Problem with VBox III Accuracy

[linuxman]

Well!!!

Where do I start. Hmmm!!

I will start in this thread because this is where the problem was identified and will do a full review in another thread when I am finished.

I got the GBox V3000 from Sadoun today as promised.

Did some preliminary work with the VBox III just so I would have data for comparisons sake, and so far I am impressed.

Hooked up the Gbox V3000 to my Unimesh dish in the same place that the VBox III was hooked at the center of my arc. Started moving the dish East and programmed in the positions one by one for each satellite and set the East limit just to the East side of 43W. Moved back to the center and started West ending just on the West side of 143W which is where 148W comes in for me with the DBS LNBF hanging off the east side of the feed-horn.

The following are some counts (pulses) as were counted by the VBox III and the GBox V3000.

This is from about 10 pulses East of 43W (IS3R) to about 10 pulses West of 143W (Echo 1,2) which is where the DBS LNBF (hanging off the side of the feed-horn) brings it in on this dish.

On the VBox III after adding the extra magnet to double the counts (pulses) the total pulses in the above arc is 597.

Switched out the VBox III and substituted the GBox V3000, and in the same arc as above, the total pulses is 1148.

Just for comparison, I have a VBox II on my Birdview dish and run it all the way from 30W - 142W with a 34:1 gear ratio motor (less pulses) and get 907 pulses with it.

The point is that the VBox III was not interpreting the pulse count correctly on a consistent basis.

Even if I hadn't doubled the counts by adding the extra magnet in the actuator, the new GBox V3000 would have given me 574 pulses which is almost as much as the VBox III is giving me with the added magnet and counts.

The accuracy of the landing points on each satellite has greatly improved, but still have a lot of testing to do before I do the final review.

But this is a good place to continue the discussion.

All current owners of the VBox III, take note of the above.

 

[gabshere]

everything about it sounds good but it still kinda leaves me in limbo , unless you think with one magnet it would make your dish accurate enough for the g-box to make a stop at each satellite accurately....... i know the v-box III won't do that ( unless i double the sensor or magnet)

The point is that the VBox III was not interpreting the pulse count correctly on a consistent basis.

i agree here and will wait for your review but so far it sound promising

 

[linuxman]

i know the v-box III won't do that ( unless i double the sensor or magnet)

I don't think you know how many counts your current setup is actually producing, if yours is behaving like mine did as the counts show.

Yours might actually be producing the correct number of pusles, and they just aren't showing up.

unless you think with one magnet it would make your dish accurate enough for the g-box to make a stop at each satellite accurately

The accuracy greatly improved after adding the extra magnet using the VBox III, so to answer, I would have to say that I was satisfied at the improvement in accuracy, so I think I would have been satisfied had I bought the GBox in the beginning.

Does that make sense?

 

[gabshere]

I don't think you know how many counts your current setup is actually producing, if yours is behaving like mine did as the counts show.

Yours might actually be producing the correct number of pusles, and they just aren't showing up.

to a point yes you are correct but i don't know which to blame the motor ( i know it would help if the motor had more pulse counts) or the v-box III ( i suspect its a little of both motor & v-box) one direction it counts XXX#'s and the other direction it counts more or less but not the same amount of #'s if that makes sense... lol so i'm stuck in the middle but so far my next unit is pointing toward the g-box

to bad the v-box can't be upgraded with filmware to count twice as much for the unit. so if it counted 400 with a tweak it would count 800 ..... lol

 

[Anole]

On the VBox III after adding the extra magnet to double the counts (pulses) the total pulses in the above arc is 597.

Switched out the VBox III and substituted the GBox V3000, and in the same arc as above, the total pulses is 1148.

I just don't see how this can be.
Pulses are pulses, no matter who does the counting.
If the Gbox is counting both edges of the pulses, then that's not really twice the accuracy, because the "pulse on" and the "pulse off" are not equally spaced (I'm assuming).
But, let's go with that idea: count when the magnet gets close, and count again when the magnet is far away.
It's an amusing idea. And it can't hurt.

One thing you should be able to do, is measure the jack when east, then again when west.
Those inches times 48 or whatever , should be roughly equal to your counts.
Or to half the counts on the Gbox.

The Gbox will have to have a larger count range.
I know that's still an unknown, but it can't have a total 2000 count range.
When ya run across any such info, let us know.

Oh, and as for the pulse doubling, I suppose we could whip out a little circuit for that.
That's sort of what I wanted to do with the hall-effect devices - multiply the counts per magnet.
Not having a box/BUD to test on limits my imagination , but I'll sleep on the idea.
Extra magnets is the right answer, but if a simple little circuit could be put between the sensor and old Vbox ...

 

[linuxman]

If the Gbox is counting both edges of the pulses, then that's not really twice the accuracy, because the "pulse on" and the "pulse off" are not equally spaced (I'm assuming).

The guy from Von Weise called me just after the new year to ask if the magnet had worked, and I told him that it worked fine. I asked him to again explain how the polarity on the magnet worked. He told me to put a magnet in each palm and press my hands together and if the magnets both pushed against each other I was to envision placing the back of my hands against the gear/brass piece that holds the magnets and install the magnets facing the sensor just like that.

I don't know how that might affect the sensor, but I would guess that the magnets push the switch away each time they cross the sensor. I don't know what pulls the switch back.

I am guessing, but I think the "counter" in the VBox is not sensitive enough to pick up all the pulses. I don't have a way to test, so if you can come up with one, that would be great.

From the Von Weise specs on their website:

The VW-24LC comes from the factory with 24 counts per inch. With a throw of 20 inches that would be 480 counts.

Doubling that with the extra magnet would be 960 counts.

From above, the counts on the VBox III is 597.
From above, the counts on the GBox is 1148.

Which is more accurate?

I'll try to go up today and measure the actual inches that my actuator is coming out of the tube for a better multiplication factor.

The Gbox will have to have a larger count range.

I know that's still an unknown, but it can't have a total 2000 count range.

When ya run across any such info, let us know.

The Gbox displays it's counts differently than the VBoxes.

On all my dishes, I start at the center at 000 (91W). On the VBoxes (89W) is counted as 980 and (93W) is counted as 020 or (1020).

On the GBox, center is 000, (89W) is -20, and (93W) is + 20.

That makes the East end of my arc -382 and the West end +766.

The manual shows a picture of resynching with zero to 2000 counts on the scale. So I have to assume it is capable of at least 2000, and might be capable of 4000. Negative 1999 and positive 1999. It can display 4 digits.

And the discussion continues.

 

[truckracer]

My v-box 4 is not as accurate as my analog receiver. If I move the dish around very much I have to go back and resync the arc.

I went with the v-box 4 because my analog receiver did not have enough position memory. It could only store around 20 satellites. I am using a superjack 36" actuator with the arm attached to the outermost edge of the polar mount ring to give the most mechanical advantage as well as give more pulses per inch. I am still not totally impressed with this v-box 4.

But before I blame the v-box 4 I am controlling it with diseqc 1.2 commands from a Pansat 9200 HD that is tempormental and full of software bugs. It could be the receiver. This receiver locks up on a regular basis when trying to blind scan or make antenna setting changes. It seems to have seizures when you try to exit out of the main menu quite a bit.

I have the latest software from Pansat and it is still a pain sometimes. Not have as reliable and trouble free as my Fortec ultra lifetimes.

 

[linuxman]

When I got back and the rain had let up a little, I went to the roof twice to measure the actuator arm at the West limit and the East limit.

The distance travelled is 12.75".

According to the manufacturer, that should produce 612 pulses at 48 pulses per inch traveled.

From above, the counts on the VBox III is 597.
From above, the counts on the GBox is 1148.

So according to what is supposed to actually being produced, the VBox III is counting correctly or close to it.

BUT, in using the GBox so far, it has landed on the Ku channels that I have used it on with much more accuracy than the VBox III. I choose Ku channels and the weakest ones, because you have to be right on them to get good SQ.

How do you account for all the extra counts in the GBox V3000?

I don't have that answer. It could be that the processor in the GBox has logic built in to provide more counts to make it more accurate.

Maybe someone else has the answer.

 

[Anole]

I went to the roof twice to measure the actuator arm at the West limit and the East limit.
The distance travelled is 12.75".

Well, right there is proof you could move the actuator out to the side of the dish, make it travel farther, and have more counts, more leverage, and less slop.

The distance travelled is 12.75".
... that should produce 612 pulses at 48 pulses per inch traveled.
... the counts on the VBox III is 597.
... the counts on the GBox is 1148.

And this is quite consistent with the facts we know.

BUT, in using the GBox so far, it has landed on the Ku channels that I have used it on with much more accuracy than the VBox III.
I choose Ku channels and the weakest ones, because you have to be right on them to get good SQ.

So, this proves that whatever the Gbox is doing, is actually helping your accuracy.
Don't argue with success!

How do you account for all the extra counts in the GBox V3000?

I believe, as suggested earlier, that the Gbox is taking a count on each half of the count pulse.

