Connect:AMP keeps cutting out...all components hardwired

One way to knock: @Corry P 

Thanks for the help everyone and chiming in so quickly. Just FYI, I intentionally post here for SEO purposes and to bring visibility to the issue for the wider community at large. Sometimes phone calls aren’t quite good enough in my opinion.

again many thanks and hopefully the issue gets resolved with updated firmware.

Hi @Gerald Harmon 

I’m not sure which update you are referring to - we haven’t released one since the 12th of December. However, I feel I should set some expectations here - although some other products are currently experiencing similar symptoms which we are doing our best to address, I do not expect the issue your Connect:AMP is affected by to ever be fixed by a software update - the issue is physical (hardware) in nature.

If you have belatedly installed the last update, I suspect the reboot that happened afterwards has delayed the occurrence of the symptoms for a small time.

Sorry to be the bearer of bad news.

Gerald, since the Sonos response is based on a look at your diagnostics one has to assume that they have a basis for saying that your issue is different from the one plaguing other Connect Amps running S2.

You can test this for yourself if you wish - move the failing Connect Amp back to S1 and see if it continues to fail. If instead it works fine under S1, Sonos may suffer another red face.

PS: I just noticed your comment about the new update. For Sonos to be correct, your Connect Amp units should still fail even after that update is installed?

Hi @Gerald Harmon 

I understand your point completely, but as the specific hardware fault in this case is related to failing flash memory, a change in how much memory is used, or in how it is used, can change things. The update you got in June 2023 - not due to it’s content - was likely the point at which enough memory was failing to become an issue for your particular device, and installing that latest firmware may have taken the memory past it’s tipping point.

Having said all that, we're close to releasing an update that fixes certain conditions with older players. I still recommend that you do not get your hopes up, however, as things will still likely vary on a device-by-device basis.

I read that Staff post differently here and that @Corry P was just referring to this users specific Connect Amp, rather than every device out there. Stating that any forthcoming fix might not resolve their own reported issue, as it appears the device has memory locations that are perhaps ‘already’ marked internally as being unusable. This can apply to S1 or S2 updates, was my thinking - In fact I wonder if the S1/S2 downgrade process might in fact make things even worse with such devices, as I guess that can ‘perhaps’ put additional demand on a devices memory🤔? 

All muddy water to us on the outside. And while a serious issue for individual users, I would think that the total number of affected Connect Amps - the ones that could and were taken to S2 - is small compared to say play 1 units. I remember reading somewhere that play 3 and play 1 units are also in line to suffer the same issue with S2, and if so it will be a much bigger problem given the numbers of those units out there as compared to Connect Amps. Makes me glad that I stuck to S1 with units whose hardware was meant for S1; and nothing really changes now in S1 for months on end and when it does, it is probably minor tweaks. Reading all these has made me decide to not apply even S1 updates to my 12 Sonos units going forward; I don’t think that the risk of being exposed to new security issues is large enough to risk breaking the units by the next S1 upgrade. Let sleeping dogs lie!

Hi @Andrew-s 

You also seem to confusing the two issues that have the same symptoms, rather like @Kumar. I gather your 3 Connect:Amps have been affected more recently? In that case - though I cannot verify without diagnostics - it seems likely to me that, in your case, a fix will be coming soon. The same cannot be said as confidently for @Gerald Harmon, however (whose diagnostics I have seen) as he is affected by a completely different issue. Technically, this makes you off-topic for this thread, but understandably so.

I hope this helps.

I must have erred in thinking that the Gerald issues were in an S2 environment. But your last sentence quoted makes it doubly clear to me that I should shut down all future S1 updates for my system. I don’t want to expose my two Connect Amps that work perfectly, to any S1 update change driven exposure in future. Or any of my other ten units including play 1 units and early years Sub, to such.

Hi @Gerald Harmon 

Apologies - there was in fact a Service Pack release for S2 last week (on the 24th), but a typo led me to believe otherwise. I can’t find any information on what it addressed, but as a Service Pack, it would not have been a great deal.

thanks @Corry P .  You’re doing great work for the company here; I know it’s tough.

The company hopefully takes notes of your engagement efforts here.  It’s good for them.

With that said, and I don’t mean to drag this out unnecessarily, even failing memory indicates an issue if that’s the case.  it’s hard for me to understand a solid state component having a high failure rate on a device that’s likely 5-6 years old.  I know we can’t have 20 year VCRs anymore, lol, but i would hope the internal memory on the C:A units would last longer than that.

Thanks again for your time and engagement! 

S2 uses more memory that S1 so S2 pushes data into areas that were previously free or used for things that the failures were not noticed.

If the update added any bytes to the firmware (we users can’t know) it could have pushed some units that were previously performing well into the problem group.

Flash memory is not durable, it has a lifetime usually given in “number of writes” which also includes erase cycles if present. Older flash had lower limits than newer flash, originally the limit was around 100 cycles! Sonos uses much newer flash but you’d have to get the chip number and look up the specs to see what the limit is, also it will vary by chip supplier and even lot number.

Newer flash added a lot of nice stuff like wear leveling, bad block mapping and such to work around the failing memory issue but all hit the same limit, not enough good blocks left to give the system the memory it needs. Some designs error out, some just give you a ‘best effort’ and possibly corrupt data when read back out.