Telephone Interface will not work with answering machines

[PSC]

Has anyone hooked up a CBTI with an answering machine and have to two co-exist successfully on the same line?

We have found that there is a problem when we connecting a stock standard Panasonic answering machine to the CBTI.

Problem -

When the answering machine answers the call the CBTI seems to grab the line even though the "line grab" LED does NOT light up. It then continues to hold the line for 60sec. So, if you try to call back within that time all you get is an engaged tone.

This setup was tested at two separate sites -

Site 1. The CBTI grabbed the line only after the answering machine tone was heard.

Additional information regarding site 1 -
Dial in - answering machine answers – cannot leave message – cannot dial back for 1 min. (Primary)

Dial in - answering machine answers – cannot leave message – press 5 for Cbus – Cbus answers & works – hang-up - can dial back in immediately. (Primary)

Disconnect power to CBTI – dial-in - answering machine answers – can leave message – can dial back in immediately. (Primary)

Re-connected power to CBTI – dial-in - answering machine answers – can leave message – can dial back in immediately. Only works first call in after disconnection. (Primary)


Site 2. The CBTI grabbed the line before the message was complete. Disconnected the Cbus cable, left the power & telephone lines connected - answering machine works fine.

A number of different configurations were tested; including removing all other telephone devices from the system etc.

Don & Ryan (CIS), I have spoken to you guys at length regarding this problem - can you please sort it...

[Richo]

Has anyone hooked up a CBTI with an answering machine and have to two co-exist successfully on the same line?

We have found that there is a problem when we connecting a stock standard Panasonic answering machine to the CBTI.

Problem -

When the answering machine answers the call the CBTI seems to grab the line even though the "line grab" LED does NOT light up. It then continues to hold the line for 60sec. So, if you try to call back within that time all you get is an engaged tone.

This setup was tested at two separate sites -

Site 1. The CBTI grabbed the line only after the answering machine tone was heard.

Additional information regarding site 1 -
Dial in - answering machine answers – cannot leave message – cannot dial back for 1 min. (Primary)

Dial in - answering machine answers – cannot leave message – press 5 for Cbus – Cbus answers & works – hang-up - can dial back in immediately. (Primary)

Disconnect power to CBTI – dial-in - answering machine answers – can leave message – can dial back in immediately. (Primary)

Re-connected power to CBTI – dial-in - answering machine answers – can leave message – can dial back in immediately. Only works first call in after disconnection. (Primary)


Site 2. The CBTI grabbed the line before the message was complete. Disconnected the Cbus cable, left the power & telephone lines connected - answering machine works fine.

A number of different configurations were tested; including removing all other telephone devices from the system etc.

Don & Ryan (CIS), I have spoken to you guys at length regarding this problem - can you please sort it...

I see you have disabled the CBTI from answering any incoming call (both modes) when using an Answering Machine, that's good. And I assume when you say that you press 5 to access CBUS, that 5 is your local ACCESS CODE? What is the Answering Machines programmed behaviour when you press 5, you need to ensure that there isn't a conflict between the 2. Have you tried different access codes (say the default of #3#)

Did you try the Answering Machine on the Secondary Phone Port of the CBTI?

During "abnormal" disconnects (Whilst online to CBTI you hang up without using the hang up menu option) it can take a varying amount of time to detect the hang up, and thus block calls that come straight in. But it shouldn't take a full minute.

I assume the phone line is a direct Telstra exchange line and not run through some sort of PABX?

I'm doing some checking of the CBTI at the moment as Version 1.13 of TICA is being prepped for release (no firmware changes though, so the behaviour will be no different). I'll see what I can do to replicate the behaviour above.

[PSC]

I see you have disabled the CBTI from answering any incoming call (both modes) when using an Answering Machine, that's good. And I assume when you say that you press 5 to access CBUS, that 5 is your local ACCESS CODE? What is the Answering Machines programmed behaviour when you press 5, you need to ensure that there isn't a conflict between the 2. Have you tried different access codes (say the default of #3#)

I have tried several codes, this was our first thought - maybe the answering machine doesn't like our codes. So we purchased another answering machine; guess what? same problem.

Did you try the Answering Machine on the Secondary Phone Port of the CBTI?