Let's imagine you had a reed switch and one magnet passing by it.
The switch is normally open, but when magnet comes close, the reed closes.
Either end of the magnet will work, because the reed switch is not polarity sensitive.
Assume the magnet is the hour hand of a clock, and the reed switch is at noon.
So, as the hour approaches 10, the switch closes and stays closed 'till 2.
Once you get toward 3, the switch opens.
As the hour hand moves past 5...7...9... nothing happens 'till it gets back to 10.

Well, if your old Vbox only knows the hand is between 10 & 2, or more likely it only registers that you're at about 10, then that is the only position you have to count in moving your dish.
But, if the Gbox makes a count out of the 10 transition, AND out of the 2 transition, EVEN THOUGH they are not equally spaced, you still have more accuracy.
It's not linear and perfect, but you have two places to stop moving the motor.

Does that make sense?

 

[linuxman]

I believe, as suggested earlier, that the Gbox is taking a count on each half of the count pulse.

Let's imagine you had a reed switch and one magnet passing by it.
The switch is normally open, but when magnet comes close, the reed closes.
Either end of the magnet will work, because the reed switch is not polarity sensitive.
Assume the magnet is the hour hand of a clock, and the reed switch is at noon.
So, as the hour approaches 10, the switch closes and stays closed 'till 2.
Once you get toward 3, the switch opens.
As the hour hand moves past 5...7...9... nothing happens 'till it gets back to 10.

Well, if your old Vbox only knows the hand is between 10 & 2, or more likely it only registers that you're at about 10, then that is the only position you have to count in moving your dish.
But, if the Gbox makes a count out of the 10 transition, AND out of the 2 transition, EVEN THOUGH they are not equally spaced, you still have more accuracy.
It's not linear and perfect, but you have two places to stop moving the motor.

Does that make sense?

Yes, and very well explained!!!

As you said above, don't argue with success.

The next question is:

Will it improve the accuracy on other types of actuators, and possibly my Birdview with the 34:1 gear ratio when the newer motors have 70:1 gear ratios?

Or on George's actuator that he is having problems with?

 

[Anole]

Will it improve the accuracy on other types of actuators, and possibly my Birdview with the 34:1 gear ratio when the newer motors have 70:1 gear ratios?
Or on George's actuator that he is having problems with?

Well, I think you are in the catbirds seat with some pretty good hardware.
You don't actually have a reed switch.
I'm going to go out on a limb here, but based on the limited views I saw inside your motor, I think you have a Honeywell / Microswitch industrial hall effect magnetic sensor.

For the guys who do have a reed switch (does anybody, or am I in the wrong decade, here?) it's possible they close only between 12 and 1, not 10 and 2.
So, without knowing more, I'd have to say it's possible that the Gbox might do no good at all.
You'd still get your count at 12, but then again at 1, and none more 'till you got back around to 12 again.
It's that long run between 1 and 12 that benefits so well from another magnet.

If we had all these motors open here on the bench, I'm sure there a number of mechanical, electrical, and magnetic fixes we could apply to get more counts out of them.
But some might be easier to mod with an extra reed switch, some might be easier to double the magnets, and some might benefit from new circuitry.
Then, followed by the Gbox, you could go for even more.

I'd say the Gbox, when used in the right instance, is ONE way to get more positional accuracy.
And while not discounting the plug-n-play value of it, we might be able to electrically double counts on the old Vboxes.
edit:
You have already doubled the magnets, to go from 24 to 48 pulses per inch
You can change the motor mount to extend to 24 instead of 12 inches, doubling your pulses.
The Gbox seems successful (on this motor) at doubling your counts, possibly counting on each edge.
That's a: 2 x 2 x 2 = 8X improvement in possible accuracy

By all means, do put it on your Birdview, and see if you get an accuracy boost.
But, just because you get more counts, doesn't mean they're equally spaced.
(as for George, I'd have to see inside his motor before even hazarding a guess)
btw, how are the birdviews for Ku band accuracy , with their HH motor?

 

[linuxman]

By all means, do put it on your Birdview, and see if you get an accuracy boost.
But, just because you get more counts, doesn't mean they're equally spaced.
(as for George, I'd have to see inside his motor before even hazarding a guess)
btw, how are the birdviews for Ku band accuracy , with their HH motor?

I will leave it where it is for a thorough testing for at least a week bouncing back and forth for normal viewing.

When that is over, I will put it on the Birdview which will require a reset. Always good to find out how they handle resetting.

