Looks like a great community I've discovered here. Hopefully someone can help me out. I will try and keep it short -

By accident, I have discovered that my well system is intefering with my DSL service. Turn on the water, and SNR as reported by the DSL modem drops from 15-16 to 5-6. As you may know, 6 is the low-water mark for DSL service... Prior to establishing this cause & effect relationship, my telco and I had been working the problem thinking it was somewhere in their cable plant. I can recreate this at will simply by turning the water off and on.

After a lot of digging around, it appears this may be a grounding problem. But I have no clue how to prove this or fix it.

The house is new construction that was completed just a few months ago. I'm the first resident. So presumably all wiring is up to code.

NID from telco is on south side of house. Electrical panel is in garage on north end of the house. Telco tells me they have to ground in the electrical panel, so they ran #12 wire from the NID to the electrical panel. I've asked around a bit, and got some advice that this should be #6 due to length. An validity to that? If so, could that be the culprit?

I am fairly confident that the noise is coming in from the A/C power - not bleeding in to the phone line. To try and prove this, I put the modem outside in a box. The telco has put a separate filtered test jack in the NID. I powered the modem from an outside receptable, and plugged in to their DSL test jack in the NID. No ethernet to the modem; just used the built-in wireless. The same problem existed. So from my point of view, this eliminates the telco wiring inside - aside from the grounding wire. As an aside, all the internal wiring is cat-5 home runs to the various RJ-11 ports in the house.

So I'm hoping someone here can give me suggestions on how to isolate the cause and/or a workaround. My family is getting tired of me grumbling because they dare to do the laundry or take a shower... :)

Thank you -

kgrimm,

The telco bond is normally a #6awg conductor, weather that is required in your case I don't know.

I suspect the equipment grounding conductor at the well pump head may not be properly

Connected, but that is just a guess.

Typically around here they run a #10 solid grey telco ground. I see NO reason for a #6.

A 200A, or eve a 400A service uses a #6 to a ground rod. Why so big for a telco ground?

There is only so much current a grounding electrode conductor will ever carry.




kgrimm, do you know if the well casing bonded? I can't see why this would matter though.

How about the pressure tank and related equipment? Everything grounded properly?

Were any filters or softeners installed? Are they plastic, and if so were they jumpered
Around with a bond wire?

Heck, it could be in the well pump motor itself, but you don't want it to be that!

Quote:

---

Originally Posted by kgrimm

The telco has put a separate filtered test jack in the NID.

---

This makes NO sense! The DSL modem should be plugged into an UNFILTERED jack. The test port on the NID, by definition should be unfiltered.

Is the test jack an auto-jack? i.e. there is no jumper to disconnect?

Can you take a picture of the NID? Both sides if possible. Just open the telco side and both should be visible. Post pic using go advanced.

You didn't use a filter when you plugged in your modem outside, did you?

The only real way it can be ground related is if there is also noise on the telco line when the pump turns on. Is there?

Do you happen to have a good surge suppressor power strip with a power line filter in it?

What power source does the modem use after the 120 V adapter? DC? AC? What value? Modem model?

Quote:

---

Originally Posted by KeepItSimpleStupid

This makes NO sense! The DSL modem should be plugged into an UNFILTERED jack. The test port on the NID, by definition should be unfiltered.

---

Sorry - screwed up my phrasing completely - that should have read "unfiltered" jack.

My telco put a "whole house" filter in the NID. So originally, the test jack at the NID was filtered. So when I first started in on this, I could not test at the NID since the NID and entire house was filtered. So they put a second jack in the NID that I can use for testing my DSL outside at the NID. It was this second jack that I am using for my tests.

Some updates since my original post.

Telco came out Thursday. They put their modem in the unfiltered :) jack in the NID, and turned on the water. They were powering the modem from a power source on their truck. Turned on the water, SNR dropped by 10.

They took the modem to the pole. Tried again, SNR dropped by 6. So the noise (my term) is making it out through the NID and up through the buried drop to the pole!

I have asked the builder (also the landlord) to put me in touch with the company that installed the well. I'd like to speak with someone there that may have dealt with issues like this before. Hopefully he'll know what to fix, or tell me what an electrician needs to do. Then I'll have an electrician in to implement the suggestions.

Also since posting, I have tested inside and outside, while powering my modem from a good UPS I brought home from work. I brought the modem up on battery power, synced up nicely. Made sure the UPS was unplugged from power, and started the shower. SNR dropped from 17 to 7. To my thinking, this proves that whatever the pump / pressure system is doing, it's introducing the noise in to the phone lines. It's not getting to the modem via the AC wiring as I'd originally thought. How it's getting all the way out to the NID is a bit of a mystery.

In answer to another poster, I am new to this type of water system. I'll take a look at the pressure tank and see if it's grounded somehow. There is no filter or softener installed. All plumbing inside is a translucent flexible pipe, similar to what I see in my RV.

Thanks for all the suggestions and input. And sorry for the less-then-clear wiring description.

First off, I think the telco guy needs to get his act together. I'm assuming one phone line.

On the telco side of the NID you have the line going to a protector and ground.

It should then go to an (Auto-Jack) with NO electronics. More reliable than the jumper version. No electronics means no 1/2 ringer.

Then it should head to the DSL splitter. One pair, usually pair #1, should be your voice circuit and the last pair should be DSL. Idealy the DSL circuit should be a home run with no dangling cables.

