Hi everyone, I never had this specific problem in my 30 years of riding, last winter I replaced my chain/sprocket on my 1987 Schwinn Prelude road bike because of chain stretch. The rear derailleur is shifting perfectly but when I change from the big chain ring up front to the small one occasionally the chain will slide off the rear derailleur pulley as shown in the photo and I have to get off the bike to put it back on. any help please in what could be causing this would be appreciated.

Bent derailleur hanger? Bent derailleur? B-screw adjustment? Wheel not centered? Greetings from Eburg, btw.

It could be a stiff link in the chain, but I suspect that the issue is a slow to respond tension cage. When you downshift the front you release lot's of slack into the lower loop. If the RD cage is sticky it fails to take it up fast enough allowing the chain to jump off the pulleys.

Normally, I'd expect the chain to stay put despite the slack, but if the Rd is misaligned the chain might use the slack to hop off.

Step 1, check and align the hanger and RD shifting to a gear combination where the RD cage is somewhat vertical. Then, using a broomstick or other straightedge held vertically against the rear wheel as a reference, see if the cage is parallel or angled in (almost always at the bottom). If it's leaning and closer at the bottom, you can have a pro align it, or buy an expansive tool, or do a crude but effective job by using a small crowbar (aka stiff screwdriver) against the hanger and using the innermost sprocket, lever the hanger out until the cage is parallel.

Now check the cage spring by pulling the lower loop down and letting go. If it doesn't snap back quickly and cleanly, try running some oil into the lower body to flush out some crud and oil the pivot.

Between the two the problem should resolve. But don't forget that bending the change moves the entire RD outward and you'll need to reset the limits and cable adjustments.

thanks guys, it does appear to be slightly crooked. I noticed the RD is the original one that came with the bike in 1987 and there is play in it.Maybe time for a new one but this never happened with the old chain sprocket set up.

The derailleur appears to be lined up outside the smallest cog, so surprising if it shifts perfectly.

RDs generally have a tab to prevent the chain from going off the jockey wheel like that- something missing or broken?

Chain threaded outside of the tab? it would be noisy and rough.

good eyes woodcraft, maybe that was a result of me purposely taking the chain off the pulley to show where the chain is going. I haven't been able to reproduce the problem on the repair stand. The tab appears as shown in this photo to be ok, could I simply bend that a little ? In the second photo you can see the alignment of the derailleur to the cogs.

Are you saying that you put the chain in its position in the first pic, to show what has happened while riding?

What 'sprocket' was changed? freewheel? chainring?

All of FB's points above, especially alignment.

yes that's what I did,(first question) It happened twice while riding this morning and twice last Sunday. I changed the freewheel and chain, the same teeth as before. I didn't make any changes to the front chain rings. The RD itself is shifting fine,only time the problem arises is when shifting down to the small chain ring up front. I made sure the chain was the same length as before. It never did this before with the old freewheel/chain

Pulley wheels worn & don't like new chain?

as mentioned, the only thing happening when only shifting to smaller chainring is that the chain becomes slack for a moment before the chain tensioner comes into play. i'd look at that, as mentioned. it might be THE opportunity for something that is misaligned to be making it's move.

and i may be way off base here, but i suspect that many problems that can't be reproduced on the stand are due to the fact that there is less slack for the tensioner to take up on the stand, in that people rarely pedal a bike with an unloaded wheel by hand as they do when propelling it along the ground with their legs. leaving slack on top and the bottom. the slack is more equitable above and below the chainstay. just a thought. could be way wrong about this.

I'll add my $.02. In post #6 there are two photos. The chain in each seems to be different, the first looks to be an older chain that has the side plates which are larger around the rollers. The second photo's chain looks to be much more like current chains where the side plates are almost flush with the roller diameter. Why does this matter? I believe that (and assuming my observation is correct) that the older chain design will keep on the pulley better then the new more flush side plate chains do. So if the pulley teeth were well worn (as Woodcraft mentioned) the teeth won't engage the chain links as deeply as they might AND not as deeply within the side plates as the old chain did. This will allow the chain to slide off the side of the pulley more easily.

