USALS for BUDs:

[Anole]

I've been thinking of this idea and dropping hints for a long time.
Nobody seems to care.

The old-time BUD guys have their analog or 4DTV receivers to move their dish.
Or maybe they've moved into the current decade, and have a Gbox/Vbox.
I've been around Ku-FTA for a number of years, and the little H-H motors are the way to go.
eg: STAB HH-90, or the SG-2100 and all its clones.
So, upon entering the hallowed halls of BUDdom, I cannot in good conscience endorse anything less than USALS for BUDs! -
The Coolsats, Visionsats, Pansats, and Fortec receivers demand it!

This thread is to discuss how to accomplish the task.
If you don't know why, then we might let that slip, too.

Some questions to consider:

What is USALS?

Which BUDs can run USALS?

How shall we power this contraption?
- LNB voltage from the coax?
- higher voltage on the coax (36v)?
- AC power supply?

What other advanced features could we incorporate, as long as we're at it?
-
- (many ideas go here!) -
-

Where can we get this motor controller?
- build it from a Kit?
- Mod to existing unit?
- Stand-alone product?


If you think the whole idea is silly please say so.
If you think the existing ways are fine with the V/Gboxes, say so.
I promise we'll change the minds of many, or have a heck of a good time trying!

If you like the idea, how much would you reasonably spend on such a device?
I'm actually trying to set up a feature-list that would make ya happy to spend $100 retail.

 

[jsattv]

Anole, I'm far from being an expert, but I'd like to comment on your great topic. There is nothing I'd like better than to get rid of my old Uniden 9900 Receiver and its Outboard Power Supply which I use with a VBox III to control my 12 Foot Mesh Dish and the Actuator which drives the dish. I use a Viewsat Extreme Receiver in the "Installer" setting for motor control. The Uniden 9900 provides the V and H Polarity control for my Corotor II and LNBs. That new V Box IV with built in Polarity control that Sadoun had previously mentioned would also likely eliminate my Uniden 9900, but does anyone know the status of this Unit, will it ever be built??

For my wind blown 1 meter (39 inch) Fortec Star Dish which is still on my house roof I used a Stab HH 120 Motor and ran it thru the Viewsat in "USALS" Motor setup. With USALS you set in Longitude and Latitude for motor control, then go to your South Bird (which in my case is 97.0W) and hit Go to Reference on the Remote Control and your done!! The Stab HH 120 motor is programmed to know where all the Satellites are and boy is it precise and accurate.

For my 12 Footer you need to check a number of Satellites before you are dialed into the Clark Belt, and I had to compromise a bit. That is I can get C and Ku Band programming from Satellites 57.9W to 139.0W as well as White Springs, yet I've had little success with Galaxy 18 / 123.0W. But with the Fortec Dish, Stab HH120 Motor, and USALS for motor control on my Sat Rx, once I setup for 97.0W, everything on the Clark Belt on Ku Band was there once the Satellites were scanned in, so you are correct that USALS is the superior mode, but as you say how the heck could you use it for 36 VDC control of an Actuator Dish mover and a Bud Dish? Or is this maybe what Horizon to Horizon motor control setup for Bud Dishes is all about?? Linuxman and many other experts in this forum surely would know. By the way I finally solved my wind shaking problems with the 12 footer by getting a welder to weld on 3 nuts on the Antenna Pole after we drilled out the Holes for the nuts with the Dish still mounted but rotated way over to the West. Hope this helps.

 

[B.J.]

I've been thinking of this idea and dropping hints for a long time.
Nobody seems to care.

The old-time BUD guys have their analog or 4DTV receivers to move their dish.
Or maybe they've moved into the current decade, and have a Gbox/Vbox.
I've been around Ku-FTA for a number of years, and the little H-H motors are the way to go.
eg: STAB HH-90, or the SG-2100 and all its clones.
So, upon entering the hallowed halls of BUDdom, I cannot in good conscience endorse anything less than USALS for BUDs! -
The Coolsats, Visionsats, Pansats, and Fortec receivers demand it!

This thread is to discuss how to accomplish the task.
If you don't know why, then we might let that slip, too.