One line, one test jack. Not one filtered and one unfiltered. A line with DSL without a modem connected is a normal phone line. A line with a modem connected is DSL.

Splitters are used in the NID and filters at he phones.

I'll be back. Meanwhile munch on what I said.

Quote:

---

Originally Posted by KeepItSimpleStupid

First off, I think the telco guy needs to get his act together. I'm assuming one phone line.

---

Thank you for the follow-up reply. I think I owe the telco guy an apology. He's done things in the same manner as I've seen elsewhere. So if you are questioning his work, I think I haven't made myself clear enough. I work for a telco (IT, and have done DSL support in the earlier part of my career). What he has done is standard practice at my prior office. So I'll try to elaborate as to what has been done. Most of this will be redundant, but perhaps I'll throw something out there that I've left out or haven't made clear.

I have a single line of phone service. My home is new construction. Inside my home, each phone jack is wired back to the southeast corner of the basement using CAT5 home runs.

The filter in my NID has two wires going in to it and outputs for two pairs - one pair that is filtered and one pair that is basically a pass-through - the DSL signal is not filtered out.

The filtered output is wired through to the CAT5 cables going to bedrooms 2 & 3, kitchen and family room. Therefore, DSL is filtered out on these lines

The unfiltered pair output on the filter is attached to the CAT5 cable going to my bedroom. The DSL modem obviously is attached to the far end of that CAT5 cable.

As I mentioned I've seen this same setup done many places. It is especially common in homes with alarm systems.

So in the NID, the original Test jack is wired on the output side that is filtered. Therefore it could not be used to test my DSL modem at the NID.

Again, at my request, they put a second test jack in the NID that is attached to the unfiltered pair.

All that said, I've spent very little time inside the telco side of the NID. I'm not sure what you mean by an "Auto Jack with no electronics" (will Google around for that next.) Can you elaborate on what that is and how it would help with the problem I'm having?

Thank you again for taking the time to respond. I look forward to hearing more... :)

Signal and noise levels underline ADSL line driver, receiver design

http://www.telephonecollectors.org/l...ral/HRInfo.pdf

The upper link discusses DSL signal levels. The lower 1/2 ringers.

In post #4 here,

https://www.askmehelpdesk.com/landli...ng-145653.html

which is my NID. The device on the far left is a 1/2 ringer and jumper type test jack. These are unreliable in the long term. That module is not used. You can get the device without electronics.

The middle picture is a auto-jack. Note, no jumper. When you insert a telephone cord, it connects to the telco side and disconnects Everything else. Similar to an RJ31x. It's available with and without electronics (1/2 ringer). In fact it is clearly marked NO ELECTRONICS. It will also disconnect the splitter which is the item furthest to the right.

I'll bet dollars to donuts that your test jack might be like the one on the far left and it includes a 1/2 ringer and you have one installed on the DSL pair of the splitter.

So, I'd like to see this fixed first or at least tested. Modem direct to protector terminals and the lines removed from the premise side.

If that doesn't work, I'd like to see the protector swapped out for a known good one.

Third, I'd look at issues with the pump.

If you want two test jacks, fine, but the DSL one probably should NOT have electronics in it. DSL works better if the 1/2 ringers are removed entirely.

Be aware that your test jack set up would not eliminate problems with the 1/2 ringer

Yes, 1/2 ringers don't cause problems when there is only one. Will two definitely cause a problem. I don't know.

Also look at ground to tip and ground to ring voltages (AC and DC) when pump is on.

You didn't determine if you here static in the phone when the pump is running.

To give you an idea on how detrimental 1/2 ringers are. I just connected one to the telco line. Nothing connected after the 1/2 ringer. Just paralleled it at the phone line side.

Speed test - no electronics kbps, down/up

1667 706
1694 711

1/2 ringer added

867 624
705 700

I used Speakeasy - Speakeasy, Inc. - Broadband, Voice and Data Services. There are better speed test sites.

A quick update...

I called the company that installed the pump and water system in the house. I told the guy what we have found. He immediately indicated that he had encountered this before. He is going to call the manufacturer and obtain some filters that he says will fix the problem.

Also... I spoke with my telco technician today. He installed the same 5Mb DSL service in another new home in this housing development. On a whim he tested with the water turned on and encountered the same result.

So for the time being, I'm going to sit tight on any changes and see what the water pump vendor can do...

Thanks again for all the hints, tips & suggestions.

It would be best if the pump guys fixed it at the pump. It could be power or RFI. In either case, it should be solved at the pump

RFI = requires shielding

Power line induced requires a filter some where on the pump/pump controller.

What kind of speeds do you get when the pump is not operating using something the Speakeasy - Speakeasy, Inc. - Broadband, Voice and Data Services In the lower right hand corner is the speed test.

Quote:

---

Originally Posted by KeepItSimpleStupid

It would be best if the pump guys fixed it at the pump. It could be power or RFI. In either case, it should be solved at the pump

RFI = requires shielding

Power line induced requires a filter some where on the pump/pump controller.

What kind of speeds do you get when the pump is not operating using something the Speakeasy - Speakeasy, Inc. - Broadband, Voice and Data Services In the lower right hand corner is the speed test.

---

I get my full 5Mb / second service when the pump isn't in use. If the water is running, the SNR drops to a point that the connection is unusable. You can watch the modem logging FEC errors and CRC errors, while the SNR falls on its face. I test my speed regularly using Speedtest.net - The Global Broadband Speed Test.

Haven't heard back from the pump guy yet. Will give him until Monday and try again.