The Birdview dishes themselves do an outstanding job of Ku, and since I eliminated all the play out of the drive train mine has stopped pretty accurately since. I do need to get some guys to help me hold it steady while I re-tighten the cap bolts with it on the best signal.

If you remember, it moved on the cap last summer during a storm, and I haven't taken the time to put it back in place using a receiver at the dish with all the connectors in place.

I may try the trick advised in another thread about grinding down the bolts to a dull point and see if they will dig in a little better. If that doesn't work, I will have the welder come out in the spring and weld the inside plate with the threaded hole for the bolt to come down from the top to hold the cap firmly clamped down on the pole.

 

[Anole]

. . . any machinist would know . . .

I was wondering about one other thing.
Maybe, during your inspection of the dish while under control of various Vboxes, Gboxes, and your 4DTV, you'll be able to answer this one.

Approaching from the same direction all the time, will result in better accuracy.
Take our example above where the switch clicked in and stayed in/on from 10 to 2...
Well, if you approach going clockwise, then you'll stop on or soon after 10.
But, if you approach going counterclockwise, then you'll stop on or soon after 2.

Probably the best way to see if they do this, would be to use the remote and go clockwise one step at a time.
Then, go counterclockwise one step at a time.
If you go one way a step, and it's smooth, but the other way, and it sorta jerks around and seems not smooth, that might be good.
So, for counterclockwise, you'd want the motor to go one step past where you need, then clockwise to the final position.
But for clockwise moves, you just move and stop. You are done.

 

[linuxman]

Probably the best way to see if they do this, would be to use the remote and go clockwise one step at a time.
Then, go counterclockwise one step at a time.
If you go one way a step, and it's smooth, but the other way, and it sorta jerks around and seems not smooth, that might be good.
So, for counterclockwise, you'd want the motor to go one step past where you need, then clockwise to the final position.
But for clockwise moves, you just move and stop. You are done.

I have not noticed any jerkiness one way or the other.

The point where the movement is most critical is on the East side of the arc especially with an actuator. After the dish reaches 87W, the pulses between satellites gets shorter because the arm is extended more and allows greater movement with each click. Once you get to 72W it gets worse, and by the time you get to 43W Ku there are only one or two clicks allowed to be perfectly on the best signal, the arm is fully extended and the dish back plate is at about 1:30 or 2:00 if pointing at the face on a clock.

Today, I swung it all the way from 125W to 43W, and it stopped right on target for the Ku signal. That's a lot of travel with the weight of the dish behind it on the downhill side, and it stopped in the right spot.

 

[Anole]

the Birdview motor

The reason I really enjoy brainstorming with you, is that you clearly document what you observe or measure.
And when I can't envision from words, you post pictures.
It's almost as good as being there! - :up

The following are some counts (pulses) as were counted by the VBox III and the GBox V3000.

This is from about 10 pulses East of 43W (IS3R) to about 10 pulses West of 143W ....

On the VBox III after adding the extra magnet to double the counts (pulses) the total pulses in the above arc is 597.

Switched out the VBox III and substituted the GBox V3000, and in the same arc as above, the total pulses is 1148.

About two to one... beautiful data to work from . . .

Just for comparison, I have a VBox II on my Birdview dish and run it all the way from 30W - 142W with a 34:1 gear ratio motor (less pulses) and get 907 pulses with it.

Okay, this info is just gold!
Above, you say for similar (not exact) travel, the double-magnet 24"+Vbox gets 600, and the Birdview gets 900
So, what did the 24"+Vbox get before double magnets?
Did someone say 300?
Geez! - - No freekin' wonder you thought the Birdviews walked on water!!!
You'd been comparing 300 count 24"+Vbox on regular BUDs to 900 counts on the Birdview.
Apparently, the 34:1 gear ratio was pretty spiffy in it's day.
Who knew!?!?

If the Gbox takes you to the neighborhood of 1800 counts on the Birdview, you'll be just insufferable with your pointing accuracy, and bragging rights!

 

[linuxman]

Did someone say 300?

About 300 was about right before the double magnet, but that was only on this actuator. The name says it all. VW-24LC. The Von Weise guy said the LC stood for Low Count. Why they wanted to produce one with low count? I don't know.

The V76-5 produces 48 counts per thread and so would be the same as this one is now, and that is what I have on my 4DTV, my family 4DTV and would have used on this dish if I hadn't picked this one up because it was almost in new condition.

It was still a set up from the V76-5s to get the Birdview and the Pinnacle because they are gear driven around a shaft with the same precision in the center as on the ends.

