Knife Edges for Crank Balancing

[BoschZEV]

I found that it is most practical to determine an existing factor... (It saves a lot of math!)

Well, since determining the existing balance factor and math both have been mentioned, and again temporarily digressing from knife edges, I posted the following elsewhere but some of you may find the general procedure I used, if not the actual results, of interest.

In 'Tuning for Speed' Phil Irving wrote that "The only source of reliable information [on the balance factor] is the parent factory who have certainly done a lot of experimentation." What he meant by this isn't that only the factory knows the balance factor, it's that the factory will have done lots of experiments to arrive at the compromise balance factor they decided works best. As Shaky Jake wrote, you can have all the vibrations for a single up-down (0% balance factor) or fore-aft (100%), or somewhere in between where they have been found to be least objectionable for a given engine/frame design and over a given range of rpm.

I wanted to determine the original balance factor used by the factory on the 1928 Ariel Model C I'll be riding in the Cannonball so that I could rebalance the engine for one of the two new, +60 aftermarket pistons I've already purchased. My Ariel came with a worn +30 piston and with the small end rebushed to accept the 13/16" gudgeon pin of this newer-style piston rather than the 1" originally used. So, in addition to determining weights I had to correct for the weights of different bushes as well as for the socket head cap screws now locking the big end nuts in place. However, other than socket head cap screws used to lock the crankpin nuts, and the possible exception of four 5/16" holes discussed below, there are no other signs that the flywheels have been altered by additional drilling or plugging since they left the factory.

Since a search of the Ariel Club's web site shows the subject of balance factors has come up more than once, along with speculation on what value Ariel might have used, below I describe in some detail the measurements, uncertainties, calculations and assumptions leading to my determination so that others can decide for themselves if they want to accept the value I found. For those who don't want to read all of this, the executive summary is that the original balance factor used in my engine was either 56% or 60% (both +/-1%) depending on an assumption about four holes described below.

A seldom discussed but fundamental issue with the static balance method as usually described is it relies on the center of mass of the flywheels being on the axis defined by the crankpin and crankshaft. If an inhomogeneity in the flywheels places the center of mass off that axis then adding weight to the hanging connecting rod (which in effect places that weight at the center of the crankpin) will draw the center of mass close to the crankshaft axis but can never precisely balance the flywheels. To achieve perfect static balance requires applying weight to the connecting rod to draw the center of mass to its point of closest approach to the axis of the crankshaft, then adding (or subtracting) additional weight from the flywheels at 90-deg. from the crank-crankpin axis.

In all there are five 1/2"-dia. balancing holes drilled on the inside rims of the two flywheels. Because of the inaccessible location of these holes I speculate that the flywheels were balanced individually prior to being assembled into a complete crankshaft. There also are four 5/16"-dia. balancing holes drilled on the outside faces of the rims. There is no way to know if these were done by the factory at the time to tweak the crankshaft into the final balance factor after assembly, or if they were done during a later rebuild to keep the same balance factor for a heavier piston, or to change it to a higher balance factor. I address the quantitative effect of these possibilities on the calculated balance factor below.

In what follows I keep the precision of individual measurements (e.g. the 10 mg of one scale) although the final quoted uncertainty largely depends on the least sensitive measurement used in the calculation. I used the following tools:

200 g balance calibrated with weights accurate to 0.3 mg. Balance reads to +/-10 mg.
6 kg balance calibrated with weights accurate to 0.1 g. Balance reads to +/-0.5 g.
6-piece set of 5-50 g balance weights each accurate to 0.01 g.
Crown-brand balancing wheels of sensitivity 1 g-cm, equivalent to 0.2 g imbalance at the radius of the crankpin.
Digital calipers.

The four "external" 5/16" holes are at the crankpin end of the crankpin/crankshaft axis and have a total depth of 3.86" resulting in a volume of steel removed of 0.296 in.³. Using 0.29 lbs./in.³ for the density of steel, the total weight removed from these four holes was 0.086 lbs. (38.9 grams)

A formula for calculating the Balance Factor can be written in the form:

Balance Factor = balance weight + small end weight / piston weight + small end weight

As can be seen, to solve this requires determining three weights as well as having a fixture for holding the crankshaft so it can rotate freely when the balance weights are added.