Some questions to consider:

What is USALS?

Which BUDs can run USALS?

How shall we power this contraption?
- LNB voltage from the coax?
- higher voltage on the coax (36v)?
- AC power supply?

What other advanced features could we incorporate, as long as we're at it?
-
- (many ideas go here!) -
-

Where can we get this motor controller?
- build it from a Kit?
- Mod to existing unit?
- Stand-alone product?


If you think the whole idea is silly please say so.
If you think the existing ways are fine with the V/Gboxes, say so.
I promise we'll change the minds of many, or have a heck of a good time trying!

If you like the idea, how much would you reasonably spend on such a device?
I'm actually trying to set up a feature-list that would make ya happy to spend $100 retail.

Well, the first thing to consider is that while USALS shouldn't be too hard to accomplish for a BUD on a H-H mount, probably 98% of the BUDs out there have actuators, not H-H motors. USALS depends on the controller being able to rotate the dish around the rotation axis by a specific number of degrees. The counts that come back from a H-H mount are directly proportional to this USALS angle, 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.
The only way I think it would be conceivable to manufacture a USALS box to work with actuator type BUDs, is to abandon the reed switch count generator that exists on the BUD, and also produce a stand alone angle counter connected at the rotation axis of the mount to detect angle rather than the hypotenuse of a non-right angle (I know it wouldn't be called hypotenuse, and I've probably spelled it wrong). However every type of BUD would have a different way that this angle sending unit would have to attach.....
............. SOOOOO ....

If you think the whole idea is silly please say so.
If you think the existing ways are fine with the V/Gboxes, say so.

Well, it'l not a "silly" idea, but I think it is next to impossible to produce one that could work on even 10% of the dishes out there.
And do I think the existing ways are fine? Yes. I generally prefer DiseqC-1.2 anyway.

 

[SatelliteAV]

One of our partner manufacturers provided a 36" sample jack which uses DiSEqC 1.2 and an external power supply. The unit is in prototype stage. We will have to wait and see if it makes it into production. Many RD and production facilities are struggling and most product development is on hold.

Overall the sample tested very well, but then you start thinking (dreaming)...... could the unit also incorporate polorotor control?

I agree. USALS implementation would be highly unlikely on a jack design, due to the mechanics involved in a polar mount. Each mount uses a different amount of throw to position the reflector depending on the placement and distance of the jack clamp and arm attachment point.

 

[Lone Cloud]

Although I do think it would be hard to justify in a marketing sense ( app 1 out of 200 homes with a BUD, maybe 10% of those actually working) I am not of the opinion that the measurement and angle variants can't be overcome. integrated circuits are too inexpensive and have so much computing power for it to be a real issue.

The improvement I would like to see, on big and small dishes, is the motor(s)that moves the dish both across the arc longitudinally and for elevation, and then computing the arc for the user, via the remote control. Although many of us have our own quick met,hod of aligning a dish, most of us went through some frustration to learn that.

Imagine aiming your dish roughly south and that being the last time you have to go out there.

You'd find a satellite with your remote. You'd tweak the elevation and longitude controls until you got the best signal, You'd lock the position and tell your controller the longitude of the satellite you just locked. After locking two or three satellites, your controller, USALS or otherwise, computes the entire rest of the arc- from the warm comfort of your living room.

 

[surfdevil83]

This may be way out in left feild but I had an idea to try to measure the spike in voltage that a hh90 motor recieves to move it. When the dish arrives at the desired sat I imagine the voltage disappears. So if you could trigger a power supply(37v) close to the bud with this voltage, the 1.2 settings could be adapted. Almost like a relay. Just a idea. Maybe thats what SATELLITEAV has.

 

[guapoharry]

USALS Info USALS is a calculation system allowing receivers to calculate the position of satellites. Another link...

If your BUD has an H-H motor, it seems plausible. For a polar mount with a linear actuator, I agree with Sat AV's last paragraph. Yeah, it's possible, but there are a lot of differences introduced with using an arbitrary polar mount and actuator. You don't have to deal with that when using a USALS H-H motor design.

I don't see how you are going to move the mass of a BUD with the power of a typical STB. If so, you are going to need a VBox style thing or something to interface the STB with thing that powers the motor.

 

[pendragon]

I have two H-H motors, an Ajak H-180 on my 10' and a Moteck HH180 on my 6'. When I set them up last year I logged the 'click' counts for each orbital position on my GBOX positioners for both dishes. I put these in spreadsheets and calculated the corresponding motor angles. I was hoping to automate this into a crude USALS-like process so I could dial up a new orbital position and hit it dead-on.

Unfortunately the number of 'clicks' per degree of motor angle varied over the range I tested for both dishes, from 37.5W to 148W. I could dig up the data, but at the time I concluded this might be useful for interpolating between known positions, but no where near accurate enough to perform absolute positioning.

I hate to burst the H-H USALS balloon, but I assume the mechanical precision of these H-H motors was not as tightly controlled as for typical USALS motors. It seems unlikely at this stage that we will see a resurgence in C-band H-H motors that would address this issue.

Linear actuators may suffer from similar mechanical variations in addition to the theoretical problems already discussed. However the latter could possibly be addressed by calibrating a small number of known positions and deducing the arm geometry from these measurements. This doesn't appear to be a insurmountable problem, but I haven't enough interest or need to make the investment. If mechanical precision is a limiting factor, it won't matter.

 

[Anole]

Thank you all for your well thought-out answers.

Some of the questions I proposed were not serious ones.
Certainly, the power to operate a BUD motor would be considerable.
Probably in the 35 to 50 watt range. Possibly more for some motors.

This enire matter is not one that just came to me.
I've been wanting USALS on a BUD since I first understood it and how BUDs were normally operated.
Quite a few years.

I also assumed that an H-H motor would be a necessary requirement.
(but you've all given me amazing insight into other motor/mechanism options!) *
Two improvements to a typical H-H BUD motor which would undoubtedly be necessary, are:
- higher accuracy (I've not calculated the number, so help would be appreciated)
- reduction of "play", or slop
I have observed how Linuxman has accomplished both these requirements with his Birdview dishes.

As satellites viewed by the dish approach the horizon, the positioning accuracy of the dish becomes more critical.
Overhead, the 2° satellites look farther apart to the observer.
As you view 2°-spaced satellites near the horizon, they appear closer together.
So, it may be a limitation, that the USALS range of a converted BUD H-H mount would be limited.
(great for the middle of the sky, but not so hot down low on either side)
This could also be a limiting factor of some of the cheaper grapefruit-sized H-H motors for Ku dishes.

In a private conversation with an industry member, I suggested that the only user-setup required to add USALS to a controller device, would be a user settable "Counts per Degree".
(or more likely, some decimal representation of "Degree per Count")
This assumes of course, that sufficient counts would be available, and that they would be linear across the positioning range.

Additionally, I feel that an anti-slop mode should be user-selectable.
This would mean any dish movement always approaches the final position from one direction.
And even if the user had the remote control and was stepping the dish, it would (or could) engage this same mechanism.

To support my interest in USALS, I have gone looking specifically for dishes/motors with H-H mounts.
I feel the two selected, a Birdview, and a nice spun aluminum perf on an AJAK 180, fit the bill. (links in my sig)
It's amazing what sort of hardware you can find when you start out with a shopping list, and know (a little about) what you are looking at.
The AJAK was a total surprise to find, but I recognized it instantly, from several years of reading this fine forum!
I'm already under way on a high precision magnet disc for the Birdview.
And have no doubt the AJAK accuracy can get a boost, too.


* I'll discuss another time some great ideas brought to light by your comments on modifying a standard polar mount & jack.

 

[pendragon]

Some of the issues here are the often confused distinctions between accuracy, resolution and repeatability. If I have a tape measure, accuracy means I can measure distance between two points on any part of the tape and come up with the same result, regardless of temperature, humidity or other independent effects. Resolution means how close the markings are on the tape; however this isn't the same as accuracy if the marks are at different spacings on different parts of the tape or if say the tape changes its dimensions with temperature fluctuations. I may be able to make high resolution measurements with mediocre accuracy. If the tape markings aren't evenly spaced, I may still have repeatability if I can find the same position each time as long as I measure from the same absolute point.

