Brass NPT coupler failures 2

[JWJDesign]

We have been building a simple coolant box that has two pipes going through it. One has an electric valve on it and one has a check valve. All the parts are brass. There are 4 brass 3/8 NPT female to female couplers, one on each end of each pipe. Early in the manufacture, we had returns due to (usually) one of the couplers splitting. They were not in service very long. We contacted the manufacturer to find out what they recommended for tightening. We had been using the standard recommended for NPT threads of that size, which is 2 to 3 turns from finger tight. We were surprised to find out the mfgr recommends 1 turn from finger tight. Ok, so we changed our method to that and we still see failures. Then we changed to a higher pressure rated fitting of the same type, still splitting. We changed to Loctite sealant instead of pipe tape, no change. Our customer says the coolant running in the system is just chilled water with some anti-fungal in it. We can't get an answer as to what exactly the anti-fungal is but I can say that it doesn't have any odor to it, thinking one of the things that attacks brass is ammonia. There are temperature swings but it is not more than 40F to 150F. The pressure is no more than 100psi. The other brass parts, the valve and the check valve are not splitting and have the same 3/8NPT female threads and are tightened to the same spec. Here is a pic of one of the split couplers

I sent two split couplers back to the manufacturer (distributor is not saying who the manufacturer is) and they said they have no opinion on why it's splitting other than we are doing something wrong or they are not being used appropriately. Very frustrating, we are not over tightening them or exposing them to anything they wouldn't see in most applications and we still have split couplers coming back. Any insight would be appreciated.

 

[WKTaylor]

What alloy [number/temper], form [rod, tube, etc] and specification [AMS, MIL, ASTM xxxx] for the brass fittings?

Where is the crack origin? Have you submitted a sample for metallurgical analysis? Any significant zinc content to the brass?



Regards, Wil Taylor

o Trust - But Verify!
o We believe to be true what we prefer to be true. [Unknown]
o For those who believe, no proof is required; for those who cannot believe, no proof is possible. [variation,Stuart Chase]
o Unfortunately, in science what You 'believe' is irrelevant. ["Orion", Homebuiltairplanes.com forum]

 

[Mr168]

What kind of lines are going to the couplers? i.e. hardline, flex lines, etc.

 

[metengr]

Send the failed fitting(s) to a local materials lab for proper failure analysis to determine root cause.

 

[Tmoose]

no marking on the couplings to ID the manufacture? Once upon a time, not that long ago, That would have been a little suspicious by itself. Maybe that is just how it is today.

I'd start over with couplings from Parker or Weatherhead while having a cracked one analyzed.

For my own curiosity I'd cut a cracked coupling apart to expose the fracture surfaces, and post a bunch of picture here.

 

[EdStainless]

Better couplings have a more generous hex, and they have one at each end of the coupling.
This provides reinforcement at the highest stress areas.

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, Plymouth Tube

 

[JWJDesign]

Thanks for all the responses.

Wil, all the information I have on the fitting is this from the distributors catalog:

Extreme Pressure Brass Pipe Fitting
Material: Brass
Fabrication: Machined from bar
Max pressure: 3900 psi @ 72 F
Specs met: ANSI/ASME B1.20.3; ASTM B124; ASTM B16; ASTM B453, SAE J476.
Not RoHS compliant

There isn't any alloy information. We dont' have the expertise to determine the crack origin, other than it looks like it usually propagates from the end towards the middle. I have not sent one for metallurgical analysis, but will as soon as I can find a lab that can do it. Any suggestions for Northern Cal?

Mr168, the lines are brass nipples.

metengr, I will send these to a lab as soon as I know of one.

Tmoose, The original coupler had no markings on it but this one has PB on one of the hex flats. I will see if I can spare on to open up and take a picture.

EdStainless, I thought we were buying premium fittings with these. In the meantime, here is a picture of the custom ones I had made. The originals are shown as well. The nipples are brass, they don't look it though.

 

[metengr]

You can easily find a metallurgical test lab within California. You can call and Fed Express the pieces to them.

 

[JWJDesign]

Correction to a bit of the info on the fitting, the original fittings we were used were marked PB, the second version that is described above and is what we have been using is marked P3. It's the 3900psi version.

 

[MikeHalloran]

Borrow an Enerpac pump and see if you can get even one of those fittings to 3900 psi without a failure. I doubt it. It looks like a generic 150 psi fitting.

The nipples appear to be 'red brass', a common alloy for pipe.

Given any number of field failures, I'd be using forged fittings, at least just for that one customer. I'd guess they're applying a moment load to the fittings, and I wouldn't expect them to admit it.

Can you show us how a moment load applied to the fittings would interact with your box?






Mike Halloran
Pembroke Pines, FL, USA

 

[mk3223]

Were the failed couplings from the same manufacturer? I suspect that the splitting is the material issue due the casting defect in the bar stock. Based on the service pressure and temperature, either Class 125 or 250 brass fittings are suitable for the service. Did you try the couplings made from other vendor since it's cheap?

 

[JWJDesign]

I have sent some of the split fittings to a lab and will let you all know the outcome.

There is no external stress on these assemblies. The two assemblies shown in the previous picture are contained end to end in a box. There are male fittings that connect to them through the box walls.

All the fittings are obtained from the same distributor and they all have the same marking on them. It appears to be the same manufacturer mark.

It will take about 2 weeks to get the test report.

 

[metengr]

Well done.

 

[JWJDesign]

This is the summary from the lab analysis:

Two assembled water line control box couplings and three unassembled couplings were submitted for analysis to investigate the cause of cracking in five of the seven couplings. Results of the examination indicated the cracks were due to stress corrosion cracking (SCC). An intergranular fracture path was found on an opened crack fracture surface and in a transverse cross-section taken through a second cracked coupling. The cracks appear to initiate at the end of the fitting and propagate faster along the outer diameter (OD) surface. Crack propagation faster along the OD surface indicated hoop stress drove the propagation.
Energy dispersive spectroscopy (EDS) showed traces of chlorides and sulfur, which can be part of compounds known to cause SCC in copper alloys. The presence of other chemicals known to cause SCC in copper alloys, such as ammonia and certain nitrates, cannot be detected using EDS.

The only thing I don't get is they are saying the cracks initiate at the end of the fitting. There isn't fluid there to cause the SCC to begin with and there is no fluid on the outside of the nipple until it is cracked and leaking. However the report is very thorough with photos showing more corrosion at the outer part of the crack

 

[moltenmetal]

First guess would be the tapping fluid used when tapping the female NPT parts- it's probably a high sulphur material. But not knowing the exposure history, it's tough to know what else might be the cause.

 

[Tmoose]

Thanks for providing some follow-up.

What is the plan for fixing the problem? OR have you already ?

Cleaning the part chemically, and then the exterior and ends mechanically, followed by shot-peening ( not just grit blasting) can offer improvements in resistance to SCC.

 

[JWJDesign]

MM: That's certainly possible but the nipples look very clean when they arrive from the distributor. We have never considered that we should clean a standard brass fitting prior to assembly. But that is something to consider, prior exposure.

Tmoose: We have discontinued the product and are only in the maintenance phase with it now. We stopped producing it because we didn't want to continue making something that would not work long term. I helped the customer find a different way to do the same thing. The only fix is when we receive a set back to repair we replace all 4 of the nipples with the custom versions we had made that are way bigger in diameter (see a few posts above). We have only done one like that so no data on fails there.

Thanks

 

[metengr]

Nice follow-up. Thanks.