Yes, originally we had it connected to the secondary as instructed. During our trial and error testing we connected it to the primary port - same result.

During "abnormal" disconnects (Whilst online to CBTI you hang up without using the hang up menu option) it can take a varying amount of time to detect the hang up, and thus block calls that come straight in. But it shouldn't take a full minute.

An abnormal disconnect would mean that we have connected in the first place? We don't even have to try to disconnect for the CBTI to hold the line for 60sec.

I assume the phone line is a direct Telstra exchange line and not run through some sort of PABX?

No PABX's.

I'm doing some checking of the CBTI at the moment as Version 1.13 of TICA is being prepped for release (no firmware changes though, so the behaviour will be no different). I'll see what I can do to replicate the behaviour above.

Buy a Panasonic answering machine from Bing Lee (or some other Panasonic dealer) and test it out for yourself...

[Richo]

Buy a Panasonic answering machine from Bing Lee (or some other Panasonic dealer) and test it out for yourself...

Thanks for the extra info.

I am waiting for Don and Ryan to get back in to the office, they had today off as they did training over the weekend. They know where the test equipment is (i.e. answering machine). They should be back tommorrow. Not sure the brand yet.

Will do some testing and contact you if I need more info, I have you number :)

[Richo]

OK.... I'm getting somewhere in my investigations on this one... more info to follow soon.

Hang in there.

[Bad Boy Bubby]

pcs you might just have easily said that you had a problem fitting a panasonic answering machine to a stock standard c-bus phone interface have you asked panasonic why there answering machine doesnt work with c-bus?

[PSC]

This is not the first third party device that does not work with the CBTI.

Clipsal have a register of mobile phones that are incompatible with the CBTI, including my Nokia 6600.

So bearing in mind that Panasonic have been around and involved in the telecommunications field for a litter longer than Clipsal; and are less likely to release a product without thorough testing (unlike Clipsal), I would strongly suggest the problem is with Clipsal and NOT Panasonic!

But thanks for the tip...

[Duncan]

This is not the first third party device that does not work with the CBTI.

Clipsal have a register of mobile phones that are incompatible with the CBTI, including my Nokia 6600.

So bearing in mind that Panasonic have been around and involved in the telecommunications field for a litter longer than Clipsal; and are less likely to release a product without thorough testing (unlike Clipsal), I would strongly suggest the problem is with Clipsal and NOT Panasonic!

But thanks for the tip...

I seem to recall my Panasonic Cordless Phone/Answering Machine dropping my modem out every 25mins back in about 1999, oh well.. no sleep lost, I bought a different brand and it worked.

Its common knowledge that getting a variety of devices working on the same telephone line all trying to offer some form of value added loop or DTMF detection is rarely straightforward, you often end up mixing and matching devices until a compatible mix is found.

Its nothing to do with a lack of testing.

[PSC]

Could you please tell me what brands / models of answering machines you tested it on?

[Duncan]

Could you please tell me what brands / models of answering machines you tested it on?


No, I write software, I dont test hardware.

[PSC]

If Clipsal product testing was carried out on interfacing the CBTI with answering machines; surely someone in tech support would have informed me by now of a brand / model that does work.

My guess is that Clipsal has not carried out testing the CBTI with answering machines...

Is this another product that CIS has been released, awaiting the CIS integrators to pin point all the bugs?

Clipsal is starting to sound like it's a division of Microsoft more and more each day.

[Duncan]

If Clipsal product testing was carried out on interfacing the CBTI with answering machines; surely someone in tech support would have informed me by now of a brand / model that does work.

My guess is that Clipsal has not carried out testing the CBTI with answering machines...

Is this another product that CIS has been released, awaiting the CIS integrators to pin point all the bugs?

Clipsal is starting to sound like it's a division of Microsoft more and more each day.

Peter,

I know for a fact it was tested in various scenarios.. alarms, faxes, answering machines, ADSL.. I dont have the specifics to hand.. but they wouldnt be much use to you as these sort of issues are often further dependent on characteristics of the individual telephone lines and exchanges and presence of things like Pair Gain Systems and then the actual method of connection overlaid on top as well as the actual configuration of the 'other' consumer device which is often capable of operating in many different modes.. This is why you are called an "Integrator" and not an "Installer" :-)

[PSC]

We are both actually...

