I was just composing this post when you posted yours so here goes.
I was reading up on RS485 and came across this, it is considered a bad way to connect RS485 multi drop network. To me it looks like one of the ways Dynon says you can hook things up
Yes, Dynon says it will work anyway you can connect the wires and it will, but it will be less noise immune....(as you know)
The daisy chain method would mean a failure in one component would fail down line units.
No...The buss can be daisy chained and connected to the terminals of each device without actually passing thru the device so if that device fails, it will have no effect on the buss. Think of it as a really short stub connection to the buss which is OK.
I have my DSAB master and all other components going into a DB25 connector. The opposite side has the top pins connected with a piece of wire soldered to each pin cup, the same with the bottom.
You are connected in a Star configuration. This will work if the arms of the Star are not very long. I assume that the longest arms of the Star are for the Roll and Pitch servos and if so, I would install 120ohm resistors at the ends where the servos connect to the buss. One on each of the long arms of the Star. Do not put more than two resistors on the buss.
This is how I connect all DSAB "A" and "B" wires. Is this considered parallel setup as in the figure above? I read about terminating resistors, I kind of think that is built into the Dynon units, but my DSAB connection scheme might be my problem.
Most likley the Dynon devices do not have resistors built into them. There should only be two resistors at the ends of the buss line. If Dynon had one on each device, the impedance of the network would be too low and dependant on the number of devices connected.
I thought my setup made the connection that Dynon wanted "each wire electrically connected". When you say star, is this kind of the same thing as what I am doing.
Dynon's advice does not agree with the industry standard RS-485 specification used by thousands of devices in the automation world. Dynon did not invent the RS-485 buss, they just used it....
How did you connect the DSAB.
I used a Hybrid method and did not use terminating resistors but I can tell you that I have seen problems with noise as well. Not the ones you are having but others. I have used my DSAB bus as a way of finding the noise sources and fixing them at the source in my plane. I plan to rework my DSAB wires to fully follow the RS-485 specification and standards. This will ensure the best chance of noise immunity.
All of my Dynon units except the servos ground at the instrument panel ground bus, this bus grounds at the forest of tabs via 5 wires that run from the instrument ground bus to the forest of tabs, you can see the ground bus below the DSAB hub in the picture below.
If all of your DSAB devices do not share a common ground, it can lead to issues with the RS-485 differential signal. This third wire must be common to all the devices on the buss. There could be a significant difference in potential from the airframe ground at the servo to the ground at the forrest of tabs. When you transmit, there most likely is also many ground loop currents flowing around in the airframe as well. All of this could be changing your 3rd wire ground reference on your devices.
My servo DSAB wires are joined together under the floor and one set of DSAB runs to the DSAB DB25 Hub, all other wires from the servos are also joined here. Do you think this might be the issue?
You are branching arms of your Star which is also not recommended.
Maybe running each servo DSAB to the hub would be the answer. Did you run your servo DSAB together like I did, the instructions said this would be OK, that the wires could be joined anywhere along the way.
Again, Dynon's advice does not follow the spec. I used the Stein bundle which twist all the wires into a single cable. This seems to work OK. Both of my servo DSAB wires come all the way to behind the panel where they become the long arms of my buss.
I don't think my issue is with the servos though because the drops I get are with the AP74 first then HS34 if transmission is long enough.
You really can't tell where the issue is by what device drops first. The software may be more error tolerant in one device vs the other. The RS-485 interface chip in the device might be different in one vs the other ect.
The HS34 DSAB wire is shielded, not sure where the shield terminates, the main panel harness was done for me by John Stark, I did the AP74 wires myself.
There is no real reason to shield RS-485 wires since this type of network operates on a differential type signal. If shielded cable is used, make sure it has an impedance of somewhere around 120ohms to help prevent issues. Ground the shields at one end only.
I have some shielded 2 wire on the way thinking this might help, but where do you attach the pigtail from the shield?
One end only, does not matter which in this case but I would recommend the forest of tabs.
You can direct email me at rvg8tor "@" comcast dot net if you would like