Loud buzz on OBi110 line PSTN

[Jon9999]

OK, voltage tests! (And I didn't get electrocuted!)  (And Shale, I was never worried about damaging the meter -- I was worried about dying. )

DC tests (red probe on the jack, black probe in the ground hole of an electric socket):
TELCO LINE 1 (the one that OBi was on):

wire      ON HOOK        OFF HOOK             ringing
-----   -----------      ---------       ----------------
RED        -0.4 VDC       -20.5 VDC     +0.1 to  -2.1 VDC
GREEN     -51.6 VDC       -29.1 VDC    -48.0 to -60.8 VDC
-----   -----------      ---------       ----------------


TELCO LINE 2 (the one that bypassed OBi):

wire      ON HOOK        OFF HOOK             ringing
-----   -----------      ---------       ----------------
BLACK      -0.1 VDC       -20.4 VDC      0.0 to  -2.1 VDC
YELLOW    -51.7 VDC       -29.2 VDC    -48.1 to -58.9 VDC     [added off hook since original post]
-----   -----------      ---------       ----------------




AC tests (red probe on the jack, black probe in the ground hole of an electric socket):
TELCO LINE 1 (the one that OBi was on):

wire    not ringing         ringing
-----   -----------      ----------------
RED         0.3 VAC           1.5 VAC
GREEN       0.2 VAC          91.0 VAC
-----   -----------      ----------------


TELCO LINE 2 (the one that bypassed OBi):

wire    not ringing         ringing
-----   -----------    ----------------
BLACK       0.4 VAC           0.8 VAC
YELLOW      0.2 VAC          91.4 VAC   [corrected from original post]
-----   -----------      ----------------

[Jon9999]

I just noticed one thing, which I don't know whether it's normal or the problem.

I re-tested all the wires with the connected phone off-hook (not ringing).  I notice that the DC voltages changed considerably by 20-30 V compared to on-hook as reported above.  I've added the new observations to the original table above. The AC voltages seemed to spike a tiny bit (a few volts) when I first took the phone off-hook, but then they settled back down pretty quickly at the same on-hook voltages.

I also tried looking at the voltages on Line 2 while taking Line 1 off- and on-hook (and vice-versa), and there was no effect. I also tried looking at the voltages on Line 2 when Line 1 was ringing (and vice-versa), and there was no effect.

[Shale]

I have to congratulate you on your testing and reporting. Marking VAC or VDC for reading could be an improvement, but I don't  think that will show anything significant. I see nothing abnormal in your measurements, unfortunately. Just check that your OBi power supply has 2 prongs and not 3 where you plug into the wall or outlet strip.

One thing that does hit me is that do you have a remote door lock control or buzzer or something shared with line 1?  Maybe moving the OBi to line 2 would be a good idea in that case. Even then, I don't know what would cause your continual buzzing.
                   _____                                             _____
                    | R |                                             | B |===TO LIVING RM JACK (L1)
TELCO/LINE 1 (R/G)---| J |           /--(R/G=L2)---[OBi]---\           |   |===TO BEDROOM JACK (L1/L2)
TELCO/LINE 2 (Y/B)---| 1 |==SPLITTER<                       >SPLITTER==| U |===TO KITCHEN JACK (L1/L2)
                    | 1 |           \                     /           |   |===TO OFFICE JACK (L1/L2)
                    ~~~~~            \--(R/G=L1)---------/            | S |===TO DINING RM JACK (L1)
                                                                      ~~~~~

<----NID-------->    <---------------MEDIA CLOSET--------------------------->  <---REST OF HOUSE---->



One further test comes to mind. Before a possible shipping back your third fried OBi, try it on a phone line at a different location. Maybe it is not frying the OBi, but that at some point your doorbell system puts a kind of interference (look up "common mode" if you are wondering what I am envisioning) on the wires that your phone rejects but the OBi does not sufficiently. The big flaw in this is why would the interference not be there at the beginning, but always be there for the same OBi110 even though you plugged and unplugged several times This is a strange situation, so I am coming up with weird ideas.

I just noticed one thing, which I don't know whether it's normal or the problem.

I re-tested all the wires with the connected phone off-hook (not ringing).  I notice that the DC voltages changed considerably by 20-30 V compared to on-hook as reported above.  I've added the new observations to the original table above. The AC voltages seemed to spike a tiny bit (a few volts) when I first took the phone off-hook, but then they settled back down pretty quickly at the same on-hook voltages.

I also tried looking at the voltages on Line 2 while taking Line 1 off- and on-hook (and vice-versa), and there was no effect. I also tried looking at the voltages on Line 2 when Line 1 was ringing (and vice-versa), and there was no effect.

