Ravenna cable run limitations

Forum to discuss your Pyramix systems, motherboards, RAID arrays, graphic cards, laptops, etc..
Forum rules
The Merging Technologies team cannot be held responsible for support queries logged on the public forums. If a support query is logged here and only here, it may not be found and dealt with by the appropriate team.
To ensure that your support issue or bug report is dealt with properly and in good time, please use the link to the tech support request form page on the Merging website.
Make sure to let us know what version you are using when you send your mail. THANKS!
klaukholm
Posts: 504
Joined: Fri Jul 29, 2005 00:36
Location: Stockholm, Sweden

Ravenna cable run limitations

Postby klaukholm » Sun Apr 15, 2012 11:58

What length of cat5 cable can we expect before channel count is reduced when using Horus and the merging ethernet card point to point.

For the sake of argument lets say I am using a cable similar to the vandamme tourcat

"Van Damme Tourcat Flexible•Meets the requirements of Category 5E TIA/EIA-568-B.2-2001.
•Recommended maximum distance for 100MBits/s data rate is 75 metres; longer runs are possible but not guaranteed"

attenuation specs etc here at the bottom of the page
http://www.van-damme.com/38.html

Also, if we were to use the hybrid cable,
Van Damme Tourcat 2 x U/UTP 4 x AES/Analogue pair multicore
would running AES or line level audio on the same cable assembly cause any relevant loss of bandwidth?

User avatar
mpdonahue
Posts: 460
Joined: Fri Jan 16, 2004 03:20
Location: Boston, MA
Contact:

Re: Ravenna cable run limitations

Postby mpdonahue » Tue Apr 17, 2012 02:28

Kjetil,
This is simply a Gig Ethernet question.
First off, I wouldn't use anything less than Cat6A and SSTP. Termination here is important. Proper termination and shielding standards will increase length. Also, you can go for heavier 23 or 24 ga cable to reduce loss over long runs.
Just FYI, stranded cable has a shorter functional range than solid core cable. we've never had any issues with stranded over single long links, but we've never tried to go over the 100m length with 1000baseT. (AES will go about 250m at 96k over the same cable.) The real problem comes with extensions and couplers. Each one marginally cuts off about 15% of total range. Something to think about.

All the best,
-mark
*********************
Mark Donahue
Soundmirror, Inc.
Boston, MA
mark@soundmirror.com
www.soundmirror.com
*********************

klaukholm
Posts: 504
Joined: Fri Jul 29, 2005 00:36
Location: Stockholm, Sweden

Re: Ravenna cable run limitations

Postby klaukholm » Tue Apr 17, 2012 07:12

Great info Mark,
Now off to find a good ruggedized 2way CAT6 cable on a cable drum

Do you know of any 1000baseT solutions over fibre?

A fibre card such as this would be good if it works well with ravenna prociding it would let me use madi spec fibre.

http://eu.startech.com/Networking-IO/Ad ... EX1000MMSC

and this in the other end
http://eu.startech.com/Networking-IO/Me ... CMGBSC15EU

User avatar
mpdonahue
Posts: 460
Joined: Fri Jan 16, 2004 03:20
Location: Boston, MA
Contact:

Re: Ravenna cable run limitations

Postby mpdonahue » Sat Apr 28, 2012 19:30

klaukholm wrote:Great info Mark,
Now off to find a good ruggedized 2way CAT6 cable on a cable drum

Do you know of any 1000baseT solutions over fibre?

A fibre card such as this would be good if it works well with ravenna prociding it would let me use madi spec fibre.

http://eu.startech.com/Networking-IO/Ad ... EX1000MMSC

and this in the other end
http://eu.startech.com/Networking-IO/Me ... CMGBSC15EU

There are plenty of fiber based solutions for 1g and 10g ethernet. The thing to remember is that MADI spec'ed fiber is the lowest rating, basically 10baseT. 1000baseT requires a much higher precision on the polish of the fiber end. (At a much higher cost). To be honest, copper will be fine for virtually any length that we need to run and the addition of a repeater shouldn't effect latency significantly. how often are you going more than 300m? If this is actually an issue, you should be looking at single mode fiber links, good for 10g ethernet out to 80km. Single mode is marginally more expensive than multi-mode fiber which is really only good out to 500m at 1g and 300m at 10g. I never use more than 200m at 1g without a repeater.
All the best,
-mark
*********************

Mark Donahue

Soundmirror, Inc.

Boston, MA

mark@soundmirror.com

www.soundmirror.com

*********************

klaukholm
Posts: 504
Joined: Fri Jul 29, 2005 00:36
Location: Stockholm, Sweden

Re: Ravenna cable run limitations

Postby klaukholm » Sat Apr 28, 2012 20:06

I have never needed more than 150meters

Anyone here have a lead on reasonably priced ruggedized cat6 on drums here in the EU?

Perfect Record
Posts: 646
Joined: Thu Apr 24, 2008 06:09
Location: St Paul, Minnesota USA
Contact:

Re: Ravenna cable run limitations

Postby Perfect Record » Sat Apr 28, 2012 20:28

Mark,
What system are you running over ethernet? Are you talking about Ravenna?

