Endshake problem solution - acceptable or a hack?

[LarryAC]

I've searched but can't find a good answer to this problem elsewhere.

I'm working on an Ansonia mantle clock movement that did not run when I acquired the clock, but it is old so I'm sure it ran at some point. The problem is the 2nd and 3rd wheel of the going train have have virtually zero endshake and the escape wheel has less than 0.001". The 2nd and 3rd wheel spin "ok" when assembled individually or with the train. The Escape wheel a little better, but if I add a 0.010" shim under the corner post near those gears, I get about 0.01" endshake and all wheels in the going train spin freely and very nicely.

With or without the gear trains installed, the plates measure near perfectly parallel along that edge and the bushes are all dead flat with the plates. There is no discernible or measurable warp or bend in the plates. I don't have a lathe to cut down the gear arbors and I can't think of an alternative solution than to either recess the bushes or use a shim under that one column. The latter is a reversible fix so would be my choice, but is using a shim an acceptable fix or is there a better way?

EDIT: The shim thickness was incorrect. It is 0.01", not 0.02".

 

[roughbarked]

You could try recessing the bushes. Though my question would be, do the bushes look like they have been chamfered on the inside? If not, do you have a chamfering tool? This can often make that difference.

 

[bruce linde]

You could try recessing the bushes. Though my question would be, do the bushes look like they have been chamfered on the inside? If not, do you have a chamfering tool? This can often make that difference.

is there anything wrong with using shims or shim-y washers? non-permanent, and .02" of endshake still sounds more precise than it needs to be.... isn't a reasonable amount of slop a good thing?

 

[roughbarked]

Russian Poljots often use balance cock shims from the factory. A reasonable amount of slop is necessary to avoid binding.

 

[LarryAC]

The bushes have no chamfers on the insides of the plates, but they also look factory original. There is clearly some amount of binding since with the 0.01" shims, and the subsequent endshake, the wheels spin much more freely - they seem about perfect.

I do have a chamfer tool and I can try that as a fix. I suppose the question is: would a professional repairer use a shim or go some other route?

 

[roughbarked]

If it is your clock, why would it worry you?
If it works with the shims then use them.

 

[KurtinSA]

Are the plates perfectly flat? If they've been bent, that will reduce the end shake if the plates are bowed towards each other.

Kurt

 

[LarryAC]

The plate measure flat by straight edge and the distance between the plate measured with digital calipers is nearly identical along the length of the plate.

It's my clock and just a learning project. The objective is to not cut corners but to learn how to repair clocks "properly". That may or may not mean shims are acceptable.

 

[bkerr]

I know you said you measured but, in order to loose end shake the plate may be warped or something else creeped up, I would also look for a pivot that is not straight. Look closely at the shape of the pivot with a loop or a microscope. The pivots should be straight no taper. Did you bush the clock? Maybe the bushings are a tad tight on the i.d.
A clock will run loose but it will not run tight.
Also, you could try to install one wheel at a time to check how it reacts. Then add another wheel and see what that looks like, keep doing that for each train. It is possible that the bushings may have been put off center if they were replaced at some point. Checking the wheels this way will show where the problems are.

Good luck

 

[roughbarked]

Well, you could actually try to solve the problems without the shims too. Twice the fun factor.

 

[R. Croswell]

I would say the important question is, has the clock been running OK for some time (years), or did this problem turn up when the clock was put back together after being disassembled of for bushing work or some other reason? It can't run if it is tight, it must have run at some point, so we need to find what changed. The shims won't hurt anything but shouldn't be needed. If the pivots are worn there may be a radius at the base of the pivot. When the radius is forced to enter a new bushing in may bind. That's the reason for the chamfer. I believe bkerr was touching on this area.

If it is any comfort, I have an Ansonia mantel clock on the bench now that seems to have the same problem. Got to clean it up first but it reportedly was running OK until the mainspring busted. Looks like the plates may be bent but there are reports that sometimes plates were bent at the factory to achieve end shake (I'm not sure if that is true or just something that's been repeated so many times that it is believed.). I've never really noticed if Ansonia movements normally have less than the expected amount of end shake. It doesn't take very much, just not zero.

RC

 

[LarryAC]

I bought the clock off OfferUp as not running for $20 as a case and movement restoration project. I don't know the clock's history except is is fairly old and and the movement was gummy ans sticky so it probably hasn't ran in years.

The bushes in the clock look original and are flush with the inner plate surfaces. Only 2 need replacing - not the ones involved with the binding gears.