[ashleigh]

The telephone interface uses a modem IC from a well known manufacturer to do all of the telephone related functions. Its ability to do magical things is limited by the modem IC's capabilities.

This has the advantage that you can use it as a modem for dial-in update / commissioning access to a site, using the "data mode". It also means that its limited by the abilities of the modem IC.

It is fairly well known that trying to do lots of functions on a single phone line is a difficult and error-prone business. Factors that can make a difference include type of exchange, distance from exchange, condition of wiring between subscriber and exchange, and the nature of the devices attached to the line.

The issue with some mobile phones is bouncing back and forth with the modem IC manufacturer. At this stage it looks like some models of mobile phones are the culprit, but there is no quick and simple resolution to this one. Mobile phones DO NOT use DTMF to do their signalling of phone numbers and so on to the base station - its all done with clever digital stuff. When the call has been established, the mobile phones then switch the keyboard to generate DTMF. It *looks* like the DTMF generation is a bit cheap and nasty - falling outside the frequency and/or amplitude limits established by Bell Labs many years ago :mad: Follow up of this is difficult and time consuming, with lots of complex measurements that need to be made locally and sent to the USA for analysis.

The issue with answering machines appears to be related to some of those machines that stop the answer / record process when you pick up the (nearby) phone handset. Again, there is no simple change possible in the CBTI, though some further measurements and experiments might reveal more. At this stage, options on this one appear to be:

. choose an answering machine where you can disable the option to automatically cease recording when phone is picked up (OK if the answering machine is near the phone - just manually press the stop button when you pick up the phone)

. put in a second phone line :mad:

Further investigation will reveal more, but there won't be an instant/quick fix.

[Frank Mc Alinden]

Hi Peter
I wouldnt say its a lack of testing on Clipsals behalf is the reason why the telephone interface works , or even brand types...I use Homevision which also has a telephone interface and some functions of the interface wont work in some installations , ithink the main cause with the Homevision telephone interface is line noise ...In my installation for instance Homevision cant detect when another phone takes the line off hook....
Frank

[ashleigh]

Detecting line going off hook is one of the biggest pains in the universe :)

[Richo]

Phew...hasn't this thread been busy over the weekend.

Peter, as Ashliegh mentioned the problem between the answering machine and the CBTI is to do with the answering machine deciding to hang up when it detects a local pickup. The CBTI does a local pickup to sniff the DTMF tones to look for the access code.

There is no fault with either unit, it's just a behaviour mismatch. The immediate answer is to use an answering machine that doesn't hangup under this circumstance.

I can't understand the tone of you recent posts. I spent a good deal of time researching the problem the moment you posted to the forums. I even pulled the very expensive Exchange Simulator out our production facility to create a clean room test environment and experimented with work arounds to try and resolve the issue. I would have thought this was all a good thing. I especially spent extra time on this because of the very valuable and constructive feedback you have provide during the development on the C-Bus Toolkit.

I have been waiting on final approval to post a information notice on this issue, and hence the slight delay in responding on the forum. Hopefully this will be approved soon as I'm waiting on the exact text prior to releasing the next verison of Tica (Which has a had a bit of time spent on it to resolve communication issues with USB <-> Serial adapters and Laptop serial ports).

[Bad Boy Bubby]

I think we could all take a leaf out of Rossco's book (different thread re effect of ellectric fence on C-bus comms) and try this stuff out before installing. I would have thought it prudent to test out first any set up that used equipment from two or more different manufacturers rather than assume that it will work assume = ass u me.

[PSC]

Richo,

Do you have any updates re the CBTI / answering machine incompatibility?

[Richo]

Richo,

Do you have any updates re the CBTI / answering machine incompatibility?

Hi Peter,

Currently the only solution to having a CBTI and an Answering Machine on the same line is to use an answering machine that doesn't hang-up when it detects a local pickup.

We are researching the viability of changing the behaviour of the CBTI, but that will be a lengthly process and there will be no short term modification available.

[Nathan]

Im with you Psc I have tried 5 sites now with Tica and the first three work fine with old telstra answering machine and the other two were lucky to have a dedicated line, they were programmed with v1.10 and have had no dramas.
Then we got two more and used 1.11 and have had none of them working with any of our nokia mobile phones 5110, 5140, 3120, or can get them working on a answering machine.

