Siemens Gigaset DA810A landline telephone with answering machine resets every minute

[Nick Nougat]

I'm posting this here since there does not seem to be a more fitting category.

The Device
The Gigaset DA810A is a landline telephone with an LCD screen, an answering machine and integrated telephone book. I have the version for the European (i.e. German) market.
It has two separate modules that can operate independently of each other: One is the basic telephony part that always works and is powered by the DC voltage coming through the telephone wire. It relies the TEA1110A telephone transmission and dialler chip for that.
The 'smart' part of the phone with the answering machine is powered via an external 7V power supply.
The system has two separate SoCs. One is built into the display unit and drives the LCD. It probably also stores the phone book. At least some of the phone controls are routed to it. That chip is covered in a black resin, making it hard to access and identify.
The other chip is on the main PCB and is accessible. It seems to be mostly specialized for DSP operations. It is e.g. connected to the room mic and speaker. That chip is identifiable, but seems to be rare or custom made. I was not able to find a data sheet anywhere. It has the label D56561AA3ALC.
The display chip seems to be unpowered when the external power supply is not plugged in, but starts up as soon as a call comes in or the receiver is lifted.

The Problem
After moving to a new home, the answering machine module has been acting up. When I plug in the external power supply it starts up fine and works, but after ~60s it resets, starts up again and repeats the same after 60s again.

What I did
I've been monitoring the voltages and oscillators for stability and have done an extensive analysis of the circuitry. So far I have not been able to find any particular faulty component and nothing seems to get hot.
I've also been trying to see interesting signals on my oscilloscope just before the reset on the DSP chip pins, but wasn't able to find anything conclusive.
I also grabbed the flash rom from the main board chip connected to the D56561AA3ALC, but the data seems encrypted, so I could not find anything useful there, easily.

My working theory is that the system waits for a resource or an event to happen until a timeout of 60s occurs, which causes an internal error condition to be flagged which in turn causes a self-soft-reset.
My problem is that I have limited knowledge about the SoCs that are built into the phone, so I can't easily tell what input they expect.

I could identify and get data sheets for almost all of the other ICs on the board. One is the MC34119, telephone amp, the others are flash chips (one for the Dsp chip, one for the display chip). The only one I could not identify was one labeled PT1103, but I'm pretty certain that this is an Amp for the room speaker.
You can find pictures of all the chips and both sides of the boards on Wikimedia, where someone was nice enough to upload all this: https://commons.wikimedia.org/wiki/C...Gigaset_DA810A
I have a slightly different board revision, but it's pretty much identical. Some parts of the board are hidden behind the electrolytic capacitors on the pictures, but most of it is visible.

Questions
If anyone here could provide me with a data sheet for the D56561AA3ALC, that would be great! If anyone has a sheet for the PT1103, I'd also be curious if my assessment was correct.
I would also appreciate any pointers to what else I could try to narrow the problem down.

Thanks a lot for any help!

Here are some of the annotations I made to the original Wikmedia images. I am happy to take additional pictures if required or provide data sheets etc.

[redwire]

I had an old Nortel phone do the constant restarting and it was bad electrolytic capacitors. They are old now. Maybe check some of them.
It is a regulatory requirement for the phone to work with no mains power. So they would put in a second, independent POTS chipset just for that, to run off 48V telephone line power.

I do see airwaves are polluted from dirty routers or cell tower and it does crash my cordless phone MCU, rarely. I think it's at 900MHz but there were times something was fighting for the channels. You could try run the base somewhere else.

"Resetting a Gigaset DA device
DA710 & DA810A
IMPORTANT: This procedure will erase all memory contents of the phone including the phonebook, redial list, etc.!"
Pick up the handset
Press the menu button
Enter 1 2 5 0
Confirm with OK
Hang up the handset

[petehall347]

IThe Problem
After moving to a new home, the answering machine module has been acting up. When I plug in the external power supply it starts up fine and works, but after ~60s it resets, starts up again and repeats the same after 60s again.
}[/ATTACH]

the clues are here . something has changed causing the problem . maybe the power supply has a problem .

[stj]

should it work without a telephone connection?
and yes, suspect 1 is the psu or the wiring/connectors from it.
but if it does use the line for something then maybe a DSL filter is causing problems
or somebody is using a stupid "internet through mains wiring" setup nearby

[Nick Nougat]

Thanks for the replies. I now realize that, because my initial analysis was some time ago, I forgot to mention that the PSU was my first suspect, too, but the behavior was the same when I hooked it to a bench PSU, Also I could not find any voltage drop or power ripple with the original PSU.
As I said in my original post, I also checked all voltage regulators for stable output and they are stable. They only briefly go a few mV higher after the device resets, but I'd say this is because it briefly draws less power then.
The reset is pretty easy to detect because the speaker voltage drops abruptly causing an audible click.

should it work without a telephone connection?

