The only shower in Denmark fell on me as I rode to the shops, but I pressed on blindly. The shortened, ratio-changing lever seemed to be doing quite well but not as well as yesterday. Then it occurred to me that I had not allowed for the diameter of the cable bolts nor their shaped clamping plates. By shortening the arm I had made it much more sensitive to measurement errors. So the ratio change was incorrect and probably much too large When the chain was on the 3-4th largest sprockets it occasionally jumped back and forth as if trying to move the chain onto the next larger sprocket. Otherwise the majority of the indexed gears seemed fine. I searched a large DIY store for suitable materials for making a new arm and pulleys without success. I think the reachable charity shops would serve me better but some aren't open every day. 23 miles.
Thursday 10th 45F, 7C, cloudy with roaring gales. Gusting to over 35mph on a 20mph base. Nearly 8.30am and still quite dark. The sky looks like ground rice mixed with blackcurrant jam. It is supposed to rain at 3pm so any ride had better start before lunch. In fact it started raining earlier than promised and the gales continued under a leaden sky. With no real pressure to go out I took yet another rest day working on the trike.
The reason for yesterday's error in gear indexing is simple enough to calculate: It requires a pull ratio change of 1:1.38 for the Campag Ergo levers to accurately control the Shimano XT11 MTB rear derailleur. 1:1:38 suggests 68 and 94mm from the pivot to avoid the clamp bolts hitting the chainstay as much as possible. These distances are what I measured and center punched for accuracy in drilling. However, these were the centers of the clamping bolts. Not where the cable entered and left the arm when wrapped around the bolts. Let's assume an error of 3mm [only 1/8"] increase in the cable entry and exit radii. This changes my original figures to 65 and 97. 97/65 = 1.49 instead of the desired 1.38. It is no wonder the changer sometimes overshot at the "fat end" of the cassette!
Let's assume I leave the top clamping bolt where it is: Measuring down from the pivot center: 65 x 1.38 = 89.7 or ~90mm. If I left the bottom bolt where it is: 97/1.38 = 70 So I can raise the bottom bolt or lower the top bolt to compensate for the cable wrap, radius error. The important measurement is where the cable enters and leaves the cable clamps. Not the nominal bolt centers. I shall have to measure the radius error next to see how far I need to move one, or both, of the clamping bolts.
The actual measurements to the entry and exit cable points were 66mm and 98mm to the pivot. I decided to move the top cable clamp bolt down to 73mm to compensate. But then I had to hang the lever from the other side of the pivot clamp for the cable clamping bolts to safely clear the chainstay. The lever pivot clamp is now as high as it will go against the underside of the front gear changer. The image shows the fully forward position of the arm in bottom gear on the 36t sprocket. The arm is now far too long but handy for tensioning the cable during initial experiments. It is probably safe to saw it off now unless I can find a "prettier" replacement.
The gear indexing is now perfect with the trike up on the workstand. The Campag Ergo lever is still proving slightly sticky on upward changes when the cable is let out. So I removed the lever and its rubber hood and examined the innards with a bright LED torch and my strongest [supermarket bought] reading glasses. There was no sign of any stray strands from the frayed cable which I had removed before beginning the project. All I could do was smear the parts and the cable run with grease and refit the lever and its hood. After watching a Campagnolo video of dismantling an Ergo lever I decided to avoid meddling.
The large pulley being used for a downtube cable guide is not ideal and I would prefer a pipe or tunnel style of guide for much less visual impact and a much better cable run. It would be better if the cable ran in a more normal position on the downtube then curved gently around the seat tube to point straight at the top clamp of the radius arm.
I am pondering how to add a suitable pipe to a clamp without it all getting very ugly indeed. I suppose a pipe could be squashed beside the clamp's tension screw. Searching for cable guides doesn't produce anything commercially available. Only old-fashioned 1960s/70s Campag guides seem to pop up on eBay and 34.9mm was unheard of back then. If I used an under-BB guide the angle of the cable with the top clamp would be extreme and alter the ratio between start and finish of the arm's movement. Completely spoiling any hope of accurate indexing.
As the image above shows a brake noodle fitted perfectly behind the screw of the clamp and was snugged up perfectly. Having removed the unwanted end pieces I even slipped the low friction liner hose back in in to help things along. The "trumpet" shape on the end will ensure it stays there. The noodle pipe was bent in place which resulted in some minor kinks but these seem not to affect cable friction. I can make a much better job with a new noodle when I find a prettier clamp. A smarter 35mm [34.9] hinged clamp would certainly help the appearance. Which I have duly found and ordered online: A plain, hinged clamp with a neat bulge for a pulley center bolt.
[I may still make two small pulleys for the arm to avoid the cable damage from being wrapped tightly around the bolts. If I did use pulleys I must allow for the radius of each pulley to find the correct distances from the pivot. The pulley rims are the active radii measured from the lever's pivot. Larger pulleys would overlap if they maintained the correct 1:1.38 radii.] We're back to swings and roundabouts again!