The used tool store had more than a dozen 1-3/4” sockets but not a single 1-7/8”. So, I bought a 1” socket to weld to my existing 1-7/8” (which I have in my BSW/BSF drawer masquerading as a 1-1/4” BSF) so it can serve as a deep-drive socket. I’ll enlarge the current 3/4” square drive hole to 1” dia. and then weld the 1” socket to it. I’ll also face the entrance end of the socket to remove the small taper to let it seat fully over the rather thin gearbox sprocket.
I also bought a 5/8” socket to be part of a special tool I’m making for Gold Stars, about which I’ll write in the GS forum in due course. But, thanks to having to spend time on other things, visiting the tool store was my sole Ariel accomplishment today.
I modified the socket needed to remove the drive sprocket, requiring all 200 Amps from my welder to do it. I should have switched to the He-Ar tank to put more heat into the weld but managed to do a good enough job with pure Ar. Not that such a large diameter is involved with the Ariel, but since the socket is for fasteners whose normal diameter is 1-1/4" I bored it to just over that diameter and welded a 1-1/4" socket to it in order to make a custom 1-1/4" BSF deep-drive socket.
With the new socket on an impact wrench the nut came off immediately. However, when I tried to slip the £65 replacement sprocket on, it wouldn't go. The composite photo shows why. This is the third aftermarket item I've purchased from Draganfly that was unfit for use. I've written to them[*] but I don't know toward what end, since the postage to return it would be excessive.
Four nuts underneath the platform and two bolts on the side hold the gearbox and once removed it was on the workbench. The photo shows the insides along with the replacement gearset underneath. However, there are enough gears, shafts, and linkages that I thought it best to read my 'gearbox' chapter before proceeding so I turned to the clutch basket.
The next photo shows the 'nut' that holds the clutch assembly together. As can be seen, it has "benefited" from precision mechanical work with hammers, chisels and punches. However, rather than continue that tradition I went back to the used tool store to buy a socket to machine into a sprag socket. I managed to get it about 3/4 done before stopping for the evening. It's not (only) that I like tools, but having all the required "special tools" along on the Cannonball could prove invaluable if I encounter difficulties on my hoped-for trouble-free ride. If the clutch basket has to come apart, the team mechanic will thank me for having it in the toolbox.
[*] They wrote back to say I now have an account with them which they've credited wit the £65.
Last edited by Magnetoman; 05/02/184:14 pm. Reason: [*]
Richard, there is a lot of interesting information in your two posts which I've tried to address below.
I hadn't seen any information on the AOMCC site about the availability of profiled spanners. I would have bought them had I been aware. But, given how tight the gearbox nut was I would have ended up making the deep-drive socket anyway in order to use an impact wrench to remove it.
I learned of the Czech club fairly recently and tried to get a Slovak colleague to translate their web pages for me. He was quite willing, but his lack of automotive vocabulary in either language was an insurmountable obstacle.
I won't be able to open my clutch hub until I finish the sprag socket later today, but it looks fine on the outside so, fingers crossed, I won't find a disaster inside. However, given that the inner hub doesn't turn smoothly I have my fears about what I might find with either the rollers or races.
Whoever made the sprockets for Draganfly simply did not know what they were doing. It must have been some amateur in a backyard shed with a broach and press who bid on the job thinking the spec. for the width of the teeth only applied to the tips.
The drive sprocket is going to cause real headaches. The Draganfly order was last October so that's when I was assembling components. Without going through old emails to reconstruct details, on the shelf is the Draganfly 19T, a 23T with correct-looking (but incorrect) splines, and an industrial 23T sprocket with a hub. Perhaps you suggested that latter sprocket at the time(?). Anyway, I suspect the other 23T came from the AOMCC and is the one I was going to use along with a spacer to substitute for the hub. However, as the attached photograph shows, the splines on it are incorrect as well. It doesn't fit on either the present or the replacement gearset.
Given the issues with pattern parts, all I can do is cross my fingers and hope the replacement 'QH' gearset is made properly. It's not unreasonable to hold out hope since detailed measurements on the dimensions and hardness of a replacement RRT2 mainshaft for my BSA Spitfire Scrambler showed it was excellent. The replacement Ariel gearset passed the only test I've been able to do on it so far -- the "original" sprocket from the Ariel fits.
