Dale Farmer wrote:

It was another approach one could use. Some venues it is more
appropriate to use one or the other. The other issue is that most cat-5
cables are rather flimsy. All it takes is one person wearing a spike heel
to step on it wrong and the cable is no longer going to have an
acceptable error rate. There are ruggedized cat-5 cables out there,
but they are substantially more expensive. There is also the 100 meter
cable length limit with ethernet, which I don't know if audiorail is
vulnerable too.

"One person wearing a spike heel" will either break/short an internal
wire, in which case you lose all 32 channels of audio in that direction,
or they won't, in which case there will be no effect. Just deforming
the twists in one isolated spot (to the extent that you could deform
them through the outer sheath of the cable) isn't going to do anything.
There are four twists per inch, so the twists are quite tight and not
easy to separate.

The bottom line is that you don't want a raw CAT5 cable of the cheap
installed grade to be in a completely exposed position such that it
could be stepped on. You need to throw a rug over it, cable protector,
or whatever, and get it over to where you can run it along the wall or
otherwise out of the way.

In answer to your question, AudioRail is spec'ed for 100 meters per hop.
We have tested to more, but we prefer to be conservative, so that we
can ensure that the error rate is zero.

Garth D. Wiebe wrote:
"One person wearing a spike heel" will either break/short an internal
wire, in which case you lose all 32 channels of audio in that direction,
or they won't, in which case there will be no effect.

This brings up a question I've been wondering about. A cat5
cable has four twisted pairs in it. Assuming things work the
way I think they work, you only need one pair in each direction
(just like Ethernet uses) to transmit the signal. This leaves
with you two extra pairs.

Has there been any thought into modulating and demoulating
that stuff twice and transmitting it redundantly, once on
each of two pairs? If you put checksums on the data, you can
of course detect which of the two signal paths had the error.
And since it's digital and both would be driven from the same
clock, it's easy to keep the two perfectly in sync and switch
between them as necessary.

Or if you are already using all 8 conductors, you could set
it up to use two separate cat5 cables, which would actually
be better in case it is the connector that is failing. (Such
as when the plastic clip that secures the connector has
broken off or -- worse -- looks good but is no longer springy
enough to do its job.)

The advantage here is that when your cat5 cable starts to fail,
the system can flash a red error light (one at each end) and
keep pressing on until the end of the show, at which time you
toss out the cat5 cable.

I'm not really in the market for anything like AudioRail at
the present time, but if I were buying it for live sound, I
would want to know that it is designed (and has tested) to
fail gracefully instead of catastrophically, like so much
digital stuff does.

- Logan

Mike Rivers wrote:

The connectors aren't so robust and wouldn't take much of the kind of
handling that live sound gear gets. Neutrik has a solution to this,
the EtherCon line - essentially an XLR shell around an RJ-45
connector, but you don't find those on Behringer gear. The only
commercial product I've seen that uses them at this point is the
Line 6 Variax guitar.

There are others. I don't know about the Line 6 product, but if it is
on stage, then I would agree that the traditional RJ45 connector won't
cut it.

Again, in our case it was a call. We're at the back of a 19" rackmount
device (not that you have to rackmount it), and alongside IEC power
cords that can just pull out when yanked or other gear with 1/4" TRS
cables that can just pull out when yanked, and so on.

For sure, if it is subject to the risk of yanking, the CAT5 cable better
be taped down or tie-wrapped down to something. If you yank a CAT5
cable, the wires could come detached from the contacts inside the
modular plug housing, which would be a very bad situation, obviously.

Dale Farmer wrote:

hank alrich wrote:


Dale Farmer wrote:


If you have someone competent to use a punchdown tool, buy
cat-5 solid cable for sixty bucks per thousand feet and just abandon
it at the end of each gig.

RANT: ON

Okay, that's exactly what is wrong with our "culture": we throw away
stuff because we think it's cheap, with no regard for the longterm
environmental impact of having made that stuff in the first place,
mining the copper and all the lovely **** attending the process of that
and its manufcaturing, making the PVC and all the **** that goes with
it. We seem to think we'll get to keep doing that kind of damage forever
and not pay a hefty price. Perhaps we take comfort in having our heads
in the sand and not stopping to think of the type of act it is to pass
the consequences on to the following generations.

