Mainspring hardening and tempering

[mbhorology]

I'm in the process of working on a set of repairs for an old car clock.

 

It's actually for by final year Record of Repairs for by BHI course.  It has requested some significant work to the lever escapement including correcting horn and guard pin shake and the pallet stones needed adjusting for correct drop etc.  All very fiddly stuff but it works lots better than before under manual power.

So moving to the mainspring, I thought some might find this interesting.  Never seen anything like this before, but its not right!  For obvious reasons the resilient hooking didn't work very well and slipped after a couple of windings.

 

 

 

I ended up breaking the mainspring when I removed it with a spring winder so it needs replacing.  It gave way near the inner hooking.  Unfortunately the spring is something between a large pocket watch spring and a very small clock spring.   It measures 2.4mm high, 0.3mm thick and I've calculated it should be about 1m long from measuring things up.  

As an experiment I fitted a pocket watch spring of 0.2mm thick and not surprisingly the clock runs but the balance is running at a hopelessly low amplitude.  It needs more power!

Spring material this size isn't easy to find.  I've purchased something from Meadows and Passmore which I thought was ideal (Eureka spring material), the size is spot on, but the material is supplied fully annealed!  Perfect for forming the inner and outer hookings, but in this state totally useless as a spring!

So to my question.  The material needs hardening and tempering; an operation I'm very familiar with for making items like screws.  But for a 1m long coiled mainspring that's something to be considered carefully.  For the spring to be reliable I'm going to need an even heat treatment especially the temper.  So has anyone done this before?

I was thinking that I could put the hardened spring on a brass plate and heat it from underneath until it hits the blue colour?

 Is this madness or should I look harder for the correct spring material?

Edited August 30, 2019 by mbhorology
typo

[HSL]

As you know the design of the barrel and the mainspring is a crusial factor for the function of the watch.
The ratio between the barrel and center pinion must have a definite relationship to the length and strength of the mainspring
to perform satisfactorily and run for the desired number of hours. 

To harden and temper it to a correct strenght might be tricky, I would personally look for a match or a near match and make a new hook on it.

I think this might be a fun excersice for the members on this site too so if it isn't a too big effort you might:

1. Count the teeth of the barrel and the center pinion (the small wheel driven by the barrel).

2. Meassure the inside diameter of the barrel.

This will give us a chance to calculate the size of the spring and make it easier to find.

[oldhippy]

My God what a mess. An easy why when it comes to mainspring size. The spring takes up 1 third of the barrel, the arbor takes up 1 third of the barrel and space left between the arbor and spring takes up a third. Just looking at that photo tells me that is not the correct spring. 

[clockboy]

This is what is needed to calculate.

 

https://www.m-p.co.uk/formulae/sprlen.htm

[mbhorology]

Thanks for the replies.  I've already estimated the spring dimensions so I should have included my calculations and train information.  These are my calculations, but I'm happy to be corrected.

 - Centre pinion = 10 leaves

- Barrel = 104 teeth

The centre is driven directly from the barrel so essentially its a big watch!

So thinking about the turns of the spring for different running periods, I started with 30 hours.

 - the centre would turn 30 times, so the barrel would be 30 *(10/104) = 2.3 turns - unlikely as the resulting thickness of spring would be very powerful

So considering again, would it run for 8 days?

 - the centre would turn 192 times, so the barrel would be 192 * (10/104) = 18.4 turns - which feels like lot!

The proportions of the barrel are quite large so going with this the barrel dimensions are:

 - Barrel diameter = 30.3mm

- Winding arbor diameter = 9.2mm

So based on 1/3 arbor, 1/3 spring, 1/3 space the spring would occupy very roughly 5mm

So if I turned a spring 18 times around the arbor the resulting thickness of spring would be 5/18 = 0.27mm

 

That feels about right given that I've got it running with a very low balance amplitude on a 0.2mm spring.

So taking it a bit further and using the M&P spring calculator it confirms that my spring would be between 970 and 1200mm

 

The spring is 2.4mm high...