Although my Ariel seems to have had very few miles on it since it was previously rebuilt there was a trail of grease down the case from the layshaft bush on the sprocket side and I found a fairly thick coat of it under the gearbox when I removed it. Given this, and your comments, machining a blind bush as a replacement seems like a very good idea.
Thanks very much for the information in your posts.
I finished machining the sprag socket for the clutch hub today. Based on the diameter I calculated that each of the three fingers needed to be 7-deg. wide, and that the 1/2" cutter I would use was 22.4-deg wide. Using these values I held the socket in a chuck on the rotary table and cut one face each at 0-deg., 120-deg. and 240-deg., then cut the other faces at 29-deg. less than these angles. The fit is perfect. Although the photograph makes it look a little loose that's because the outer portions, in particular, of each of the three slots have been hammered out of shape.
Holding the basket in the vise the 320 ft.lb. impact wrench slowly turned the nut off. I can't imagine how much I would have had to beat on it with a hammer and chisel to do it without the socket. As can be seen there are 24 1/4"x1/4" rollers inside. I removed the rollers, cleaned them and the races, and found no damage when I inspected them. I re-greased everything with Mobil 1 synthetic having a 550 oF drop point. This is the grease I use on my Gold Stars' wheel bearings since it has a drop point as recommended by Michael Morse of Vintage Brake. He found greases with a lower drop point showed signs of running when used under racing conditions.
After greasing the rollers I torqued the nut to 60 ft.lbs.(limited by my ability to hold the assembly, not by the torque wrench) then hit it for a few sec. with the impact driver. I re-punched two of the punch marks to lock the nut to the threads.
The rotation is now quite smooth. My fear was that I'd find Brinelling of the races, but the previous roughness must have been caused by a small piece of grit.
While I had the rotary table on the mill I took the time to mill a hex in the Gold Star special tool I mentioned earlier. I hope to find out before too long if it does the job I made it to do, which would save quite a bit of time on a certain repair. I'm not trying to be cryptic, but this is the Ariel forum and all will be explained in due course on the Gold Star forum (unless the tool doesn't work...).
Other things came up today that took precedence over the gearbox so the only progress was to open a package containing the 7/16" carbide burr that I no longer need for the pushrods. Oh well, I can use it on the next Ariel...
I agree. I have sets of them for steel and for Al (wider spacing between teeth to reduce clogging). Unfortunately, neither set had the 7/16" ball needed for the pushrods. I suspect using this burr in a high speed grinder would have left the surfaces ready to use as-is, or at least significantly reduced the time required with grinding paste. Now I won't know until the next 1928 Ariel I rebuild.
I was being facetious when I wrote "the next 1928 Ariel I rebuild," because the odds of stumbling across another one like it in the U.S. seem pretty small. However, I certainly hope this isn't the last bike I rebuild.
Turning to the gearbox, not wanting to take any chances on being able to reassemble it, I set up a camera on a tripod looking down at the area around the gearbox. A remote shutter (in a baggie to keep my greasy fingers from touching it) ensured I would have a large set of photographs to refer back to, if (when) needed.
Things went smoothly for a while, then I hit a roadblock when I had all the nuts removed but it still acted like there was something holding the cover on. Tapping the cover and both ends of the mainshaft with a Cu hammer didn't help. Whenever tempted to get a bigger hammer and 'whack' rather than 'tap' it's time to step back, so I re-read the "Gearbox" section of my manual and this time noticed the words "after which the Kickstarter Ratchet itself must be unscrewed with a tubular spanner." As the photograph shows, hiding at the base of the ratchet is a hex. I couldn't budge it even with a 3/4" breaker bar and chain wrench on the sprocket at the other end, but the impact wrench had it off in a few seconds.
Unfortunately, removing the ratchet didn't make any difference. After very closely inspecting everything, the only thing it could be was the kickstarter shaft, so I pressed it out. After that the cover came off without a problem.
Things get a little odd at this point. I wanted to document the number of teeth on each of the gears in my 'QL' Low Standard gearbox as well as on the aftermarket 'QH' High Standard gearset I bought from the AOMCC to replace it. All are the same. I've emailed the guy I bought the gears to confirm what is apparent, i.e. that someone already had replaced the gears with a QH gearset.[*] Sigh... Oh, well, it's only money.
