SPA112 reboots after hangup and ...

[fredericb1]

I just bought a SPA112 and set it up for voip.ms

I've noticed that it's rebooting itself quite often and I have pin pointed one of the cases when it happens and I can reproduce it.

1. Dial a voip.ms extension  from the Line1, makes my MicroSIP ring on my desktop.

2. Answer on the desktop.

3. Hangup in MicroSIP.

Expected behavior, that happens sometimes: Line1 gets informed of hangup, turns sound off and then goes to Dial Tone.

Experienced behavior: Sometimes the expected behavior happens, but usually, the line turns off, it tries to get the dial tone off, but the SPA112 just reboots.

Side note: Sometimes the SPA112 just reboots for ever (looping)

This behavior also happens in other cases:

 - I pickup the phone from line1, and it reboots.

 - I pickup the phone from line1, hang up not long after, and it reboots.

 - During a phone call that was 37 minutes long, it suddenly rebooted.

- After ending a phone call with a telephone (but I assume the other user hanged up before me, thus creating this behavior)

- Sometimes it just randomly reboots, but it's been better than my 1st SPA112 that I exchanged which simply froze... (no need to go into that one)

 - I receive a call, I pickup and it reboots. The other party keeps on ringing on his side.

If I hangup the phone on line1 before the other user, I seems to work normally.

It basically seems to reboot when it switches from CC_CST_IDLE to CC_CST_DIALING after the other party hangs up.

I'm attaching a Log file that includes this behavior toward the end.

Log notes:

 - I changed some strings to hide my username and phone number.

 - Line 697: The connection is opened (people can talk)

 - Line 698 to 833 (inclusive) The other party hung up on me and by the time we get to line 833, this is when the SPA112 just rebooted.

 - Line 834: SPA112 rebooted.

Handset Note: I'm using a VISTA 100 BELL telephone from Northern Telecom (without a power supply connected to it)

If anyone can help me figure the problem it would be greatly appreciated. It's the 2nd SPA112 that I go through and it's driving me nuts. (The 1st SPA112 didn't have the same issue, it was worst and would loose the LAN connection while keeping the LAN light ON and the SIP registration would get lost, power cycling wouldn't always make the device boot at all).

Hope the logs help, but naturally since the device seems to reboot, I'm not sure it's including what caused the issue within those logs.

[fredericb1]

Oh, I forgot, if I disable the CPC, I never get the Dial Tone after the other party hangs up. If I set it to 0.001, I get a dial tone. And with 0.001 (vs 0.5's default) the device doesn't reboot when the other party hangs up. So I'll leave it to 0.001.

[Dan Lukes]

Well, you may consider it acceptable workaround.

By the way, you should not get Dial Tone again after other party hangs up with no hang up from your side. Despite it may work with CPC, it's not expected mode of operation on POTS line.

[fredericb1]

You are right, I just double-check with my land-line and when the other party hangs up, it goes on Busy Signal (or some fast beep). In the old days, I believe it would go back to a Dial Tone, but maybe my memory is failing. So, I've disabled it.

[Dan Lukes]

It's rather "Congestion" (a.k.a "fast busy") than standard "Busy" signal.

By the way, I remember analog phone switches based on stepping relays (it seems I'm older than you). Call can be terminated by caller only here (called phone has got fast busy signal then). Called side has no way to terminate the call at all - even on-hook, the call path has not been dropped.

But it's history ;-)

[fredericb1]

Interesting. I just discovered that if I pickup at anytime that the ringer is going, it reboots. If I pickup between rings, it's fine. Mysterious. Does that give you any more ideas on what might be causing this ?

EDIT: I changed the Ring1 Cadence, so that it rings only .5/4  and it obviously helps, since it's harder to pickup the phone while it's ringing. I tried picking up right after 0.5 and it works, I tried calling during the ring and it reboots as it always does.

Now this is probably the last issue that I gotta solve.

PS: I tried this with more than one phone and it always behaves this way. The two phones I tried with are both not powered and wired direct-line phones.

I tried changing the "Ring Waveform" to Sinusoid, that didn't change anything. I also tried changing the voltage to 90 instead of 85 (default), and still hasn't changed anything. I also tried "Synchronized Ring", but that obviously changes nothing. I haven't tried changing the "Ring Frequency", but I have a feeling it won't change anything either.

Sounds like the same issue as I found here: https://community.acanac.com/acanac/viewtopic.php?f=17&t=27394 ; This guy describes it exactly as it is.

I don't believe the plug the power supply is plugged in is underpowered, I've tried it standalone on it's own breaker.

Is there a way to maybe introduce a delay between the ring stopping and the line turning on, so as to draw less power from the transformer ? Is this even the cause?

I also tried with two phones, using both FXS ports. Both devices ring synchronized and use different SIP sub-accounts that are tied by a ring group. I just can't answer on a device that is ringing. If I un-synchronize the rings, I can pickup the phone that isn't ringing at that moment without issue, while the other phone is ringing at that time.

Any ideas ?

[Dan Lukes]

Well, we have just few SPA112 deployed as I never considered them to be stable enough for production deployment. We are using old gold PAP2T instead. But even with those few SPA112 we have no such issue reported so far. In advance I failed to reproduce the issue in my lab (but I have no end phone with old mechanical ringer, just those modern with electronic ringer, so  it's hard to pick up during incoming ring signal as they doesn't ring in sync).

