Mk3 starter sprag failures

I've heard of using the kick start to ease the motor past compression, before hitting the button, so the starter has a chance to build some momentum (or inertia?), to reduce strain. Might help?

Is ignition timing advancing and causing kick back due to the voltage drop caused by the starter? If you have replaced the mechanical parts with no success, it sounds like the fault is elsewhere. Modern electronic systems allegedly do not advance the ignition if the voltage is low, but earlier Boyer did.

When checking the drive gear, sprag, engine sprocket assembly, assemble the components in your hand without the large washers. Then looking through the engine sprocket puller holes from the engine side of the sprocket you can see the dogbones in the sprag and see what angle they are at. You may have to move the sprag slightly in the engine sprocket to get one of the dogbones in view. If they are almost upright (as following a radial line from the center of the sprocket) it is in danger of flipping, going over and crunching. If they are are over at an angle, say 30 degrees-ish from the vertical  radial line, it should be ok. 

I used to sleeve the drive gear to increase its diameter 0.007" above the standard size. I now check the angle of the dogbones and just sleeve the drive gear to suit each individual set of sprocket,sprag,drive gear. It doesn't matter if its a 14  or 18 count sprag if the angle is wrong they will fail. My own Mk3 is still on its 14 count sprag. 

The surface finish of the drive gear can be too shiny as well causing failure to engage. A fault i found in a Kawasaki Z1300 one time!

A horrible clattering noise that is sometimes attributed to the sprag failing can be the backfire device. especially if it has been incorrectly set.  

A belt drive in a Mk3 is a big No-No. The sprag needs some lubrication and a belt drive run dry risks you seizing the sprocket, sprag, drive gear. If this happens you will destroy your starter motor.

I can understand belt drives being fitted to the Dommies and the like with their tin chaincases that are very hard to keep oil tight, but Commandos have one of the best chaincases to keep the oil in. Hearing (and seeing) the tales of broken drive belts that happen and riders having to carry spare belts just in case seems a little silly

At the Austrian rally in 2018 there was a Commando with its primary drive in bits in the middle of the rally field with what was left of its drive belt for everyone to see, presumably the owner was trying to source a new one from somewhere. 

Describe what noises it makes, fails to engage but makes a clatter? doesn't engage but just spins? seizes up and won't disengage?  'still throws sprag' means what exactly? sprag going over center? dogbones broken? what kind of damage is left after a 'sprag throw'

Photos would be good to help diagnose whats wrong.

What kind of starter, battery, cables, solenoid are you using? is the charging system delivering the goods? More info needed. 

Peter as always is right. Unless you have removed the sprag and found that it is falling to pieces it may not be the sprag at all. The fact that there is less of a problem having fitted the tri spark ignition may be that there is insufficient battery power getting to the starter. As big a battery as you can squeeze in together with greatly improved positive and negative leads to earth and the starter will probably solve your problem. If not check the setting of the backfire device although getting this correct can be tricky.

I am putting a 3s delay relay into activating the ignition, so that the flywheels are spinning quickly before the first spark.

Looks ok to me. What was the sequence of events when a 'sprag throw' occurs? Pressed the green button and the starter system makes a horrible clattering sound? press it again and the bike starts? Or did you go straight for the kickstart after it makes the horrible clatter. Its clear that you have still been able to start the bike so I suspect that your backfire device has been incorrectly set and that is the source of the 'sprag throw' noise.   

In reply to Steve Marshalls idea of a delay relay, all that will do is drain your battery for three seconds.  

Is the starter motor turning the right way? When you assemble the drive gear into the sprag do you check that it engages properly? before you fit anything else.

Have you had one or more of the dogbones drop out of the sprag when you have removed it? If so make sure they go back fitted the right way. Even if the offending dog bones are removed it will still engage.

What starter motor is fitted? thicker cables fitted? Battery used? More info needed. 

That looks fine to me also.When you put the outer gear in can you turn it by hand ? It should bite immediately when turned anti clockwise. Presumably you bought all three parts from a reputable source ? 

If the sprag teeth go over centre it can lock up rendering the whole thing as scrap. 

As regards electricity, it helps to replace the standard solonoid with something a bit beefier. A BL MINI one fits in ok.

When I think back of the kick-backs I experienced on kickstart Commando's, they were mostly due to a somewhat unconcentrated weakish push on the kickstarter. A "good kick" usually does the trick. Me thinks the secret there is that a good kick gives the crank the required rotational speed,  momentum (allthough "kick" starting a well sorted Commando by hand (!) has been demonstrated to be possible..) and carburation for starting.   If one applies this phenomenon to e-start models, could the delayed ignition Steve Marshall installed make sense? Ensuring the crank has achieved a minimum even speed before ignition is switched on, preventing kick-backs? The starter has to work the hardest getting the crank in motion at all and past compression (highest amperage and voltage drop), once the crank is spinning and has some flywheel effect, it becomes easier for the starter (much less amperage). The delay is actually a feature found in some after market ignitions, like for high combustion engines and for Harleys.  Steve, 3 seconds looks a bit long, how did you establish?   

