Today’s welding challenge

[Kingcreek]

Old Sanborn “Comet” air compressor. Was dad’s and might be as old as me. Made in America with a Westinghouse motor and cast compressor. The stylized comet decal on the side looks 1950ish to me but the serial number might be coded for 1970. I knew it had a leak and was only getting worse but hadn’t messed with it. Had a flat on the truck and a tractor tire was low and I got fed up enough to say today is the day I fix it. Wouldn’t go much higher than 45 pounds and the compressor never shut off. Sure I could replace it but this is almost a tool museum piece.
I knew welding a pinhole leak would be a challenge since where there is a rust thru pinhole there is thin metal around it. I found the leak and cleaned the area with a flap disk. Set up the mig and started at the hole and spiraled out about the size of a dime. Let it cool and pressured up but now I have 3 leaks at the margins of my weld. I hit it with a grinder and decided one more try before I fall back on cutting a patch to weld on. Now I have weld as big as my thumbprint.
One more time with pressure. Went right up to 95 and an hour later is still holding same. Shot some rattle can rustoleum over it. Looks like I might get a couple more years out of it. Better than a new one.

[Brad Johnson]

Rust visible on the outside is likely only a small portion of the entire involved area. Every time you hit it with the welder you're just welding into more rust you can't see.

Also, the fact that it rusted through means the tank has been structurally comprimised. That's bad. an air tank pressurized to 125 psi is a lot of potential energy. Chances are any failure will be little more than another leak but there remains s the possibily of a catastrophic event.

Retire it and get a replacement. You can get a replacement tank new or, with a bit of patience, source a compressor being sold cheap because of compressor/motor failure. Swap on your good parts.

Brad

[Fly320s]

I agree with Brad.  Keep the compressor, replace the tank.  You can even paint it to make it look original.

[dogmush]

Third guy going to be a downer.  Compromised pressure vessels are very bad juju. Find a good tank and put you classic compressor on it.

[zahc]

Very bad idea. Never weld a pressure vessel.

[zxcvbob]

Long time ago (30 years?) I bought an old Quincy air compressor at a garage sale.  It worked but had a small leak in the tank.  Upon further examination, it looked like maybe it had been dropped and broke off one of the feet, and somebody brazed it back on.  The brass had infiltrated the steel and compromised it.

One of the engineers I worked with said he thought he could fix it.  I don't remember all the details; I think he boiled out all the brass with a TIG welder, then built up a patch with rows of stick welding beads.  The important thing is we pressure tested it when he was done.  Filled it almost full of water, then pressurized it to 300 PSI using SCUBA tanks or something.  (operating pressure was 150 so we doubled that for the test)

I used that compressor for years, until the compressor threw a rod and self destructed.  I should have kept the tank.

[Kingcreek]

Thanks guys but I’m not too worrried about a catastrophic failure. It resides in a storage space and is only powered up when I need air. Hoses long enough that it never leaves the storage space. Yeah I know it will probably spring another leak at some point but I’m going to get what I can out of it.
I’ve seen tanks welded, brazed, and even drilled and tapped for bolts.
Main thing is it’s working again. It’s a 75-100 tank and switch. I have the reg set at 85 now. No bleed.

[HankB]

Very bad idea. Never weld a pressure vessel.

Why not? I just saw a TV shooting show (Brothers in Arms) where the guys did some welding on the tanks of a WWII flamethrower as they were restoring it.

Perfectly safe to take a rusty vintage antique, fill it with gasoline, strap it to your back, and pressurize it while touching off a flame at the nozzle.

They DID pressure test it, but even so . . . 

I'd like to see this live . . . from a couple of hundred yards away.

[slugcatcher]

I don't see a problem welding up a 100psi air tank for a pinhole leak.  Not enough pressure to worry about.   

[Brad Johnson]

I don't see a problem welding up a 100psi air tank for a pinhole leak.  Not enough pressure to worry about.  

Presuming an old-school wheeled shop compressor of ten gallons or so.

Ten gallon cylinder will be roughly 12" diameter / 30" length
= 1131 sq in cylinder surface area
= 226 sq in end cap surface area (113 x 2)
= 1357 sq in total interior surface area
@ 100 psi = 135,700 lbs total
@ 100 psi = 67.85 standard tons

Yes, it is something to worry about. We're discussing a tank with enough internal rust to manifest as an external blemish. Rust that's penetrated from the inside out is like an iceberg. What you can see is a small portion of the total. There is significant involvement of the area around the detectable leak. That means the structural integrity of the entire cylinder is compromised.

A tank failure could easily launch a fair sized chunk of metal with enough velocity to do serious bodily harm. That much pressure contained in a cylinder with obvious structural integrity issue = Nope. For reference, a champagne cork experiences roughly 80-90 psi. The slow-ish pressure release of carbon dioxide from its liquid host also imparts a relatively short pressure pulse, much shorter than a container of compressed gas. Finally, a cork is a relatively forgiving projectile in terms of both weight and density. Even so, plenty of folks walking around with eye patches.

Now imagine the same scenario, but with a razor sharp shard of rusty steel at greater velocity. Ditto having a wrench or hammer blown back at you.

Brad

[zahc]

Exploding air tanks can also compromise a structure due to the low-pressure pulse. Structures are not built to withstand non-equillibrium pressures of even a tiny amount. Pressure tanks letting go can fold up a shop like a house of cards. Buildings that house boilers or large refrigeration systems sometimes incorporate flaps or louvers to quickly release pressure changes without popping any seams.

[MechAg94]

If you have to repair it, at least weld on a big patch over the bad area instead of trying to build up the thickness with an overlay.  At least you have a chance of covering the bad area.  Bad part is you don’t know how bad the rust is.  

It would be safest to just replace it.  Might just be one rust pit.  Might be pitted all over and heat affected zone around your welding area will rust at a faster rate.

[lupinus]

There are some things better left replaced than repaired, pressure vessels are one of them....

Sent from my Pixel XL using Tapatalk

[p12]

Hey Brad

Remember Harmon Tank Company?

Good example of pressure vessel failure.

Semi sized propane tank being pressure tested.

Result. Large building 150-200 foot square. Nothing left but a clean slab and red iron. City condemned about 20 or so homes around the business.

It dislodged some of the dust on the support framework on the roof of the dealership that was about 2 miles away.


Sent from my iPhone using Tapatalk

[Brad Johnson]

Forgot about Harmon. Yeah, that was a big "whump"! I was my office on the opposite side of town, six or seven miles as the crow flies, and felt it.

Brad

[Kingcreek]

First, thank you all for your concern for my safety.
Here's where I am currently.
The weld looks good. The area surrounding the weld seems solid enough with several punch strikes under no pressure.
It's holding pressure now and sitting in a storage area of my 1909 barn (in what used to be an oat bin). I have enough hose to reach my work area without removing the compressor.
The weld is on the bottom back edge of the tank and even if the entire tank rockets it's not likely to take out the yellow pine walls. Any short impulse is not likely to be exacerbated by enclosure in structure, the worst that could happen is some barn boards blow loose. The hay mow above it is empty.
In my opinion, failure at some point in the future is inevitable but far more likely to be another bleed vs a burst. Any catastrophic burst failure would probably not endanger myself or anyone else in its current location.
What I will do. I will continue to use the compressor on those occasions when needed. I will also watch for a good price on a new compressor to replace it.
Thanks all.