It is normal for the  off-hook voltage to drop, although the voltage usually drops by a lot more than what you are observing. Off hook voltages that I have seen have all been less than 15 volts across an off-hook phone line.  I will do some looking to see what this might indicate. It may be normal.

[Jon9999]

Thanks, Shale. There's no doorbell or other service on either line -- just POTS. The only difference between L1 and L2 is that L1 has Distinctive Ring service, so when it rings, it rings with a different pattern than L2. But it's a US telco-standard (Verizon) ring. Both lines have Call Waiting, Caller ID, and all that -- but no other kind of other input from the outside, not even DSL service. On the inside, I do have a fax machine on L1 and not L2, but I don't think the problem is coming from the inside as I said previously because the damage (if it really is electrical damage) is on the OBi jack that faces the outside.

I'll add "VAC" and "VDC" respectively above.

[Jon9999]

It is normal for the  off-hook voltage to drop, although the voltage usually drops by a lot more than what you are observing. Off hook voltages that I have seen have all been less than 15 volts across an off-hook phone line.  I will do some looking to see what this might indicate. It may be normal.

The voltages I reported above were DC from the phone wire to ground. So the voltage across the line would be the difference between red-to-ground and green-to-ground, for example, which is indeed less than 15 volts (8.6 V). That's right, right?

[Shale]

The voltages I reported above were DC from the phone wire to ground. So the voltage across the line would be the difference between red-to-ground and green-to-ground, for example, which is indeed less than 15 volts (8.6 V). That's right, right?

I agree.

[Jon9999]

I'm tearing all my wires (and all my hair) out...

I've noticed by using a RJ11/RJ45 tester that on some of my telephone cords, the wiring goes straight through (pin 4 <-> pin 4 and pin 5 <-> pin 5), while on other of my telephone cords, the wiring is crossed over (pin 4 <-> pin 5 and pin 5 <-> pin 4). The cables that came with OBi are in the latter category, but my phone cords are pretty much split evenly between the two patterns. Similar for pins 3 and 6. The in-wall cables are all straight-through.

Does this matter? I'm assuming not, since R (pin 4) and G (pin 5) are interchangeable, but I'm trying to find SOMETHING before committing involuntary obicide again.

[Shale]

I'm tearing all my wires (and all my hair) out...

I've noticed by using a RJ11/RJ45 tester that on some of my telephone cords, the wiring goes straight through (pin 4 <-> pin 4 and pin 5 <-> pin 5), while on other of my telephone cords, the wiring is crossed over (pin 4 <-> pin 5 and pin 5 <-> pin 4). The cables that came with OBi are in the latter category, but my phone cords are pretty much split evenly between the two patterns. Similar for pins 3 and 6. The in-wall cables are all straight-through.

Does this matter? I'm assuming not, since R (pin 4) and G (pin 5) are interchangeable, but I'm trying to find SOMETHING before committing involuntary obicide again.

Using 1...8 numbering, yes,  what you describe for phone cables is normal for phone cables 4-5 and 5-4 for pair 1 plus 6-3 and 3-6 for pair 2. Straight through would work fine. Ethernet cables are always straight through, although the twisting in them is important for ethernet... but not so much for phone for phone.



Normal.

[SteveInWA]

Sigh.  I did understand what you meant about how and when the buzzing started.  Based on your additional information and testing, my assumption is that, after some (variable/unknown, as you now point out) amount of time, the OBi gets "fried" by something other than normal telco voltages on the line jack.  Given that your basic voltmeter test results are normal/nominal, then I suggest that you do as Rick said:  plug a known-good analog telephone (preferably one without its own power supply/AC adapter) into the phone jack, plug the OBi's line jack into the NID, and let it "cook" at least overnight.  What's the rush?  Take the time to prove that the device itself isn't just a crappy production run, before adding more complexity to your situation.

If it dies, call Obihai, and demand to get some help from them -- suggest that they have a big problem on their hands, and they, not Newegg, RMA-swap your box and send you another one, and ideally, they test the failed box to see why it failed.

I would also recommend that you call Verizon repair service again.  Tell them that you have been frying some equipment when plugging into their circuits, and ask them to send a truck out to your location to test the NID for a possible ground fault.  If the NID isn't properly, tightly connected to a reliable, true ground per the National Electrical Code (like a cold water pipe or a ground rod in the earth), then you can get ground loops or static discharge on the phone lines.  

