After lots of input from the Ariel community, I now understand how the kickstart spring cover is supposed to work.
There are four parts in play, illustrated by this page from the parts book:
LS254, the kickstart lever. I think I have the right one
LS262, the kickstart bush. This comes all the way out of the end cover up to the end of the shaft, LS253. The bush is 1 3/8" OD
LS263, the distance piece. Andrew H of the AOMCC tells me that this is 1 1/2" OD, so the wall thickness must be something like 1/16".
LS264, the spring cover
The big reveal is that it is the distance sleeve that keeps the spring cover in place, and that it is itself kept in place by the kickstart lever. This means the spring cover, LS264, must fit tightly enough around the bush LS262 to be retained by the 1/16" thickness of the distance piece LS263.
The other thing I've seen is that the top of the spring cover is not flat, but follows a nice radius. The former has been back to the lathe for reshaping:
I've prepared another bit of sheet, by cutting it to size, squaring the ends and removing the paint:
Before we start on that, I found a piece of exhaust pipe the perfect size for the distance piece. I made a wooden former so I could cut it to length in the lathe and form a short radius on the end.
The radius prevents it sliding towards the gearbox quite successfully and it fits rather well behind the kickstart lever.
I formed the spring cover in the same way that I did last time but there was no way it would shrink far enough to get to the required inside diameter for the kickstart bush, so I ended up cutting darts in it and welding them up
There is a bit more work to do, but all the bits fit together. I think the spring cover is too long but I always knew I would have to trim it and I need to finish the welding. Lastly there will be a hole for the spring to pass through.
Before we get into that though - this is my latest invention for welding - an arm rest! Makes controlling the torch so much easier:
The arm rest made finishing those tiny welds much less stressful. The welds were backed with a curved copper strip which as I have shown before minimises the dressing required on the back side and gives you much more control when welding this thin material. It's been welded, ground out inside, straightened, ground outside and so on...
Next job is to drill a 1/4" hole near the edge, and file it oval:
Then I've used a drill to form the material such that the spring can pass through:
The spring is wound in from above:
In the video, the spring cover has been trimmed to length and fitted:
It's not perfect but it's nearly there. I shall raise the edge of the cover a little more since the whole assembly has to account for end float in the kickstart shaft, and there is a bit more clearance than I would like.
It's a beautiful day in Norfolk and after doing a lot of routine maintenance it's time for a quick run out to warm up the oil prior to changing it. We make a 20 mile round trip to Cromer to do some shopping.
On the way back we stop at home to drop off the contents of the panniers and find something interesting - leaving the bike running on the field stand reveals that it smokes on the field stand - and drips oil from the decompressor.
The reason is obvious when you think about it - the rocker box is drained down the pushrod tubes and on the field stand the whole rocker box is higher on the push rod side and will fill up with more oil than it would when standing vertically. Consequently oil goes down the valve guides and out of the decompressor...
The picture above is the giveaway to this theory - here the bike has been standing outside Morrisons for 10 minutes on the field stand but not running. There's barely a drop of oil on the ground and what there is is on the timing side. It comes from the magneto to chaincase connection I think.
Still, the decompressor shouldn't leak so we'll have to have a look at that yet again.
I wondered whether the main nut was square to the rocker box and was successfully making a seal there but it appeared that the decompressor body thread was loose enough in the rocker box to accommodate any out-of-flatness. What wasn't so good though was the thread - it's 5/8" BSW but it has an undercut down to something like 9/16", so the 3/8" BSP Dowty washer I was using is way too big. Since I don't have a 9/16" sealing washer, I used a 13 x 1.5 mm O ring and a 5/8" fibre washer - we'll see how that goes whilst we order the correct parts!
This is a job I've been intending to get around to for years.
These Chinese mini lathes have a saddle to bed clearance set by two M5 screws providing a clearance between the saddle itself and it's retaining plate. These screws are more or less on the same centre line as the bolts retaining the plates and thus the plates are never parallel to the bed and the contact between the plate and the bed is more of a line than a flat area - something you can see in one of the pictures below.
This leads to poor control of the saddle clearance to the bed which can spoil finish or make your saddle stiff to move. The fix varies - since you don't adjust this very often, I think the easiest way is to shim the plates but various people have shown other options such as a sliding wedge operated by a screw which has the advantage that you can adjust the clearance from outside without having to strip the lathe down.
As I said we don't do this very often so we will go with shims. The first job is to measure the thickness of the saddle accurately and this is something around 24 mm - actually 24.08 mm.
The next job is to measure the distance between the bottom of the bed and the top of the saddle which will be similar to the first measurement plus a few hundredths of a millimetre (This lathe is Chinese, so we will be talking in metric for this post!)
The difference between these measurements is 24.33 - 24.08 = 0.25 mm.
I don't have any shim stock but I do have various bits of sheet including a chunk of oil can which is 0.025 mm thick - easily cut with the Wiss snips. Making holes however is extremely difficult - I drilled these between a piece of wood and the saddle plate but it really needs a punch.
It turned out that I had already done the apron side but with a 0.15 mm shim - but I had left the adjusting screws in for some reason. I took them out and replaced the plate with the same shim and found that the saddle wouldn't move at all, so I put a 0.25 mm shim in and reassembled the lathe.
The saddle seems to move smoothly but not have any up and down movement - but it is very hard to tell. Perhaps I'll put a DTI on it and we'll see if we can measure movement.
When you have a few old bikes, and you use them, and especially if other people use them, you must expect a few diversions from your ongoing project whilst you fix the other bikes in the fleet. The Square Four has absorbed a bit of time recently and it's now become the daily rider again while the hunt must have takes a little holiday in Norwich.