Back in the day there were pulleys that were perfectly toothless, just round disks. (millions of the Huret ders used this design). And for their time they shifted well enough, or more to the OP's problem, didn't allow the chain to derail off the pulley easily. But ALSO BITD the chains had much more significant side plate dimension. (It was only when the Sedis Sport chains came about that this side plate dimension changed dramatically IIRC. And by them most all pulleys had some sort of tooth profiling).

A side note is that for decades the der cage didn't need a tab between the pulleys to help keep the chain on the pulleys. IIRC it was the advent of the mountain bike that changed this. Which happened close to the same time as the new slim chain dimensions.

I wonder if the OP installed new pulleys, of the correct width, and or changed back to an older chain design if the problem of derailing off the pulley would stop.

As you said, you could be wrong here ---- and are.

With a working freewheel, the upper and lower loops are totally isolated from each other. The upper (drive) loop would work identically if there was a mile of chain behind the bike, and the chainrings spooled used chain onto the ground. The tension there is 100% the result of driving forces.

So the lower loop is simply a return loop to return the chain the chainrings are spooling forward back to be recycled. The tension in the lower loop is purely from the RD idler cage spring. If it weren't for the RD, the lower loop would be slack and sag under it's own weight (as it does on single speed bikes) regardless of how easy or hard one pedals.

However there is a real difference between the stand and streets. That's the bouncing around that bad pavements and potholes create.

my post doesn't assume that when pedaling a bike on a stand people are always exerting a force. many times i've seen them just chasing the freewheel with the crank and changing gears at the same time. in fact, IME, it constitutes the majority of the time when testing a RD on a stand. mostly, i suspect, it's because it is so easy to spin out the wheel when there is no load on it. one only has to overcome the inertia of the wheel and in order to maintain the speed, the drag of the bearings and the minor air resistance.

this has made me wonder, if anybody that did or does RD tuning a lot, like a mech, for instance, if they have had reason to create some sort of artificial resistance device to more accurately mimic real road conditions and forces.

Same thing, upper loop tension is rider controlled, lower loop tension is constant based on RD takeup spring.

i was pretty much expanding on the concept brought up in the first paragraph of the second response in the thread. and why it happens on the road and not on the stand.

as an experiment, i just tried it on my geared bike. what i did, and what most people do, i think, when testing a DR is to spin up the rear wheel and then test the shifting while the freewheel is, well, freewheeling. the freewheel is turning sure enough, but not as fast as the wheel.

and the chain sags both above and below the chainstay while shifting. if anybody is interested try it.

Andrew, the photos were taken just minutes apart and within a few minutes of making that post last night

this is the position of the RD hanger with it both of the smallest ring.sprocket. Tension on the hanger seems to be strong.

Did you use a narrower chain than was originally spec'ed on the bike?

Huey, for what it is worth, I have gotten so that I set up my geared bikes on the trainer to dial in shifting, especially after dealing with my Campy build. I could not get it to shift as smoothly as I wanted with the wheel freewheeling, and was going to seek professional help, when I decided to plop it on the trainer and do it with a loaded wheel. Ever since, I have done all of my tuning that way, Shimano and Campy alike.

I have not noticed a significant difference on the single speed, though.

interesting.

BTW, i've sought professional help in the past... my friends tell me it hasn't helped.

They can only work with what they are given.

Changing out the rear derailleur, even temporarily for trouble-shooting is a good idea. That eliminates (or confirms) a problem with the derailleur itself. Could be that the cage is bent. Could be that something is wrong with the guide pulley (you didn't take it apart and mix up the pulleys, did you?). Best way to narrow it down is to swap out parts until you pinpoint the problem (and taking measurements).

this is possible that I used a narrower chain, the chain does say narrow on it. This bike has always been a 6 speed freewheel one. If that's what's causing it can I close the tab on the derailleur a bit to prevent it from passing by it ?

From your photo in post #6 , I see why not bent the tab just make sure the chain and/or pulley is not rubbing on the cage when you do. It wouldn't hurt to replace the derailleur to test the drivetrain . If the chain is too narrow it could cost your problem , replace the chain with one not so narrow , like for a 5,6 speed .