A mystery to be investigated further tomorrow is the sleeve gear bearing is an SKF D207-2RS1 which the catalog says is metric, with a 72 mm OD and 35 mm ID. However, although I measure the OD to be 72 mm the ID is 34.84 mm, i.e. ~0.006" smaller than the catalog value. In Imperial units the ID is 1.3717", which is ~0.003" smaller than 1-3/8". Maybe more careful measurements of the ID will solve the mystery. Luckily, the bearing seems to be in perfect condition so I won't have to deal with replacing it.
[*] He responded this morning confirming my QL-stamped housing contains a QH gearset and offering to take the replacement back. My hope against hope was that he would say he also sells a QL and that from the tooth count he realized he had sent me one of those by mistake. But, QH it is.
Last edited by Magnetoman; 05/06/184:36 pm. Reason: confirmation of QH
Congratulations, MMan---over 50,000 views in about 9 months for this thread. Must be some kind of record...
Things picked up because the thread passed 100,000 views after only another 3 months.
I degreased the case and gears and then inspected everything closely. Everything is in good condition, although there are a few signs of previous repairs. As shown, the stud where the kickstarter spring mounts has been welded at some point.
I found that the other bearing in the gearbox also is metric. It is installed with the identifying information on the inside where I can't see it but has a 60 mm OD and 25 mm ID. The end of the shaft that goes in it also is 25 mm, of course, so almost certainly it was ground to fit the bearing at some later time in its life.
To check on the gearing ratios, as well as make sure I understand the power transmission routes, I made the attached composite. The values calculated from the gears correspond well with those in a Burman catalog for the QH "High Standard" gearing that I want. Interesting, there is disagreement in the third decimal point of the ratios for 1st and 2nd, no doubt due to long division errors made by an apprentice 90 years ago.
I had used a wide angle lens on a high resolution camera on a tripod so I wouldn't have to re-aim the camera each time. I spent an hour with Photoshop to crop only the areas of interest in 32 of the photographs I took as I disassembled the gearbox. I then pulled them into a document in reverse order, four photographs per page, to create a detailed illustrated assembly manual for Burman Model Q gearboxes.
I discovered the cause of the difficulty I had removing the outer housing. The kickstarter spindle has a slight step in it that doesn't allow it to pass through the steel piece in the outer housing that mounts the end of the kickstarter spring. This is why the spindle had to be pressed out of the main case from the back (i.e. from the left in the photograph) before the outer housing could be removed. All to be duly noted when I annotate my illustrated Burman assembly/disassembly manual.
I'm not rushing the assembly process since I don't want to overlook anything. So, to be continued...
Enjoying your chronicle (as always), Check the (bolted in) kickstarter stop for cracks (and the fit) where it passes through the outer housing. Some like to fit the later steel-backed rubber bump stop (3338-33) over the pin.
I had to make a new kickstarter spindle; wear similar to yours, possibly due to a loose cotter pin and the kickstarter moving on the shaft..
Draganfly have the Main gear bearing in stock (35-Q), although I purchased from the AOMCC. From memory a special bearing, ground a combination of imperial and metric sizes.
You have been so methodical; please don't take a chance on the old bearing. At the very least, carry a spare in your kit; if it fails you are unlikely to get one on route in enough time to regain the lead!
Golden rule; bearings are cheap, never re-use a bearing.
Draganfly have the Main gear bearing in stock (35-Q),
Richard, again thanks for your helpful comments which sent me back to the garage for additional measurements.
The Burman catalog shows 35-Q as the bearing at the kickstart end of the gearbox (which Burman calls the 'Mainshaft Bearing') and 21-Q ('Driving Gear Bearing') for the sleeve gear. Draganfly lists the 35-Q but not the 21-Q. Instead, for the 'Bearing, sleeve gear' they list part number 74X which when the link is clicked on is then called the 'Bearing, sprocket.'
Unless you know otherwise, I very much doubt if the gearbox had metric bearings when it left the factory, which means my cases have been modified. I will make one additional careful check, but it appears for the 35-Q that both the OD and ID of my bearing are metric as is that portion of the mainshaft (no doubt, turned down to make it fit the bearing) so I can replace that one with a standard metric bearing. The current bearing is unsealed, allowing a direct path for grease through it, down to the kickstart spindle, and onto my foot. For that reason alone replacing it with a bearing with a seal on one side seems wise.