I've cleaned all the parts and examined the pivots and bushes. The pivots all all look good - straight and no grooves by both lighted magnifying glass and a laboratory dissecting microscope.

The plates are flat with no measurable signs of bends or warps. Spacing between the plates without or with gears is constant along the length of the plates. I have tested each gear in the going train individually and as a set - no difference in effect. The train spins really well with the 0.01" shim, and "ok" without the shim.

Only the going train has this problem with the 2nd, 3rd and escape wheels.

There is no visible space between the shoulders of the arbors/pivots and the plates and the bushes in the plates are not chamfered on the inner sides of the plates.

I would like to fix the problem without a shim, but that seems to mean either bending the plates, recessing the bushes, possibly chamfering the bushes, or cut the arbors back a bit (which I don't think I can do without a lathe ).

I could just use the shim and be done, but that doesn't seem professional to me. I'm not looking for the easy way out, I want to learn how to "do it right" (assuming a shim isn't "doing it right").

 

[roughbarked]

You do need to give some endshake to those wheels and apart from shims, recessing the bushes a small amount may be all that is needed.
I haven't seen pictures of the clock but the Ansonias that I've worked on never had bushes installed until someone decided they needed them.
Do the bushes already protrude from the back of the plate? Can they be pushed through a bit?
Sometimes the wrong height bush is fitted. These may look flush but a check with a fingernail may reveal that they are not quite flush.

 

[R. Croswell]

I could just use the shim and be done, but that doesn't seem professional to me. I'm not looking for the easy way out, I want to learn how to "do it right" (assuming a shim isn't "doing it right").

I think the professional way would be to track down the cause of the problem and remove it. With a history like this clock's, there are many unknowns. I agree, if this clock has bushings installed, they were not installed at the factory. I think the problem is likely related to the bushings in some way. I believe someone already mentioned this, but if a bushing is fitted too tight in the ID it may appear to have no end shake. If the arbor has little or no "tilt" when the pivot is standing up in the bushing (each plate separately), consider broaching the pivot hole just a tiny amount and chamfer the opening, use the aforementioned "finger nail" test to make sure the bushing really is perfectly flush with the plate and check end shake again.

Whoever sold this had a reason for selling. Perhaps a do gooder previously flattened the plates and couldn't figure out why it wouldn't run?

RC

 

[bkerr]

To have a old clock, maybe 100 years or or better not require any bushings is a bit unusual. Pictures are worth 1000 words. So like stated above you have several choices. Bend the plates, add a shims on the posts or modify the wheels. What I mean by that is to increase the length of the pivot, cut back on the arbor a bit to give you end shake. This will require a lathe. Your last post explains what wheels are effected. Have you tried one at a time to narrow that down? If all three show the same condition and you have checked the pivots then to me, it points back to a bent plate or both plates bent. It does not take much to get end shake. One other thought, I am guessing that this movement has trundles? Double check these as well if nothing else eliminate that possibility. Keep looking and working, you will find and fix and that will be a good day!

 

[Willie X]

I've seen factory installed shims, assuming you are speaking of a thin washer on a post. Your clock may have had a washer that fell out, could be as simple as that ... no need to go on a quest on this one.

Bent plates are out there too, but I've only seen that on old American clocks.

Willie X

 

[LarryAC]

Here's some photos. The first photo is of the outside of the top plate and the second photo is of the inside of that plate. The back plate is similar. The markers point out the locations of the 2nd, 3rd and escape wheels. I cannot feel the bushes, if that's what they are, with my fingertip - they are dead flat. The third photo is of the 2nd wheel showing there is virtually no space/gap between either end of the arbor and the plates. The 3rd wheel looks the same when assembled like this. The escape wheel has maybe 0.001".

The 2nd and 3rd wheel have about 5 degrees or side-to-side tilt in either single plate. The pivots of both move just a bit in their bushes when the plates are assembled - pretty much as I think i want.

Based on my limited interaction with the guy I bought this from, I got the impression he bought it to try to fix it but gave up. There is one badly bent tooth on the 2nd wheel I am going to try to repair (that will be another thread - I have no idea how to do this). There's a crack at the base of that tooth that looks fresh (bright brass). I suspect he was the cause. He might also have removed a shim or straightened a bent plate that was necessary to let the wheels run. The trundles on the adjacent 3rd wheel are also bent and will be replaced.

I plan on trying chamfering this afternoon. If that doesn't work I'll go with the shim and call it a day on this problem and move on to the broken gear tooth.

