OTHER Improving the DM2800 Opticum H-H Motor (removing the mechanical slop)

[sat_cat_25]

The DM2800 H-H motor is capable of handling a 1.2m GeoSatPro offset dish, but it may take some work to get it to point reliably. The large amount of mechanical lash, backlash or "slop" in the positioner will make it difficult to aim properly. I struggled with this for a few weeks before deciding to tackle the root problem once and for all.

The slop allows the dish to tilt up and down and rotate side to side. A stiff wind can easily blow you off target. At max elevation (due south) the slop makes peaking the dish unrepeatable. And near the ends of the arc, the weight of the dish will take up the gear slop and point you further E or W than you expect to be based on the USALS calculated position.

The manual mentions a simple screw adjustment that will make the worm gear mesh tighter with the plastic bull gear on the output shaft. This may make some improvement, but you will quickly find that it increases friction to the point that the little motor either 1)won't turn the 1.2m dish or 2)won't pull it back up from E or W. Most of the slop is NOT due to this gear interface anyway.

Below are a few things I did to completely remove all mechanical slop or play in the mount.

The above made a huge improvement. But it still wasn't rock solid. I was missing the hidden adjustment below.

The pole mount bracket that this thing comes with is also junk. It makes it hard to make small azimuth adjustments, and the simple act of tightening the mount will move the azimuth. I took an old 1.2m Prodelin mount and bolted on the side plates that came with the DM2800 bracket.

This thing is rock solid now. Enjoy.

 

[Hyper Casey]

Excellent post sat_cat_25. Backlash or "slop" on a H-H motor is so annoying, and has been a problem on so many models for years. It is too bad that this DM2800 Opticum motor needs to taken apart and modified like this, but on the other hand, there is not a lot of choices out there as far as H-H Motors, especially for one that will TRULY move a 1.2M dish. The first 2 steps are easy enough, but the last thing, most folks wont have a 1.2m Prodelin mount sitting around, so they will have to get creative.

 

[thatnewguy]

This post was very helpful, but I made some slight changes:

1) Instead of using JB weld, I used a thin shim of aluminum cut from a Red Bull can (they're thinner than most soda cans.) It was about 20 cm by 10 cm. Place it before you mount the arm, and made sure it's staying in place and you're not driving it into the bearing when you mount the arm.

2) Properly tightening the through-bolt CAN work, as the mounting arm is split. The shim helps push the halves apart, so proper bolt tightening actually results in compression of the arm halves into the soft shim and onto the drive lug. You'll draw the two halves closer together - BUT it has to be done properly. The drive axle is threaded, so first you must tighten the through bolt into the drive axle threads using a 6mm allen wrench. That draws one half flush with the drive lug. THEN you must tighten the nut onto the bolt using a thin walled 13mm socket - that draws the other half flush. Improper tightening in the wrong order will definitely lead to slop. This method allows you to remove the mounting arm from the drive axle in the future if you need to service it yourself.

3) I didn't need to tighten the long adjustment screw to draw the worm gear assembly ends together. I didn't find slop there.

Using these methods, I was able to reduce the slop to less than I could perceive by manual handling and twisting of the motor. No more slight 'click;click' as I twist the mounting arm.

A side note: I found the plane of rotation of the motor didn't match the etched markings for the elevation angle on the supplied mount. I found I needed to reduce it by about 0.8 degrees more than what the markings showed.

 

[a33]

A side note: I found the plane of rotation of the motor didn't match the etched markings for the elevation angle on the supplied mount. I found I needed to reduce it by about 0.8 degrees more than what the markings showed.

to reduce it by about 0.8 degrees more... so less elevation, or more elevation? It is formulated somewhat confusing to me...

By the way: 0.8 Degrees less axis elevation might also partly be caused by the needed forward axis tilt angle of about 0.6 degrees (for latitudes between about 30 and 50), of the so-called 'modified motor angles'.
With those you can follow the Clarke Belt better than with the 'traditional motor angles'.

In that case you have aligned the angles very well.

Greetz,
A33

 

[thatnewguy]

to reduce it by about 0.8 degrees more... so less elevation, or more elevation? It is formulated somewhat confusing to me...

