussed calculated positioning of a polar mount with a linear jack and it certainly is doable, but the calculation would be unique to each system and not universal like with HH KU motors.

USALS positioning is popular because it only requires the user to input the install location coordinates to position a properly aligned dish mounted on a HH USALS motor. Doubt that most casual users would seek the benefit of calculated positioning, which is mount and actuator specific.

Please keep sharing your project, it may inspire another hobbyist to experiment with USALS on with their linear jack mount.

 

[wvman]

Interesting thread, but my solution was to install a dish for each satellite desired using dual output Titanium LNBF's. Most people don't have the room to do what I did, nor the desire to have that many BUD's sitting around. Even with an installation like mine is not without its challenges.

A couple years ago, I had trouble with certain receivers that interacted when hooked to the same LNBF. After trying different combinations of receivers, I eventually stumbled onto the problem. Wasn't the receivers at all, but rather two electrical outlets on different breakers. Another problem was different signal qualities in two different rooms off the same LNBF.

I had a couple channels that would stay locked in one room, while the same channel in another room had a lower signal and broke up a lot. I found that to be a processor problem rather than a problem at or between the dish. I fixed a couple other problems after bonding all the grounds in the house. Being new to FTA, it was a interesting journey and I still find a head scratcher once in a while.