Without repeatability in a H-H motor, all bets are off. With both of mine, repeatability seems to be a non-issue. If I drive the motors either direction to a known 'click' position, I see virtually the same signal conditions whether it be on a spectrum analyzer, SQ on a STB or the various measurements available on a computer tuner. This is more meaningful on a 3 m dish than 1.8 m dish because of the narrower beamwidth.

I have decent resolution; about 0.055 degrees per click on the 3 m (C & Ku) and 0.9 degrees per click on the 1.8 m (C only). Given the beamwidths and all the electrical inconsistencies in a beam pattern, I doubt that having any more resolution would result in better or more repeatable reception. The variation in a given satellite's orbital position over a day might even be greater than the Ajak's resolution. Unless I employ a loop tracker, having more resolution won't help.

What I lack is accuracy (see attachments). The x-axes show the motor angle with zero being true south and negative-east and positive-west. The y-axes show the degrees per click from the previous orbital position. These graphs are raw data from when I first set up the dishes and are probably only representative because I have refined my techniques for measuring the best position since that time. I'm sure some of the orbital locations were not perfectly determined and this likely causes some oscillations from position to position as is shown on parts of the plots. After I plotted these I realized I could have done this differently to minimize that effect but it's not worth the time to do them over

What annoys me are consistent slopes seen in parts of the plots and regions with very different, but consistent degrees/click. I'm guessing this has nothing to do with the magnet wheel or how many clicks it makes per revolution. It probably is more likely variations in gear pitch, wear and play. The Ajak motor is about 15-20 years old while the Moteck was brand new out of the box. I'm not claiming these results are typical for every setup, but am only providing the data as a warning that putting in a lot of effort to increase the number of clicks over the arc may not result in a lot of reception improvements or absolute positioning accuracy.

 

[Lak7]

I kinda have an idea, but it's just beyond me...
What about a "device" that is attached to the Axis of the Dish, like the guts of a little Ku Motor, the part the does the counting. The Dish would be physically moved by the usual means, but instead of counting motor revolutions, the "device" would receive the commands and start and top the motor.

 

[Anole]

more advanced that I planned:

Lak7 -

That's a good idea.
Back when we first discussed making our own magnet wheels (2 years ago?), a member came up with a sensor capable of 2000 or 4000 counts per revolution.

It was the recent discussion of using one of the jacks to motorize a dish, that made me think of this sensor again.
If it could be applied to the dish rotation, it could measure directly the movement, and motor-counts could be ignored.
At the time, it actually had too many counts, but in this application, it might be useful! - :up

For now, I'm concentrating on the existing magnet-wheels buried in the Birdview and AJAK motors.

 

[pendragon]

I kinda have an idea, but it's just beyond me...
What about a "device" that is attached to the Axis of the Dish, like the guts of a little Ku Motor, the part the does the counting. The Dish would be physically moved by the usual means, but instead of counting motor revolutions, the "device" would receive the commands and start and top the motor.

What you're talking about is partially accomplished with a rotary or shaft encoder. I haven't played with one, however they are nice toys. But they can cost and may not be the perfect solution. On a quick web search I found this:

RE36 - RLS d.o.o. rotary and linear motion sensors (encoders)

The specs say 8192 counts per revolution, which would seem to put it in the ballpark, except the accuracy is only +/- 0.3 degrees. That's worse than any C-band positioner I've seen, but it only sets you back 300 clams. There are some fancier units I came across that had far more counts per rev, with accuracies in the 1 arc-minute range. That would beat my factory standard Ajak with about 3 arc-minute repeatability, but I couldn't get an easy quote. Probably more than pocket change.

I haven't thought much about other error sources, but I would be concerned that slightly out-of-round shafts and slight off-center coupling to such an encoder might cause a sizable deterioration in overall accuracy.

