positioning accuracy

[LocutusOfBorg]

I figured that I'll start a new thread on this subject.

I just installed some angle iron but I only bolted it on for testing. I found that I got an additional 1.8 pulse counts per degree, only. That's an additional 3.6 pulses between sats. That's not very impressive to me. The shaft was clamped down an additional 13 inches after the angle iron was installed. I could adjust it out an additional 13 inches but I can't justify the additional effort just to get an additional 1.8 pulses per degree. For 2 degree spacing of the sats, this would give me an additional 7.2 counts separation of sat positions (if I were to adjust out an additional 18 inches). And I didn't see an improvement in pointing stability, either. On the positive side, I got a greater view of the arc and the angle iron didn't bend.

After just checking, I keep getting pulse count errors of 10 or more. Everything on the dish is secured so I'm considering replacing the reed. There's got to be a way to get more accurate pointing without it costing "an arm and a leg". I wonder if there's a reed that I can get that'll double the counts. Or maybe I should go find a better actuator? I don't recall seeing any reviews on actuators in our forum or on the web.

I once alluded to pulse-modulating the motor power, but nobody seemed to care.
Having experimented with model trains and my own pulsed supply, I can attest to their usefulness! -

We crept along the engine below the visible motion threshold.
And at full torque!
Only way to see it move, was with through closed-circuit video, with the camera aimed at a dust mote on one drive wheel!

The problem with moving dishes, is that there is currently no controller using PWM technology on the market.
And making one from scratch, would be a tall order.

edit:
I'd like to go one step further.
For an H-H motor setup on a BUD, I'd like to use USALS.
It'd just take setting up one variable: the number of steps per degree.
The rest is all left to the receiver to calculate (if I remember correctly)!

Maybe pulse modulation would help with improving pointing accuracy. If there's something available that has been thoroughly tested - I'd be interested in hearing the details. Maybe next time I'll consider buying an HH mount. Would that be more accurate?

 

[gabshere]

After just checking, I keep getting pulse count errors of 10 or more. Everything on the dish is secured so I'm considering replacing the reed. There's got to be a way to get more accurate pointing without it costing "an arm and a leg". I wonder if there's a reed that I can get that'll double the counts.


Or maybe I should go find a better actuator? I don't recall seeing any reviews on actuators in our forum or on the web.

a reed just makes it count no double .... the magnet wheen makes the reed count ... but i have though but never tried it .... is placing a second reed switch in there so it will double the count ... this might work

Maybe pulse modulation would help with improving pointing accuracy. If there's something available that has been thoroughly tested - I'd be interested in hearing the details. Maybe next time I'll consider buying an HH mount. Would that be more accurate?

no , no , yes anything that has more counts is more accurate so a magnet wheel with 6 magnets is less accurate than one with 12. and so on but there is a point to where its too much ... so i would bet 30 to 42 clicks / counts per inch is where your at now and you want 48 to 60 clicks / counts per inch ( all this is from memory and i've slept a few times since i worked it up )

 

[1captain]

but there is a point to where its too much ...

Please tell me what is too much, I would like to know.

all of my home brew reed set ups, I do try and strive for a high count, and I thought the same thing, but never have reached that point as of yet. now when I set up my ChannelView last year, and was test running it I thought it was going to be too much with almost 8000 counts from T-12 to AMC-8. but after it has ran for close to a year and I have never needed to touch up any of the satellite locations, and it has always been spot on. so what is too much?

 

[gabshere]

well i was talking about too many counts per inch . where movement is so small that it would take a counter / mover forever to count. And i don't know what the limitation of the counter ( say it as a v-box , g-box remembers the count locations)

so if the v-box has three # positions and goes from 000 - 999 thats a thousand count positions . I know the g-box which has 4 number positions 0000- 9999 so that 10 thousand count positions. I don't know what the limitation of others are but you would be limited to that many count locations on those . or am i thinking to far out there

 

[B.J.]

I think that more important than the number of counts per revolution are (1) free play in the gears or the mount itself (2) lost counts or extra counts due to poor connections or poor wiring or due to the reed itself sticking or bouncing and giving extra counts.
I'd spend more effort trying to make sure your wiring is good and your connections are clean and dry and to removing free play in the mount, and I wouldn't spend much time trying to re-design the pulse mechanism. Most actuators and H-H mounts are capable of sufficient accuracy provided everything is working right. I think my current dish is about 1/10 degree per count, but I'll still end up a degree or so off much of the time due to other factors.

 

[1captain]

Thanks for the reply gabshere, I just thought that you might of known the magic number, my G-box handles it fine, and so did a DSR922 when I was setting it up with that high of a count.

and Yes B.J. you are right on with all that you mention. now their are several areas that a person needs to look at when it comes to poor accuracy of a H-mount or actuator and you cover them all real good.

 

[LocutusOfBorg]

I think that more important than the number of counts per revolution are (1) free play in the gears or the mount itself (2) lost counts or extra counts due to poor connections or poor wiring or due to the reed itself sticking or bouncing and giving extra counts. I'd spend more effort trying to make sure your wiring is good and your connections are clean and dry and to removing free play in the mount, and I wouldn't spend much time trying to re-design the pulse mechanism. Most actuators and H-H mounts are capable of sufficient accuracy provided everything is working right. I think my current dish is about 1/10 degree per count, but I'll still end up a degree or so off much of the time due to other factors.

I guess you could say that I was getting .092 degrees/count (10.85 counts/degree) versus what I'm getting now with a re-positioned actuator arm which is 0.079 degrees/count (12.65 counts per degree). Anyway, it seems that my dish has stabilized since I haven't had to re-peak since yesterday afternoon. I don't know what changed. Maybe I'll go ahead and extend the actuator mount point out as far as is possible to get best possible performance. But, since I don't know how to weld, I'll have to have someone come over to the house and permanently weld the angle iron onto the mount ring.