Bicycle Mechanics - New chain, newish cassette, still skips

I have about 2,500 miles on an LHT purchased six months ago.
On a ride last weekend, I experienced a little chain skip and auto shifting while mashing up some small but steep hills.
I figured in was time for a new chain. I purchased a SRAM PC971 9-speed chain, the same one that came on the new bike. I was prepared to cut it to fit, but right out of the box it was just slightly shorter that the chain I had removed -- not quite a link shorter.
I installed it and test rode it, and it still skips when I mash the cranks on the middle chain ring and the 11 or 13T cog on the cassette, a Shimano CS-HG50.
I'm sure the chain is threaded right. I see no obvious wear on the cogs.
What are your thoughts?
Thanks!

If it's skipping on a cog, it's the cog not the chain. A new chain will be more likely to skip on a worn cog than will a worn chain. The amount of wear that causes skipping is pretty small and difficult to see by eye. Autoshifting is a separate problem unrelated to the chain, IMO.

I don't know what an "LTH" is but it seems odd that the new chain was shorter rather than longer and that 1 link shorter (1") was OK. Does an LTH have unusually long chainstays?

Did you ever adjust the rear derailleur?

Edit. As the derailleur cable stretches it throws off the shifting as every shift is a little behind; this will first show it self with the small cogs because they have fewer teeth to hold on to the chain. There is a barrel shifter on the end of the derailleur where the cable/housing is attached; this needs to be truned counter clockwise (IE unscrew) to adjust out the stretch. If this is too basic hopefully you will give up some additional information..

If those are the gears you use the most suggest you measure the chain you junked. If its past the 100% wear mark chances are you've damaged the cassette sprockets. If on the other hand its below the 100% wear mark - suggest you check the cassette for proper torque. It slways surprises me how earily too many of them come off. The tightening torque is indicated on the locking ring of every brand and its a MINIMUM recommended value - usually 40 to 45 Nm.

I don't know what an "LTH" is but it seems odd that the new chain was shorter rather than longer and that 1 link shorter (1") was OK. Does an LTH have unusually long chainstays?It's "LHT" and is short for the Surly Long Haul Trucker. Indeed the LHT was very long chainstays (460mm) and requires 116 links; the SRAM PC-71 chain comes in either 114 links or 116 links; the OP apparently bought the former..

. the SRAM PC-71 chain comes in either 114 links or 116 links; the OP apparently bought the former..
I didn't know that.
The new chain was just barely shorter than the old. Not two links shorter.
Burton, what is the 100% wear mark?
Onespeedbiker, I did adjust the RD at about 1,200 miles for initial cable stretch.
Thanks again.

I didn't know that.
The new chain was just barely shorter than the old. Not two links shorter.
Burton, what is the 100% wear mark?
Onespeedbiker, I did adjust the RD at about 1,200 miles for initial cable stretch.
Thanks again.

There are several ways and tools to measure chain wear. Any of them should get you the answer you need:
See the section ln "measuring chain wear": http://sheldonbrown.com/chains.html

You have to separate skipping under load, (which is usually (but not always) because of chain or sprocket wear), from auto shifting or related issues, (which are usually because of poor trim, bent hanger, or related RD problems) .

Skipping under load, especially on smaller or the most used sprockets is classic, and most often shows up when the chain is replaced, but not the cassette. However, it's very possible to be fooled into thinking the skipping is in the rear when actually it's happening at the chainring since they feel and sound the same. My favorite way to diagnose chain skip takes two people.

Put the bike against a wall and have a friend attempt to ride through it. This lets you carefully observe the chain under load but not moving. On a bike prone to skipping the chain will move outward and appear to shift back on the chainring (actually the sprocket is moving forward) as the crank is loaded. At the cassette the chain moves out and forward slightly.

"slightly shorter that the chain I had removed -- not quite a link shorter"

If it less than a full link shorter that could be explained by "stretch" or wear of the chain. it is not uncommon for a chain to have gained an inch or more from wear. Did you match up some the chain rivets as you compared length?

"Mashing" on an 11 tooth cog is flat-out a bad idea. Due to the small number of teeth you will wear the cog down very quickly. You should be in the larger cogs on the larger chainwheel that have similar gear ratio. In my view there is only one time to use an 11 tooth gear, and that is when one is already at speed and you wish to lower your rpm's.

Good advice from CNY-B.