 

The closest I've been able to get is the M&P 2.4 x 0.3 x 1000mm annealed!  I didn't realise it was annealed until after I purchased it and I don't fancy risking my own heat treatment and a spring breaking.

 

My difficulty is finding a supplier for this spring material.  I need something close to 2.4 x 0.27 x 1200mm.  As well as M&P I've checked all the springs from Cousins and no luck!  Any suggestions on a supplier?

 

Edited August 31, 2019 by mbhorology
typo

[nickelsilver]

It's marked 4 days on the dial, so figuring one day reserve that's 120 hours, at 10.4 hours per turn of barrel you need 11.5 turns. That's a lot, most available springs are made figuring on around 6 turns. This site has some really great info on mainsprings, with good calculators. According to them you need a spring 0.225mm thick, and betweed 1420 and 1550 in length to get 11.5 turns.

 

As for hardening and tempering a mainspring, I don't know how it can be done without specialized equipment.

[mbhorology]

I'd totally missed the 4 day duration staring me in the face of the clock!

So yes I agree the calculations can be revised:

 - 4 days + 1 = centre rotates 120 times

- Barrel rotates 11.5 times

<update from earlier comment>

I need to understand how that calculator works...

 

Edited August 31, 2019 by mbhorology
1

[HSL]

I see Nickelsilver alredy answered on your post but here we go with some alternative theories.. (Harold C Kelly, A Practical Course in Horology,2013)

Let us calculate the thickness of the spring first. 

Calculating n= b/c n=104/10 gives us the ratio of how many turns the centre wheel moves on one rotation of the Barrel. n = 10.4.

Next we calculate the supposed runtime. 4 Jours (4 days) = 4*24 = 96 Hours.

Calculate how many turns to wind your spring. 96/10.4 = 9.23

Your barrel inner diameter is 30.3 mm we divide this with 12.5 which gives us a parameter value of 2.424

Next we calculate the thicknessof the spring. 2.424 / 9.23 = 0,262

The ratio I usually use is that the arbour covers one third of the space and one half of the remaining area is covered by the spring.

Putting this into the online calculator says your "Length should be between 999 and 1249", it is using the general formula.

SpringLenght = ((BarrelInnerRadius^2 * π) - (ArbourRadius^2 * π))/ (2* SpringThickness) = 1249,4 mm
The 80% is what they call a minimum 1249.4 * 0,8 = 999.52 mm

NAWCC and me use:
http://www.nawcc-index.net/CalcMainspringLength.php

SpringLenght = (1 / (SpringThickness * 2)) * (((π*BarrelInnerDiameter^2)) / 4) - ((π* ArbourDiameter^2)) / 4)) Which is same as the first but using the diameter instead.

In the back of my head I have some recommendation of number of turns one should have in the barrel should be around 11 to 13 so I always use 12 turns. (Looking in your barrel it seems to be 12 too.)
To calculate the lenght depending on this one can use the formula:
SpringLength=π(BarrelInnerDiameter * Turns −Turns^2 * t)

The recommendation falls well in place with my humble calculations which give you a mainspring with the dimensions:
2.4 X 0.26 X 1020
Remember the slightest deviation in the number of decimals you use will make an noticable impact in your calculations

So even if we calculate slightly different the height of this Mainspring will be a challange. For these 4 day car clocks they usually 
were 2.4 or 2.5 mm heigh. 
I'll attache an image so one can see every number collected
 

[vinn3]

17 hours ago, mbhorology said:

I'm in the process of working on a set of repairs for an old car clock.

 

It's actually for by final year Record of Repairs for by BHI course.  It has requested some significant work to the lever escapement including correcting horn and guard pin shake and the pallet stones needed adjusting for correct drop etc.  All very fiddly stuff but it works lots better than before under manual power.

So moving to the mainspring, I thought some might find this interesting.  Never seen anything like this before, but its not right!  For obvious reasons the resilient hooking didn't work very well and slipped after a couple of windings.

 

 

 

I ended up breaking the mainspring when I removed it with a spring winder so it needs replacing.  It gave way near the inner hooking.  Unfortunately the spring is something between a large pocket watch spring and a very small clock spring.   It measures 2.4mm high, 0.3mm thick and I've calculated it should be about 1m long from measuring things up.  

