I've had a faint smell of gas in my utility room since I purchased my home almost a year ago. I had to replace the faulty Honeywell gas control valve on my AO Smith gas water heater shortly after I purchased the home. That was easy and got the water heater working properly, but did not fix the gas leak. I've bubble tested every single gas connection in the home several times (including all lines for pinhole leaks), but never found the leak. I finally purchased an electronic combustible gas leak detector and found there is gas leaking at the water heater vent pipe that's peaking the meter! This is why I was unable to detect the leak with the bubble method. It's important to note that the burner has a strong clean blue pilot flame and combustion looks ~normal when the water heater turns on. The vent pipe appears to be drafting normally with smoke tests, both with utility room door closed and open. The air intake screen at bottom of water heater is clean and unobstructed. My question is why am I detecting a gas leak in my water heater vent if the pilot is lit and it's burning and drafting properly? Is it possible the burner assembly has a leak somewhere? I'm concerned as there is clearly gas detected at the vent. Any suggestions would be greatly appreciated. Anyone else ever experienced this?
Gas Leaking At Vent
- Thread starter JonCurl
- Start date
Help Support Plumbing Forums:
You may have a small leak inside the unit where the pilot tube connects to the burner. I’d check that last.
Your water heater creates a draft in the room so it’s pulling air from the surrounding area and exhausting it through the vent. So the leak could be outside the water heater and being drawn in.
Make sure you’re using a thin soapy solution and watch for bubbles for 10 minutes.
Your water heater creates a draft in the room so it’s pulling air from the surrounding area and exhausting it through the vent. So the leak could be outside the water heater and being drawn in.
Make sure you’re using a thin soapy solution and watch for bubbles for 10 minutes.
Thank you, I really appreciate it.
I understand any gas leak in the area would most likely sink and be drawn in through the air intake at the bottom of the water heater. However, I assume it would also be burnt up by the pilot and not detected at the top vent?
I've already checked all external gas lines and connections in the utility room with soapy solution, though I didn't watch them all for ten minutes. I also checked them with the electronic sniffer and only get a reading at the vent. I'll check again tomorrow.
I know some natural gas leak detectors will also pick up carbon monoxide, but not the other way around. I've never seen a carbon monoxide detector that will register a combustible gas leak. So, I'm wondering if the high (peaked) meter reading I'm getting at the vent is possibly a combination of some unburnt natural gas and some carbon monoxide?
There's also no gas backdrafting from the furnace, which is the only other appliance connected to the vent. I checked this by cutting gas at the furnace and also testing for leaks at the furnace.
I highly suspect the internal gas line connection to the pilot assembly is leaking. I don't see a safe way to test this connection at the pilot. I suppose I could switch the control valve to pilot only and pull the burner assembly while the gas is still on and pilot is lit, and then soap test the ferrule connection with the whole burner assembly outside the water heater? I've observed the flames through the glass window, both pilot and burner, and they look normal. Do you think it would be safer to first turn the gas off, then pull the burner assembly, turn gas back on, then soap test the pilot tube connection without actually lighting the pilot?
I understand any gas leak in the area would most likely sink and be drawn in through the air intake at the bottom of the water heater. However, I assume it would also be burnt up by the pilot and not detected at the top vent?
I've already checked all external gas lines and connections in the utility room with soapy solution, though I didn't watch them all for ten minutes. I also checked them with the electronic sniffer and only get a reading at the vent. I'll check again tomorrow.
I know some natural gas leak detectors will also pick up carbon monoxide, but not the other way around. I've never seen a carbon monoxide detector that will register a combustible gas leak. So, I'm wondering if the high (peaked) meter reading I'm getting at the vent is possibly a combination of some unburnt natural gas and some carbon monoxide?
There's also no gas backdrafting from the furnace, which is the only other appliance connected to the vent. I checked this by cutting gas at the furnace and also testing for leaks at the furnace.
I highly suspect the internal gas line connection to the pilot assembly is leaking. I don't see a safe way to test this connection at the pilot. I suppose I could switch the control valve to pilot only and pull the burner assembly while the gas is still on and pilot is lit, and then soap test the ferrule connection with the whole burner assembly outside the water heater? I've observed the flames through the glass window, both pilot and burner, and they look normal. Do you think it would be safer to first turn the gas off, then pull the burner assembly, turn gas back on, then soap test the pilot tube connection without actually lighting the pilot?
Call a plumber and get the leak repaired. This really isn’t the time to learn.
Read this.
https://www.cunninghambounds.com/blog/2007/october/wife-describes-explosion-that-killed-husband/
Read this.
https://www.cunninghambounds.com/blog/2007/october/wife-describes-explosion-that-killed-husband/
Last edited:
Same thing happened with my most recent water heater install (AO Smith). The inspector came out pulled the "inspection port", found the leak, but because he wasn't allowed to fix it himself, he told me to put a wrench "there" and turn. The smell went away and he passed the installation.
As was mentioned, get it fixed fast!
As was mentioned, get it fixed fast!
Thanks guys. I understand the concern and danger of gas leaks, so I've turned the gas to the water heater off for now. I had an inspector come out. He tested all gas lines and connections in the entire house from the water heater out to the meter. He confirmed what I already know, there are no external gas leaks, but I have gas leaking from the vent.
I called a plumber as Twowaxhack suggested. All he did was confirm what I already know and offer to replace the entire water heater. He said I'm probably right that the internal pilot light connection is leaking, but wasn't sure why it would start leaking. I explained that a contractor jack hammered the foundation under the heater to install drain tile. They moved the heater over, but I did see one worker actually leaning on the heater while he was breaking up the foundation with a jack hammer. So, I assume the extreme vibration must have loosened the internal connection.
I'm not going to replace the entire water heater due to a loose connection that can probably be fixed by tightening as BlueSkyHigh did. What I may do is order a new burner assembly for peace of mind, install it myself, and see if I can get a plumber or inspector to just check my work as BlueSkyHigh did. The hard part is finding a professional to check my work.
I really don't understand why many "professionals" want to replace an entire appliance, when it can be easily fixed? Is there some hazard I'm not understanding here where the entire water heater needs to be replaced?
I had a similar experience years ago with my wife's electric clothes dryer. It stopped heating. She called a "professional" repairman who offered to sell and install a new one. I unplugged it, took it apart and cleaned out all the lint and ductwork, and replaced the blown thermal fuse, which cost ~$10.00. That was in ~2012. It's been working great since then.
I called a plumber as Twowaxhack suggested. All he did was confirm what I already know and offer to replace the entire water heater. He said I'm probably right that the internal pilot light connection is leaking, but wasn't sure why it would start leaking. I explained that a contractor jack hammered the foundation under the heater to install drain tile. They moved the heater over, but I did see one worker actually leaning on the heater while he was breaking up the foundation with a jack hammer. So, I assume the extreme vibration must have loosened the internal connection.
I'm not going to replace the entire water heater due to a loose connection that can probably be fixed by tightening as BlueSkyHigh did. What I may do is order a new burner assembly for peace of mind, install it myself, and see if I can get a plumber or inspector to just check my work as BlueSkyHigh did. The hard part is finding a professional to check my work.
I really don't understand why many "professionals" want to replace an entire appliance, when it can be easily fixed? Is there some hazard I'm not understanding here where the entire water heater needs to be replaced?
I had a similar experience years ago with my wife's electric clothes dryer. It stopped heating. She called a "professional" repairman who offered to sell and install a new one. I unplugged it, took it apart and cleaned out all the lint and ductwork, and replaced the blown thermal fuse, which cost ~$10.00. That was in ~2012. It's been working great since then.
![MEISTERFAKTUR drain snake 2.0 [50 FT] - with drill attachment - Ideal plumbing snake for sink and drain unblocking - Solid drain auger for real DYIs! (50 FT - 1/4 inch)](https://m.media-amazon.com/images/I/41VwmTiOsgL._SL500_.jpg)
I repair water heaters on a regular basis. Find the right plumber.
I know there are the exceptions, but some plumbers probably find more profit in a complete replacement and associated service contract over the warranty period of the tank they installed.
Since you know how to pull the burner out and then your going to replace it why don't you pull the burner out
and take the pilot tube off of the burner assembly and put a little dope on the threads and retighten it up and
put it back together and see if you smell gas anymore.
and take the pilot tube off of the burner assembly and put a little dope on the threads and retighten it up and
put it back together and see if you smell gas anymore.
Thanks for the suggestion Tom. I considered either mechanically tightening the connection as BlueSkyHigh did, or using a dope/sealant. There were a couple things holding me back.
First: I don't know how to safely test the line connection at the pilot for leaks, as it's internal. I wasn't able to reach the connection with soap while it was lit/running. I could only use the sniffer to pick up gas at the vent on top. The burner and pilot lines are somewhat flexible, but I don't want to kink them trying to run the pilot outside the tank to bubble test. I'm also not sure if that's even safe/recommended. Unfortunately, I can't just tighten the pilot line connection or add a bit of dope, re-install the burner assembly, and then test it at the vent without lighting the pilot. If the pilot's not lit it's going to leak gas there anyway. So my main question is how do you safely test this internal pilot connection?
Second: The only gas rated sealants I have on hand are RectorSeal "T Plus 2" pipe thread sealant, Blue Monster PTFE Thread Tape, and some generic yellow tape. None have any temperature ratings listed, and I couldn't find any info online about safety for use near open flame/high temp areas. So, I'm not sure if there is a specific thread sealant/dope required for high temperature? I don't have any formal training working with gas, so I just don't know. I'd hate to have it seal fine for a while, but then degrade over time after being exposed to heat from the burner.
