[simpits-tech] More on USB

[Matt Bailey]

On Mon, 15 Sep 2014 06:56:37 -0700 (PDT)
geneb <geneb at deltasoft.com> wrote:

> On Mon, 15 Sep 2014, Matt Bailey wrote:
> 
> > Just ran across this article while researching USB interference via
> > inductive loads (relays, motors, etc), very interesting:
> >
> > http://www.controlanything.com/i.htm
> >
> > Basically, they are saying to avoid USB for ANYTHING that involves
> > any sort of inductive load. I knew USB was finicky, but I had no
> > idea there is exactly ZERO ERROR CHECKING/HANDLING for the USB
> > spec! No wonder we have so much grief with USB-based devices!
> >
> > Amusingly, they recommend going to good ol' RS-232 as one option for
> > controlling inductive loads.
> 
> Inductive loads shouldn't matter - if you're not isolated from the
> USB interface, you've got more problems than just that.
> 
> Note that I've never had any issues with the USB stuff I work with.
> 
> g.
> 

As I understand it, there are two different situations going on here
where inductive loads OR high voltages can wreak havoc with USB:

1. The USB device is being used in close proximity to an inductive
load, such as a relay or motor. I have only observed this once, when I
was activating a relay. One (and only one) of my Arduinos would kick
offline when the relay was switched on and off (I don't recall if it
happened on energizing, or de-energizing). The Arduino was NOT connected
to the relay, that was being controlled by a Phidget I/O board (which
did not kick offline). Presumably, sufficient EMI was produced by the
arc when activating the relay, or perhaps by the magnetic field
collapsing when the relay was de-energized, to interrupt the Arduino's
USB connection. Your point about isolating the USB interface from the
load is a good one, but I don't believe that is what was going on here.

2. The USB cable, or wiring connected to the USB interface, is near
some high voltage AC stuff, such as 115v 400hz aircraft power for
lighting. I have not observed this myself as I am not dealing with
400hz yet, but I have been told by other folks that it's a nightmare.
My understanding is that the magnetic fields produced by the AC
voltages in the aircraft lines pass voltage to the wires connected to
the USB interface, delivering potentially high voltages to the device. I
believe this would also be considered voltage via inductance, no? Or
just general electrical noise, at the least.

My experience has been that most sim folks using only low-voltage COTS
devices shouldn't have too much trouble from inductive loads. But as
soon as you start getting into aircraft voltages or a lot of custom
electronics, the USB interference becomes an issue.

Even on my sim, the only time I've seen genuine, reproduceable USB
issues cause by EMI is when I had that relay hooked up, as most of my
circuits are low voltage, and I am only using a couple of small relays
now. But even when everything is working properly, I still hate dealing
with USB enumeration order, virtual device nodes, and having to keep
USB "sessions" alive (and rebooting X-Plane or my instrument software if
a USB session is interrupted for any reason).

I realize ethernet connections are susceptible to EMI as well (though
not nearly as bad, it seems). The difference, of course, is that if I
drop a few UDP packets, or even a bunch of packets, nothing bad
happens. There is no host, no one network device depends on any other
network device. And if you need absolute data integrity and can't drop
any information, that's what TCP/IP is for. :)

-- 
Matt Bailey

Keeping It Real:
Sabreliner 60 Flight Simulator
Serial # 306-61 - N1JX
http://sabrelinersim.com