Yes. As I said, it operates like an old "dumb" phone and gets its power from the landline. The LCD and phone book then only work after a second or so which it takes to boot the system up.
I'm pretty sure it's a regulation here that devices mustn't draw power except when a call comes in or the receiver is lifted to reduce load/costs for the operators.
Of course these days the power usually comes from a VoIP router anyway, which is also the case for me.

but if it does use the line for something then maybe a DSL filter is causing problems
or somebody is using a stupid "internet through mains wiring" setup nearby

I don't have a DSL filter since I get my internet via TV cable, but I think we can rule that out, too, because the device does the reset routine even when not connected to the telephone line.
I would also tend to rule out that it's some kind of mains modulation, since given that I tried it on two different PSUs and in different rooms I'd assume that one of the PSUs would filter correctly or differently.
I also don't have problems with any other device in the household.
OTOH, I can look at the voltages again with my newer, somewhat better oscilloscope. When I last looked, I only had a very basic Fnirsi device that can only go up to a few KHz, so I might have missed something.
I'll report back on that.
What I would expect, though is more sporadic or random failures if the problem were caused by ripple/jitter, but it's always and repeatedly after 60s, which seems a suspiciously "human" number.

I'll add the missing facts to the post, so that it's easier to catch up.

[petehall347]

you tried master reset ?

[redwire]

I had an old Nortel phone do the constant restarting and it was bad electrolytic capacitors. They are old now. Maybe check some of them.
It is a regulatory requirement for the phone to work with no mains power. So they would put in a second, independent POTS chipset just for that, to run off 48V telephone line power.

I do see airwaves are polluted from dirty routers or cell tower and it does crash my cordless phone MCU, rarely. I think it's at 900MHz but there were times something was fighting for the channels. You could try run the base somewhere else.

"Resetting a Gigaset DA device
DA710 & DA810A
IMPORTANT: This procedure will erase all memory contents of the phone including the phonebook, redial list, etc.!"
Pick up the handset
Press the menu button
Enter 1 2 5 0
Confirm with OK
Hang up the handset

[Nick Nougat]

Thanks for the replies. I checked the voltage of the PSU again with my new oscilloscope, a Hantek 6022BL and I found that the base noise level of the oscilloscope is quite high - but it's the best I have atm, so the result is a bit inconclusive.
Re-measuring with my multimeter and my Fnirsi DSO-TC3 I found fluctuations only in the mV, so I'd rule this out for now.

Also thanks for the hints about resetting the device - I wasn't aware that this was possible. I'll try to back up the contents of the display side flash chip before resetting - I assume that is where the address book is stored, since I could not find anything on the other EEPROM.
And thanks for the capacitor hint - I suspected something like that already, but was skeptical because of the "human" seeming 60s, and I wanted to exhaust other plausible options before unsoldering anything. I'll go and unsolder one or two of the electrolytics that are close to the Dsp chip and will check them.
I'll report back once I've done that.

[stj]

i'm curious, you said the speaker voltage drops during reset.
find the amplifier chip and trace it's power back - see if it's actually switched by the controller or not.

[CapLeaker]

I had an old Nortel phone do the constant restarting and it was bad electrolytic capacitors. They are old now. Maybe check some of them.
It is a regulatory requirement for the phone to work with no mains power. So they would put in a second, independent POTS chipset just for that, to run off 48V telephone line power.

I do see airwaves are polluted from dirty routers or cell tower and it does crash my cordless phone MCU, rarely. I think it's at 900MHz but there were times something was fighting for the channels. You could try run the base somewhere else.

"Resetting a Gigaset DA device
DA710 & DA810A
IMPORTANT: This procedure will erase all memory contents of the phone including the phonebook, redial list, etc.!"
Pick up the handset
Press the menu button
Enter 1 2 5 0
Confirm with OK
Hang up the handset

Yes, the old Nortel phones had this exact problem and also the audio quality was bad giving a bunch of hizzing / buzzing noise. I also fixed many of these by changing capacitors inside of them.

For this one I haven't got much of a clue. I would look with a thermal camera at it and scope the pins of the ICc to see if I could find the reset pin. I also would try to freeze the IC's and other components to see if that has an effect on the duration the phone resets.

[Nick Nougat]

OK, I finally can report back. I have unsoldered two of the larger 470 uF capacitors and tested them out of circuit. They are all OK and actually read around 510 uF with an ESR of around 0.28 Ohms. Of course, the pads on that board were extremely sensitive and I managed to detach them for one of the caps - that was a first for me.
But, well, I'm still learning and that gave me an opportunity to practice board repair... I managed to bridge the gaps and attach the cap legs to vias close by...
Next I tried the hard reset after making a backup of the secondary flash, and..... drumroll.... that fixed the problem!
In retrospect I should have noticed that the clock was always showing 23:59. My latest hypothesis is that the date had some invalid state and when the minute was over and the day changed, it triggered a software exception causing a reset. Then after reboot the system returned to the last known state, which was 23:59 again. Rinse and repeat. As a software developer, I should have guessed that something like this could be the root cause, but oh well - at least I learned quite a lot about analyzing circuits. Probably me moving the device to a new home caused some freak bit flips or a similar data corruption leading to that state.

So thanks a lot redwire and petehall347 (I guess that's what you meant with master reset?) for the solution. And of course thanks to everyone else for your contributions. I'm pretty sure this won't be the last repair project I'll be struggling with

[petehall347]

yes full reset . its a bit like turning it off on and on properly .