cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
764
Views
0
Helpful
7
Replies

IAD 2431 - Nailed up FXS-FXO tops out at 16kbps

randal
Level 1
Level 1

Howdy all,

We are using a Cisco IAD 2431-16FXS with a 4-port FXO card in the VWIC slot.

We are attempting to cross-connect or "nail up" a modem call from one of the FXS ports to a FXO port that is in turn connected to the PSTN. We can dial consistently, but are getting very poor connect speeds, topping out at 16.8k.

For debugging purposes, if we plug the modem directly into the same phone line that's on the FXO port, we get consistent 50.6k connects.

Any advice on why connection from FXS to FXO would cause such a quality loss? Any ideas on how to rectify it?

Thanks,

Randal

7 Replies 7

Hi Randal,

In theory the DSPs won't be involved at all since this should be a local-bypass scenario (pots-pots on the same router).

voice-card 0

local-bypass (default)

You can try forcing the DSPs in and using modem passthrough:

voice-card 0

no local-bypass

voice service voip

modem passthrough nse codec g711ulaw

And then on the ports you can try adding this:

voice-port 2/x

input gain -3

output attenuation 5

This will reduce the amount of echo on the call since we turn off the ECAN for high speed modem calls.

hth,

nick

Nick,

Thank you very much for your informative post. We tried everything listed above, and verified that the calls went from type "tele" to g711ulaw once the DSPs were put back into play, all with the same results -- nothing above ~16kbps.

Presuming the FXO was to blame or the PSTN is broken or ???, another engineer took the IAD home and tested it there. He had the same results, except for it topped out at 26k, even though he could connect at 49.3k without the IAD in the way.

Of note, no matter how ridiculously absurd we set all of those settings -- including wrecking the impedance settings - they all had absolutely zero affect. (Yes, we shut/noshut the voice-ports)

Any other ideas?

Thanks,

Randal

Hi Randal,

Unfortunately getting high speed faxes and modems to work on our DSPs isn't something promised. I've tried working through issue likes this before, but 14.4 is the fastest the DSPs are supposed to do.

Technically if you have pots-pots and local bypass is on, and it is by default, our DSPs shouldn't be involved.

Plus, if our DSPs aren't involved, there is almost nothing we can change.

-nick

Nick,

We're testing this with the onboard FXS and a VIC2-4FXO if that makes any difference.

The -real- question is whether or not this kind of behavior will persist when we replace the 4FXO card with a VWIC-2MFT+PRI. Any input on that one?

Randal

Hi Randal,

You may have a better line quality with a PRI, but again, the router can't promise speeds above 14.4. I think you'll have a better chance, but who knows.

-nick

Nick,

Sorry to be a pest, but is there a platform that is >>> the IAD for this kind of fax/modem PSTN termination? We've considered something like an AS5x00, but our proof of concept is to make it work on an IAD (which is what we have laying around), then invest into some better gear.

The background for this is that we're passing a SIP handoff from a remote IAD to this IAD, then dropping it off onto the PSTN -- the modem over ip & fax over ip work fantastically, but we're having troubles getting the fxs-fxo-PSTN link working right (the follow-up is of course SIP-fxo/pri-PSTN).

Thanks!

Randal

Do you get better results if the FXS and FXO are on different devices? This shouldn't be the case and if this is happening there may be something wrong in the router.

If you can test two different IADs, and you get better results, open up a TAC case. You could even email me the SR and I can try to take a look.

But if you get the same result, then it may just be a limitation of the DSPs at this point.

As far as the platforms are concerned, I don't know of any differences on a DSP level that would allow better performance. Things like a NM-HDV are going to work different than a SPE on a 5400, which is going to be different than a C5510 in an ISR. Since it's all best effort, I don't know of any data supporting which one might do best effort the best.

hth,

nick

Getting Started

Find answers to your questions by entering keywords or phrases in the Search bar above. New here? Use these resources to familiarize yourself with the community: