Congreve Clock Repair & Rebuild Project

[raynerd]

Many of you know I'm in the process of moving house and so I've no workshop and I am already fed up doing nothing. I've been looking for a project I can work on at the dining room table with little tools and last week came across this small non-running congreve clock at a good price for the work that has gone into it. A congreve as always been on my "maybe but probably not" list of build projects. They look a stunning clock but there is a heck of a lot of effort that goes into them, a lot of material and they are poor time keepers - there are certainly clocks I'd prefer to make with the time. However, as a project, this was too good to miss. Apparently it has run but with problems and also, if I can get it running, there are many improvements I can make to improve the clock as well.

Here is the clock when first purchased:










These clocks are apparently incredibly sensitive to dust, any dust on the bed can stop the ball. This bed doesn't have dust, it was caked in oil and crud!!

As you can see in the video, one of the first major steps would be to find the plans from which this clock is based. John Wilding has plans for a congreve, but I'm sure this isn't based on those plans. Not only is Wildings a 30 second bed, as with all other congreves I've seen, they have the two dials on either side of the main clock face and this clock wouldn't be able to have this due to the position of the arbours - they just don't match Wildings design.

Any thoughts or suggestions, it would be great to hear anyone's thoughts! There are not many clock repairers that have worked on a congreve so everyone's opinion is more than welcome! I think the issues purely are to do with the bed & ball section of the clock and the gearing looks fine and runs ok! 

[raynerd]

So I was curious what a quick wipe of the bed would do to remove some of the rubish...

This video is misleading as it would stop every few 3-5 minutes. It just showed how sensitive the bed is to dirt on the bed.

I think I have a deeper issue of how the bed is built but I'll post more on this later.

Next step I think is to fully strip down the bed and get it in the ultrasonic cleaner.

[raynerd]

I've cleaned and inspected the bed and I think the problem lies here.

The bed in my opinion is too complex in its construction. It is a single sheet of thin brass with lots of additional brass triangles cut. The brass triangles are screwed down to the plate and the gaps between the triangles form the rolling bed. The edge pieces that are cut into semicircles make the balls jump over and onto the next channel.

I see two issues.

1. The triangles are screwed down in the centre and glued along the bottom. The thin ends of the triangles are lifting on some, which means the channel goes uphill before the corner and the ball stops.

2. The corners where it swaps to the new channel is based on the ball being guided around by the end piece which forms a barrier edge. What this means is that the ball needs to have enough momentum to jump up off the channel and fall into the new channel. In other congreves, the ball runs through a continuous channel so there is no jump to make for the ball to get out of the channel - I think that is the case!







I have a few options now. I think for a start all the brass triangles need removing but the old glue will be a pain to get off. I then would need to re glue but then am I setting myself up for the same to happen in a few years time when they start to lift again! Sadly, this project was meant to be a table top project while I wait for my new workshop - i think the best option is to make a new table but I haven't the equipment available. I could draw it up and see if I can have it laser cut or see if anyone would cut me a new table. 

Your thoughts would be really welcome.

[awemawson]

Thicker slab of brass and cut the channel with a ball endmill on a CNC milling machine.

(Personally I'd make an aluminium test piece first to prove the design due to the price of brass.)

Always fancied a Congreve having seen one in action in the South Ken Science Museum as a teenager MANY years ago, but I certainly don't have the patience to make one.

[raynerd]

Yes, I agree.

And so comes the end of the dining room table project lol  unless I can find someone with a CNC!! 

I've got the brass. I don't think you'd need it too thick due to the weight of it, just enough for the groove, but to be honest, like in other congreves, should it be a ball end groove or cut right through the sheet?

[awemawson]

OK I'll give you a fiver for it 

[raynerd]

Haha.. I think I'll keep hold of it for now!

The bed is only 6" x 7" - I'd prefer to have a CNC grooved bed made but if it allows me to keep going with it, I might find out the cost of a laser cut through bed. Gutted, because my little CNC machine might but up for the job!! - but it's tucked away in storage!

[Noitoen]

I'm just curious. When I look at the groove by the curves, doesn't the ball have to jump from one groove to another? I would think that the tips should be rounded to allow the ball go around the corner without having to jump. Maybe I'm seeing it wrong.

[raynerd]

You are absolutely right!!!!!!! This is what I mentioned in point number 2 in my previous post with the pictures. The ball needs enough momentum to actually spring out of the channel and loop around the edge to be guided in the next channel.
I think like you are saying, if it was going to follow this design, the triangles should end sooner so that the groove was continuous and it stayed in the grove as it went around the corner.

[Noitoen]

Even if you use a cnc to cut the groove, you should use a regular mill and cut the groove deep enough so that the ball rides on the edges and doesn't touch the bottom. A half round groove would create more drag.

[raynerd]

I'm curious know as to how much to take off. I think I'll go down this route as it means I can continue without a workshop! I can't imagine it is too critical as the ball should hug the cut out edge piece as it goes around the corner.

[chipenter]

I that last video the triangles look to be uneven one is wider than the rest , how good are you with a dremmel to cut the ends , it won't matter iff you need to replace the table cutting the ends off by eye will maybe get it going .