The 21-Q seems to be a standard metric bearing as well. It happens to work with the nominal 1-3/8" OD (34.925 mm) sleeve gear thanks to 35 mm being such a close match to the Imperial dimension, plus an SKF chart shows bearings of this size have bores undersize by 0.012 mm, i.e. the bore should be 34.9888 mm = 1.3775". My sleeve gear has an OD of 1.3770" and that from the new QH set 1.3778". Consistent with the SKF chart my gear slips into the bearing but the new one would have to be pressed.
So, after I re-reconfirm the dimension of the 35-Q my gearbox will be treated to a pair of new metric bearings. Since they both seem to be standard bearings this should only delay progress on the gearbox by a couple of days.
Unless you know otherwise, I very much doubt if the gearbox had metric bearings when it left the factory
Burman for some reason best known to themselves used mixed dimension bearings (imp/mm) for the high gear bearing in most of their gearboxes ???? Off hand I cannot remember if the bearing on the other end of the Burman mainshaft is metric or imperial BSA and most other Brit manufacturers used a lot of metric bearings, (and mixed them ) BSA for example used 6207high gear and LJ3/4 on either end of the gear box shafts, Your GS's have 25mm and 1 1/8in bearings on the same drive side mainshaft !!
Burman for some reason best known to themselves used mixed dimension bearings (imp/mm) for the high gear bearing in most of their gearboxes ????
I know that's the case for modern bikes like BSAs, but is that true even for gearboxes like mine that date to the Napoleonic wars? Evidence that my gearbox was at least partially metrified later is that, compared with the new gearset, the inboard end of my layshaft is ~0.015" smaller to make it 15.5 mm rather than 5/8". The outboard end of both layshafts is 5/8".
It took me over an hour on the web to find what I needed but I finally have new bearings on the way. Apparently the 12 mm thickness of the outboard 25x62x12 mm bearing doesn't leave enough room for a shield (ones that are 17 mm do) so I'll have to devise some sort of oil/grease baffle of my own to fit between the bearing and the ratchet gear.
The outboard end of the layshaft easily slides into a bush in the main housing, which means there's plenty of clearance for the oil/grease mixture to ooze out. So, I'll be making a closed-end replacement bush. The layshaft is pinned at the other end so it doesn't rotate so I'll probably make the bush from brass rather than from some exotic bronze alloy. I have a set of metric reamers but only in 1 mm increments, and reaming all the way to the bottom of a blind hole is problematic anyway, so I'll be turning the ID on the lathe rather than reaming it. This will give me something to do while waiting for the bearings to be delivered.
but is that true even for gearboxes like mine that date to the Napoleonic wars?
A couple of years ago I rebuilt a Burman box from an OK Supreme 250 Flying Cloud The original sprocket bearing was a mixed dimension item, not available off the shelf I machined up a sleeve and fitted a standard double sealed bearing My experience with 50's Burmans is that they have imperial / metric dimensions on the sprocket bearing
Quite a lot of modern gearboxes have double sealed bearings fitted, it I believe keeps metallic particles out of the rolling elements??? Lubriant can migrate into the bearings past the seals on heat cycling as the seal lip(s) are fitted to keep the lubricant inside from coming out
The output (sprocket) gear bearing is a NTN 2607LLU (T11408).
I think you transposed two of the digits. I have on order an NTN 6207Z. It has metal shields which will be more difficult to remove than rubber, but it's what I was able to find. It is supposed to be delivered tomorrow.
Originally Posted by Richard Kal
The main shaft is quite long. Mine was bent (clutch end).
I haven't put mine on the bench centers yet but it's on the to-do list.
Originally Posted by chaterlea25
A couple of years ago I rebuilt a Burman box from an OK Supreme 250 Flying Cloud The original sprocket bearing was a mixed dimension item, not available off the shelf
Very interesting. Luckily, I was able to get my bearing situation sorted out with relatively little headache.
While awaiting tomorrow's arrival of the bearings, I spent today on issues of oil.[*] The layshaft is well below sea level and is held by through holes in the case and cover. I don't know whether or not it was originally this way, but the hole in the case is filled with an open-ended steel bush with a ~0.002" clearance that doubles as an escape route for the oil. The hole in the cover has no bush and it probably used to be a tight press fit, but now it is more of a slip fit, also allowing oil to ooze.