I'd rather just fix the leak on the existing burner assembly. I certainly don't want to buy a whole new tank. I assume the pilot connection on a new factory assembled burner assembly would be pre-tested for leaks at the factory and would be a safe bet without spending an arm and a leg? It was already running for a long time with the leak, so I assume it would be safe to remove the burner assembly and either tighten the pilot connection or dope it and re-test at the vent. I'm really not sure, so I'd rather be safe than sorry. Another issue is that I'm rural, so it's almost impossible to get a plumber who works with gas to even come out here. My choices seem to be DIY or pay some dude to replace my entire water heater.
First: I don't know how to safely test the line connection at the pilot for leaks, as it's internal. I wasn't able to reach the connection with soap while it was lit/running. I could only use the sniffer to pick up gas at the vent on top. The burner and pilot lines are somewhat flexible, but I don't want to kink them trying to run the pilot outside the tank to bubble test. I'm also not sure if that's even safe/recommended. Unfortunately, I can't just tighten the pilot line connection or add a bit of dope, re-install the burner assembly, and then test it at the vent without lighting the pilot. If the pilot's not lit it's going to leak gas there anyway. So my main question is how do you safely test this internal pilot connection?
Second: The only gas rated sealants I have on hand are RectorSeal "T Plus 2" pipe thread sealant, Blue Monster PTFE Thread Tape, and some generic yellow tape. None have any temperature ratings listed, and I couldn't find any info online about safety for use near open flame/high temp areas. So, I'm not sure if there is a specific thread sealant/dope required for high temperature? I don't have any formal training working with gas, so I just don't know. I'd hate to have it seal fine for a while, but then degrade over time after being exposed to heat from the burner.
I'd rather just fix the leak on the existing burner assembly. I certainly don't want to buy a whole new tank. I assume the pilot connection on a new factory assembled burner assembly would be pre-tested for leaks at the factory and would be a safe bet without spending an arm and a leg? It was already running for a long time with the leak, so I assume it would be safe to remove the burner assembly and either tighten the pilot connection or dope it and re-test at the vent. I'm really not sure, so I'd rather be safe than sorry. Another issue is that I'm rural, so it's almost impossible to get a plumber who works with gas to even come out here. My choices seem to be DIY or pay some dude to replace my entire water heater.
Although not a substitute for a formal sign-off (if that's what you are seeking), your nose is a pretty good indicator.
Thanks BlueSkyHigh. That's how I found the leak in the first place, by using my nose. I confirmed it with the electronic sniffer, and it was subsequently confirmed by the gas co. inspector. My main question is how would I safely test the ferrule/compression fitting at the pilot for leaks without lighting the pilot? If the pilot is not lit it's going to smell like gas anyway. However, if I re-light the pilot and there's still a leak, it's not an entirely safe situation.
If there was some type of bubble putty I could apply around the pilot connection that doesn't just drip off like the soap solution that would be ideal. That would allow me to tighten and/or dope the threads of the connection and put everything back together while still being able to bubble test without having to light the pilot.
Am I being overly cautious about re-lighting the pilot to sniff test for leaks again? My old neighbor Joe used to leak test with a f&%king lighter. It was already running with a leak for quite a while, so maybe tightening/doping the pilot connection and re-lighting the pilot to test for leaks isn't unheard of?
If there was some type of bubble putty I could apply around the pilot connection that doesn't just drip off like the soap solution that would be ideal. That would allow me to tighten and/or dope the threads of the connection and put everything back together while still being able to bubble test without having to light the pilot.
Am I being overly cautious about re-lighting the pilot to sniff test for leaks again? My old neighbor Joe used to leak test with a f&%king lighter. It was already running with a leak for quite a while, so maybe tightening/doping the pilot connection and re-lighting the pilot to test for leaks isn't unheard of?
I am not sure and hesitate to advise you any further for obvious reason. The pressure in a natural gas line is exceedingly low at maybe a half a PSI or less I've heard, compared to the pressures in, say, a home water plumbing system so the likelihood of a leak in a properly tightened and doped system, if you want to dope the connections, is exceedingly low. I personally would have no problem putting the system back together and, if I couldn't smell anything, going ahead and lighting off the pilot and firing up the water heater. After that and smelling nothing, I'd consider the job completed and no further action necessary. But that's just me and I've been accused of going where Angels fear to tread.
Last edited:
Please don't test it with a lighter. I know you won't. Had an old time plumber years ago test for gas leaks in
houses with a lighter. He was told about that. Well, anyway he went up with the empty house as it blew up
to nothing. If you feel uncomfortable fixing the pilot tube please don't. Buy the assembly like you wanted.
I have been doing this for a long time and have my ways of testing the pilot tube but would not ask you to
try it. I wouldn't do it my way unless I thought it was safe for me. Not for the DIY'er.
houses with a lighter. He was told about that. Well, anyway he went up with the empty house as it blew up
to nothing. If you feel uncomfortable fixing the pilot tube please don't. Buy the assembly like you wanted.
I have been doing this for a long time and have my ways of testing the pilot tube but would not ask you to
try it. I wouldn't do it my way unless I thought it was safe for me. Not for the DIY'er.
And they never hear from him again..... Jesus!, I would never test with a lighter.
Thanks for the advice Tom. I'm sure there are ways to test the pilot tube, but, as you said, the most prudent thing to do in my situation is to just buy a new burner assembly. I don't feel comfortable doping a connection that's designed to be purely mechanical.
I pulled the burner assembly and it is extremely corroded. The tank isn't leaking. I'm in Alabama. It's very humid here, so everything rusts fairly quickly. I already ordered a new burner assembly. With all the corrosion I figure having new metal on all the electrical and sensor connections is an added bonus.
Hopefully the new burner assembly will arrive within a few days. While I'm waiting for the assembly I'll install a new anode rod for added insurance. I'll let you know how the install goes. I appreciate all the advice. This forum is amazing.
Thanks for the advice Tom. I'm sure there are ways to test the pilot tube, but, as you said, the most prudent thing to do in my situation is to just buy a new burner assembly. I don't feel comfortable doping a connection that's designed to be purely mechanical.
I pulled the burner assembly and it is extremely corroded. The tank isn't leaking. I'm in Alabama. It's very humid here, so everything rusts fairly quickly. I already ordered a new burner assembly. With all the corrosion I figure having new metal on all the electrical and sensor connections is an added bonus.
Hopefully the new burner assembly will arrive within a few days. While I'm waiting for the assembly I'll install a new anode rod for added insurance. I'll let you know how the install goes. I appreciate all the advice. This forum is amazing.
Don't forget, we here love pictures!
Maybe check with Twowaxhack to see if he can come out and repair it for you. I think he's in the Mobile area?
New burner assembly arrived yesterday. Installing the new burner assembly completely fixed the leak. Tested for leaks yesterday and again today. No gas at the vent at all. I'm still not sure if the leak was due to the pilot line working loose due to vibration from all the jack hammering, or severe corrosion. As you can see from the pics, the old burner assembly was severely rusted.
I double checked the date on the water heater and it's actually from 2003! If I had the bread I would have replaced it. I need the $ for the holidays. It's in a ground level garage space with nothing but concrete, so not too worried at the moment.
Anyway, I pulled the anode rod and flushed the tank while I was waiting. It flushed fine & anode rod was actually less than 1/2 worn. I assume the rod was original. The anode rod was really on there. I hit the threads with penetrating oil a couple days before removal. I still had to use a breaker bar with a six foot cheater pipe & brace the tank's feet with boards to keep it from spinning. I ended up having to brace my furnace as well, as the tank was moving the furnace over. Maybe I should have used the impact driver?
Maybe the rod was in such good shape because we have really hard water in this part of Alabama? I bought the factory replacement rod and it matched what was in there exactly, so I don't think it's ever been replaced. The house is all copper, but whoever installed the heater used ~two feet of pex to tie it in to the copper on both hot and cold. I assume the PEX is a much better dielectric union than the standard unions. I'm sure it's not code, but maybe the PEX actually helps keep the tank in good shape?
Big thanks to everyone here who helped out with advice. I've never seen an internal leak like this, so I was at a complete loss. This forum is amazing!
I double checked the date on the water heater and it's actually from 2003! If I had the bread I would have replaced it. I need the $ for the holidays. It's in a ground level garage space with nothing but concrete, so not too worried at the moment.
Anyway, I pulled the anode rod and flushed the tank while I was waiting. It flushed fine & anode rod was actually less than 1/2 worn. I assume the rod was original. The anode rod was really on there. I hit the threads with penetrating oil a couple days before removal. I still had to use a breaker bar with a six foot cheater pipe & brace the tank's feet with boards to keep it from spinning. I ended up having to brace my furnace as well, as the tank was moving the furnace over. Maybe I should have used the impact driver?
Maybe the rod was in such good shape because we have really hard water in this part of Alabama? I bought the factory replacement rod and it matched what was in there exactly, so I don't think it's ever been replaced. The house is all copper, but whoever installed the heater used ~two feet of pex to tie it in to the copper on both hot and cold. I assume the PEX is a much better dielectric union than the standard unions. I'm sure it's not code, but maybe the PEX actually helps keep the tank in good shape?
Big thanks to everyone here who helped out with advice. I've never seen an internal leak like this, so I was at a complete loss. This forum is amazing!
Attachments
Nice job!
That burner assembly looks like it could have been under water. Really rusted.