[RussellT]

Does the ball run on the bottom of the groove or the edges - if it runs on the edges then there is a risk it will fall into a hole if you cut away the end of the triangles.

Soft soldering the triangles might be better than gluing.

Russell

[vtsteam]

The ball jumping is one problem.

The ball bearing against both the groove and the raised semicircular rail at the turn is another. The ball must, when contacting both, be in sliding contact, not just rolling.

It is skidding in the turn. That's a friction entry as well as a jump

And that semicircular rail looks rough, too. So it's skidding on a rough surface when entering the jump.

I like Noitoen's point about two bearing points as a track - that would minimize sliding contact, and area of contact -- though technically there is a small amount of sliding contact in a curved portion of a railed groove.

It is possible that the centrifugal force at the turn is too great for a flat curved track to contain, and the ball might leave the track without a bank.

In that case the raised sheet metal turnarounds were probably designed to prevent the ball leaving, as well as direct the turn and force the jump.

One way of mimicking a banked turn would be to make the track of two pieces of sheet metal -- one for side A of the groove and one for side B. these would intermesh, like two combs. The root edge of the combs would be bent up to form the banked outside track at the turns.

I don't actually see a reason for a bottom, since the ball would ride on the rails, not the rail bed. Except for looks of course, and maybe stiffness.

[SwarfnStuff]

Not a clock person but after my looking at the photos I tend to agree with Noitoen as I see the Vee points extending under the turn around thus demanding a jump. Seems to me that you have nothing to lose by seeing where the ball jumps and shortening the point to just before that spot. I think the inertia would keep that ball on track to contact the turn around the continue on its merry way. Worth a try as you are contemplating re-cutting the table anyway.
John B

[raynerd]

Hi guys, I appreciate the input.

The ball is definitely rolling on the bottom as you can see a tiny wear on the track where the ball has been touching. vtsteam, you are also correct in that most congreves I have seen do not have a bottom solid sheet - the tracks are cut right through the sheet. I`m not clear on what you mean about skidding and sliding ? - could you explain a little more.

The only think I have noticed on other congreves is that the grooves come to an end shorter so the ball comes to the end of the groove, lifts out and slides around the rail to fall back in the new groove. It doesn`t jump - it rolls out of the groove and follows the turn on the edge. Others, like in the video below have a continuous track. None seem to have to jump up over the edge like on mine.

     

Do you agree then if I shorten the ends of the triangles it will improve the situation.

Chris

[chipenter]

If the ball is rolling on the bottom try a larger ball first .

[vtsteam]

If the ball is rolling on the bottom try a larger ball first .

I like that! 

Could be the original ball was replaced at some time.... It might work better with a larger ball.

.................................................................................

Chris, re. sliding contact -- If the ball is rolling on the bottom (or even sides of the lower track) the axis of rotation is horizontal -- parallel to the plane of the bed -- and perpendicular to the direction it is rolling --- just like a wheel if it had an axle. The axle is on the axis of rotation.

But when it encounters the U shaped raised sheet, the axis of rotation wants to change to vertical -- to accommodate the new edge it is riding against. Yet it is also still riding along the lower track -- so the two axes of rotation are fighting each other -- and either one turns to slip and the other remains rolling, or they both partially slip. But a ball can't roll purely in two axis at once. It can have only one axis of rotation.

Since this clock seems so sensitive to a free rolling ball, I'm guessing the friction due to slip is probably significant, where it might not be for another kind of mechanism or equipment .

Further, the slip/friction is not consistent, but happens at the turns mostly. And if the turns aren't consistent either (due to roughness of sheet metal, and the strips loosening) I imagine the timekeeping is inconsistent, and in this case it even stops.

Maybe it wasn't meant to "jump" the strips but ride on them until it hits the U turn sheet, where it rides only on that, lifting above the strips. But it would really need a fine adjustment, and the right sized ball for that.

If it rode only on the U sheet, above the track, the single axis of rotation would be satisfied.

[RussellT]

Hi Chris

I've just been looking at the latest video you posted.  There's a bit at 1m 23s that shows the turning action clearly and it shows that the triangles stop quite well clear of thesemicircular turning "wall of death".

If the ball rides on the edges of the grooves then as it comes out of the turn it has to get back up on the edges of the grooves, while if it rides on the bottom of the groove then it only has to make the turn rather than change height as well.

A larger ball would certainly raise it up on to the edges of the grooves but if I am visualising it correctly it would collide with the corners of the wall of death.

I think the answer is probably to progressively shorten the triangles (or a dummy one made of ally) and observe the effect on the action.  If it has to get back up on the edges then the triangles might need to be tapered in the horizontal plane too.

Given your workshop limitations it might be possible to cut aluminium that thickness with the kitchen scissors.

Russell

[tom osselton]

Interesting clock!  My take on it is the triangle piece would be the key and should be away from the wall of death by at least the radius of the ball bearing if it extends more than that it will hammer and bounce trying to go in two directions at once rather than using momentum.

[raynerd]

 Hi guys, thanks for the replies! My daughter was with me when I made the video and wanted to say something - sorry 

Cutting the ends makes a MASSIVE difference!!

[RussellT]

Well done.  That looks much better.  It looks as though it should adjust OK too.

Russell