I had an intermittent autoshifting hassle upon upgrading to 9spd I was having a hell of a time diagnosing, thinking it must have been previously negligible cable hassles as a result of the finer tolerances and shorter cable pull...

Turns out it was freehub body with excessive runout... took me a while to spot.

When chains are specified as being 114 or 116 links, it's really 1/2 links they're counting; counting inner and outer links separately. If you think of a chain as being made up of identical links that repeat, then a link comprises a set of inner and outer plates as well as two pins and two bushings. Both senses are used in cycling which can be a bit confusing. When you shorten (or lengthen) a chain, you can only do it in complete link increments, +/- 1 inch, not 1/2 links.

(Yes, depending on the size of the chain, there may be "half links" available, but bringing that up only further complicates the discussion without being relevant to the topic at hand)

If those are the gears you use the most suggest you measure the chain you junked. If its past the 100% wear mark chances are you've damaged the cassette sprockets. If on the other hand its below the 100% wear mark - suggest you check the cassette for proper torque. It slways surprises me how earily too many of them come off. The tightening torque is indicated on the locking ring of every brand and its a MINIMUM recommended value - usually 40 to 45 Nm.
Burton, the old chain was less than 1/16" long at the 12" mark and still around 1/16" at 36", so I surmise that it had less than 100% wear. The 11T and 13T cogs still look sharp and chiseled.
If I have to check the cassette torque, that's a LBS job for me.

However, it's very possible to be fooled into thinking the skipping is in the rear when actually it's happening at the chainring since they feel and sound the same. My favorite way to diagnose chain skip takes two people.

Put the bike against a wall and have a friend attempt to ride through it. This lets you carefully observe the chain under load but not moving. On a bike prone to skipping the chain will move outward and appear to shift back on the chainring (actually the sprocket is moving forward) as the crank is loaded. At the cassette the chain moves out and forward slightly.

FB, you may have named it. While the rear cassette shows no wear to my eye, the cogs on the middle chain ring do look a bit worn.
To me, that seems premature at 2,500 miles and without overly stretching the original chain. What do you think?
I haven't been able to try your two-person test yet ... maybe later today.

just buy a new cassette, keep an eye on chain wear indication.
and prematurely replace the chain..
Rohloff makes a cog wear indicator.. if you want more tools.
KMC makes a digital chain wear indicator..

FB, you may have named it. While the rear cassette shows no wear to my eye, the cogs on the middle chain ring do look a bit worn.
To me, that seems premature at 2,500 miles and without overly stretching the original chain. What do you think?
I haven't been able to try your two-person test yet ... maybe later today.

If you've already replaced the cassette, then that's virtually eliminated, so it's either chainring slippage, or the other class of possibilities relating to how the RD is feeding the chain. Even without the test, one indicator pointing the chainring is slippage that happens on one ring, but not the other. In the mtb world it's usually the inners (mostly smallish alloy middle rings) that go first. I've watched mtn bikers go nuts and replace everything but, only to find that the 34t ring was the problem.

If the ring does check out, then you're down to stiff links, hanger/RD alignment, & B screw adjustment (only if it skips on mid and larger sprockets).

As far as it being premature, it is early, but if the bike is ridden at high chain tension - heavy loads, hard climbing - it isn't impossible. Recently on this forum someone asked about chainring wear, and posted a photo of an unbelievably worn Campy inner, where the tooth edges were rolled to the side by the chains rollers.

12" steel machinist's rule is probably the best way of determining when to replace chain. 1/16" elongation over 12" is the usual spec but I replace at 1/32" to prolong the life of the cassette and rings. I just replaced a chain showing just under 1/32" elongation after 4200 miles.

Chain gauges add in twice the roller wear which does not figure into chain elongation. Elongation is the critical determination that needs to be made and these gauges will have you replacing chains long before they need to be. I brought this up with the manufacturer of one of these gauges and they agreed.

"Mashing" on an 11 tooth cog is flat-out a bad idea. Due to the small number of teeth you will wear the cog down very quickly. You should be in the larger cogs on the smaller chainwheel that have similar gear ratio. In my view there is only one time to use an 11 tooth gear, and that is when one is already at speed and you wish to lower your rpm's.

Am I reading this wrong, or is this not how gearing works? Wouldn't it be larger cogs on the larger chainwheel for a similar gear ratio?

Typo/brainfart. It's not you.

whew.

Typo/brainfart. It's not you.

keerect - fixed typo