I replaced the bush with a closed-ended one I machined from brass. Since the layshaft doesn't turn I could have used Al to somewhat disguise it, but I like the look of brass. That addresses one of the weak points in the Maginot Line of marauding oil. The other end of the layshaft I sealed, I hope, with Permatex when I installed it in the cover. That leaves the two bearings as the remaining possible sources of leaks.
The sleeve bearing has a seal on one side of it that should block that path, but the bearing at the other end of the main shaft is too thin to allow a seal. However, as the second photograph shows, the 1.45" OD of the inner race is larger than the 1.30" ID of the hole in the case, largely blocking that route other than the thin space between the race and the case. Unfortunately, if oil does make it to the inner race the fact the race is spinning will "pump" that oil to the outer cavity. The gear at the left in the third photograph is the one in question, with a raised portion that sits against the bearing race but whose OD is smaller than the ID of the hole in the housing that is shown in the second photograph.
As can be seen from the third photograph the gear allows easy access to the bearing for any oil that wants to escape. So, to make that escape path less enticing I machined a spacer from teflon to fill in the gaps between the backside of the gear and the face of the case as well as between the OD of the raised portion of the gear and the ID of the case. This teflon spacer forces the oil to make two 90-deg turns before reaching the bearing race.
These anti-oil measures seem like they should be effective but, like the original Maginot Line, road tests will be needed to determine whether or not they actually are.
[*] As Richard Kal points out in the following post, Burman specified use of a "semi-fluid" grease rather than oil. Their attitude toward lubrication was that if the gearbox leaked it meant it had been overfilled and once the grease "found its normal level, no further escape will occur." In other words, if your gearbox leaks its your fault because you overfilled it. Anyway, I will be using Morris K400EP which that company specifically recommends for these gearboxes. However, people have noted that this grease still is solid enough that the gears cut channels in it, reducing its ability to reach the spinning surfaces to lubricate them. For this reason it seems common to recommend adding "an egg cup" of gear oil as well (in non-Imperial units I converted this to 50 cc). This is a case where the recommendation may be valid or not, but it will help the gearbox if it is and won't do any harm if it isn't. For this I have Miller's Classic Green Mineral Oil 90 GLI. It's like a modern gear oil except without the Extreme Pressure (EP) additives.
Last edited by Magnetoman; 05/09/185:03 pm. Reason: additional text explaining the teflon spacer and the lubrication
Hi MM, The gearbox lubrication spec is for grease. The old - timers seemed to accept the box oozing grease. Grease from around the roller bearing on the kick starter end of the main shaft would have oozed through into the kick starter quadrant chamber. There was a tip from a contemporary motorcycle rag explaining that the amount of grease leaking out from behind the tin cover over the quadrant could be reduced by inserting a piece of felt between the cover and the kickstarter shaft (hole cut in the middle for shaft to pass through). Think I saw that in Dave Berkshire's "Black Ariels" book (in the last pages).
I wrote 'oil' for simplicity, but I'll be filling the gearbox 1/3 full with Morris K400EP along with 50cc of 90W gear oil rated that's rated API GL1. That's the mixture whose escape the various seals and baffles will try to slow.
Originally Posted by Richard Kal
...in Dave Berkshire's "Black Ariels" book (in the last pages).
His book was one of the sources I used when compiling my workshop manual. Starting from having no information at all about Black Ariels when I got the bike a year ago, every relevant tidbit I could find anywhere ended up in the manual. It's now up to ~600-pages and in two 3-ring binders.
The composite image below shows how this 3-speed gearbox works. There are three pairs of gears that are in constant mesh and changing the gears involves sliding the middle pair one way or the other to engage or disengage dogs. The red arrow at the right of the top image shows that a saucer-shaped projection on the rather wide and complex middle mainshaft gear engages a slot on its mating wide and complex layshaft gear to force it to do whatever the mainshaft gear does. The arrow at the left shows the slot in the mainshaft gear into which the gearchange mechanism fits to do the forcing.