 

[Cham]

Great idea. Don't see a major issue in designing something to work in an HH-dish combo. The problems I do see coming up is modifying existing old dish mounts to work.
The electronics can be a carbon copy of what is out there already (several Ku HH motors out there!). The physical part just needs to be bigger and much stronger. 36v motor with 115vac supply built in would have lots of torque and could speed up the movment when not dealing with 100' of buried cable and line resistance.
But... Then there is the cost! Something like this is going to retail in the $1g range+ for sure in order to capture development and manufacturing costs.
-C.

 

[Cadsulfide]

Rotary quadrature encoders can be expensive, but a cheap 256cpr encoder coupled by a timing belt and pulleys to the dish could easily provide a multiplier effect of several hundred "ticks per degree".

 

[B.J.]

...

Without repeatability in a H-H motor, all bets are off. With both of mine, repeatability seems to be a non-issue. If I drive the motors either direction to a known 'click' position, I see virtually the same signal conditions ...

Back when I had an actuator driven BUD, I got pretty good repeatability, however my current H-H is not very good at all, and is getting worse, probably due to the big semi-circular gear or worm gear wearing out. Both with H-H and actuator, though, I've always noticed a difference in the positions depending upon which direction you're coming from, particularly when you go past center. Ie when you go past center, you are pushing from one side of the free play in the gears, and when you come back, you are pulling from the other side of the free play (hope that makes sense). Anyway, in general, I haven't been that satisfied with the repeatability of any the 3 BUDs I've owned over the years.

However the other issue, besides free play in the gears, is losing or gaining counts. Back when I had the actuator system, I only had a 100' cable run, and I very rarely had to re-sync the actuator. With my current H-H, I'm always having to re-sync, and it's getting worse and worse. I think part of the problem is that I have a very long cable run now, somewhere around 250' or so, and while this doesn't bother the rf sat signal, it does seem to affect the DC on/off signals from the reed switch. Either that, or perhaps the distance between the magnets and the reed switch isn't proper. Also, the reed switch on my H-H is pretty exposed to the elements, and it gets very wet. Darn thing has grease all around it, which somehow seems to keep the moisture from draining properly, so I think the grease makes it worse. I bought a replacement reed switch, but it didn't help.
In any event, I haven't been very happy with the repeatability of any of my BUDs. My 90CM dish on the SG2100, on the other hand, has been quite repeatable.

What I lack is accuracy (see attachments). The x-axes show the motor angle with zero being true south and negative-east and positive-west. The y-axes show the degrees per click from the previous orbital position. These graphs are raw data from when I first set up the dishes and are probably only representative because I have refined my techniques for measuring the best position since that time. I'm sure some of the orbital locations were not perfectly determined and this likely causes some oscillations from position to position as is shown on parts of the plots. After I plotted these I realized I could have done this differently to minimize that effect but it's not worth the time to do them over

You say "motor angle" above is this what you really mean, or are you talking satellite longitude along the Clarke belt? Ie I'm curious how you determined this? I'm kind of assuming that you don't mean motor angle, even though that's the only angle that you'd really expect to be related to the counts. Since you mention oscillations from position to position, I'm assuming that you determined the angle by peaking on the satellites. I'm wondering if you could have smoothed out your plots by using a sat tracking program with the newest keps, to get the actual orbital position of the sats at the time, since they vary by amounts consistent with the oscillations you observed. It would take a lot of work to do this, but it would be interesting.
I have a spreadsheet that I've set up, which I use to relate counts to orbital position of the sats. This has been very useful to me in terms of being able to locate sats that I don't have programmed, particularly sats over the Atlantic, when my last programmed sat is like PAS9, or whatever it's called now. I've always been able to pop in the longitude of the sat I was looking for, then it would tell me what count that sat shoud be at, and when I'd go there, I'd be close enough to see the signals from that sat. However this spreadsheet was also using orbital position angles rather than motor angle. I'm thinking of re-doing this spreadsheet, and instead of orbital location, use the USALS angle for each sat. I'm guessing that it will result in much better results. Since I have a program that capable of spitting out both current KEP based orbital locations and USALS angle, I'm tempted to try modifying my spreadsheet to use USALS angle as well as the semi-real time sat positions.