By the way: 0.8 Degrees less axis elevation might also partly be caused by the needed forward axis tilt angle of about 0.6 degrees (for latitudes between about 30 and 50), of the so-called 'modified motor angles'.
With those you can follow the Clarke Belt better than with the 'traditional motor angles'.

In that case you have aligned the angles very well.

Greetz,
A33

Somehow, english is no longer my first language when I've been working long hours.

The answer is - it depends on which scale you use (they're opposite on each side of the mount.) I'm at about latitude 36, but I read the 54 degree scale because it's the one I can reach easily without maneuvering around to the other side. So '54' was in my head and small brain produced the words 'reduce more.'
I have it set at 53.2. That means the other side should read 36.8, if I bothered to worm my way around to the other side. That's really a bogus number, because the scale is much more coarse than that. Claiming that I can read tenths on those thick lines is silly.
What I can confirm is that the plane of the rotation of the motor is correct, and it doesn't match the stamped scale exactly. I strongly suspect that someone else using the same mount may find a different error depending on how well the scale stamper was aligned on that day. I should also note that the plane of rotation of the motor is not parallel to the main surface of the case - it's sloped. It's parallel to the case half join lines.
In any case, get out your digital level and confirm.

The dish is in a non-permanent location without good H-H visibility, so I'm not sure how well I'm following the belt. I'll find out when I get it up on the roof next month and have true H-H visibility and I can measure angles at opposite ends of the dish arc.

There is still a very small amount of slop, but that's just part of the rest of the dish mount. I added a fine-tune addition to my dish movement script, and it checks each side of the expected position in 0.1 degree steps for the best signal. Before I took out the slop, the correction could vary by 0.5 degrees from one movement on one day to another. Now I find it's consistent within +/- 0.1 degree. I'm estimating the 0.1 step; it's likely a bit smaller than that. Every little bit helps.

PLEASE LOG IN TO GET RID OF THESE ADS!

 

[sat_cat_25]

I added a fine-tune addition to my dish movement script, and it checks each side of the expected position in 0.1 degree steps for the best signal.

Script? Tell me more about this controller. Home built? Arduino or ?

I recently built three Arduino based dish positioners that control the actuators, take feedback from az and el absolute encoders, and has a web interface to monitor/control it all.

 

[thatnewguy]

Script? Tell me more about this controller. Home built? Arduino or ?

I recently built three Arduino based dish positioners that control the actuators, take feedback from az and el absolute encoders, and has a web interface to monitor/control it all.

Sorry, I thought it was in this thread...but it's in the other thread.

It's a TBS6909X in a linux box that is waaay overkill, but the server does a lot of other things for me as well. It had an empty slot and I could relocate it to be within a 8m cable run of the dish, so it now also runs the dish.

I've reused some open source dvb code but a significant chunk of it has been rewritten to serve my particular needs. The algo for the fine tune is pretty straightforward:
Go to your reference point using USALS.
Move 0.1 degree farther from center. I used timed pulses. At 2.9 degrees per second with 18V, I decided on a 25ms pulse. You can feel the motor emit a 'click' (1/40th of a second!) when it moves, but you can't really see the dish move. But on weak TPs, you'll see the difference at each step.
Move away from center until the signal on your favorite TP degrades (which may be immediately, but keep going until it does.)
Once the signal drops, start moving back up in 25ms pulses. The signal should start increasing on each step. As soon at it drops on the other side of peak, move back 25ms and stop.
It takes a few extra seconds, but I can get better locks in worse conditions, and it's just built into my script to move to a new bird. I don't notice it.
I haven't built an elevation controller yet, but I'd like to.

I'm building the web page monitor/controller for the whole system this weekend.

 

[thatnewguy]

Also, the shim was 20mm by 10mm. Duh. Funny how the brain works at 5am: "I made a typo on that post 2 days ago..."

 

[a33]

What I can confirm is that the plane of the rotation of the motor is correct, ...

How did you check this?

Because it looks like you are not sure yet that you are following the Clarke Belt correctly (which would be the ultimate proof that you have the axis-elevation/rotation-plane set optimally):

... I'm not sure how well I'm following the belt. I'll find out when I get it up on the roof next month and have true H-H visibility

Did you check with an inclinometer or so?
Checked against 'latitude angle', or against 'modified latitude angle' (see my post about the 'forward axis tilt', above)?


BTW. Nice, this precision alignment!

Greetz,
A33