Small end weight:
With the bushing to reduce it for a smaller gudgeon pin it weighs 267.5 +/-1 g. However, from this subtract 7.4 g for the "excess" weight of the bronze (see 'sidebar' below for details) so it originally would have weighed 260.0 +/-2 g.
-- Current small end weight = 267.5 +/-1 g
-- Original small end weight = 260.0 +/-1 g

Piston weight:
The "piston weight" is that of the complete assembly of piston, gudgeon pin, circlips and rings. Although it doesn't enter into the calculations shown here I'll note that the weight of the additional Al used in a, say, +30 piston is not negligible. It can be calculated from the annular volume of a piston of stock diameter and one 0.03" larger than that.

-- weight of +30 piston assembly that was currently in my bike 467.5 +/-0.5g

I was lucky to find two people with original piston assemblies for the Ariel. The one in Australia is used and weighs 507.2 g and the one in Canada is new and weighs 503.5 g.

-- weight of original piston assembly weight (average of above) = 505 +/-2 g
-- weight of aftermarket +60 Gandini piston assembly 516.5 +/-0.5 g
-- weight of aftermarket +60 Omega piston assembly 435.0 +/-0.5 g

Balance weight:
I hung balance weights and washers from a wire attached to the small end until the crankshaft was in balance and weighed the final total mass. It took 196.59 g plus 10 g on the rim at 90^o. Taking into account the off-axis imbalance I estimate the uncertainty in balancing the crank using only weights hanging from the connecting rod and none at 90^o is +/-3 g.

The weight of the heads of the two 1/4" cap screws pinning the big end is 2 x 2.74 grams = 5.48 g. Without the cap screws it would have required that much additional weight to balance the crank originally, offset somewhat by the 7.4 g "excess" of the current bronze reducing bushing, i.e. 196.6 + 5.5 - 7.4 = 194.7 grams.

-- weight to originally balance crankshaft = 195 +/-3 g

If the 5/16" holes were added sometime later the original weight required to balance it would have been 38.9 grams less.

-- weight to balance crankshaft without the four 5/16" holes =165 +/-3 g
-- total weight to balance crankshaft in its current form = 196.6 +/-0.1 g

Original factory balance factor:

If the crankshaft in its current form (less the cap screws) is how it left the factory, the original balance factor was:

294.7 + 260.0 / 505.0 + 260.0 = 455 / 765 = 60.2 +/-1%

If the four 5/16" holes were added later it would have required 38.9 grams less to balance it originally. In this case the original balance factor would have been

165 + 260 / 505 + 260 = 425 / 765 = 55.6 +/-1%

For comparison, a 1960 BSA Service Bulletin shows 60% for the 250cc 'C' series, 58% for Gold Stars, 55% for the essentially identical 'B' series singles in the same frame as the Gold Star, and 55% for both the 500cc and 650cc 'A' series twins, also in the same frame as the Gold Star. A 1930s Vincent Comet used 66% (claimed weight 390 lbs. vs. 290 for the Ariel) but this had to be reduced to 61% in a lightweight speedway frame.

Current Balance Factor:

With current piston in it:
196.6 + 267.5 / 467.5 + 267.5 = 464.1 / 735 = 63.1% +/-0.3%

With Gandini piston in it:
196.6 + 267.5 / 516.5 + 267.5 = 464.1 / 784 = 59.2 +/-0.3%

With Omega piston in it:
196.6 + 267.5 / 435.0 + 267.5 = 464.1 / 702.5 = 66.1 +/-0.3%

To reduce the balance factor to 60% in order to use the Omega piston in it would require reducing the required balance weight by 43 g which in turn would mean removing roughly half that weight from the rim of the flywheel. This could be achieved by, for example, drilling two additional 5/16" holes approx. 1" deep each.

If the Omega offered a significant advantage over the Omega I would modify the crankshaft accordingly. However, since the Gandini piston could be used with the crankshaft as-is, I'm particularly interested in any experiences people have with these aftermarket pistons.

[TechNoir]

Hi Cotten, thanks for posting the picture of your rig.

You have confirmed what I had intended to do which was make my rig high enough to hang a rod and piston under it. Tommo had mentioned something similar previously so thats what I am going with.

From the axle to the end of the rod on my 20F is about 12 inches and I would obviously need a bit more to allow for a piston. I have a 1950's Matchless single that will need attention in the foreseeable future but that crank/rod assembly is a couple of inches less than the F. I am guessing (without checking the engine specs) that a JD would be a bit more than the J Harleys? I would also assume that a Big X or an Indian would be similar dimensions to a J or JD and that all of those would be more than most British singles? So, if you don't mind me asking, what height would you go for from the top of your base to the top of the knives if you were to do it again?

