Check my math, please
 

I had posted earlier about how to measure chain stretch and found out (thanks to urban assault) that the stretch limit is 1/16 of an inch (.0625 inches) for every 12 inches of chain. As Svloid mentioned, 12 inches is hard to measure with calipers.

So my question is, to calculate the max amount of stretch for any given legnth of chain is the equation

(.0625/12)(any given chain legnth being measured)=max stretch? Thanks.

Yes.

Thanks. By the way if anyone is going to use this, convert to metric! Tried doing it in inches and the caliper reads in fractions of an inch (including vernier at 128ths in.!) and its a real PITA to convert all thos fractions.

That's what wives are for. I have my wife pull the chain tight by holding the rear wheel and preasuring up on a pedal. Then I measure a foot of chain with a good steel ruler.
The Park chain checker works OK once you get the feel of it. You take the slack out and avoid too much pressure on the tool.
There is no magic amount of stretch we should tolerate. The 1/16th rule is a rule of thumb. It really depends on how much chance of cassette wear you want to take. I try to catch mine before the 1/16th.

Al

Thanks for the props, but, as I think I mentioned in the original thread, credit should go to Mr. Sheldon Brown for the well written explanation.

You can read 1/16" with a steel rule or even tape measure. It may be preferable to measure 12 links instead of one. Measuring 12 is like using statisical average for 12 links. Measuring one just gives the length of one link and assumes all links wear the same.

FWIW, Park tool sells a go / no go chain gauge. If the tool can fit into the chain, then it's time to change.

I know some people don't want to buy another tool, but the chain-checker is really simple to use.

Thanks all. What I'm doing is using a cheapo vernier caliper (I think it was less than 10 bucks at home depo) that reads to .1mm. Using the inside measurement part of the vernier, I set it to 106mm and put it between the links, as I open the vernier, it stretches the chain automatically and then I just set it till I get a little friction as I remove the vernier, much like using a feeler guage. The stretch limit on the legnth I'm measuring is .55mm, so the cheapo vernier is more than accurate enough.

Beware that, if I'm actually reading this right, what you're measuring isn't actually a pin center to pin center measurement. You're measuring inside to inside, including play in the rollers. It's an important distinction that should be reflected in the pass/no pass algorithm that you're using.

Yes, I'm measuring inside to inside. I'm assuming that once you take the slack out of the chain, you can measure from any two points and compare that same measurement from the same points on a new chain and you will get an accurate reading on the stretch. Even if you measure pin to pin, you must still remove slack (play) to measure. Is my assumption wrong?

It should not matter much if I understand your concern DiegoFrogs. I think you are saying that the measurement could be inaccurate if the rollers have a difference in the amount of play/wear, meaning the rollers have worn unevenly. From Sheldons explanation, what is being measures is not physical stretching of the plates but mainly the amount of wear at the pivot points. The difference in wear of one roller compared to another should be statistically small. If in doubt, you would probably be better off changing the chain. Letting it go could mean a more extensive drivetrain overhaul and replacement.

In pre-index days I used to tighten up my shifting by removing some of the deflection from the chains. I would do this by pressing the sideplates slightly tighter. I did do it a little to much on occasion but it was readily obvious when I had problems getting to the higher deflection gears. I don't have time to mess with that anymore, plus index shifting is much better than I need at my aged level :).

I think your caliper technique is OK since you are comparing the worn chain to a new chain. My experience has been that my chains do not wear evenly but if you check various links around the chain, and use the greatest stretch as the determinate, you should have the right answer.

Al

I use the caliper technique, measuring inside to inside. I found that the unstretched dimension over 10 links (10? 11?) is 5.2 inch. I found this by measuring the diameter of the rollers new (0.3 inch), and offseting the pitch by the diameter (5.5 inch minus .3 inch = 5.2 inch. Given that, I calculated what 0.75% stretch and 1.0% stretch would amount to, and I made a little table showing % stretch vs. actual measurement. Here it is:

0.50% 5.226
0.55% 5.229
0.60% 5.231
0.65% 5.234
0.70% 5.236
0.75% 5.239
0.80% 5.242
0.85% 5.244
0.90% 5.247
0.95% 5.249
1.00% 5.252

So I measure the chain in 3 or 4 spots, take an average, and look up the % stretch.
I toss the chain if it runs rough no matter what the dimension say.

I think I'm doing essentially what you did. I took a brand new chain and found that x no of links (I don't remember the no) was 106.5mm and adding the calculated stretch limit of .55mm, the wear limit of a used chain is 107.05mm. My road bike was at 106.9 and my mountain bike at 107.7(!). So I put the road bike chain on the mountain bike and the new one on the road bike. I was surprised the mountain bike didn't jump, not even the smallest cog.