Dynamo belt tension

[MG]

Since I don't feel too confident about running the belt in either oil or grease, and since I'm positively sure it was not over-tensioned, and since some others run those things for 15000+ miles without problems (hi groily  ), I wanted to get at the bottom of this, so today got Conti's design tool and fed the geometry and power data into it to see whether the belt is basically up to the intended purpose.

So for those interested, that's the calcs and results:

belt type: Conti AT5 GenIII, 420mm length
pulley1: driving pulley (intermediate pinion at half engine speed), 46T
pulley2: driven pulley, 17T

The calculations are based on a power of 100Watt (0.1kW) on the driven pinion (assuming 60Watt electrical power output and 40% power loss in the dynamo, which I think is quite reasonable for a DC brush-operated design) at roughly 1900 generator rpm (this being the design rpm stated in the Lucas service sheet), thus resulting in the highest driving torque over the operating rpm range.

The results show a maximum transferable power of 0.52kW on the driven pulley, so we are facing a safety factor of 5.24 here. The belts are designed to work in temperatures up to 100°C, reduction of maximum power above 50°C is not taken into account in these calculations, but with a safety factor of more than 5 there is plenty of room there.

So the good news is that these belts should be more than up to the task and failure actually is very unlikely.
What is interesting though is the required pre-tension force in the belt, which is 18.5N (about 4.1pounds) in each span of the belt. This is quite some tension, imagine an 8.2 pound weight hanging on the belt, and the "easy 90° twisting-thing" from the mounting instructions misses the point here imo.

So I'm led to believe that my belt failed because it wasn't tensioned sufficiently, I shall give it another try with a new one and this time rather trust my own judgement than the instructions (let's face it, probably t'was my fault this time)  

Cheers, Markus

edit: Lacking equipment to properly measure belt deflection or frequency of oscillation, this could be a neat idea to get the pre-tension about right: Hang an 8.4pound weight on the belt using two discs similar to the pulleys in diameter and twist the belt by 90° to get a feeling for the necessary force, and then apply same to the mounted belt.  

[groily]

Brilliant analytical work Markus, as ever!
But I'm embarrassed to say mine is indeed going strong on at least 15000 miles. Probably rather more, but speedometer repairs have never been a priority for me.
Not ridden it at your or Musky's 7000 rpm though, it has to be said, just normal boring backroad 55-65mph cruising with blasts up to 75 or so. It's only a cooking iron engine after all and it's done who knows how many miles. I am running an extra tooth over standard at the gearbox to keep things sweet on main roads, but that makes no odds to the dynamo. Had a look at the belt a few weeks back when doing some mag swapping. First time I've had the cover off in ages and the belt tension seemed the same as when I put it there. Those with long memories will remember a few years back I had nowt but trouble with the pulleys, but having resolved that by re-using the original drive sprocket's steel centre in a re-machined SRM pulley, no problems.
I've been darn lucky, this very tatty bike's bullet-proof - or has benefited from some sorcerer's sprinkling of magic dust. It can't run for ever, natch, but it continues to surprise me.

[Dynamo Regulators Mike]

Hi Bill
Good to hear of your 15,000 mile and going strong modified SRM belt drive. At the Loreley International I sold a Belt drive kit to one chap, whose dynamo chain broke the next day. And a Danish guy next door to our tent replaced his dynamo chain on arrival. There may well have been more. Then at our Hampshire BSA camp a member missed the run due to noise from timing cover which proved due to the eccentric sprocket issue which is very familiar.

Interesting guide calculation from Markus via Mulco. The Gen III belt is good for 100 C continuous so de-rating from 70 C is in order. So even more safety factor to allow for shock loads through the dynamo. I think it important to note that over-tensioning is warned against strongly by those who should know. Better a small amount of slack rather than too much tension. And the tension does increase markedly with the engine hot. Erring on the tight side when setting cold will lead to excessive strain on the belt. It certainly would be useful to come up with a better measure of tension that does not rely on expensively derived 'feel' as the belt is twisted.   

[MG]

Thanks for the comments, manormike. I shall adjust at operating temp to be on the safe side.
I agree, proper measurement of the belt tension would be at order. Anyone got a good (i.e. cheap) source for those small frequency counters? 

[Dynamo Regulators Mike]

A forked rod slipped over the belt, with a force balance at a set length to allow torque measurement at a given twist. Might work to take away the subjectivity in setting tension.

[MG]

yeah, but then there is the problem of translating the required pre-tension force (or oscillation frequency for that matter) into the appropriate twisting torque, i.e. some sort of calibration will be required.
I have found a post on another forum, where one of the guys wrote a LabView program to measure the oscillation frequency of car engine timing belts, using a laptop and microphone, just like those small hand-held devices do it. Unfortunately the link is broken. 
Anyone here good at Fourier analysis and software programming?   

[RichardL]

Markus,  et al,

Not that I am anyone to challenge the belt manufacturer's or your (Markus's)  mechanical engineering calcs,  but I have a question. In the power calculating program,  the number of teeth in mesh on the driven pulley is given as seven. The drawing would appear to support this,  however,  it seems to me that only half that many could be working a any given time,  as the top run of the belt is in tension while the bottom is trying uselessly to push. Then,  I wonder if more than even the first couple of meshed teeth are truly working, the work having been mostly done before getting to,  say,  the third meshed tooth. If correct,  would this change the results for maximum power that the belt could tranfer?

On the question of oil or grease in the case, i can't see how it would hurt, as the belt can't possibly be depending on static friction from the face of the teeth to avoid slipping around the pulley.  At the same time I don't see how it would help.

Richaed L.

[MG]

Hello Richard!
You are right about the bottom of the belt being more or less uselessly "pushing" the pulley, hence the need for pre-tension in the belt to keep as many teeth as possible in mesh and avoid slipping of the belt. The higher the load on the belt, the higher the pre-tension (and width of belt). These effects are taken into consideration in the calculations, should be at least, that is the problem with such design tools, you can't be sure what it is doing in fact.
Gates' tool (called design flex pro iirc) for example also offers the possibility to distinguish between constant and periodic/intermittent loads, thus reducing the maximum transferrable power by 50-100%.

That is also the main concern I have about grease or oil in there, reducing friction in critical situations, like insufficient tension plus a torque impulse or resonance effects.