Well, I'm out of advices.

It's either SPA112 malfunction. It may not affect just one box, it may be related to particular hardware revision, or particular lot.

It may be hardware/firmware bug/race condition that depends on complex combination of configuration values and/or environment as well.

You can try contact Cisco SMB support. Even in the case you have no active support contract. But as I claimed above, I have no good experience with them. Their list of response is so short. They will tell you it's configuration bug on your side, with no further details how to analyze or solve. Or they will offer you SPA112 replacement (if under warranty). It may or may not help ...

[fredericb1]

I'm so baffled, it suddenly worked just fine earlier this morning. All was perfect. Now things are not working at all. When it's supposed to ring, it reboots often.

It's probably easy to blame this on environmental (phone / power source / temperature / humidity) in this case... The phones would probably be the first, but than, those are standard phones, and I would have imagined this SPA would handle normal voltages properly, just like my Land line provider supports.
 

[Dan Lukes]

The line is powered by SPA (SPA for this line is the same as Telco's PBX for  land line). Phone device is power consumer only - it must newer power the line.

Basic phone functions (e.g. incoming ring, outgoing dial) can be maintained even with no local power supply. Just line-supplied power is sufficient  (in some countries it's mandatory to maintain basic phone functions even during blackout of local power supply)

Voltage used by SPA112 is not configurable. Just line impedance. Most of phones can handle wide range of line voltages, but I have experience with phone device incompatible with PAP2T (SPA112 predecessor) as well.

[fredericb1]

Do you think that modifying the line impedance would change anything ?

[Dan Lukes]

Anything ? Yes. Wrong line impedance may cause sound distortion.

No, I see no way the improper line impedance can cause reboots.

[fredericb1]

Ok. It's set at 600 (default) and I'm in North America and the phone line is short, so I assume I shouldn't change it.

I've returned the device at the store. I have a credit and I'm contemplating what to do next. Either get another SPA112 or a HT702. It seems the SPA112 has more options/features, but I can't tell. Don't know which one is better, and this is probably not a conversation to have in this forum :oP

[Dan Lukes]

I never signed "no criticism to Cisco's product" agreement. And I claimed already we considered SPA112 so unreliable for production deployment even in this thread. Despite of it, we have few SPA112 deployed (as we depleted our PAP2T stocks).

The most unique feature of SPA112 is the embedded client certificate (it allow secure provisioning and unattended deployment even in untrusted or hostile environment). Also, format of XML configuration file is stable enough. So configuration file can be used even with different firmware version with no modification (most of time). And even more - it can be used on different device type (e.g. SPA122 or SPA232D). Or on devices from other product line (like SPA50x or so).

On the other side, firmware of SPA112 is somewhat unreliable sometime. And there is little or no client support from Cisco side.

I have no experience with HT702 so I have no opinion on it. But unless you require a feature unique to Cisco's device, you should take the HT702 into consideration.

[fredericb1]

Thanks for the update on this and all your time. I'll try a HT702 for now (I'll get it in a week or so) and I'll see how that goes for me. I don't actually need anything like provisioning and embedded client certificates. I only want to use it for voip.ms at home and I hope it'll work with my old'ish phones. We'll see.

I got suggested the HT702 because a friend of mine does some deployment work for different contexts and he uses those often enough and he hasn't seen any issue with it. But obviously, my situation might be out of the ordinary.

Nevertheless, the SPA112 is a neat box and I'm sure it works great for most of the cases.

[Dan Lukes]

Is there a way to maybe introduce a delay between the ring stopping and the line turning on, so as to draw less power from the transformer ? Is this even the cause?

It's not possible. The ring signal (AC signal superimposed on line's DC potential) is supplied by switch/PBX (SPA112 here). The DC potential is here all the times (not counting optional CPC signal on the end of call). "off hook" is claimed by low impedance put on line by end phone. E.g. ring signal and seize signal are originated on opposite ends of line. So no configurable delay is possible here.

According specification the ring signal must be removed from seized line as soon as possible. If I remember correctly, it should not exceed 0.1s.

[fredericb1]

Randomly, without changing any settings, it's working today. I've tried all combinations I could with the two phones I have, switching ports, using only one or both. And it's always working. I even increased the ring cadence to have it ring 2 seconds again (as default) and it's working fine. I'm baffled.

I have a feeling this won't last, but I'll cross my fingers.

It is interesting that in one of my earlier tests, after a full physical reset button factory reset, I had 3 tests that worked flawlessly, until I switched the NAT mapping and interval to "yes" and that after setting it back to "no" i still had issues. Very strange.

If ever this bug happens again, I'll make sure to do a physical button factory reset and never ever set the NAT and test that again.

PS: Oh joy, the SPA is now on a reboot loop.... And now it finally eventually decided to boot.... while on the reboot loop, at some point it would reboot and reboot within half a second many many times. Strange. I think I need new phones.

Note: Now that it has rebooted by itself for no apparent reasons (I wasn't touching/using the phones), now the bug is back. :o/

I now even get situations where the phones won't even ring and the device just reboots instead, but that happened just once. In general it's bugging when I pickup and it rings at the same time.