As an experiment I fitted a pocket watch spring of 0.2mm thick and not surprisingly the clock runs but the balance is running at a hopelessly low amplitude.  It needs more power!

Spring material this size isn't easy to find.  I've purchased something from Meadows and Passmore which I thought was ideal (Eureka spring material), the size is spot on, but the material is supplied fully annealed!  Perfect for forming the inner and outer hookings, but in this state totally useless as a spring!

So to my question.  The material needs hardening and tempering; an operation I'm very familiar with for making items like screws.  But for a 1m long coiled mainspring that's something to be considered carefully.  For the spring to be reliable I'm going to need an even heat treatment especially the temper.  So has anyone done this before?

I was thinking that I could put the hardened spring on a brass plate and heat it from underneath until it hits the blue colour?

 Is this madness or should I look harder for the correct spring material?

      anealing and hardening is an art.  a "spring makers " aprentiship was availeable at one time.  much of the wire (or flat stock)  un heat treated.  as i recall.  "water or oil quench".   the art is ( on a existing spring)  aneal an end to make the hook,  and then harden it.   vin       p. s.,  what kind of a car did it come from?

Edited August 31, 2019 by vinn3
add a question

[mbhorology]

Regarding the question about the car it came from, I'm trying to find out.  Currently the clock is mounted in a wooden mounting to hasn't been in the car it came from for many years.

Thanks also for the spring calculations above.  I've been backing this up by reading the section in George Daniels Watchmaking book.  

So I'm now in a position to know what I need, but where can I get 2.4mm height x 0.26 spring material from?  M&P sell it annealed and I can't see anything from HS Walsh or Cousins.  The closest I can get is my pocket watch mainspring which I got from Cousins...  it is however far too short and not thick enough.  It runs for a few cycles at very low amplitude.  I feel if I had the correct spring as calculated above it would work...

[JimmyD]

I read this and watches are not my thing, however looking from the clock view you said;

14 hours ago, vinn3 said:

I ended up breaking the mainspring when I removed it with a spring winder so it needs replacing.  It gave way near the inner hooking.

How close to the inner hook? I would think it would be a lot easer to make a new inner hook hole on the old spring as it is the right spring and the loss of a few inches would not make that much differance. Trying to temper a spring, I doubt it can be done, getting the tension right/even over the distance without the right gear I would say near impossable?

[vinn3]

7 hours ago, JimmyD said:

I read this and watches are not my thing, however looking from the clock view you said;

How close to the inner hook? I would think it would be a lot easer to make a new inner hook hole on the old spring as it is the right spring and the loss of a few inches would not make that much differance. Trying to temper a spring, I doubt it can be done, getting the tension right/even over the distance without the right gear I would say near impossable?

 

7 hours ago, JimmyD said:

I read this and watches are not my thing, however looking from the clock view you said;

How close to the inner hook? I would think it would be a lot easer to make a new inner hook hole on the old spring as it is the right spring and the loss of a few inches would not make that much differance. Trying to temper a spring, I doubt it can be done, getting the tension right/even over the distance without the right gear I would say near impossable?

      the bove post "22 hrs ago vinn3 said"        i did not say !  vinn3

[mbhorology]

Thanks for the further replies.  Just as an experiment I made a new eye on the mainspring which came out of the clock.  It is 0.28mm thick and the clock is now running at a fair amplitude.  This spring is however far too short (it measures about 600mm when it needs to be about 1000mm).  I also fear it breaking again and causing damage.  It is also set, when removed from the barrel, it doesn't open up much like you'd expect a new spring to be.

I think I'll give up on the idea of hardening and tempering the annealed material I got from M&P.

Does anyone know of a supplier where I might get hold of spring material of this size?  I've tried my usual sources of M&P, Walsh and Cousins.

2.4mm is in there now, I might get away with 2.5mm.  From calculations above and my own experiments needs to be 0.27 to 0.3mm and around 1000mm long.