Titanium Satellite Polarity control case

[KF4YLM]

Here's a fun one.

Linkbox 9000iLocal
Ti ASC1
250' of cable
Uniden 10' w/ VonWiess linear mover
Chapparal C/Ku servo-driven feed

Worked great up until a few weeks ago. Started losing servos -- or apparently so -- polarity control would disappear.

Swapped three of them. Then I took a pull and hooked it directly to the ASC1 terms. It was pulsed into gear strip! Tore the gearbox cover, re-indexed. Worked fine.

Fixed the other two this way, swapped the one on the feed. This time, to aid my diagnosis, I connected both the servo on the antenna and one in the house -- in parallel. This allowed me to observe the servo's behavior.

It turned out that moving the linear actuator (changing the dish longitude) developed a slow-decay parasitic voltage on the servo's pulse wire! A quick bump east or west simply bumped polarity like a windshield wiper. A long travel (more than 5 Reed clicks) sent the servo into gear-strip.

Wires on this system are a mix:

Leg 1: Cat6 shielded takes all signals from the ASC1 to the junction / TVSS block in my basement. All signals (m1; m2; s1; s2; gnd; pulse; +5vdc).

Leg2: motor wire m1;m2 ride AWG 10 solid dir bur pair. S1; S2 ride Cat6 shielded WBU/BU. Servo gnd on WOR/OR parallel for reduced loss. Servo pulse on WGN/GN parallel. Servo +5vdc on WBR/BR.

Leg 3: At the antenna mast, the leg2 wires hit a junction. new cat6 shielded wire 1 runs to the motor, carrying M1;M2;S1;S2. New Cat6 shielded wire 2 runs to the feed, gnd;servo;5vdc.

Placing the servo at the dish mast junction didn't change the phantom pulse voltage. Placing it at the basement junction, or directly behind the ASC1, oror dissconnecting the servo all together squelched it.

Removing motor wire from ASC1 and applying E/W command pulse squelched it as well. Disconnecting all servo wires from ASC1 and measuring the wires showed linear motor action deffinitely created a slow decay voltage... 1.2v was observed after a 10 second motor run, and with only the servo itself connected to the wires at the antenna, with no connection back in the house, measuring the wires, the voltage took almost 7 seconds to fully discharge. Servo exhibts about 20k ohms between gnd/pulse wires.

It was apparent that something capacitive was going in between the motor wires and the servo wires. The loading of the servo pulse circuit isn't enough to squelched it alone. Since we don't want to disturb the pulse waveform, I elected to add some serious inductance to the motor wires at the point where they connect to the ASC1. Thought being, if there is excessive stray C (reactance) involving these wires, we should be able to cancel that reactance with series L. THIS GREATLY REDUCED THE PHANTOM PULSE! There is still enough getting on the pulse line to "bump" the servo. But it comes right back. Motor performance doesn't seem to suffer with this added L in the M1;M2 lines.

Hoping to get some further testing underway (including TDR and impedance measurements on my power control lines to my antenna). Will update when I have more data. Theories abound, but I don't want to invest a lot in them until I get more data on what is going on here.... and why this condition very suddenly appeared.

99% SURE It does NOT have anything to do with the ASC1.... I call it out above on because many of us use this controller with our antennae for longitude + fine polarity control (and it works). Indeed this problem would only show up on a system so or similarly equipped. I suspect a wiring change or motor behavior change to be at fault.

 

[Brct203]

I can think of several potential problems here:
- Cat6 is too thin for the motor wires. I'm not sure that it has an impact on this specific problem though.
- you have the servo and the sensor wires alongside each other without shielding between them, and in Leg1, you also have the motor wires. That might cause enough electromagnetic noise from the pulse sensor wire to cause problems on the servo wire.
- you are hinting to the fact that Leg2 is maybe buried - you might have some damage on that section of Cat6 that is causing problems
- Cat6 is twisted pairs, that might indeed be causing some inductance problem

 

[Titanium]

Not so fun.... but easy to solve.

Agree 100% with Brct203. Note in the ASC1 manual that shielded servo and shielded servo wires as well as much heavier gauge motor wire is recommended. Maybe the M1/M2 wires overheated in the Cat 6 and started melting insulation or ??? Thin gauge wires under load make great heating elements!

The motor power in your system is coupling with the servo circuit within the unshielded servo distribution wiring. Servo distribution shielding is a very common problem in robotics and there some great Google referenced articles on servo shielding and grounding.

Cat 6 is not suitable for BUD operation. The M1/M2 wire gauge will yield excessive voltage loss and lack of servo and sensor distribution shielding isolating each run will cause many problems. If the Cat 6 shield is connected to GND, it is not shielding, but rather it is capturing the motor power and sensor cycle noise inside the distribution cable. If you want to continue to use the cat 6, maybe use only for the servo circuit or only for the sensor circuit distribution?

For M1/M2 distribution, I would run an appropriate gauge low voltage (example): www.homedepot.com/p/Southwire-100-ft-12-2-Black-Stranded-CU-Low-Voltage-Landscape-Lighting-Wire-55213443/202316562

For S1/S2 and Servo distribution (example): www.homedepot.com/p/Southwire-By-the-Foot-22-2-Gray-Stranded-CU-CL3R-Shielded-Security-Cable-57572199/204725140

For S1/S2 distribution, connect the shield only at the ASC1 GND terminal and leave open at the actuator. For the servo distribution, use the shield as the GND and connect on both sides.

In addition, you may also find that a capacitor across the motor terminals at the actuator may suppress motor run noise.

 

[KF4YLM]

I hear y'all on ragging me on the wire choice. . No prob. Here is some more detail... It is not quite so ugly as 100% cat.

Leg 1 is about 15' (ASC1 TO BASEMENT TVSS JUNCT) .as described 1 run Cat6.


LEG 2 is abt 225' (HOUSE TO ANT MAST). Here, motor hops off to a larger gauge -- direct burial Romex, #10(first post was wrong).

Leg3-1 is longer to the servo (20'). Separate leg3-2 to the motor is shorter.

Everywhere Cat6 carries motor power (except leg1) it's on two pairs. Measuring 38vdc at ASC1, 34.6vdc at the mast junction box. Wide open R on all conductors when open at antenna. Less than 2 ohmswhen paralleled at antenna.

So the only spot where M(x) and Servo(x) are /not/ shielded from one another is Leg1. Place a servo here at leg1-leg2 junction, parallel one upstairs, behaves.

Sensor wires do follow the servo wires all the way to leg3... So I can see a C path maybe here..... Sort of. But leg1 would be an easy enough run to break M(x) out on separate shielded conductors. Agree that is a more appropriate plan. The XYL is used to the wires now so she probably wouldn't mind +1.

Bugger is that this wiring scheme worked great for over a year ...

Seriously, thank you guys for your input here. And yes.... Problems like this are fun to me.

 

[KF4YLM]

Following up.

Ran new Cat6 STP dedicated to servo action 100% from ASC1 to servo. Pairs were doubled for each line (5VDC = WhBu/But, Pulse = WhOr/Or, Gnd = WhGn/Gn) to keep voltage drop in check over this long run. Super clean action now.


Thanks Ti and EST for the the on this one.