What was that about eating the seed corn? Seems smarter to use stranded
and employ it for a long time.

RANT: OFF


It was another approach one could use. Some venues it is more
appropriate to use one or the other. The other issue is that most cat-5
cables are rather flimsy. All it takes is one person wearing a spike heel
to step on it wrong and the cable is no longer going to have an
acceptable error rate. There are ruggedized cat-5 cables out there,
but they are substantially more expensive. There is also the 100 meter
cable length limit with ethernet, which I don't know if audiorail is
vulnerable too.

Given the diameter, price & availability of CAT5 cable I think it might make sense to permanently install in venue(s) where you planned to operate.

Logan Shaw wrote:
Garth D. Wiebe wrote:

"One person wearing a spike heel" will either break/short an internal
wire, in which case you lose all 32 channels of audio in that
direction, or they won't, in which case there will be no effect.


This brings up a question I've been wondering about. A cat5
cable has four twisted pairs in it. Assuming things work the
way I think they work, you only need one pair in each direction
(just like Ethernet uses) to transmit the signal. This leaves
with you two extra pairs.

Correct for 100 Mb Ethernet technology, which we are currently using.
When you go to 1 Gb Ethernet technology all four pairs are used.

Has there been any thought into modulating and demoulating
that stuff twice and transmitting it redundantly, once on
each of two pairs? If you put checksums on the data, you can
of course detect which of the two signal paths had the error.
And since it's digital and both would be driven from the same
clock, it's easy to keep the two perfectly in sync and switch
between them as necessary.

Redundancy is one option. Others say, why don't you double up, use all
four pairs of the cable, and get 64 channels in each direction.

Or if you are already using all 8 conductors, you could set
it up to use two separate cat5 cables, which would actually
be better in case it is the connector that is failing. (Such
as when the plastic clip that secures the connector has
broken off or -- worse -- looks good but is no longer springy
enough to do its job.)

The advantage here is that when your cat5 cable starts to fail,
the system can flash a red error light (one at each end) and
keep pressing on until the end of the show, at which time you
toss out the cat5 cable.

I'm not really in the market for anything like AudioRail at
the present time, but if I were buying it for live sound, I
would want to know that it is designed (and has tested) to
fail gracefully instead of catastrophically, like so much
digital stuff does.

There is not going to be a "starts to fail" scenario, like with analog,
where you "press on until the end of the show". With digital audio, you
start dropping data bits, and they can be as easily the "most
significant data bits" (MSB) as they could be the "least significant
data bits" (LSB). What you end up with is garbage audio, and not some
sort of sonic degradation that you can EQ or trimpot your way out of,
like *sometimes* you can with failed analog lines.

Unfortunately, this is the nature of "digital stuff". I don't know that
there is a solution in the audio industry that provides true redundancy
in terms of data error detection, correction, and transparent failover.

Garth D. Wiebe wrote:

Mike Rivers wrote:

The connectors aren't so robust and wouldn't take much of the kind of
handling that live sound gear gets. Neutrik has a solution to this,
the EtherCon line - essentially an XLR shell around an RJ-45
connector, but you don't find those on Behringer gear. The only
commercial product I've seen that uses them at this point is the
Line 6 Variax guitar.


There are others. I don't know about the Line 6 product, but if it is
on stage, then I would agree that the traditional RJ45 connector won't
cut it.

Again, in our case it was a call. We're at the back of a 19" rackmount
device (not that you have to rackmount it), and alongside IEC power
cords that can just pull out when yanked or other gear with 1/4" TRS
cables that can just pull out when yanked, and so on.

The IEC cable rarely leaves the rack. Most of the ¼" TRS cables probably don't leave the rack or the immediate area.

The CAT5 will definitely leave the ares.

Garth D. Wiebe wrote:

There is not going to be a "starts to fail" scenario, like with
analog, where you "press on until the end of the show". With digital
audio, you start dropping data bits, and they can be as easily the
"most significant data bits" (MSB) as they could be the "least
significant data bits" (LSB). What you end up with is garbage audio,
and not some sort of sonic degradation that you can EQ or trimpot
your way out of, like *sometimes* you can with failed analog lines.