I found that the first time the answering machine will work then the second ring in it grabs the line and locks it up and wont release it for a minute.

I also found that if I disconnect the answering machine I get to the Id password stage and it hangs up on me at random times with various phones.

Another time a got through the id password stage then it sounded like the voice recording was slowing to a stop and would then hang up.

I have tried this at several sites and several projects but cant get nothing that works.

The customer is getting annoyed now and so am I can someone sort this problem out???? :confused:

[ashleigh]

I'm working on this right now.

Answering machines issue has been referred to the manufacturer of the modem IC used in the CBTI.

Mobile phones issue has been tested with a range of about 10 phones, with about 5 known to work and about 5 known not to work. The modem appears to be doing the right thing and reporting the DTMF.

Why it is dropping the calls is still a mystery - lots of firmware to wade through, and I did not write it :mad:

Please accept our apologies, we are doing all we can. Answering machines issue is *not* likely to have quick resolution, the modem manufacturer has historically been quite slow to respond.

If a mircale occurs I *may* find why the mobiles are dropping out before Christmas :)

[ashleigh]

For those who find the line help for 1 minute after hangup: this is normal. I guess you want the period reduced?

Can you also confirm, who is the Telco provider?

If Telstra, then the hangup *should* be detected causing the CBTI to release the line. Other carriers might not send the busy signal when the remote end hangs up.

Remote hangup detection is another of those things which is *extremely* difficult to do...

It would help a lot to know answers to the above, ie:

. do you want the period to be shorter (if so, how much?)
. which telco is being used

Thanks

[Nathan]

Telstra were the carriers on all occassions except our trials in the office which was through a pabx analogue ext with still the same problem.

Hang up time would be preferred about 15 sec.

Also some other Additional Tica changes would be welcomed for eg

Can we get a range of inputs for each lighting device such as a swicth gui as on/off doesnt cut it for dial in control especially when most customers reqiure pulse triggering for A/C, Security Systems, Garage Doors and access control. If we could get a pulse or timer trigger this would save us having to trigger CA group adresses that trigger scenes in Schedule or Ctouch and then having to reset the trigger group adress to give status updates.

If you could also change the report style status acknowlegdement from just on/off or percentage to Open / Closed or even ability to add active voice and inactive voice for report style

The Menu to number command (ie press 1 for this 2 for that) doesnt keep regular format through the different modes so if you are getting formilar to jump ahead it changes through various menus.

And finally the library needs to be extended a little even if you can email a suggested library adds and import them as you need. I need words like Spa,Deck,veranda,audio,bar,watering,sprinkler,hell o,sorry,we,are,moment,later,will,be,with,outside,p ump,table,games,area,party,scene,multi,letter box,run,running,heating,cool,and cooling just to name a few.

Last one is to change the ring in time to over 3 rings say up to 10

One last grouch I just recieved another wireless gateway and the Learn dip switch wasnt soldered down so it wouldnt learn so I repaired it and I know it voids warranty but shoudnt this be picked up in QA testing??? :confused:

Hope this helps

[ashleigh]

Telstra were the carriers on all occassions except our trials in the office which was through a pabx analogue ext with still the same problem.

Hang up time would be preferred about 15 sec.
OK, thanks. (Try this - get somebody to call you, and when they hang up, make sure you do not. You should hear the engaged signal). So when the remote party hangs up, the CBTI should detect this busy tone and also hang up. The 1 minute timeout is there in case the hangup is not detected. In some countries the Telstra behaviour does not exist, so it has to work this way anyhow.

If I reduce the timeout period, you will see that for no activity (if you don't key any DTMF to the unit for 15 seconds) it will hang up. In the event that remote hangup is not notified by the exchange, the only other alternative is an inactivity timeout....

Doing 15 seconds won't be too hard... We shut down here for Christmas after COB Wednesday, so I won't be able to have a firmware / TICA update for you until about mid Jan. Hope that will be OK.


Also some other Additional Tica changes would be welcomed for eg

Can we get a range of inputs for each lighting device such as a swicth gui as on/off doesnt cut it for dial in control especially when most customers reqiure pulse triggering for A/C, Security Systems, Garage Doors and access control. If we could get a pulse or timer trigger this would save us having to trigger CA group adresses that trigger scenes in Schedule or Ctouch and then having to reset the trigger group adress to give status updates.

