cyccommute 

Again the chlorinated solvent, which is a very different beast than a chloride ion, isn't going to do anything to the chain unless you could liberate the chloride atom. But that's not going to happen until you reach very high temperatures and/or use some very fancy...and very hazardous...chemical techniques.

A chloride ion in an aqueous media, which the Green Klean mineral spirits appears to be, will do all kinds of funky stuff to metals.

I've started a test tonight to see if I can duplicate the cracking. I won't be using Green Klean spirits but I will be using a 6% solution of sodium chloride in water. This amount of chloride should speed up the process nicely. I'll post results and pictures later.

joejack951

Now we're having fun

joejack951

I used to have a few short pieces of Shimano and SRAM chains. If I can find them, I'll toss them in too.

I started the thread both to warn others and because I was interested in what caused the cracks. If you made your statements with no other information, I'd question you as that has not been my experience with the same chain, nor others that I know who have used those chains (or any other chains). Chains don't typically break and if they do it's usually an assembly issue which is what I'd suspect was your problem (unless you posted pictures that showed otherwise). No one else has posted about their experience with long term storage of a chain in GOMS so there isn't much data to go on other than what I've presented and facts about how steel typically reacts in such a solution.

Hillrider has posted achieving close to 8000 miles on a chain (well he posted 75000 but I believe that was a typo and should have been 7500). I'm a relatively lightweight guy (160 lbs.) and I'm also a spinner. 80 rpm feels really slow to me. 90-110 is more like it. I don't shift under (much) power and I try to keep my chain lubed (using Chain-L). Summary: I'm pretty light on chains and do decent maintenance. As DaveSSS pointed out, I'm also using a 9 speed chain which is going to offer longer life than a 10 speed chain.

I'll gladly test a section of donated chain or you can do like cycocommute and test it yourself. That way there's no reporting bias nor can it be potentially exposed to the gremlins living in my garage with teeth hard enough to crack chain links

Yup.

joejack951

Definitely "green" odorless mineral spirits, possibly with a hint of DelDOT road salt.

The pins aren't moving inside of the outer plates (maybe microscopically). And this picture shows a link crack that looks more like the chain was used in compression rather than tension:



If all of the cracks had been towards the exterior of the link I could have believe that my cycling-added stress contributed to the failure. But this oddball crack (which I witnessed only once) seems to indicate that corrosion is randomly attacking the links.

Your final statement is very true. Any little defect in the link gives a spot for the corrosion to start and once it starts it gets worse quickly as the crack gets bigger aided by the existing hoop stress of the riveted pin.

dedhed

I'd be sending those pictures to KMC and asking them what happened. Probably a free chain in it too.

jim hughes

I was very interested to find out that Klean-Strip isn't even pure mineral spirits, but some sort of magic water-based emulsion. Because that's what I've been using, and I don't think it's as effective as good old mineral spirits. Also (and this might be my imagination) it seems to evaporate slower, making it harder to get rid of when I'm done.

So I went to Home Depot and guess what: all they have is Klean-Strip, you can't even get the real thing any more.

joejack951

Klean Strip makes normal, non-emulsified odorless mineral spirits. It's what I've switched to using. What you need to avoid is the "green" odorless mineral spirits which has the water added.

Kimmo

Ignore the sh|tpost. This is one of the most interesting threads here.

deermouse

My chain broke by pulling a pin out of the side plates (and I was sprinting up a hill). Shimano doesn't use a master link and their 10 speed chain has special (use once) pins that you insert when you assemble the chain. Also on 10 speed chain the pins don't extend past the edge of the side plates, so they may separate easier. In my case it probably was an assembly issue. I have no reason to assume that if I got another Ultegra chain it would fail in the same manner, however I did replace it with a Sram chain (with a master link).

My point exactly.

I agree that 9 speed chain will probably hold up longer than 10 speed chain, but not 8 times longer. After my chain broke I did a general scan of other forums and threads pertaining to chain life. The general feeling seemed to be that chain should last between 2000 to 4000 miles with fewer miles for 10 speed chain. If your weight and riding style allow you to get more miles, that's great. That's another variable that goes into the equation.


I'll think about doing it myself. I will probably forget about it and I'm not sure I can get the exact solvent you are using. The gremlins in your garage are also part of the equation. I will be glad to send a section of chain if you are interested. Maybe both of us should test it if I have enough.

Bill in Houston

Sorry to resurrect this, but I gotta say, JoeJack, those are the best photos of a mechanical problem I have ever seen on a discussion board. They really got this thread off on the right foot by clearly communicating to us what the exact failure was. Well done.

tru

taking a chain off for cleaning is silly...and a total PITA. I used to do that until I got lazy...and guess what? my chains last just as long as they did before.


just take a clean rag to it (I use old t-shirts)...spin the cranks until the chain doesnt dirty up the rag.
re-apply lube. one drop per link.
wipe off excess lube after a couple hours. I usually wait at least two hours...or overnight.
do this every so often...once a week, every other week...whatever, depending mileage & weather conditions.

works just fine.

reddog3

It is the internet after all. Five pages trying to prove that a defective chain isn't really defective?