Depending on the age of your residential 110 AC wiring and the competency of the person who installed it, you may also have one or more outlets that are not properly grounded, or the neutral (white wire) isn't properly bonded to ground at the electrical service panel.  You'd need to hire a licensed electrician to test that, or at least buy an outlet tester at the hardware or big-box store.  This isn't likely your problem, but since it's been brought up repeatedly, you might as well get some peace of mind.  And, speaking of house wiring, it is not unusual to find that the phone wiring, or even the AC wiring, has been punctured somewhere inside a wall by a nail, drywall screw, wire staple or rodent teeth.  My house was built in 1989, and it has a flaky line 2 pair of an 8-wire bundle, somewhere inside a wall, likely caused  by a puncture.

As an aside, regarding voltage testing and safety:  the better/safer way to test phone wiring (vs. the "sewing needle" method) is to cut the modular plug end off of a phone cord, strip off a couple inches of the grey jacket, and strip each of the 4 wire conductors.  Then, use an alligator clip to attach your voltmeter leads to your new "phone wiring test cable".  On-hook voltage is, as you found, about 50 VDC.  This won't kill you, or even hurt much.  Off hook is much lower, since there is now an impedance across the line that lowers the voltage.  It is the RINGER voltage that will hurt, since it's ~100V AC, and, as Thomas Edison used to demonstrate by electrocuting small animals, to try to scare people into fearing his competitor's AC power, 100V AC hurts more than DC, but only during the brief ring cycle.  You wouldn't die, unless you were sitting in a bathtub, though...it's just one of those OWWW $#&%$! moments.

RE:  phone cord wiring:  all telephone cords should have the 4 wires crimped to the RJ-11 jack on one end in the opposite direction from the jack on the other end.  Throw away any that aren't built like that.  Ethernet RJ-45 jacks are wired straight through, unless, of course, it's a "crossover" cable.

So, using the photo of the RJ-11 jack in this thread, one end should be 1-2-3-4 = B-R-G-Y and the other end should be Y-G-R-B.  Modern telephone equipment doesn't care about polarity, as did the old Western Electric phones, but you never know if some weird phone device DOES care, and the cords are so cheap that there is no reason to keep any cords that are wired bass-ackwards.  If you are a cheapskate, or you want to make custom-length cables, you can buy a crimping tool and a bag of RJ-11 jacks and crimp them accordingly.  If you have a HP OfficeJet or similar fax machine, be sure it is connected to the phone line with a 2-conductor, not 4-conductor cord.  I don't understand why, but this has always been a requirement by HP, and I assume they are doing something non-standard with the outer conductor pair.

[Jon9999]

Sigh.

One last question before I bite the bullet and plug the new OBi back in:  Could it have something to do with Call Waiting?  What happens, voltage-wise, then a CW signal comes through?  Is it like a ring, with high-voltage AC?  Could this be something that might damage the OBi?  I tried testing with the voltmeter, but I couldn't coordinate getting the voltmeter on the contacts at the same moment that the beep came through.


As for my building wiring beyond (outside) my NID, it is definitely quite a mess. I live in an old building. The Verizon patch cabinet in the basement is a load of spaghetti, with loose wires sticking out all over the place.  The wire for my apartment comes up from the basement in an elevator shaft, goes through more loose-wire/spaghetti in a junction box on my floor, then through my neighbor's apartment before it comes though a wall into my apartment, where the NID is located.  The NID is just a phone-jack-like box -- nothing appears grounded to anything.  The cable that comes into the NID is a really old, fabric-covered bundle with 4 thick-gauge, solid copper wires inside (R/G/B/Y). So it's a bit of the Wild West there, and I wouldn't be surprised if the signal of the Oprah Winfrey Show sometimes finds its way onto the Verizon wires!

The electrical wiring, sockets, and circuit breaker in my apartment was all completely replaced from the risers a few years by a licensed electrician during a gut renovation. The installation was inspected by the city and passed without problems.

[Shale]

Sigh.

One last question before I bite the bullet and plug the new OBi back in:  Could it have something to do with Call Waiting?  What happens, voltage-wise, then a CW signal comes through?  Is it like a ring, with high-voltage AC?  Could this be something that might damage the OBi?  I tried testing with the voltmeter, but I couldn't coordinate getting the voltmeter on the contacts at the same moment that the beep came through.

Since 100 vac is normal ringing, the Line input to the OBi should be robust.

Call waiting is not a high-voltage event. There may or may not be momentary reversal of the polarity on the two phone wires, but no ringing voltage is involved.

I have another off-the-wall thought. There are a lot of old DSL filters sitting in people's drawers. I wonder if this might have some minor protective properties if placed between the OBi and the phone line. It is just a low-pass filter that passes frequencies less than 15 kHz. But it might add a little impedance to keep some mystery surge from doing damage. Getting OBi involved sounds useful. They might have some insight.

