Thursday 24 January 2013

Front Saddle Mounts


As mentioned last time, I've spent my shedweek making the front mounts.

 I've never seen front mounts like these before. 
They match the catalogue illustrations, so clearly the original saddle and worth copying.


These are tricky little blighters with lots of angles and sticky out bits and holes and stuff. After a little planning (read trigonometry) I calculated that I could just get them out of a piece of 1 1/8" bar stock if I was careful and had a bit of a tailwind. The mounts have angled holes and faces in multiple planes which means that I need to machine the containing shape out accurately before drilling and tapping the relevant holes. Only when this is done can I then smooth out the angles and make them appear more organic. The originals would have been castings with very little clean up.

My new milling machine got an extensive trial this week, I'm still learning the best way to use it with the tooling I have available to me. Also I spent quite a considerable period of time making the thing run a bit more quietly. This is a known issue with the Mk1 Dore Westbury mills. They have an epicyclic gearbox that uses straight cut Myford change gears that are noisy in use.

Initially I machined the angled flats by using a Myford swivelling vertical slide, but after a period of time I realised that I could get the same results with more rigidity if I just used a machine vice and set up the work with a combination square with a protractor head.

A series of photos will illustrate the process as usual.

 Mill flats on either side of the bar stock.

 Mill the angled faces using the swivelling slide.

 And again.

 Before I ditched the swivelling slide due to clearance issues 
and just used a combination square to set the angles.

 Mill another flat on the side.

 Mark out and cut the two halves. 

And now mill faces in the other plane.

Hacksaw off some excess...

... before milling to size.

 All the faces are now done, so I can drill and tap the
 holes at the correct locations and angles.

 Starting to clean up with a file.

Beginning to take shape.

A little more  thinning is still needed before I'm happy.


Next week I'll start on the saddle tensioning components.

In other news, the whanau have left me and gone to visit relies in Australia for the week, Whilst they have gone we have had 30+ degree days and bush fires. Honestly what has 'stralia got that you can't get right here in Canterbury. A few weeks ago I even saw a wallaby at the side of the road as I was cycling home. It was just at the Tram Road and Northern Motorway intersection. I know that they occur down in South Canterbury but I never seen one this far North before. I know for certain that it was a wallaby because I was able to watch it for a long time without it running off. In fact I would have been greatly surprised if it had run off since it was dead. I looked it up and apparently it is a Red Necked or Bennett's Wallaby. It is of course possible that it was killed down South and inadvertently or otherwise carried North. It wasn't there the next day so it was probably just resting or something.

Tuesday 15 January 2013

Saddle Rails


The next part of the saddle to get my attention is the rails, the front to back members that bolt to the neck at the front and the cantle plate at the rear. I cross checked measurements of the original facile saddle with a couple of original Brooks of the period and the wire is 9/32" or very close to 7mm, if you prefer. This is quite a substantial gauge to have to bend by hand, even more so when you consider that the rails are made from spring steel which must be bent cold. Spring steel is hard to shape since it is designed to bend back (obviously), it is tough stuff.

A quick trip around the corner to CMI Springs and I was able to get some straight lengths of 9/32" spring steel wire. They don't normally sell small volumes to the public but I chatted the receptionist up and she caught the boss on a good day. This was just before Christmas and they were all looking forward to the holiday. I bought a few spare lengths at the same time.

Most of the bends are very straightforward, the hard one is the eye loop at the rear of the rail which bolts to the cantle plate. I don't like hitting things, I prefer to use measured force so I made a little jig to assist in the process.


First make a loop in the wire as tight as possible by gripping the end very firmly
 in a large vice and bending back on itself so that it forms a coil. 

This was about the limit of my physical strength and I ended up 
on my arse a few times as the wire slipped in the vice. 

When the basic loop has been formed, the vice can be 
used to decrease the size of the loop and also to flatten it. 


Then carefully cut off the end that has been gripped and 
again squish it in the vice to flatten as much as possible. 

I needed to use various spacers to push the coil too far the other way so that when it sprung back it was flat. Cutting spring steel is also a bit of a mission, hacksaw blades will dull very quickly. An abrasive cut off wheel is probably the best option.

The basic loop can now be squished into the jig...

... and bolted up so that it can't move. 

This is then held in the vice and the return bend completed. 

The jig lends a certain amount of uniformity to the shape of the eye.

With the eyes completed, the remaining bends seemed easy in comparison. The only difficulties were ensuring that both sides remained identical mirrors of each other and that the straight, front section of both rails were parallel to each other. This is where the saddle tensioning hardware slides back and forth.


According to the head guy at CMI, all I need to do now is soak in my oven for an hour at 200 degrees Celsigrade to remove the residual stresses from bending. No other heat treatment is needed. I'll have to be a bit careful about when I do this, I don't particularly want my lovely wife to find out.

I have been slightly surprised by the length of the saddle on this bike, from the front tip of the rails to the back of the cantle plate is 16.5". This is much longer than the usual Brooks B70s found on contemporary penny farthings. The reason becomes clear when you consider the rationale behind the bike, the whole idea being to get the rider lower and further back as a safety feature to help prevent headers. Remember that the bike was in direct competition to Starley's new rear driving safety and of course we know which design became dominant. Over the course of the 6 years that the geared facile was in production, the bikes became less upright and the rider was placed further back. It is very fortunate that the original I am copying is so complete, even down to the saddle leather. Had I not had an original saddle to copy I would have made it to the same dimensions as a conventional Brooks B70 International saddle.

The original saddle from the geared facile.

Next week I'll make the front clamps that bolt the rails to the neck, the ones on a facile being particularly complex (of course).