You'd been comparing 300 count Vbox on regular BUDs to 900 counts on the Birdview.

Actually comparing 600 count VBoxes on regular BUDs to the 900 on the Birdview. If the Birdview had the 70:1 gear ratio motor, there would be even more counts.

The big winner is, of course the AJAK H-H mount on the Pinnacle which boasts a whopping 1300 counts from limit to limit. See Pic:

 

[gabshere]

i haven't ordered parts yet but , i'm tempted to oder the g-box first to see if it will track acurately for me. I questioned Sadoun about the pulse counts and he's supposed to be checking on the 24" & 36". But the post by truckracer doesn't sound promising he has a 36" superjack.

even with ordering the g-box i need it for another dish. i can always go back and attempt to add the second sensor to the 18 inch actuator. I haven't tested it but my other actuator is a von weise V76-5 and has the correct amount of pulse counts to be accurate.

I will also have to re-read this thread several times to get all the techno-discussion to sink in lol

 

[linuxman]

edit:
You have already doubled the magnets, to go from 24 to 48 pulses per inch
You can change the motor mount to extend to 24 instead of 12 inches, doubling your pulses.
The Gbox seems successful (on this motor) at doubling your counts, possibly counting on each edge.
That's a: 2 x 2 x 2 = 8X improvement in possible accuracy

I just read back through the posts for today and finally saw your edit Anole.

I was thinking about your extension theory earlier myself. It won't be too hard to put an extension on the backing ring because it has two points to attach the actuator. One on the East side and one on the West side. It will be easy to put a flat piece of steel all the way across from the East to the West with a 2 or 3 inch extension on the West side for more travel, and more counts for accuracy.

A three inch extension will give me how many more inches of travel? I am not very good at math, so can someone run the numbers?

Here is a picture of what I am proposing.



Hey George,

I will also have to re-read this thread several times to get all the techno-discussion to sink in

There's no techno-discussion on my part. I am the guy who gathers data and makes observations.

Anole, Caddata, and some of the rest of you are the techno guys.

 

[Anole]

experimental is good, too

Without knowing how far apart your existing east and west motor hookup points are...
...and how far from that white line the motor base is connected, I couldn't hazard a guess.

Let's say it's 18 inches across the mount points on the back of your ring.
I'd make my bar that long, plus half again (18 + 9 = 27")
Then, I'd put mounting holes for the end of my motor about every 3/4 inch out to the end of the bar.
But, that's just shooting from the hip.
Use your own best judgment.

edit: oh, I just thought of another consideration
Minumum and maximum extension of the motor may require relocating where the heavy end of the motor attaches.
We sort of need to really know what part of the motor throw you're working with now.
In other words, what is the minimum length and maximum length now, when you are east/west.?
...and how much more throw does the motor have at each end of that run??

 

[dem0nlord]

A little more data guys, as I've been fighting a similar situation here.

Currently with the 24" superjack, I have 913 counts from just on each side of 37.5W and 139W. That comes out to 38 counts per inch.

Previously I had a new 36" Thomson-Saginaw on there, main reason being wanting to try to get over to Hispasat. I never did look to see how many counts it was for the whole arc, but it did seem to have a bit of a problem keeping accurate positions, I attributed this to it probably just being too fast attached to the dish in the same position as the 24" was. This was when I had the DSR922 driving it. It wasn't terrible, but sats did need readjustment occasionally.

Then I got rid of the 4DTV and went strictly FTA. The first thing I noticed is no matter what I have hooked up, none positioned the dish as accurately as the 4DTV did. Some things like the Manhattan were downright terrible, I know it was missing counts, and I attribute that to the 200' cable run, it worked fine hooked up out at the dish, however the 4DTV always worked right, so that's a Manahattan limitation. Then I switched to this Satcruiser, and it does an interesting thing. In any case that it misses the targeted stopping point, it actually reverses the motor and comes back to it. This is the only box I've ever seen do such a thing. Things still weren't great, and switching back to the 24" superjack helped a ton. Big joke there, cheap thing like that outperforming the $400 Thomson-Saginaw, but again, I think it probably needed slowing down, and I had the same thought about the dish attachment point Linuxman has, as we have the same dish. Unfortunately the Thomson-Saginaw ended up bending the tab on the west side some. Some would say you idiot why didn't you set the limits, well you can't, it doesn't have adjustable limits, only a clutch which is supposed to disengage it before doing things like bening parts.

So I've been following with interest about the 2nd magnet and the results, I just think the receiver or vbox seems to be equally important, and nothing as repeatable yet as the 4DTV and old school analog boxes.