[Jon9999]

OK, gentlemen, I've connected the new (#3) OBi110.  I put a 2-line splitter onto the NID itself and hooked the LINE port of the OBi directly into the Line 1 half of the splitter, right there at the NID, placing only a telephone surge protector in between. I hooked the PHONE port of the OBi directly into a stand-alone phone (it's a cordless one -- sorry, I don't have an old phone on hand that doesn't need to be plugged in). Both the OBi and the cordless phone are plugged into a surge protector that has built-in indicator lights showing that the AC circuit and grounding are OK. I'll let it cook a while and will report back. If it survives the night, I'll move the OBi into the media closet first, and finally I'll isolate both the PHONE and LINE ports with phone surge protectors before connecting the PHONE side back into the bus that connects to the rest of my phones.

[Jon9999]

24 hours in, with heavy use...  So far, so good.

I've added a DSL filter between the NID and the LINE port, as Shale recommended earlier, in addition to the surge protector, as a "belt-and-suspenders" thing. (I don't have DSL service.)

I still haven't had the courage to connect the PHONE port to the rest of my internal phone network. It's still hooked up directly to just one cordless phone base, but I do have the first 2-line splitter between the NID and the LINE port (actually, between the DSL filter and the line port).

NID==[DSLFILTER]==<splitter>--<surgeprot>--[LINEPORT]OBi110[PHONEPORT]--(cordlessphone)

I think I'll give it another couple of days and then bite the bullet. I'll leave the DSL filter and the surge protector on the LINE side, and I'll add another DSL filter and surge protector on the PHONE side too before running back into the 2nd splitter and thence into the internal phone network bus.

NID==[DSLFILTER]==<surgeprot>--<splitter>--[LINEPORT]OBi110[PHONEPORT]--[DSLFILTER]--<unsplitter>==<surgeprot>==NETWORKBUS

Any other thoughts before I do this?

Unofficially: Do any of you have any thoughts on whether there is any chance at all that the problem came into the OBi from the PHONE side, since the effect (damage?) seems to have been solely on the LINE port? Can you see any way that hooking the PHONE side back into my phone network would have been the source of the problem? I'm sort of thinking that whatever happened must have come from the PHONE side, so I should be out of the woods. Right? (Why don't I believe myself?)

[SteveInWA]

With regard to your question about the LINE side vs the PHONE:  You didn't have Verizon come out and inspect/confirm that your NID is properly grounded.  Without doing that, you can't be sure that some voltage surge/spike/lightening strike might damage your OBi, your attached phones, and/or yourself.  If some other part of your house phone wiring (what you've been calling the "bus") is somehow carrying voltage it shouldn't be (as I discussed, via a damaged spot of wiring touching AC), you can create a path for current to flow.  You want that current to flow to ground instead of into your equipment or into you.  I'm just repeating this as a disclaimer.

That said, make sure you don't have a wiring error on your "bus", such that your jacks are still carrying your L1 signal from the NID.  If so, and you plug in the OBi's PHONE jack into your bus, and the Verizon POTS L1 line rings, it'll send the 100+ VAC into the OBi's PHONE side and may fry it.  The only way to be sure of this is to plug in a known-working, single line phone, into a jack on your bus and make sure it is dead, or use a tester.

Keep in mind that there are two types of RJ-11 splitters, that can look identical to the casual observer.  One kind simply duplicates all 4 conductors on every jack, and the other kind breaks out L1 and L2.   Be SURE you have the right kind, or you will be feeding L2 back into L1 or vice versa.

If you don't already have them, I'd strongly suggest buying two items of test equipment:  an AC outlet tester (a plastic plug with LEDs on it, that show properly wired outlets vs. reversed polarity or bad ground), and a phone jack tester (a plastic dongle with LEDs on it, that shows live phone lines, reversed polarity, or dead lines).  The first is available at any hardware or home warehouse store.  The second used to be at Radio Shack, but seems to be gone.  You can still find them online, for example:

http://www.amazon.com/Gardner-Bender-TT6200L-Phone-Tester/dp/B000GL95JA

[Jon9999]

You know, I've been thinking. Maybe it's wrong to hook the OBi in serial between my PSTN line and my internal phone network. Is it within the unit's design parameters to be used that way? Or should I perhaps connect the OBi on its own, isolated branch of my phone network, coming off a 2-way splitter at the NID, with its own separate telephone instrument? I know that means I'll have to use that particular phone when making a GV or SIP call and I won't be able to "flash" into the OBi from my other phones to do transfers, etc., but at least I'll be able to use OBiON, which is really the main point for me. It would be a trade-off, but maybe worth it if it means the unit won't fry again.

[SteveInWA]

There is nothing wrong with connecting your OBi as POTS-->LINEonOBI-->PHONEonOBI-->all phones in house, as long as your are damn sure that your POTS is not also connected to the bus on the phone side.  That's how I have mine connected, and it's worked fine for years.

[Shale]

You know, I've been thinking. Maybe it's wrong to hook the OBi in serial between my PSTN line and my internal phone network. Is it within the unit's design parameters to be used that way?

I see no problem with your current plan.

[Jon9999]

You didn't have Verizon come out and inspect/confirm that your NID is properly grounded.

The refused to come. They said they tested the line from the "automated equipment" at the central office, and that since it tested out okay, they would not send someone to my house. I pushed, jumped up and down, asked for a supervisor, etc., but they still said they would not come because they're certain there isn't a problem. Then they told me to have a nice day and thank you for using Verizon.

That said, make sure you don't have a wiring error on your "bus", such that your jacks are still carrying your L1 signal from the NID. ... The only way to be sure of this is to plug in a known-working, single line phone, into a jack on your bus and make sure it is dead, or use a tester.

Done. Dead. The bus, incidentally, is a professional panel meant for this purpose (with the Security System bypass feature switched off), not a homemade breadboard.

Keep in mind that there are two types of RJ-11 splitters, that can look identical to the casual observer.

I know. These are definitely the line 1/line 2 variety (actually, L1/L2/L1+L2, with the 3rd spot unoccupied) and not the 4-connector to 4-connector-times-2 variety. I've triple-checked.

If you don't already have them, I'd strongly suggest buying two items of test equipment:

Actually, I have an electric surge protector with the ground/circuit test lights built in, and all my sockets check out. I also have an RJ45/RJ11 tester (similar to this, and I've run all my internal wiring through it to make sure all the pins line up everywhere between the jacks and the patch board in the media closet. Check.

This is why this is so maddening. Everything seems to check out fine all around.

[Jon9999]

There is nothing wrong with connecting your OBi as POTS-->LINEonOBI-->PHONEonOBI-->all phones in house, as long as your are damn sure that your POTS is not also connected to the bus on the phone side.  That's how I have mine connected, and it's worked fine for years.

Sigh. Well, it was a thought!

I'm damn sure indeed that Line 1 (the one that OBi uses) wasn't connected to the bus or anything beyond as well. I'm damn sure that only Line 2 (non-OBi) went directly to the bus. Could there have been "leakage" from Line 2 on the "bus" side back into Line 1? Well, maybe. I've tested that by probing Line 1 at the bus while making Line 2 ring with Line 1 disconnected at the NID, and I detected only a minuscule change in voltage, if any.

BUT... one slightly strange thing I have noticed this afternoon is that with Line 1 currently disconnected from the bus (it's going straight into the OBi at the NID, remember) and with only Line 2 coming through to the rest of the house, when I got an incoming call on Line 2, the "Line 1" light on BOTH of my 2-line phones also blinked. It didn't register a "ring," but the line selector light lit up momentarily. I don't know -- maybe that's expected because there's no voltage "holding" Line 1 in the on-hook state.  When Line 1 was connected normally into this 2-line phone, Line 1 would never flash when Line 2 was called. One of these particular phone (AT&T Model TL86109) has a feature that causes a call-waiting-like signal to come into the handset in you're on a call on one line when the other line rings. Maybe the light flashing has something to do with that? Come to think of it, could the whole problem have something to do with that???


Edit: Indicated that the Line 1 blinking thing was on BOTH of my 2-line phones (AT&T and an older Panasonic), not just the AT&T. On the Panasonic, I can tell that it's not the same kind of lamp-blinking as a ring. The ring-blink on Line 2 is a very fast oscillation blink, while what I'm seeing on the Line 1 lamp is a slow, steady, on-off-on-off-on-off flash, maybe twice a second.

[SteveInWA]

No way to be sure, but I doubt the AT&T phone is the culprit, unless it it (or any of your phones, for that matter), are defective, and have a short internally.  It sounds like they don't, based on your metering.  I'd leave all but one physical telephone disconnected and go for it.  If the OBi survives, and I bet it will, then you can add one more phone in at a time and retest, until you are happily finished, or in tears because it died again.  That's the only way I know to proceed at this point.

*highly unlikely, but for the sake of OCD completeness:  look at each phone's manufacturer label or documentation.  Note the REN number.  Make sure they all add up to less than 5.0, per:
http://www.obihai.com/docs/OBi110DS.pdf