Some of this is targeted to a future firmware release.


If you could also change the report style status acknowlegdement from just on/off or percentage to Open / Closed or even ability to add active voice and inactive voice for report style
Difficult. I did not write the firmware so this requires a lot more digging and changes.... No promises.


The Menu to number command (ie press 1 for this 2 for that) doesnt keep regular format through the different modes so if you are getting formilar to jump ahead it changes through various menus.
The menu structure has been set up so that it will expand easily for future changes (eg adding scene triggering).



And finally the library needs to be extended a little even if you can email a suggested library adds and import them as you need. I need words like Spa,Deck,veranda,audio,bar,watering,sprinkler,hell o,sorry,we,are,moment,later,will,be,with,outside,p ump,table,games,area,party,scene,multi,letter box,run,running,heating,cool,and cooling just to name a few.

There have been a number of requests for new words. We plan on adding support so you can add/change words in future. However are not currently planning to updatethe current libraries. We are currently supporting English, Mandarin, Cantonese & Malaysian. All words are studio mastered and then format converted to suit. Its a long and expensive process, and the current word set only just fits into the available memory.

Suggestions are being accumulated and we may do a monster big word library update in future, but again, no promises.


Last one is to change the ring in time to over 3 rings say up to 10

Yep. This one is a pain. Each country imposes different requirements on the number of rings to answer. At the moment we are driven by the need to be compliant everywhere. Plan on changing this the ring limits can become a property of the selected country of installation. Unfortunately, this is far from a trivial change.

Same goes for blind dialling...

Hope to have these addressed either partially or completely in an update early next year.


One last grouch I just recieved another wireless gateway and the Learn dip switch wasnt soldered down so it wouldnt learn so I repaired it and I know it voids warranty but shoudnt this be picked up in QA testing??? :confused:

Sorry about that. Yes, it should have been picked up. It would have been preferable for the unit to be returned so that QA know about it and can track back through the manufacture process. I know its not always convenient to return, though.

If it happens again we want to know about it, because its not good enough to ship items that are not built properly.

[Richo]

Im with you Psc I have tried 5 sites now with Tica and the first three work fine with old telstra answering machine and the other two were lucky to have a dedicated line, they were programmed with v1.10 and have had no dramas.
Then we got two more and used 1.11 and have had none of them working with any of our nokia mobile phones 5110, 5140, 3120, or can get them working on a answering machine.

Are you saying you notice a difference between programming a CBTI with Tica Version 1.10 and 1.11?

Did you find a change in behaviour between these 2 versions?

[Nathan]

Programming is nearly indentical between the two versions just different skin and voice styles it was just a weird fact that one site that was programmed V1.10 all our mobiles tried work fine.

Same exchange up the road V1.11 and all our mobiles dont work it just hangs up at random stages but only a few times can we get to the controlling of devices where the voice seems to get distorted then hangs up.

Then I took it back to the office same thing this is off a analogue pabx ext.

The only thing a noticed that was different was a modem firmware update text line happening when the download was being done.

I have one more tica hear new in its box any sugestions????

[ashleigh]

Here is a late update before Christmas.

It *looks* like some mobile phones do have lousy DTMF generators, and the effect of this is that the CBTI detects what it thinks is a BUSY (ie engaged) signal interspersed amongst the DTMF.

The moment it sees what it thinks of as BUSY, it hangs up.

The plan, after returning from the Christmas silly season, is to try and change the code to make it look for several busy indications in a row, and to generally fiddle the modem driver code to make the DTMF / busy detection a bit more robust.

There won't be any more progress here until early January, and depending onthe number of crises that spring up there *might* be a cure by about the end of Jan.

Sorry about the delay on this, its not code I wrote and I'm having some difficulty understanding what it does. But I'm getting there...

If possible, I'll roll in some of the other changes that have been asked for also, but more on that when or if its known what I can do.

[ashleigh]

Another update.

I'm making progress on some fixes!

I've figured out what to do about the mobile phones.

I've figured out PART of what to do about the issues of holding the line after a call (need more info to turn this into a really good fix).

I reckon I can sort out the Telstra / Homemessages 101 / stutter dial tone / blind dialling thing.