What annoys me are consistent slopes seen in parts of the plots and regions with very different, but consistent degrees/click. I'm guessing this has nothing to do with the magnet wheel or how many clicks it makes per revolution. It probably is more likely variations in gear pitch, wear and play. The Ajak motor is about 15-20 years old while the Moteck was brand new out of the box. I'm not claiming these results are typical for every setup, but am only providing the data as a warning that putting in a lot of effort to increase the number of clicks over the arc may not result in a lot of reception improvements or absolute positioning accuracy.

There are so many things that would affect the clicks and angle where you peak on a sat that I'm surprised that your results look as good as they do. I'm a bit confused though with respect to why you chose to plot the degrees per count between sats, when you knew that the distances between sats is always changing? And if both adjacent sats are changing the affects are sometimes doubled, probably why every other one seems to be up/down/up, etc. Again, using the keps to use the actual real time positions might help here, but that would require re-peaking on all the sats, which is a pretty time consuming effort.
Last question, is how you got the "degrees/count" for the Motec motor? Only thing I can think of is that you used DiseqC-1.2 and went one single pulse movement at a time until the next sat was peaked? I sort of did this when I first set up my arc using DiseqC-1.2, and was locating each next sat by estimating how many single moves it took to get to the next sat. However it seemed like it wasn't as accurate doing this as I though, I'm guessing because the motor might occasionally miss a command (whereas when you're sending a goto type message, it sends it several times to make sure). Anyway, I'm curious how you got those counts with the Motec motor?

Interesting study, but with my BUD, I'm just happy if I end up close enough to see the signals so I can re-peak.

 

[pendragon]

Back when I had an actuator driven BUD, I got pretty good repeatability, however my current H-H is not very good at all, and is getting worse, probably due to the big semi-circular gear or worm gear wearing out. Both with H-H and actuator, though, I've always noticed a difference in the positions depending upon which direction you're coming from, particularly when you go past center. Ie when you go past center, you are pushing from one side of the free play in the gears, and when you come back, you are pulling from the other side of the free play (hope that makes sense). Anyway, in general, I haven't been that satisfied with the repeatability of any the 3 BUDs I've owned over the years.

Even though my Ajak motor is old, it doesn't appear to have much wear. Still, I worked on both my H-H motors to reduce any play before they went up. So far I've been unable to find any meaningful difference in positioning based on the direction I approach the desired point. A couple of months ago I tried to measure this error on my 3 m with a combination of my spectrum analyzer and a scrap of Linux code I wrote that ran a USB box receiver. Even with long integration times and measuring the beam pattern, I wasn't able to determine if there was any difference.

However the other issue, besides free play in the gears, is losing or gaining counts. Back when I had the actuator system, I only had a 100' cable run, and I very rarely had to re-sync the actuator. With my current H-H, I'm always having to re-sync, and it's getting worse and worse. I think part of the problem is that I have a very long cable run now, somewhere around 250' or so, and while this doesn't bother the rf sat signal, it does seem to affect the DC on/off signals from the reed switch. Either that, or perhaps the distance between the magnets and the reed switch isn't proper. Also, the reed switch on my H-H is pretty exposed to the elements, and it gets very wet. Darn thing has grease all around it, which somehow seems to keep the moisture from draining properly, so I think the grease makes it worse. I bought a replacement reed switch, but it didn't help.
In any event, I haven't been very happy with the repeatability of any of my BUDs. My 90CM dish on the SG2100, on the other hand, has been quite repeatable.

My first Moteck HH180 had a similar repeatability problem out of the box. I bought a replacement as a spare, and before sending the bad unit back, I swapped sensors. That solved the problem. Just before doing that I had put a scope on the sense lines at the controller box and noticed the pulses were rather ragged. I have nearly 200' of run to the Moteck mast. I breadboarded a quick differential amplifier so the two sense lines could float and played with a bit of low pass filtering. I also tried a monostable multivibrator. Both cleaned up the pulses, but all this proved was pulses were still being lost/added. I ran a benchtop electronic counter in parallel to the controller and both saw exactly the same thing. With the new sensor the pulses were cleaner and I don't use any preconditioning before the controller. I mention this because you might be seeing either or both effects.

You say "motor angle" above is this what you really mean, or are you talking satellite longitude along the Clarke belt? Ie I'm curious how you determined this? I'm kind of assuming that you don't mean motor angle, even though that's the only angle that you'd really expect to be related to the counts.

I calculated the motor angles by the nominal orbital position using the same technique you and I previously discussed in the USALS Notebook thread.

Since you mention oscillations from position to position, I'm assuming that you determined the angle by peaking on the satellites. I'm wondering if you could have smoothed out your plots by using a sat tracking program with the newest keps, to get the actual orbital position of the sats at the time, since they vary by amounts consistent with the oscillations you observed. It would take a lot of work to do this, but it would be interesting.
I have a spreadsheet that I've set up, which I use to relate counts to orbital position of the sats. This has been very useful to me in terms of being able to locate sats that I don't have programmed, particularly sats over the Atlantic, when my last programmed sat is like PAS9, or whatever it's called now. I've always been able to pop in the longitude of the sat I was looking for, then it would tell me what count that sat shoud be at, and when I'd go there, I'd be close enough to see the signals from that sat. However this spreadsheet was also using orbital position angles rather than motor angle. I'm thinking of re-doing this spreadsheet, and instead of orbital location, use the USALS angle for each sat. I'm guessing that it will result in much better results. Since I have a program that capable of spitting out both current KEP based orbital locations and USALS angle, I'm tempted to try modifying my spreadsheet to use USALS angle as well as the semi-real time sat positions.

If I had seen better results I had planned to go along with an outline similar to what you describe. My initial analysis on the 3 m was the step resolution at the controller was about half of the station-keeping requirement for the orbital positions I checked. I therefore elected to neglect the instantaneous positions of the satellites, because I was going to measure a lot of positions; if the dominant error was the off-center satellite location I would see a lot of +/- 1 count variations from a nominal value for degrees per click. However the data showed locally consistent variations of larger magnitudes consistent with mechanical errors. I did my best to peak each position with two independent measurement techniques, but I could do better if I did it today.

There are so many things that would affect the clicks and angle where you peak on a sat that I'm surprised that your results look as good as they do. I'm a bit confused though with respect to why you chose to plot the degrees per count between sats, when you knew that the distances between sats is always changing? And if both adjacent sats are changing the affects are sometimes doubled, probably why every other one seems to be up/down/up, etc. Again, using the keps to use the actual real time positions might help here, but that would require re-peaking on all the sats, which is a pretty time consuming effort.

I agree using the orbital elements would have eliminated the effect you describe. My supposition at the time was the mechanical errors would dominate and I think that turned out to be the case. I also found the on-line calculators for satellite position didn't all agree. Sorting out the rights from the wrongs was out-of-scope to the level of effort I wanted to invest. Realize I collected the data for a quick look and hadn't planned to use it as a basis for any detailed studies.

Last question, is how you got the "degrees/count" for the Motec motor? Only thing I can think of is that you used DiseqC-1.2 and went one single pulse movement at a time until the next sat was peaked? I sort of did this when I first set up my arc using DiseqC-1.2, and was locating each next sat by estimating how many single moves it took to get to the next sat. However it seemed like it wasn't as accurate doing this as I though, I'm guessing because the motor might occasionally miss a command (whereas when you're sending a goto type message, it sends it several times to make sure). Anyway, I'm curious how you got those counts with the Motec motor?

The Moteck HH180 is a traditional BUD motor with 36V motor lines and separate sense lines. It does not do DiSEqC 1.2 and requires a controller like a GBOX to run it.

Interesting study, but with my BUD, I'm just happy if I end up close enough to see the signals so I can re-peak.

I'll count myself fortunate that all my dishes, whether DiSEqC 1.2 or BUD controller-based hit the same spot each time. Otherwise the other family members, who believe this is all child's play, would have my head on top of a pole when they wanted to switch to a different channel and had to peak the position

 

[pendragon]

Rotary quadrature encoders can be expensive, but a cheap 256cpr encoder coupled by a timing belt and pulleys to the dish could easily provide a multiplier effect of several hundred "ticks per degree".

