random faults with isolated RS-232 hub. Grounding issue???

[newbFTE]

Hello all, this is my first post and it is a long one, sorry.

I am having an issue with the use of an optically isolated RS-232 to USB four port hub.

I am using the hub on an aircraft to monitor RS-232 traffic transmitted from three separate RS-232 devices. The data is sent from the devices to the aircraft EFIS. Theses devices are hard wired into the aircraft electrical harnesses and have a power and ground wire for each device, but they only have the serial Rx and Tx wires, there are no other wires on the devices to connect. So each device has power, ground, Tx and Rx with power coming from the aircraft avionics bus.

The Tx of each device connected to an Rx pin on the isolated hub. The Ground pins for each port are connected to the EFIS as that is all that is available for ground in the connector that we are tied into. The longest RS-232 runs is about 15 feet.
This works, I have repeatedly received the same data as transmitted by the devices. My tie in does not cause the loss of any aircraft bus traffic.

The isolated hub is mounted to the airframe with rubber mounts (for protection from vibration, all devices are mounted this way), and is powered only by the USB cable. The USB is connected to an Ultra Mobile PC (UMPC) which is strapped to the flight test engineer's leg. For various reasons (think risk assessment here), the UMPC runs off an internal battery and is no way connected the the aircraft (meaning no grounding what-so-ever).

We are getting random and sporadic errors. Some of these seem to be alleviated for a short time by disconnecting and reconnecting the USB cable, but then the fault returns. The software on the UMPC (LabVIEW compiled program) does not error out and does not loose the connection with the com port. It keeps pulling data out of the buffer. However, I get alarms such as random failed checksums (about 10% failure rate), and I am receiving garbled bytes that cause data spikes when decoded, yet in these rare cases, the checksum passes (about 1% of the data). Oddly enough, when the fault has occured, it has only been on one device at a time. I may loose GPS data one day, and I may loose AHRS data the next. (by now you may have guessed that this is an aircraft bus monitor that I are trying to build)

I am wondering if this error could be grounding related. The isolated side of the Hub is connected to the EFIS ground, but the computer side of the hub is not grounded in any way. It is free floating.

I can ground the UMPC (I think) by plugging a dummy cable into the power socket and connecting the ground from the power socket to the airframe ground. Should the isolated hub be grounded as well? How could I go about doing it? Do you have any other advice as to the source of the bad data frames?

I am hoping this is the right forum to post to. I imaging instrumentation engineers would have the most experience with RS-232 devices.

-Chris.

 

[djs]

Another issue may be how the USB hub multiplexes the four serial lines. As a test of this and also potential ground issues is to isolate your serial devices one by one. In other words, disconnect three of the Rs232 devices and run with just one and see if you have problems. Then disconnect that device and connect the next and so on.
Also how is the USB hub isolated? Can you do testing on the ground or does it have to be in the air? If on the ground, you might connect an o-scope to the RS232 leads and take a look at the signal levels. Which brings up another issue, that is RS232-C signal levels. Many modern devices do not have a full voltage swing as specified in the RS-232-C standard. This includes many comerical computer products.

 

[newbFTE]

The hub has built in optical isolation on each of the ports. Problems in flight are sporadic. I have not yet seen this problem on the ground, but knowing what to look for, I can try to identify it on the ground.

There is no way to access the hardware in flight, it is out of reach in a cramped cockpit. The UMPC runs XP and the acquisition program is controlled using the touch screen. It would even be a challenge to access hyperterminal in flight to view or record raw data. I thought of mirroring the data on UMPC to a virtual com port and running a logger on that to capture raw data, but I am that would slow down the acquisition too much, that has been my experience.

I'll check with the device vendor for the output voltages on the RS-232 line, but I am guessing it is pretty low.

 

[itsmoked]

If the hub is really optically isolated the grounds on either side of the hub should in no way need to be referenced to each other.

On the other hand unless this is some wide-body I would expect there to be zero problem with using a non-isolated hub. There is no reason for those various RS232 output instruments to have widely varying ground potentials.

Replace the isolated unit with a non-isolated unit on the ground and try it. If it works,(and it will), use it on a flight. If you have the same problem then you need to look elsewhere.

Keith Cress
kcress -

http://www.flaminsystems.com