Unfortunately, this is the nature of "digital stuff". I don't know
that there is a solution in the audio industry that provides true
redundancy in terms of data error detection, correction, and
transparent failover.

I think this is the good news. When digital stuff like this starts getting
flakey, it sounds really horrible. Or, it goes dead. Clicks, pops, mutes,
gravelly sounds, etc. Nothing subtle about how it is broken.

Actually, for digital audio, it doesn't HAVE to go bad all at once. If you are
willing to sacrifice bandwidth, you can send extra error correction bits down
the line such that even if it has a horrendous error rate you can reconstruct a
perfect data stream.

However, for THIS application, its being run all over one cable with 4 little
wires. Each wire is crucial to keeping the signal streaming properly. There's
no redundancy in the wires themselves. This is likely not to fail a little bit
at a time, its likely to be either works great or no sound at all -- not a lot
of middle ground. This is not like a clocking issue where you can get the
clicks/pops, etc. This is a "signal present" or "Signal not there" type of
situation. If the signal is present, its likely to be fine.

-lee-

In article writes:


Again, in our case it was a call. We're at the back of a 19" rackmount
device (not that you have to rackmount it), and alongside IEC power
cords that can just pull out when yanked or other gear with 1/4" TRS
cables that can just pull out when yanked, and so on.

I wasn't thinking about the plug getting yanked out once the system
was set up. That's something that the person setting it up should deal
with - for power, audio, and Ethernet cables. But you can step on an
IEC power connector, XLR, or TRS plug, or bash it against the side of
a packing crate, or have it hang out the side of a milk crate and
smash against another crate and it will survive. The locking tab on an
RJ connector is quite fragile and when that's gone, I wouldn't trust
the best gaffer tape in the world to hold it in for the duration of a
show.

For what it's worth, I feel almost the same about TOSLink connectors.
And don't even get me started on mini phone jacks.



--
I'm really Mike Rivers )
However, until the spam goes away or Hell freezes over,
lots of IP addresses are blocked from this system. If
you e-mail me and it bounces, use your secret decoder ring
and reach me he double-m-eleven-double-zero at yahoo

Leoaw3 wrote:

Actually, for digital audio, it doesn't HAVE to go bad all at once. If you are
willing to sacrifice bandwidth, you can send extra error correction bits down
the line such that even if it has a horrendous error rate you can reconstruct a
perfect data stream.

However, for THIS application, its being run all over one cable with 4 little
wires. Each wire is crucial to keeping the signal streaming properly. There's
no redundancy in the wires themselves. This is likely not to fail a little bit
at a time, its likely to be either works great or no sound at all -- not a lot
of middle ground. This is not like a clocking issue where you can get the
clicks/pops, etc. This is a "signal present" or "Signal not there" type of
situation. If the signal is present, its likely to be fine.

-lee-

It has been previously established that CAT5 is awfully cheap. If it
costs four times as much it's still cheap, so run four CAT5's.

And perhaps AudioRail can put two or more RJ45's on the box, then run a
small DC current (called "sealing current" by TPC) to verify cable
integrity and fail over to the standby one, allowing you to swap the
primary out.

S O'Neill wrote:

It has been previously established that CAT5 is awfully cheap. If it
costs four times as much it's still cheap, so run four CAT5's.

And perhaps AudioRail can put two or more RJ45's on the box, then run a
small DC current (called "sealing current" by TPC) to verify cable
integrity

You don't even need special circuitry to run the DC current. You
can just use a checksum, parity bits, or some other similar form
of redundancy in the data stream. Then, decode the data at the
receiving end and you can tell in software (or in hardware if you
are building a custom chip) when the data is bad, because the
numbers won't "add up". Then you switch over to the other
connection (assuming it has good data, of course -- some things
could cause both to fail...).

- Logan

Logan Shaw wrote:

S O'Neill wrote:

It has been previously established that CAT5 is awfully cheap. If it
costs four times as much it's still cheap, so run four CAT5's.