I know its taking a while, but as pointed out before, I did not write the firmware in the 5100TAU and its taking me a while to get my head around it.

To get a *really* good set of fixes will still be a few weeks.

Any of you fellas want an interim release that knocks over some of the problems, or do you want to wait a bit longer?

[marklittlefield]

Can anyone tell me if they have successfully install a Telephone interface in line with an answering machine, if so what is the brand and model number of that answering machine. I have a very irate client who has been waiting for a fix for about 4 months.

[Richo]

Can anyone tell me if they have successfully install a Telephone interface in line with an answering machine, if so what is the brand and model number of that answering machine. I have a very irate client who has been waiting for a fix for about 4 months.

Hi Mark,

The CBTI currently works as designed. We are endevouring to change the designed behaviour to make the CBTI more compatiable with a greater range of answering machines, but this is considered a new feature rather than a fix. It looks like it will require sourcing a different modem chip from our suppliers and therefore is a significant change to the product. This means that it may take many months or longer. There is no immediate timeframe to change the current behaviour.

Unfortunately I don't have a compatiability list available.

[ashleigh]

I have just tested a Uniden 7500AI and it works fine. This is a digital answering machine.

I've done a quick internet search and I *suspect* (though don't have any to test) that pretty much any digital answering machine with 2-way voice recording will be OK or can be set up to be OK. There are a few brands readily available.

[ashleigh]

Oh yeah - and all being well we should have a new TICA Beta out in about 2 weeks, all going well....

[Caroline Springs]

Ashleigh,

I'm about to install a Telephone interface in my project and I see that there is a list of problems known and kind of known posted by different people in different situations.

Given that this is an expensive device, I would like to see a list of the current problems that have been reported. So, I'll be aware of them, install and purchase additional components accordingly. For example, is there any problem with PABX systems? what is it? But please don't answer this single example, help me with the problems I will be facing.

Thanks,
Oscar.

[Richo]

I'm about to install a Telephone interface in my project and I see that there is a list of problems known and kind of known posted by different people in different situations.

Given that this is an expensive device, I would like to see a list of the current problems that have been reported. So, I'll be aware of them, install and purchase additional components accordingly. For example, is there any problem with PABX systems? what is it? But please don't answer this single example, help me with the problems I will be facing.

Hi,

A list of some of the problem that have been fixed in Tica 2.0 (CBTI firmware 6.0) is here http://www.cbusforums.com/forums/showthread.php?t=1129 . This version of Tica will be available soon as a Public Beta.

Tica V2.0 (CBTI firmware 6.0) fixed most known issues with the CBTI.

The only outstanding issue (that I'm aware of) that may cause problems is how the CBTI does DTMF snooping. This has been documented well on this forum, but basically can be summerised as the CBTI needs to actually pickup the line to snoop DTMF tones. This snooping is for 2 purposes:
- Detecting Local Access code entry (either from local handset or whilst a user is dialled in)
- Recording the phone number that someone is dialling, this is then published on C-Bus for other devices to consume if they want.

The problem with the CBTI having to pickup the phone line to do this is some other devices *may* react to this pickup. This can be a problem with *some* answering machines that decide the user has answered the call and therefore stops recording messages. This can be avoided by using a compatible answering machine (one that doesn't detect pickups, often marked as a machine that can record conversations). This has also been observed where someone is using a CBTI on the same line as a dialup Internet connection, where the modem detects the pickup and thinks someone is using the phone and won't dial out.

This problem is prevented for *some* users in Tica v2.0 as there is now the ability to turn off detection of Access Code, and therefore the CBTI doesn't do a local pickup. However, this of course, removes the handy feature of being able ot jump into the CBTI whilst dialled in to other devices on the phone line.

This problem is part of the actual design of the modem we use, which is supplied from a 3rd party supplier. Unfortunately they are unwilling to actually change the behaviour, but have given us the source code to the modem so we can fix it ourselves. This however will take quite awhile as the source code is thousands of lines of assembler language which is quite difficult to work with when it's someone elses code. So, basically, we are planning on changing this behaviour, but it will *not* happen soon.

Hope this helps with your evaluation.

[Richo]

Everything you need to know is here:

http://www.cbusforums.com/forums/showthread.php?t=1136