In other news we've been having more weather this week, we had 38 degrees Celsigrade in my garden on Saturday, it was so hot that we all went down to a local swimming hole at the river and messed around for a couple of hours. The braided rivers in our part of the world are excellent for making dams and other contrived hydraulic civil engineering projects. Much fun can be had diverting the river about. Of course after the next big rain, all our efforts will be washed away as the river shifts again.

It has also been too hot to do much machining, but after seeing Tweed Pete's spotless shed a few days ago, I have forced myself to have a tidy up in mine. I've found all sorts of things that I'd forgotten I had, hidden under several feet of swarf. I may have to do it more regularly.




Wednesday 9 January 2013

It all comes together


As mentioned last time, I wanted to get all the outstanding brazing work completed before I went back to work in the new year. I'm now back at work but I'm pleased to say that I did manage to find half a day to go and hire the oxyacetylene and get the bits stuck together.

I'd accumulated quite a pile of bits ready for brazing and it was good to get them all attached to each other at last. There are 5 sub assemblies to be brazed:

The new bolts into the cantle plate.
The bolts into the spring carrier
The saddle post clamp
The lever pivot castings to the end of the fork legs.
The individual handlebars into the head and attaching the brake lever pivot.

I only made jigs for the handlebars and the lever pivots, the remaining assemblies being too simple to require holding, this proved to be a mistake as I'll explain later.


The handlebars and brake pivot are held in the correct locations with some very special frame building pins - old nails. unfortunately, I drilled the holes very slightly too large and the pins have a tiny amount of movement, this meant that the individual handlebars could move relative to each other and could end up not being aligned. I solved this by zip tying the bars to a thick sheet of MDF to maintain the alignment.


The lever pivot jig needed to be more complex and accurate, the pivots need to be concentric, parallel and correctly spaced relative to each other. They also need to be parallel to the main bearing housings. I recycled the old jig used to braze up the main bearings and with suitably machined spacers I was able to hold everything securely in the correct locations. I had to be careful brazing these castings as I had already had them nitrided and I didn't want to affect the hardness of the bearing surfaces. My idea was to use a large lump of steel as the spacer, this will double as a heat sink and hopefully prevent the inner races from getting too hot. I also tried a new brand of 45% silver solder this time, this has a lower melting point and should flash in quite quickly. I intended to go in quickly with a hot flame and spend as little time as possible heating the castings.


As always the first job is to thoroughly clean and degrease the surfaces to be brazed and then flux and assemble into the relevant jigs. The actual brazing went really well and the lever pivots are firmly attached and still hard after the heating. I've only cleaned up the handlebars and lever pivots so far, the saddle components have yet to be done.



For the first time I can now assemble all of the parts made to date. The sun gear still has a lot of material to be removed and the planet has nothing to attach to yet. The total weight at this point is 33lbs 14oz (~15.4kgs) which is on target as per an original.



It is great to have the basic bike together after all this time. As I mentioned before, I didn't use a jig to hold the saddle post mount and consequently it has come out at the wrong angle. It's only 5 degrees off but it will affect how and where the saddle sits and to my eye it just looks wrong, it not being parallel to the fork legs. You may remember that I wasn't completely happy with it when I machined it up. I'll make another one at some point.

Next week, I'll continue with the saddle.

In other news, we've had a little weather this week. It's been quite varied with snow on the Lindis Pass, very heavy rain on the West Coast that took away a few bridges, a tornado in Ashbuton and very hot Nor'West days like today. It's almost like we've had 4 seasons in one day or something. Hey, somebody should write a song about that.

Thursday 3 January 2013

Cantle Plate Cheating


For the first time on this project I have cheated, I have taken a short cut, I have not made something from scratch. I'm pretty certain that life will go on so I'm not going to loose any sleep over it. I could have made the cantle plate from scratch but as I had a suitable donor saddle handy, it seemed silly to be pedantic about it. Similarly I have found some springs that I can use for the saddle. The springs on this bike are much longer than usually found on Brooks saddles. I found them in a drawer in Scottish Dave's shed and he generously donated them to the cause.


The donor saddle is an old Brooks something or other and the spring mounting bolts are too wide at 4" centres, I need to modify the plate to chop out a section from the middle and rejoin it so that it still maintains the same radius, this requires a little thinking before getting the hacksaw out. This is in stark contrast to Mr. Middleton who is very fond of his welder. Mr. Middleton has been bragging that he has finished his bike before I have. Pffff.

First remove all the old paint and rust with a dunk in some acid.

Then make a little jig to hold the cantle plate while it is tweaked.

The drawing illustrates the section that needs to be removed 
and still maintain the correct radius and get the bolts at the correct centres.

The section duly removed.

Then mount on a suitable scrap piece of angle...

... and give it to the excellent young fellow Tweed Pete to weld up for me. 
The original rivet holes were too close after removing the section 
so I asked Pete to weld them closed at the same time.


Finally clean up top and bottom and drill the two new rivet holes.

I've also removed the old bolts as the threads were a bit cack and also a very odd size. I'll braze some replacement bolts in using my beloved 26 tpi cycle thread. In fact I'll try and get everything brazed up before I go back to work next week.


In other news, we have been away for Christmas, we went to Wanaka and spent a very happy week doing all the usual family holiday stuff. We took a couple of mountain bikes and explored the local tracks which were fabulous. I can highly recommend Deans Bank Track and the long loop along the north bank of the Clutha crossing at Luggate and coming back along the south bank. My mountain bike is almost an antique (naturally), it is a 1993 Cadex CFM2 and still going strong. It amuses me to ride 'gnarly trails' on a bike that I should have upgraded several decades ago. "You just rode that on that. Awesome dude"

And for the record, my lovely wife did buy me a Christmas present, she bought me a new T-shirt that I'm not allowed to wear in the shed. I bought her some earrings, thus maintaining marital harmony.