I've concluded that using a shim to fix this problem won't get me banned form the forum

 

[roughbarked]

Are you sure that these are bushed?

 

[LarryAC]

Not at all. They look bushed to me from the outside, but the inside is not so clear.

 

[roughbarked]

Which implies to me that they are not bushed. So your problem comes down to the plates and the wheel axles.
If as you say, the plates are not bent, then something else is not as it was when it left the factory. Perhaps the wheels are wrong?

 

[LarryAC]

That's a possibility.

 

[Ken X]

Just out of interest you could use your vernier to measure the height of the plate posts and then the length of the problem arbors shoulder to shoulder to see what end shake you should have in theory if all is parallel.
If adequate, you have local narrowing of the plates/Bushes or the radius at the pivot base is not matched by the chamfer in the bush. If too small then more can be generated by whichever means you are happy with.

 

[LarryAC]

Great idea. The 2nd, 3rd and EW arbors measure 34.72, 34.6 and 34.6mm The 1st wheel, that runs freely with some endshake, measures at 34.56mm. All 4 posts measure between 34.75 (the post in question) to 34.8mm. The distance between the plates right at the post in question measures 34.71 and the plates measure 34.73 close to where the 2nd and 3rd wheels insert, but this is a harder measurement to get accurately.

 

[R. Croswell]

What are those purple splotches?

There is one badly bent tooth on the 2nd wheel I am going to try to repair (that will be another thread - I have no idea how to do this). There's a crack at the base of that tooth that looks fresh (bright brass). I suspect he was the cause. ......... The trundles on the adjacent 3rd wheel are also bent and will be replaced.

If there is an actual crack in a 2nd wheel tooth, do not try to straighten it, it will have to be cut out and a new tooth inserted. It takes a good bit of force to bend a tooth like that, and it takes a good but of force to bend the trundles in a lantern pinion, typically enough to bend the arbor. If the arbors that appear to have zero end shake are just slightly bent, the pivots may fit the pivot holes just fine but when the arbor is placed between the plates the pivots will be cocked in the pivot holes and bind. It seems like this movement has be subjected to some type of trauma, Perhaps some of the wheels are from a different movement? Is this the original movement for the case that you have? If not, you might consider looking for another movement that has less wrong with it.

RC

 

[LarryAC]

RC,

Here's some photos of the broken tooth and the corresponding bent trundles.

It's an Ansonia mantle clock and the movement is an Ansonia #4. I have no idea if the movement is original, or the gears in the movement were replaced.

I am very new to this and these clocks are solely learning project. I am purposefully buying clocks that appear to need serious work - the worse the better. I seem to be succeeding in that effort. I figured out bushing and trundle replacement on the last project, an Ansonia Africa clock with a mis-matched movement, so I'm motivated to try to replace a tooth - I have nothing to lose and valuable experience to gain in trying.

I have read some of Penman's book and all of, Lloyd's and a couple of Conover's books that have been very helpful, but I get the most helpful guidance when I get stuck from the folks on these boards, so thanks for all the suggestions.

 

[shutterbug]

FYI, often the front plates from some manufacturers were pressed, and can look like bushings when they are not. If the "bushing" doesn't look like one from that back of the plate, then it isn't one.

 

[Ken X]

Great idea. The 2nd, 3rd and EW arbors measure 34.72, 34.6 and 34.6mm The 1st wheel, that runs freely with some endshake, measures at 34.56mm. All 4 posts measure between 34.75 (the post in question) to 34.8mm. The distance between the plates right at the post in question measures 34.71 and the plates measure 34.73 close to where the 2nd and 3rd wheels insert, but this is a harder measurement to get accurately.

Again just for comparison, I happen to have an Ansonia Gingerbread mechanism on my bench awaiting some pivot wire for a heavily worn fly lantern. A quick vernier around shows me the posts are around 34.80 and arbors are around 34.70. The endshake is therefore around 0.1mm or 4 thou in old money. I had previously individually fitted each wheel and got a nice click from all of them when flipped so I'm happy with that.

As has been said, the presence of damaged teeth and trundles does indicate a bit of trauma has occured. I would be eliminating slight arbor distortion as a cause of the stiffness. Rolling them on a flat surface may reveal a slight wobble.

 

[Willie X]

Again just for comparison, I happen to have an Ansonia Gingerbread mechanism on my bench awaiting some pivot wire for a heavily worn fly lantern. A quick vernier around shows me the posts are around 34.80 and arbors are around 34.70. The endshake is therefore around 0.1mm or 4 thou in old money. I had previously individually fitted each wheel and got a nice click from all of them when flipped so I'm happy with that.