Been working for lots of people (me included) for many years. If it could ever be proven that the chemical in question was the cause of failure, I'd just quit doing it. Maybe the "green" thing came back to bite you in the butt- as it normally will if not applied properly.

joejack951

Thanks, Bill! I'm still working on my macro photo skills but I learned a few things from this exercise (in addition to my learning about some chemical interactions). I'll be sure to post some more photos of my next great bicycle part failure

Bill in Houston

I'm looking forward to it?

Staggerwing

Joejack linked to the following site earlier, but to the galvanic corrosion page. I think the stress corrosion cracking is a better fit for this particular issue. It is NOT a defective chain.

If may not be clear, but the peening process would put a significant radial stress on the side plates, and this stress exists as the chain is laying free on a table top.

As an aside, some years back I did an engineering co-op rotation at an aircraft propeller manufacturer. One of the last steps in aluminum blade processing was to induce a compressive pre-stress in the surface zone of the blade root by a simple rolling process. Even a single hunk of metal, loaded only only modestly by gravity, can be significantly stressed, either by design or accident.

Thanks to Cyccommute for enlightening with the chemical origin for the corrosion.

man with bikes

A fatigue failure of the chain would not look like the failure shown here. The crack would propagate through the chain link at the point where the "C" is stamped. this failure is either stress corrosion or hydrogen embrittlement.

Scroll down this link to see that the naysayers are wrong. Clearly a chemical issue here.
https://www.hitachi-metals-techno.com...roller/05.html

RollCNY

I just stumbled on this thread, and was thinking hydrogen embrittlement from the first photos. Chain links are through hardening steels, electro plated, and can be very susceptible. Mineral spirits are hydro-carbons, so if you had an initial issue, the assembly is under constant stress (compression on pin, tension on link), and you provide more fuel for increased loss of strength.

It isn't corrosion, it is a plummet in the base strength of the steel. And as Cyccommute says, the area where waist meets end radius is a large potential stress riser. I have never manufactured bicycle chain, but nine years with industrial roller chain showed me some things. Some old dies used separate waist punch from the blanking punch, so mis-match and stress riser was easy at that point. Even with one piece blankers, that transition point was easy to flub, until we started using wire EDM on all of our blank punches. By design, that area is also weakest point, and planned failure area (smallest cross section in tension), so any stress riser just moves it closer.

Also, chain is a commodity, with a high % of material cost. We had to constantly search for cheapest material, and links had lots of acceptable alternates in strip steel, so we would vary for availability and cost. Different materials would interact in the plating process differently, and hydrogen embrittlement was a large concern. My memory is foggy at this point, but I seem to recollect grades that our metallurgist insisted that we did not plate, but unfortunately, that was awhile ago.

cyccommute 

The Great Chain Corrosion experiment is finally over! On 23 October, 2011, I placed a piece of chain in a saline solution (~6% sodium chloride). Here's a pictures showing the proportions and the clean chain in the solution




The chain I used was completely free of defects as shown in this picture



I checked the chain by removing it periodically and inspecting it for the kinds of cracks that JoeJack experienced and, frankly, I was prepared for failure. This morning, 9 December 2011, I checked it again and here's the result:

What the mixture looks like



and the chain




You can clearly see two cracks in the chain that weren't there before I placed the chain in the solution. Both cracks look very much like JoeJack's chain and are in the stamped parts of the chain where I would expect the highest stresses during manufacture of the chain. I'm both amazed that it took so long and astounded that it happened in only 40 days.

My wife will be glad to get the damned jar of rust off her counter. And I think we should all thank her for putting up with my silly bicycle science experiment.

jack002

Very interesting.

Is this not related to the failure of the hull of the Titanic? I thought I'd heard that the cold salt water reacted with the steel/iron and that this type of steel/iron was made really brittle in the water. I'm not a chemist, but that's how I recall the story I heard.

bobn

The OP states that the damage was done by mineral spirits. See #1. Correct me if I am wrong but he mentions no other chemicals, salts or concoctions being involved.

Looks like apples and oranges?

Snydermann

I'd have never guessed this could happen, wow. Thanks for running that experiment.

Nerull

Reading the rest of the thread too difficult?

bobn

Nerull,
How about checking #22

Spoke to Roy from Klean Strip, Barr Products Memphis Tn. He said he has never heard of their products damaging steel.
Roy said it was OK to use his name and will accept calls. 1-800 398 3892.
He asked for pictures and any information that he could forward to their engineers/chemists for examination.
There you go. We are not talking about all kinds of other concoctions. We are talking about what the OP originally used.
Want to settle it once and for all, call the manufacturer and ask the question. Send in the proof.

Mondoman

Actually, the OP states he did not use traditional mineral spirits, but rather some "green" substitute, which apparently is an emulsion and thus contains significant amounts of water. This aqueous phase may have salts in it as it comes out of the can, but the OP also mentioned that there was some amount of DelDOT road salt on the chain as well.

Thus, we've got a water phase with some kind of salt in it, according to the OP.

Booger1

The cracks appear around the pins,where there is the least amount of room.

My guess is that the salts are re-crystalizing between the pin and side plate and needs more room,so it makes room,much the way loctite works or sorta like making rock candy on a smaller scale.

I'm guessing those cracks would not appear if the fluid was kept moving and crystals couldn't form.

Pull a link apart with just a hole,then put it in a salt solution.....I'm guessing the hole will be fine.

Gotta admit,it is interesting......