And perhaps AudioRail can put two or more RJ45's on the box, then run
a small DC current (called "sealing current" by TPC) to verify cable
integrity


You don't even need special circuitry to run the DC current. You
can just use a checksum, parity bits, or some other similar form
of redundancy in the data stream. Then, decode the data at the
receiving end and you can tell in software (or in hardware if you
are building a custom chip) when the data is bad, because the
numbers won't "add up". Then you switch over to the other
connection (assuming it has good data, of course -- some things
could cause both to fail...).

Well, sure, that's the usual way. The DC way is faster and more
specific than a checksum; also failover could possibly be completed
mid-packet with no loss.

S O'Neill wrote:

Logan Shaw wrote:

S O'Neill wrote:

perhaps AudioRail can put two or more RJ45's on the box, then run
a small DC current (called "sealing current" by TPC) to verify cable
integrity


You don't even need special circuitry to run the DC current. You
can just use a checksum, parity bits, or some other similar form
of redundancy in the data stream. Then, decode the data at the
receiving end and you can tell in software (or in hardware if you
are building a custom chip) when the data is bad, because the
numbers won't "add up". Then you switch over to the other
connection (assuming it has good data, of course -- some things
could cause both to fail...).


Well, sure, that's the usual way. The DC way is faster and more
specific than a checksum; also failover could possibly be completed
mid-packet with no loss.


TPC knows just a little bit about getting longterm reliable performance out of unshielded twisted pair cable in realworld environments...

"Garth D. Wiebe" wrote:


You do have a point here. I suppose presenting a turnkey system as a
sales option is a valid presentation method, and then any savvy user
"with a brain" will go buy the parts himself, rather than wait for us to
do it for him and charge him extra for it.

Think of the turnkey system in the same way that semiconductor
manufacturers sell evaluation boards. These boards may actually find
their way into some bits of low volume kit but most manufacturers
selling reasonable volumes will design their own once they've proved
their concept with the evaluation board.

Cheers.

James.

Garth D. Wiebe wrote:
Arny Krueger wrote:

So I would see your "ship a whole package in a box" proposal as
going way out on a limb, unless you're aiming to redefine our
business.

How big of a limb? You don't even have to buy the first new-in-box
ADA8000 until you get your first order. IME ZZounds and MF ship
today with delivery in about three days. In a bit over a week you
can deliver a turnkey system.

Zero out-of-pocket investment other than the web page, until you
book your first order. Then you are on the hook for the parts for
that first turnkey system. If it sticks, your take from the first
one finances the second turnkey system, etc. I told you to mark the
turkey systems up, right? I mean mark it UP, just this side of
gouging.

I think some of these skeptical installed sound guys who are
cautious about AudioRail, might just buy a turnkey system, lay it
out on the floor and watch it work in their warehouse, and then eat
the Behr parts on the grounds that they had educational value. Look
at the cost of the competitive systems!

If you are proposing that we don't *stock* the turnkey systems, then,
yes, this makes the proposition a lot less of a financial risk.

That's right, keep no stock on hand. I think I pointed out that some
regular Behr dealers generally deliver quick enough to not cause problems.

BUT, supposing there was a step by step grocery list on the website,
including even hyperlinks to the click-to-buy URLs of Musician's
Friend, SameDay Music, ZZ, and etc.?

I'm of the do-it-now presuasion. As I said before, turnkey systems
don't and IMO shouldn't be your only mode of operation. However even
just a complete description of your turnkey systems(s) will drive
the point home to a lot of the purchasing public. Anybody with a
brain will look at them and say "I could do that myself for a
heckuva lot less money", and run right out and do it. You still sell
a pair of AudioRail boxes, which is what you wanted to do, right?

This could be a success for you even if you never actually sell one
turnkey system, right?

You do have a point here. I suppose presenting a turnkey system as a
sales option is a valid presentation method, and then any savvy user
"with a brain" will go buy the parts himself, rather than wait for us
to do it for him and charge him extra for it.

Exactly. Look at your markup (if it ever comes to that!) as a tax on people
who are risk-adverse. That's a legitimate business principle, most often
known as "Insurance".