As has been said, the presence of damaged teeth and trundles does indicate a bit of trauma has occured. I would be eliminating slight arbor distortion as a cause of the stiffness. Rolling them on a flat surface may reveal a slight wobble.

 

[Uhralt]

RC,

Here's some photos of the broken tooth and the corresponding bent trundles.

It's an Ansonia mantle clock and the movement is an Ansonia #4. I have no idea if the movement is original, or the gears in the movement were replaced.

I am very new to this and these clocks are solely learning project. I am purposefully buying clocks that appear to need serious work - the worse the better. I seem to be succeeding in that effort. I figured out bushing and trundle replacement on the last project, an Ansonia Africa clock with a mis-matched movement, so I'm motivated to try to replace a tooth - I have nothing to lose and valuable experience to gain in trying.

I have read some of Penman's book and all of, Lloyd's and a couple of Conover's books that have been very helpful, but I get the most helpful guidance when I get stuck from the folks on these boards, so thanks for all the suggestions.

View attachment 756668 View attachment 756669

Yep, that tooth and the trundles need to be replaced.

Uhralt

 

[Willie X]

Normally you would spin the arbor where it lives, or between cone centers, or between your thumb and forefinger.

To straighten a trundle you will need a small screwdrive that's just the right width to go from the root of a good trundle to the center of the bow on the bad trundle. Get the screwdriver placed just right and gradually twist it to get the results you need.

Willie X

 

[Ken X]

Normally you would spin the arbor where it lives, or between cone centers, or between your thumb and forefinger.

To straighten a trundle you will need a small screwdrive that's just the right width to go from the root of a good trundle to the center of the bow on the bad trundle. Get the screwdriver placed just right and gradually twist it to get the results you need.

Willie X

Thanks for the wisdom, I always find it worthwhile to read the advice of those with years of experience.
Rest assured I stick to non-running charity shop cast offs where I can do no harm. Trouble is the house is gently filling with running clocks and it's catching the eye of OC Domestic.

 

[R. Croswell]

I would not fool around with trying to straighten the bent trundle. Just snip all of them, remove them, and replace the whole bunch. You said you want to learn, s here's a good opportunity to learn something that you will need to do many times.

RC

 

[LarryAC]

I agree. I plan to replace the bent trundles. I replaced them all on the last clock except for the fly/balance wheel and it was not particularly difficult.

I double checked arbors and pivots on at least the wheels in questions and they are straight.

 

[Willie X]

A single, not worn to much, pinion wire would get straightened by me. Moderate wear, or more than one bent pinion wire would be replace all.

If you don't already have one, here is a photo of a simple pinion cutter that will make the job a lot easier. Add a vise and a jeweler's saw any your ready to go.

Slightly worn pinion wires get no
attention from me. Willie X

 

[LarryAC]

Willie,

I'm not sure I see how this tool is used. Can you point to a description somewhere or maybe post a photo of it in use?

I bought a cheap guillotine cutter online and it sort of worked for cutting pinion wire, but the cutting lever/blade was not "hardened" steel as advertised and quickly deformed and became useless. I'm hoping to find a piece of hardened steel I can replace it with - maybe an old file? So a better tool would be welcome.

The pinion in question has several bent trundles so I'll replace all on that pinion. The other pinions are pretty good and I agree it is probably not worth the effort to replace them. The trundles on my previous clock where in very badly worn condition. Some had nearly worn through. Since I wanted to learn the process, I replaced them all.

 

[Willie X]

These American made clocks did not have hardened steel trundles. Nearly all of the steel in the clock is very close to a modern-day nail, which makes a good replacement and you can cut it with a jewelers saw.

If you are bent on using so called 'pinion wire' or 'blue steel wire' or 'piano wire' that can be best cut with a Dremel, or Fordam rotary tool, with a cut off wheel.

Whatever you use should leave a clean flat end, and all pieces exactly the same length.

A piano wire cutter/sheer will leave two lopsided ends and both ends will need to be trued up with the just mentioned tools, or you can cut them a bit long and true them up with a stone, or vertical sander, worked across the end of the trundle tool.

Anyway, this little tool can help you come up with a set of identical pinion wires in not to much time.