The portion in yellow at the bottom is the linkage to the outside world. If the lever on the tank is, say, pulled back, toward the rider, a rod running from it pushes the yellow linkage back. This rotates the shaft and forces the yellow lever on the inside the case down. The diagram doesn't show the third dimension but that inside lever, denoted with a red circle, comes out of the screen.
When the yellow lever inside the case moves down it pushes one arm of the orange mechanism down with it, which causes the other arm of the orange mechanism to move to the left. A shaft attached to the mechanism, denoted with a red circle, that goes into the screen fits inside the slot on the mainshaft gear as described in the first paragraph. This moves the mainshaft gear to the left which in turn forces the layshaft gear to move to the left with it.
The spring-loaded green arm has a projection that fits into detents in the orange mechanism to lock the gearbox into the selected gear (it's shown in 2nd, with 3rd the final detent to the right).
The slots at the left of the complex center layshaft gear can be seen to be engaged with a set of dogs under it. Not apparent is there is an internal cavity in that gear that disengages from those dogs when it is pushed further to the left, into 3rd gear, allowing it to spin free in that position. In 3rd none of the layshaft gears do anything other than spin and power is transmitted directly from the mainshaft to the output gear via the dogs on it.
I didn't have much time today, but the gearbox is now back together, with the exception of the shift mechanism. That should be fairly easy to assemble as long as I don't drop any of the pieces and have to fish them out from the bottom of the case.
seemingly has no way of unlinking them, and partly like building a ship in a bottle. The 'C'-shaped selector piece has a ball at one end that wants to fall out of the slot, and a rectangular pad at the other end that wants to fall off the shaft. Further, the arm that connects to the ball is too long and too fat to connect to the ball in a straightforward way. Despite all this, everything does go together with enough patience along with a magnet to fish pieces from the bottom of the case. Note that grease is helpful for holding the ball and pad in place, but for the same sticky reason it's a hindrance when those items fall to the bottom of the case despite the grease and you need to get them to move to a place where you can reach them with the magnet.
The Morris semi-fluid grease isn't quite a liquid, but despite how it looks in the photograph it definitely is more fluid-like than the grease that was in the bike when I got it. Ariel's instructions call for filling the case to one-third but it's nice to have actual volumes to work with so I measured it. For the record "one-third," judged as best I can by looking down into the case, turns out to be approximately 750 cm3. To this I also added 50 cc of the 90W gear oil mentioned in a post a day or two ago.
The gearbox is now completely back together without so much as a washer left over, and I have it loosely bolted in the bike. I had thought to reassemble the clutch and try to take it for a test ride this weekend before rebuilding the magneto, but the gearbox sprocket is in the way. I have the 19T sprocket that came with the bike, but I intend to use a 22T sprocket on the Cannonball. Although I have a 22T aftermarket sprocket, as mentioned in an earlier post the splines on it are miscut so it would have to be modified in order to work. To do that requires having access to the gearbox's sleeve gear in order to test the modification so I can't assemble the bike for a test ride. Unfortunately, that entire sprocket was hardened to 65 RC so it wouldn't be easy to machine.
I have alternatives to work with, such as a 19T aftermarket sprocket that doesn't fit either. However, it's a much more reasonable 25 RC. If I were to use it I would machine the teeth off and replace them with a ring of 22T that I would cut from an industrial sprocket. I have to figure out how I'm going to proceed with the sprockets because I'm stuck until I do.
Originally Posted by Richard Kal
There was a tip from a contemporary motorcycle rag explaining that the amount of grease leaking out from behind the tin cover over the quadrant could be reduced by inserting a piece of felt between the cover and the kickstarter shaft (hole cut in the middle for shaft to pass through).
That suggestion is from a 1928 'The Motor Cycle' article that says "The only grumble I have with the box is that grease is very prone to leak out of the kick-starter housing; however, this leakage can be reduced to a minimum by the insertion of a felt washer, obtainable from the makers, between the kick-starter pedal and the sprocket." However, as the photograph below shows, even if the clutch actuating housing and tin cover were completely sealed, by the time the grease rose to the level where a felt washer would do any good it would be like trying to slow water spilling over the last bulkhead in the Titanic.
I didn't cut a washer from felt to put there but I can always change my mind and have one installed in a few minutes. I'm placing my hopes in the teflon spacer/baffle I showed in a previous post.
Last edited by Magnetoman; 05/11/186:24 am. Reason: discussion of felt washer