User avatar
Paulo M
Posts: 1445
Joined: Wed Jun 01, 2005 16:19
Location: Portugal

Re: Ravenna cable run limitations

Postby Paulo M » Thu Jul 19, 2012 01:29

Mark,
What system are you running over ethernet? Are you talking about Ravenna?


Of course, what else? The title and first post says it all. :)
Best regards,

Paulo M

Pyramix 7.1 Masscore
VCube XE 3.1
MB5 Dual & X50 MADI
Win XP SP3
Intel Q9600/Gigabyte X48 DS5 Motherboard/ASUS 4350 Graphic card

Perfect Record
Posts: 646
Joined: Thu Apr 24, 2008 06:09
Location: St Paul, Minnesota USA
Contact:

Re: Ravenna cable run limitations

Postby Perfect Record » Thu Jul 19, 2012 04:03

Paulo M wrote:Of course, what else? The title and first post says it all. :)


I ask this because
A) Mark immediately addresses this as a generic Ethernet question
B) Ravenna systems are still quite new and rare here in the US, and I don't know if Soundmirror has a system yet
C) There are other Ethernet systems around, and Mark's firm has been doing high channel count recording systems for years.

Chas
Posts: 2
Joined: Fri Mar 19, 2010 03:29
Location: Philadelphia, USA

Re: Ravenna cable run limitations

Postby Chas » Tue Oct 16, 2012 03:04

mpdonahue wrote:
klaukholm wrote:Great info Mark,
Now off to find a good ruggedized 2way CAT6 cable on a cable drum

Do you know of any 1000baseT solutions over fibre?

A fibre card such as this would be good if it works well with ravenna prociding it would let me use madi spec fibre.

http://eu.startech.com/Networking-IO/Ad ... EX1000MMSC

and this in the other end
http://eu.startech.com/Networking-IO/Me ... CMGBSC15EU

There are plenty of fiber based solutions for 1g and 10g ethernet. The thing to remember is that MADI spec'ed fiber is the lowest rating, basically 10baseT. 1000baseT requires a much higher precision on the polish of the fiber end. (At a much higher cost). To be honest, copper will be fine for virtually any length that we need to run and the addition of a repeater shouldn't effect latency significantly. how often are you going more than 300m? If this is actually an issue, you should be looking at single mode fiber links, good for 10g ethernet out to 80km. Single mode is marginally more expensive than multi-mode fiber which is really only good out to 500m at 1g and 300m at 10g. I never use more than 200m at 1g without a repeater.
All the best,
-mark


Hello,

This post is a bit old now but I am currently designing a system using Horus/Ravenna and I thought I might add some observations:
- with recent Horus demo I tested two point-to-point links: 1. OM1 fiber @ 250m via 1GbE copper>fiber transciever and 2. Cat5e UTP (riser) 24AWG @ 300m with three $35.00 GbE routers used as repeaters.
- both setups passed 24ch, 24bit/192kHz with sixteen return Ravenna channels active to the Horus, without any apparent issues.
- due to some other factors I did not have a chance to ping the systems, so I can't speak about measured latency at this point.

I was a bit amazed that the copper link worked as we exceeded length recommendations on both sides of the midpoint using vanilla grade Cat5e (the only 300m spool on the shelf at the time...) plus really cheap routers.

Re: diminished channel count, I don't think this is part of the picture... if payload is (x) channels, this is what should be delivered... unless I am missing something about Ravenna. The packet delivery would depend on the switch configuration and hardware, if the hardware delivers 1Gb per port and the switch is properly configured the throughput should be 1Gb from Horus to receiver, with due respect for ethernet physical link recommendations.

Re: the copper-fiber transceiver mentioned above, this is a PCI solution, there are relatively cheap stand-alone converters available to do the same task.

Re: the fiber spec, it's my understanding that OM1 fiber is fully capable of GbE within length limitations, and generally is considered 10baseT only in long haul (i.e. >2km) applications. Maybe someone with greater network experience could chime in. So if 300m or less you're probably OK with OM1, but why not use OM2 as it's negligible extra cost and will give you much more distance and/or bandwidth. SM is a big step up and will allow you to go several times around the block, there are also relatively cheap transceivers for this. Concerning the polish on the tips, personally I have not had any bad experiences with off the shelf OM1 fiber assemblies doing 32ch 24/96 MADI links @ 300m and I know others who regularly exceed this on MM fiber, both OM 1 and OM2. The MADI bandwidth is well below Gb, but then so is mostly any channel count you might want to transport over Ravenna... of course I don't know this for sure, maybe you need hundreds of channels!

Hope this helps, so far I like this system after working through some bugs. I have yet to "acid-test" the Horus, i.e. loading it in a 1/10GbE multi-switch network where there's lots of traffic, I hope to do this soon as this is the primary appeal of the system for me.

Best,
Charles
_______________________________
Masscore 7.1/128ch/Mykerinos AES
Native 7.1/RME HDSPe MADI

Bernhard Guettler
Posts: 885
Joined: Thu May 16, 2002 13:35
Location: Berlin, Germany

Re: Ravenna cable run limitations

Postby Bernhard Guettler » Sun Sep 01, 2013 15:48

FYI,
I'm currently running 130 m in location recording sessions, 3 cables with passive connectors, 20m+90m+20m, all CAT6 SSTP, without any glitches for the last three days.
Horus, Macbook Pro Win 7 Bootcamp.
Cheers - Bernhard

DJS
Posts: 456
Joined: Wed May 26, 2010 05:26
Location: Brisbane, Australia

Re: Ravenna cable run limitations

Postby DJS » Sun Sep 08, 2013 00:49

FWIW I am using a 50m roll of Cordial CCAT5 terminated with Neutrik Ethercon and two Ethercon NE8FF adapters either end and it has been working well.
http://katalog.cordial.eu/catalog1,2,31,mw65.html
David Spearritt
Classical and Acoustic Music, BNE, Australia

MorganN
Posts: 31
Joined: Thu Mar 16, 2006 00:24
Location: Norway
Contact:

Re: Ravenna cable run limitations

Postby MorganN » Mon Dec 01, 2014 20:24

Looking at getting the 100m CAT7 version of the Cordial roll, but swapping out the connectors with something like these RJ45's:
http://www.telegaertner.com/en/info/cat ... duct=15047

As Mark pointed out, a coupler or extension reduces the total range, so I'd rather go from device to computer without any additional couplers.

Anyone have any experience with CAT7 and Ravenna?

User avatar
mpdonahue
Posts: 460
Joined: Fri Jan 16, 2004 03:20
Location: Boston, MA
Contact:

Re: Ravenna cable run limitations

Postby mpdonahue » Thu Dec 04, 2014 23:27

MorganN wrote:Looking at getting the 100m CAT7 version of the Cordial roll, but swapping out the connectors with something like these RJ45's:
http://www.telegaertner.com/en/info/cat ... duct=15047

As Mark pointed out, a coupler or extension reduces the total range, so I'd rather go from device to computer without any additional couplers.

Anyone have any experience with CAT7 and Ravenna?

Don't waste your money. Cat6a is all you need and more for 1000baseT. We buy it by the box and do our own terminations on site. A 1000' (300m) box of Cat6a patch (Stranded) is about $175 and is actually over specified for Gig Ethernet.
Additionally, shielded cable is only a benefit if the cable is properly terminated and both the sending and receiving devices are dealing with the shielding correctly (Which is actually far rarer than you'd think. )
Also, Ethercon connectors are only valuable if you have devices that have the connectors. Having to use adapters at either end to get back to RJ45 connectors is far more likely to cause problems than a properly booted RJ45 connector alone.
Also, we find the ruggedized versions un-necessary. For the cost of 1 piece of the ruggenized cable you can get between 5 and 10 pieces of the stranded patch CAT6A. The only thing rugged about the cable is the outer jacket. The inner pairs are just as likely to to be crushed and pushed out of position. Also, the only ruggedized cable I've found has solid core conductors, which kind of defeats the purpose.
If we were talking about 10 Gig Ethernet and 1000' runs this would be different.
All the best,
-mark
*********************

Mark Donahue

Soundmirror, Inc.

Boston, MA

mark@soundmirror.com

www.soundmirror.com

*********************

MorganN
Posts: 31
Joined: Thu Mar 16, 2006 00:24
Location: Norway
Contact:

Re: Ravenna cable run limitations

Postby MorganN » Fri Dec 05, 2014 19:36

Thanks Mark!
I've already ordered the CAT7 cable. Got a good price at Thomann (actually half the price compared to buying the Ravenna CAT5e cable locally).

David J at Merging gave me this response:
CAT7 cables should work fine, but we have only tested CAT5E and CAT6.
Therefore we cannot guarantee 100% that it will work.


As CAT7 is backward compatible I would find it strange if it doesn't work.

Good point regarding shielding. The Telegärtner plugs ensure that at least the plug is shielded and I'm sure the Merging units are as well.
When it comes to i.e. the Thunderbolt-Ethernet connector, I'm not so sure.

Regarding type of cable, I'd rather have a touring grade cable than a stiff installation cable, but I will keep a roll for backup just in case.
Does the patch cable you use have shielding? To me it seems like the shielding also gives a little bit more protection than just the jacket.

I will give the Cordial CAT7 a go and report back.

otto_fosch
Posts: 14
Joined: Wed Mar 28, 2012 23:08

Re: Ravenna cable run limitations

Postby otto_fosch » Sun Feb 01, 2015 20:06

Dear all

FYI
My new horus works smoothly with this 50m rugged CAT5e cable http://www.mtiagnv.com/mte-ee-cat5e-kw.html?SID=gc092t02fvpsn7nrs2u3ombs83&___store=us with Neutrik NE8MC-1 on a Schill drum GT310.RM. The connection on my fight case is a Ethercon NE8FDP RJ45 feedthrough receptacle, D-shape metal flange with the latch lock http://www.neutrik.com/en/ethercon/d-series/ne8fdp

Best

Frank