DTMF / touch tones do not work when roaming in the US

[chrishammond]

Hey everyone,

I have a very strange, but entirely repeatable, problem when roaming with my EE plan in the US. I use a Google Pixel 7 Pro, but I remember experiencing this with a Pixel 6 Pro too back when I had that device.

Whenever I am Stateside, any calls that require touch tone / DTMF dialling to navigate an IVR (e.g., "for another twenty options you didn't ask for, press 1") do not work. I am currently roaming on AT&T, and it does not seem to make a difference whether I am on 4G or 5G. Sometimes my phone will drift over to T-Mobile (US), but this is usually only for a short while so I haven't had the chance to test this network yet.

Not only have I observed this, but I can actually reliably reproduce it using Twilio. I have a trial account with them, so any time I get a call I'll receive the warning message that I'm on a trial account, followed by a prompt to "press any key to run [my] code". No matter how many times I press any key (or which key I press), Twilio will time out and terminate the call. Calling my Skype number, on the other hand, will work just fine and I'll be able to run my code by pressing a key that way.

I have also observed this with Google's hardware support team. When you request a callback, you have to press a key to speak to a human; otherwise, you just get cut off. Again, no matter how many times I tried to do this, it wouldn't work and I had to use an alternative phone that is not roaming in the US.

It should go without saying, but I have no issues at all with DTMF / touch tone dialling when I am in the UK and no longer roaming, so there is definitely an issue here.

Any ideas on how to fix this? I saw in another post that some phones have the option for a longer tone, but I do not see that option on my Pixel. Also, I found a thread on Apple's forums in which an iOS user had the same problem, so this seems to be a network issue rather than a handset one.

Thanks in advance!

Chris

[Katie_B]

Hi @Tony_Jef. 

Thanks for keeping us posted. 

I'm sorry to hear you have not yet heard from the agent who was supposed to be in touch. This is not what we would expect when call backs are arranged.

As we have no account access here on the community to be able to look into your account the next step would be completing the online form and a member of our complaints team will be in touch to discuss this further.

Katie

[Tony_Jef]

Hi @Katie_B 

Thanks for that, but I'm still in touch with the handset TL that I spoke to and will be having a convo with him on Monday, so not yet going down the "complaints" track!

I've now replaced my "legacy" DECT phone/answering machine setup and installed EE Home Plus phones, and I've discovered something interesting.

The 1571 service allows the user to retrieve voicemail messages remotely, just as my old answering machine did. You call the landline number from any remote phone, and when the 1571 system kicks in and plays the outgoing message, you dial "*" on the phone you are calling in on, 1571 prompts you to enter your PIN, and you can then retrieve any voice messages. Sound familiar at all? It is EXACTLY the same sequence as I used to use to retrieve messages from the old answering machine. The difference is that now, when I press "*", either from my EE SIM or my GiffGaff SIM, the 1571 system responds to it as it should.

Conclusions:

1) This was (as I had already concluded and discussed with the handset TL) never about the behaviour of the SIMs or the handsets; it is all about how DTMF codes are encoded and transmitted within the network(s), how they are consumed, and by what.
2) It is now pretty clear (at least to me!) that in the case of the EE/BT/Openreach infrastructure, the assumption is that DTMF codes will be consumed by the 1571 system, and therefore they are being deliberately filtered out from the feed to any digital voice adapters that are used to attach legacy equipment. The fact that my legacy answering machine worked when I called in via a Giffgaff SIM would seem to suggest that DTMF codes are handled differently in the O2 network - maybe transmitted "in clear" as it were, as part of the encoded voice channel, rather than being encoded in out-of-band signalling.

There's an implication here that this behaviour potentially obsoletes any legacy equipment out there that relies on DTMF, and that is something that users will be blissfully unaware of. Particularly of concern - the online shop on the EE website carries a number of models of (landline) phone - some (like the ones I have bought) are compatible with Digital Voice and integrate with 1571, others (and some very similar to the set that I have just disconnected) can only be used on a Digital Voice system by connecting its base station to a Digital Voice adapter or directly to the legacy phone socket on the back of the router. Unfortunately this is a major trap for the unwary - the shop doesn't make it at all clear which models are DV compatible and which are basically legacy analogue devices. Consequently, it seems likely that there are DV customers out there that will purchase such equipment and later will discover that it doesn't function as they expect it to. This is NOT good.

[Tony_Jef]

I've been in touch with the TL I was in contact with, and updated him with the above. 

I have performed one final test - I plugged the old analogue answering machine into the back of the router, then called the home number using my EE SIM, but instead of using the keypad on the phone to press the "*" key I used a DTMF tone generator (actually a DTMF sounder app loaded onto an old Android phone) held close to the phone's mic, thus forcing the "*" tone to be carried across the network as voice (as opposed to control codes).  This worked fine - i.e., the "*" DTMF code reached the answering machine and it asked for my access PIN, which is as it should be.

This IMHO absolutely confirms what I have said in the previous post, that when you use the phone's keypad the DTMF is being encoded and carried as control codes across the EE network, and they are being intercepted and diverted to the 1571 system at the receiving end, rather than being converted into voice and passed to any attached legacy analogue device. I'm sure this will be absolutely no surprise to the engineers that put the system together!

Hopefully, this information will filter back to the engineers responsible for designing the network; whether they do anything with it is another question. In the meantime, the "workaround" hack is to deploy an old phone as a DTMF tone generator as I did. It used to be possible to buy DTMF sounders back in the day, but they don't seem to exist online anymore. Progress...