The body of the tool is 1/2" steel with a 1/16" hole. It has to have the business end reground as necessary. This one started out at about 1 3/4" long. WillieX

 

[R. Croswell]

It is important that the replacement trundles be real close to the original size. Music wire, pivot wire, and drill blanks are available in a wide range of sizes. True, most trundles in American clocks were made of mild steel, which is why they are subject to bending (and why they can be snipped for removal). It is unlikely that you will find an off the shelf nail that is an exact match. Other small diameter wire from a coil will not be perfectly straight. You have a choice (if you can find a size match), but I prefer music wire or pivot wire. It won’t easily bend and is best cut with a Dremel cutoff wheel or similar.

The trundles should be uniform in regard to wear. There is some disagreement over whether common lantern pinions were intended to have the trundles “roll” or rotate, experience has shown that some do, some do not, some may have at one time but no longer, and the worst case, some in the same lantern do while others are stuck. Resulting in uneven wear. Always best to replace all the trundles if there is any uneven wear or damage. I recommend cutting all the replacements from the same stock.

One thing to look out for is if the teeth on the gear are worn (mostly on 1st and 2nd wheels) and worn trundles are replaced in the mating pinion, sometimes the new trundles will bind on the “shelf” worn in the side of the wheels teeth. Unless the wear is bad, it is often best to leave wheel and pinion alone. Or flip the wheel and replace all the trundles. New on new, or worn on worn, but new on worn can be a problem.

RC

 

[TEACLOCKS]

There is A FORMULA for EVERY part on any clock, Except End shake ?
Hard to believe.

 

[TEACLOCKS]

These American made clocks did not have hardened steel trundles. Nearly all of the steel in the clock is very close to a modern-day nail, which makes a good replacement and you can cut it with a jewelers saw.

If you are bent on using so called 'pinion wire' or 'blue steel wire' or 'piano wire' that can be best cut with a Dremel, or Fordam rotary tool, with a cut off wheel.

Whatever you use should leave a clean flat end, and all pieces exactly the same length.

A piano wire cutter/sheer will leave two lopsided ends and both ends will need to be trued up with the just mentioned tools, or you can cut them a bit long and true them up with a stone, or vertical sander, worked across the end of the trundle tool.

Anyway, this little tool can help you come up with a set of identical pinion wires in not to much time.

The body of the tool is 1/2" steel with a 1/16" hole. It has to have the business end reground as necessary. This one started out at about 1 3/4" long. WillieX

View attachment 756733

Do you file the wire end to length ?

 

[R. Croswell]

There is A FORMULA for EVERY part on any clock, Except End shake ?
Hard to believe.

I feel sure there is a formula for end shake but it is probably different for each maker and would be in the form of x + .005" - 0.000".

RC

 

[R. Croswell]

Do you file the wire end to length ?

This was Willie's question so I expect we will hear from him.

I have an old vernier caliper that I set to the length I need. I place the stock across the jaws and mark with a felt marker. Put the trundle in a pin vice and use a Dremel to cut off at the mark. Then hold the trundle in pliers and use the caliper as a go/no go gauge. Just tap the end of the cut trundle against the side of the Dremel cut off wheel a couple times until it "goes". After a few times it gets pretty easy. I'm sure there are faster ways but speed is not important to me. If you expect the trundles to "roll" the ends need to be clean with no sharp edges. Personally, I do not attempt to make them roll in most cases because they probably won't after awhile anyway.

RC

 

[shutterbug]

There is A FORMULA for EVERY part on any clock, Except End shake ?
Hard to believe.

I'm pretty sure it's "about this much"

 

[Willie X]

Could be about 'Dat' much. You know ... Dis-n-Dat.

How-a-bout 'Yay' much? I think you need hand gestures to use "Yay much" though??

Willie X

 

[Willie X]

Sure you could file, sand, stone, whatever you like. As for me, I'm standing there with a Fordom cut-off tool, so I use that. Willie X

 

[roughbarked]

There is A FORMULA for EVERY part on any clock, Except End shake ?
Hard to believe.

If it rattles, it is good?

 

[TEACLOCKS]

Sure you could file, sand, stone, whatever you like. As for me, I'm standing there with a Fordom cut-off tool, so I use that. Willie X

What is this tool you are talking about ?

Could be about 'Dat' much. You know ... Dis-n-Dat.

How-a-bout 'Yay' much? I think you need hand gestures to use "Yay much" though??

Willie X

Give or Take

 

[Willie X]

A Fordom tool is kinda like a Dremel tool on steroids (up to 1/4 HP) and a #31 Jacob's chuck.

A Dremel with a thin cut-off wheel will do just fine.

Willie X