If the starter, battery, cables, solonoid, ignition and charging system etc is all in good order, the engine will readily start. To fit a 'ignition delay unit' is introducing another component that could go wrong. These bikes don't need one at all, My SU equipped Mk3 will be running within one and half seconds. An Amal equipped bike most likely even quicker as you can flood the engine with the tickler. 

Say you did fit a 3 second delay, a good starter motor will turn the engine over at approx 1000 rpm, that will mean the engine has turned over approx 50 times for no reason, just to drain the battery. You are then giving your charging system more work to do to top up the battery for the next 'drain before start' episode.

Dare i say it 'A non starter'

Peter, the starter achieving a cranking speed of approx. 1000 rpm looks a bit optimistic to me. With the starter gear reduction on a Mk3 being 1:14,2, that would mean 14.200 rpm for the starter motor.

Suicide. Even at only 1,5 seconds.

My guess is something around 2000 rpm (nominal, depends on a the load, etc.) for the starter, or 140 revs for the crank.

Anyway, the length of ignition delay I could think of would be in the order of the equivalent of a couple of revolutions of the crankshaft, so no longer than needed to ascertain that the crank is actually spinning at some minimum speed. 

There are programmable ignition units  with advance curve just for starting, the default setting is once over 200rpm it changes to running advance curve. You can change the 200rpm but it shows the expected rpm under a starter alone. This confirms Bennie's valve are in the right ball park.

I revise my statement that a good starter motor would turn the engine over at a 1000rpm.

Its more like 500 to 600 rpm having just tried it with the kill switch off. Starter motor, Two pole prestolite freshly overhauled with rolling element bearings.

Doesn't alter the fact that these bikes don't need any ignition delay  

Peter, I have just begun replacing my standard Prestolite motor bushes with needle roller bearings.  I have opening it the drive-end cap to 7/8” with an adjustable reamer, which worked okay, if a bit laborious.  However, having achieved a suitable ID for the B118 bearing, it is apparent that the land the armature butts against is now pretty narrow.  With the spring loading of the armature from the dished washer at the commutator end, the drive end cap land will likely wear pretty quickly.

I am considering two options :

1. Add a thin steel cap to protect the land.

2. Replace the dished washer with a spacer or shims to avoid the load it creates, as I suspect the dished washer was just used for production expediency?

I guess you have previously recognised this consequence of opening up the diameter to fit the needle roller?  With hindsight, because the bearing does not need the full available bore length, perhaps I could have got a 7/8” reamer and opened up the diameter from the oil seal end without going all the way through, leaving the original diameter at the armature end. I do have a spare drive-end cap, so it is not a crisis - just an unexpected headache.

Andy

Andy

Don't worry you can still use the drive end cap that you have modified but use a 6203 2RS bearing instead. This does require the use of a lathe though.

I've sent you a PM

regards

Peter

Andy

The photos are of the 6203 fitted to the prestolite drive end.

The spacer ring shown in one photo is if you are using a four pole harley frame (either original harley prestolite or 5780BN frame, they are the same thing) and fitting it to the norton prestolite drive end plate. 

regards

Peter

The design from the start is seriously flawed and Norton left no margin to correct, had they done so, it would be a cheap fix. Original engine sprockets are too soft, now they have tougher cores. There seems to be massive variation in success, some are using the latest 18 sprags in original sprockets without any issues, some use 14 sprags in sprockets ground for 18 sprags without issue. Some suffer more failures than others. I have looked at fitting the 20 sprag which is slightly wider, the width is fine but the outer wall thickness which is already under size would be a issue. 

Tri Spark is the best option for both the MK3 and Alton, decent leads and battery are also needed. If a battery is suspect just replace it.  A sprag needs oil, plenty of it, sadly the amount in the chaincase is not enough, yet another Norton oversight. When used with a belt drive (why on a MK3) expect it to fail sooner, but then swapping it out is a lot less messy. 

Use a belt drive on a Mk3 at your peril. The sprag and the needle roller bearing in the center of the drive gear need oil. You really don't want the drive gear/sprag/sprocket seizing together, especially at speed. At least an exploded starter motor, at worst it flings you down the road.

If you are adamant that you have to run a belt drive with a working starter motor system, make sure the belt can be run in oil. 

After going through a number of sprags i bit the bullett and bought the CNW kit, running with the standard starter with a  four brush conversion, faultless every time!