Therefore, the Model A has taken a bit of a back seat but it's time finally for the next job. This is the tin cover for a kickstart spring.
To start the job off, we can measure the machined register that it fits on the kickstart cover and then we can work out how long it needs to be by measuring the spring.
I'm going to fold the cover up from a strip of 0.9 mm thick steel around a wooden former which I will turn on the lathe from a chunk of sapele, part of an old door.
I've screwed and glued this former to a suitable piece of timber such that I can hold it in the vise
Here's the first try, cut from a bit of sheet with the aviation snips to a suitable width and rough length. With it held in place on the former some tie wraps, I can mark the end and trim.
We'll do a little trial fit on the job:
We need to bear in mind that this cover is supposed to turn with the spring. The spring fits nicely inside it but I suspect it's rather tight on the gearbox end cover.
The only way to find out to tack it together. A couple of tacks show that it fits nicely will turn on the register on the gearbox end cover, so I fully weld it and dress it on the linisher.
The next step is to work out how long it needs to be. I believe that I will need to fold it over such that the inner diameter just fits the kickstart shaft bush.
A little session with the Wiss snips sees it cut to approximately the right length.
Attempting to spin the end over on the lathe was an abject failure - the wooden former just flew out of the chuck so I returned to doing it with a hammer.
That worked reasonably well, and it fits:
There's something fundamentally wrong though:
With a lot of help from the AOMCC, I've learned that LS263 holds LS264 in place and is itself held by the kickstart lever.
Amidst a couple of days of routine maintenance, greasing, oiling, checking stuff, chain adjustment, cables etc. I finished a couple of outstanding jobs on the dynamo.
One was the missing felt pad from the commutator end bearing;
The other was fitting the missing cable boot.
It's starting first time, though there seems to be a lot of piston slap and it's quite rich.
There's also a lot of oil around. On the drive side it's clear the primary case is leaking; on the timing side we have oil coming from the back of the magneto chain case. I suspect it's not pulled up tightly enough...
For the record, let it be known that I have today fitted a second battery to the system on the Square Four. This returns it to the layout that we ran for the first 7 years or so of its time in service. We now have 19 amp hours of capacity extended from the 12 amp hours we've been running for the last year.
We are back to running these VRLA batteries again:
A few miles run later in the day showed one-kick starting, there is no fuel leak and the idle is fine. The ammeter showed a charge; the battery monitor also showed a charge and did not go out of the green zone. Since most of the run was quite slow, perhaps I should not be surprised that the battery charger spent a little time returning the batteries to full charge at the end of the day?
Of course, a broken motorcycle can’t be left for too long, especially when your daily rider has been persuaded from your grasp.
There were a few problems to investigate, possibly associated:
There was a carburetter leak, or possibly flooding
The battery indicator was showing far too much red, considering the 60 miles the bike traveled yesterday
The bike refused to start twice yesterday, needing pushing to the top of a hill and roll-starting
Putting the CTEK on the battery had it charging normally, but it was very flat:
But why so flat? According to the ammeter, the battery does charge and indeed, it was charging on the way home. I did notice the ignition had been left on as Tom had complained the engine wouldn’t idle and it had died when he arrived. I currently believe this has more to do with the battery than anything else.
There is no leakage from the battery with the engine not running and you can measure normal currents with lights on, though with the LEDs fitted to this bike these are minimal. The ignition circuit draws 400 mA without the engine running - I think it would be a good idea to try this with the engine running.
I've also realised that the red spot ammeter doesn't show very much deflection at all - I calibrated the one on the Huntmaster against my bench power supply and there is a lot more movement. I wonder if the Square Four has a 30-0-30 ammeter.
A simple test with a 35 watt bulb revealed that the ammeter appears to be 15-0-15 - while I was doing this test the multimeter was reading 5 amps.
As a first step towards sorting the fuel leak, I removed the drain plug/main jet holder and finding no new fibre washer of the right size, sealed it with a wipe of Threebond 1215. That fixed the leak, and there is no evidence of flooding.
Amelia started first time like she always does.
As to the idle, I don’t know yet as I haven’t got it warm enough to close the bi-starter.
So what's next? With the kickstart done and the gearbox back together, my plan is to reassemble the primary drive to get the bits off the bench, and move on to the front forks - a big unknown. They didn't have bushes in 1930, so we are looking at a machining job at home or away. To get there though, I have rebuilt the clutch and next is the primary chain and covers, which just need paint. I'd like to reassemble the dynamo and it's drive, but it's toast. I could either:
leave the dynamo in a box for a few months and get on with other things
reassemble it mechanically (the body needs replating)
Fully restore the dynamo
I've got no nickel plating chemicals, and I have a long list of FH bits that need plating which I don't really want to get into right now (and as of this afternoon, my son has taken the Huntmaster and left me with the Square Four which won't idle...); I don't want to leave the dynamo bits in a box for months, so we will go for option 2. I've also seen at least one picture of an E3 dynamo with a painted body - and I've seen another picture of a 1929 bike which made me realise that you can't really see the body when it's mounted on the bike!
Here's a photographic survey of the dynamo:
In the picture above you can clearly see the damage done to the armature - it's thoroughly bent and unrepairable. Notice that this early E3 has the magneto type open ball bearings and the taper fit for the driving sprocket along with the male thread for the retaining nut. Characteristically, the armature is 157 mm long. Putting all this together enables us to purchase a new armature for this 36 watt dynamo.
I've ordered this repair kit from Rex's Speed Shop which will include the proper bearings:
As I may have mentioned, this is just going to be a photographic survey for future use. Here are some views of the dynamo body and the commutator end casting:
Close up, we can see that the other dynamo and casting has been a victim to the side swipe that this bike has suffered - there is a crack in this end as well. It looks repairable though:
