Hello DYIers!. Long time reader, first time poster here.

Just wanted to share my experience and maybe get some feedback.

I have 2004 Weil-McLain oil fired 125K BTU/h boiler providing hydronic heat and hot water (via heating coil) which I converted to NG by replacing burner head with Carlin. That had been set up with wide flame pattern based on firebox shape and Carlin recommendations. As the boiler is significantly oversized for the house it is also underfired at ~110K BTU/h (boiler expects to be fired at ~145K BTu/h). As I dont have combustion analyzer access, in the past I would set up the boiler by the trace of smoke (zero smoke actually) and that worked quite well with oil. For some reason I decided post conversion smoke pump will be unusable and promptly sold it.

However now I have dilemma as I am not in position to buy and maintain combustion analyzer.

I do have generic car O2 sensor (heated kind -4 wires) and decided to try using that. I plugged it into the chimney pipe on top, attached to 10A 12V power (old computer PSU) to two of thewireswith thesame color and let it warm up for 15 minutes. Once warmed up attached my digital multimeter to two other leads. That shows 10 mV or less. Now opening air slits to 50% and starting the boiler. Voltage up to 30 mV but we are on flat part of the curve with lot of excess air.

Now slowly closing air slits on the burner and observing sensor voltage. At some point it jumps to ~900 mV indicating no free oxygen. Moving air slits one way and other found location where it jumps between 50mV and 900 mV. The curve is very steep so it will be hard to hit exact location to see output anywhere in the middle. But I can get the spot, on my burner it is at 35% open.

To obtain best burning efficiency I must be at 10% excess air, perhaps bit more to be sure. That means widening the air slit by 10% or 3-4 marks so I now set it to ~40% open. O2 sensor reads ~50mV

I verify it using following method: I have 4 air slits. I cover 1/3 of one slit (1/12=8%) and O2 sensor voltage is low - excess air. Then I cover 2/3 of one slit (16%) and O2 sensor voltage jumps to 900 mV - no free oxygen, insufficient air. Which means I am providing no less than 8% excess air but no more than 16%. Good enough for me.

Am I missing anything?

 

i think what you're doing is dangerous. without measuring co you don't know if it's burning safely.

i'm no expert on this stuff but 6 to 9% o2 in the exhaust is recommended for gas furnaces/boilers. i don't know if that applies to conversion burners.

i think 6% o2 is way more than 16% excess air.

it's a guess what the oxygen sensor output really indicates.

call someone qualified with a combustion analyzer to set up the burner.

It may cost you a couple hundred dollars but could save your life and the boiler.

 

You’ll need a proper combustion test to set that burner up.
What’s the overfire draft on the new burner?

 

user 10 unless I am not reading right Carlin recommends 3-6.2% excess O2 (see attachement), That matches with other sources I find. In my head I was thinking O2 level in air is 1/3 but now realize it is actually 1/5 so that's 15-31% excess air, not 10% like I thought initially. I corrected the gate by opening to 44% and re-checked using the same method with the car O2 sensor covering whole air hole (1/4 of the flow).That puts me into ~25-30% excess air range. Appreciate the correction.

I also agree with the 'dangerous' statement, but let me assure you, all necessary precautions were taken. After installing and pressure testing gas line (70 ft in total - all inspected by a town inspector) converting the boiler was like walk in the park. And I do have powerful community helping me correct my mistakes. Don't agree with "call someone qualified", had really bad experience with some. Wish I could borrow combustion analyzer somewhere for few hours, would made many things easier.


roughneck77 my Dwyer shows 0.023 inch over the fire and 0.030 the chimney one foot past boiler, flue gas temp 300F. I know draft is bit over the spec (guide calls for 0.010-0.020) but this is best I could get after putting maximum weight on the flap and partially blocking flue passage with a brick quite some time ago.

 

how do you know the o2 sensor is accurate? do you know exactly what percent o2 the voltage corresponds with?

this style burner seems to run much lower o2 than non-conversion ones.

if you insist on tuning this without measuring co shoot for the maximum o2 allowed to reduce the risk of high co production.

[quote]had really bad experience with some.[quote]

good residential techs seem to be difficult to find.

 

user 10 I dont need to know if O2 sensor is accurate deep in the lean area. Because it is heavily dependent on the temperature and my multimeter is not class 1 either. What I am doing with this procedure is finding the point in which (or close) where all O2 is burned and no excess is introduced AKA stoichiometric ratio. It is registered very precisely via sharp voltage change after tiny air adjustemnt. From there I add ~25-30% air on top and confirm that point by covering 25% of air vents - once again via sharp voltage reaction on O2 sensor. The thinking being once I open additional 25% of vents it will correspond to ~30% or so excess air.

good residential techs seem to be difficult to find.
Hence the DYI. And also I like working with these things. I see that generic wide band O2 sensors are now available at around $30, maybe at some point I will get one and build Arduino based O2 analyzer everyone can replicate for under $50

 

By the time you have no o2 left in the exhaust, the burner will be producing massive amounts of carbon monoxide.

How much to increase the air after you reach the point of having no o2 left is a guess.

You could have too much or too little.

 

That is true, I am riding on assumptions here that my flame is shaped correctly and air/fuel is mixed well so combustion is more or less complete. If I were to confirm CO level under 50 ppm as per Carlin manual, would that be a sufficient proof of complete and proper combustion?

I am thinking of getting this little guy to build on: https://www.amazon.com/Solu-Monoxide.../dp/B00UYSOUL8

Detection range 10-1000 ppm.

 

That type of device looks like it's designed to detect ambient co, not co in exhaust. it may not be able to take the heat and could have a long lag time.

maybe this would fit the bill for measuring oxygen ->https://www.amazon.com/Oxygen-Concen...0558932&sr=8-1

i have no idea if it's decent.

 

Found similar on aliexpress at $50. Great idea. Perhaps I will get one.

 

it could be junk.

if you can get a dedicated stack co meter and the o2 thing is decent, it will be almost like doing it with a combustion analyzer.