This is effectively the same principal used by H-H C-band motors to generate clicks. A magnet wheel spins at a higher rotational rate than the polar axis through gearing, chain drives, etc. I have no issue that any arbitrarily high 'clicks per degree' could be generated to increase the resolution of the dish position measurement. My first concern, based on an admittedly limited amount of data, is the absolute positional accuracy of such systems appears far worse than the measurement resolution. That makes USALS useless without a calibration table. My second concern is there seems to be a common view that increasing the number of clicks per degree will automatically improve positioning accuracy and therefore SQ will be also be correspondingly improved.

Putting a rotary encoder directly on the polar axis shaft gives the possibility of measuring an absolute position to high accuracy, IF the coupling is also accurate and one is willing to bear a high cost. Other solutions, such as the one you suggest, cannot provide an absolute position of the polar axis shaft and may suffer from the same limitations I've conjectured for existing H-H motor sensors.

 

[AcWxRadar]

I've been thinking of this idea and dropping hints for a long time.
Nobody seems to care.

Anole,

I am not into the C-Band too much (yet), but I definitely do care. I understand your point and direction.

I got started with satellite TV through Dish Network and through my brother. When my DN 301 remote went bad on me, I went to Ebay to find a cheap one instead of a $70 DN replacement (that is what DN wanted to charge me). I found my remote, but I also happened upon the Coolsat 5000 and the true FTA realm. I have slowly gravitated towards FTA ever since, dropping subscription channels to save money to buy new toys and being able to watch plenty of good TV.

I know that FTA programming does not compare to DN or DirectTV when it comes to the overall selection and quality. However, it suits me just fine! I can save lots of money! I don't need to watch some garbage on HBO and pay dearly for it.... I can wait a few years for it to come out on a DVD!

I can get the imporant things (news) from FTA at no charge (Al Jazeera and ONN are my two consistent favorites).

If more people in the US and Canada and Mexico realized what they could get on true FTA (For FREE!) they might be angry that no one like us had ever told them! Europe is much more advanced (so I have heard) in FTA satellite availability. I think we already have this as well, but not many folks understand the technology.

This fact is sorry, in my opinion. I would probably be in the same boat if not for the Coolsat 5000. I would be paying a premium rate for watching commercials! I don't think so! I cancelled my subscription twice, maybe three times. Started it up again and cancelled it again. I work nights, so when I get home, guess what is on DN TV that I used to pay TOO MUCH for? Freaking paid advertisements!

I don't care all that much about watching TV, but in the winter, I need somethig to do when I get off work late at night. Chatting with you guys and gals became my wintertime hobby and I love it! Screw the TV! LOL

Nowadays, I don't even care too much if there is a good show on TV or not. I cannot wait to get home from work to chat with all of you!

I highly appreciate the opportunity to be able to chat with all of you, much more interesting, informative and fun than watching anything on TV. I am sure that you all understand exactly what I mean! Thanks to all of you!

Radar

 

[pendragon]

I am not into the C-Band too much (yet), but I definitely do care. I understand your point and direction.

I care a lot, too, because I also really wanted to make this work when I installed my 3 m and compared it to running USALS on my 1.2 m. Unfortunately I've found nasty issues that make it difficult to accomplish a USALS-BUD to the accuracy I would like at an affordable cost. Along the way I ran into a number of quandaries even with fundamental USALS operation. Some, if not all, USALS-certified receivers calculate the motor angle with increasing errors the farther one gets from true south. They do not appear to take into account any declination adjustment. And my USALS motor has a measurable angle offset when it's supposedly centered.

I can get all of this to work with traditional alignment techniques, but I can also measure off-peak performance the farther I go from true south. By methodically stepping through each symptom, I've identified the fundamental problems and been able to either workaround them or fix them. With all that time invested the only thing I've accomplished is accurately positioning the dishes nearly down to the horizon. When I started I assumed this was more straightforward. Compared to what I wanted originally (a USALS-like capability for my BUDs directly run off a microcontroller/server), I don't see my efforts as terribly successful. As a learning experience it's been good, but take my cynicism as a warning before getting in as deep like I did. I don't think this problem is as insurmountable as getting C-band on 18" dish, but it is tricky and likely expensive.