John

[T. Cotten]

Folks,

Without getting into the quagmire of factor and rpm yet,
I found that it is most practical to determine an existing factor of an assembly when the 'edges' are suspended high enough to accept installed rods, and perhaps a piston.
(It saves a lot of math!)

The edges should ideally be easily leveled and trued, and their supports rock solid, and as insulated from floor vibrations as possible.
My assembly is over-kill, with a 2" baseplate and 1 1/4" hex supports, but that is what was handy, and a ground granite block was not.
The 1 1/16" planer blades have a .0625 flat ground on their edges, but will still bow under the weight of flywheels if placed too far from a support.
Six inches between supports would be better.
(I would wack six off the height of my supports as well, but I no longer have access to the equipment...)

.....Cotten

[TechNoir]

In 'Tuning for Speed' Phil Irving wrote:

I have a copy of Irvings book and it is very good and I have learned a lot from reading it.

This thread has transformed from how to make knife edges to a discussion on motorcycle design theory. I hope you don't mind! I for one find it very interesting, and as always I enjoy the level of experience and intelligence that exists on this forum.
.

No need to apologise Kevin. I am a virgin as far as engine balancing goes but I am keen to learn. Therefore I very much appreciate everyones input to this thread and replies to my questions.

If nobody minds I will go "off topic" and back to the original question. As I mentioned yesterday, today was errand day. Having pondered my options last night I decided that knife edges would be simplest in the short term and also I need an accurate scale. So first off I ordered a pair of straight edges from an engineering supplies place and I also ordered a 3 beam balance from scales company. The balance is specified to measure between 0.1g and 2,660g to an accuracy of plus or minus 0.1g. I figure that is probably good enough for my purposes. I could have gone for all sorts of fancy electronic devices which output to electronic files to connected devices via bluetooth but I figure a simple balance has to be about as reliable as you can get, the physics doesn't get much simpler.

Later on this afternoon I managed to get round to visiting the steel stockist. I loaded up what I think I need for my apparatus and went into the office to pay. Whilst waiting for the guy to crunch numbers I browsed the off-cuts section near the counter. I asked the guy how much a piece of 2 1/2 inch diameter 304 stainless round bar would cost because the sharpie marker that was marking up all the off cut prices was smudged off this particular piece. (and no it wasn't me, it was already smudged but at this point I was just casually thinking that if I had been intending on making a stand more like Bosch's rather than knife edges then this was the sort of stuff that might be suitable to make the rollers from). The guy tapped at his keyboard and then said that his computer wasn't showing any prices on 304 stainless so I could have it as an early Christmas present.

So I am now, maybe, looking at both options. I fully intend to go with knife edges but now that 4 pounds of free stainless has landed in my lap maybe I will have a look at rollers as well as knives.

Now what was it that Bobbie Burns said "To a Mouse"?

John

[Shaky Jake]

Yes - thank you guys for bringing up the importance of chassis design in deciding on a balance factor.

To help us think about what balance factor means, take a simple example of an engine with a single vertical cylinder and no balance shaft. You have to realize that whatever mass you add to the flyweights to balance out the reciprocating mass of the rod and piston, is totally unbalanced in the direction PERPENDICULAR to the cylinder. In other words, increasing the balance factor of that engine decreases the up-and-down shake, but INCREASES the fore-and-aft shake.

You could make the flyweights heavy enough to completely balance out the reciprocating weight of the rod and piston, which would be a 100% balance factor. But if you did that, the flyweights would generate an imbalance in the fore and aft direction that is equal to the entire weight of the piston and rod.

When I think about it that way, it seems to me that chassis design is possibly the most important consideration in deciding what the balance factor should be. Is the frame more rigid up and down, or fore and aft? How is the engine mounted and how will the forces be transmitted from the engine to the frame? What affect will those forces have?

In the case of my old Indian with the high balance factor, why didn't it seem to shake more than other similar bikes with the correct balance factor? It might be because up-and-down shake, which was decreased, is more noticeable than fore-and-aft shake, which was increased. And yes, even though it cruised at around 2,000 RPM, of course there were occasional excursions to 3,000+ RPMs, but you expect there to be more vibrations at those times, so it didn't seem unusual.

Food for thought.

This thread has transformed from how to make knife edges to a discussion on motorcycle design theory. I hope you don't mind! I for one find it very interesting, and as always I enjoy the level of experience and intelligence that exists on this forum.






Kevin


.

[BoschZEV]

I believe that chassis design is also a factor.

In 'Tuning for Speed' Phil Irving wrote:

"There are in fact so many considerations involved that it is impossible to quote one figure as being ideal, since it varies with every type of engine, and even for the same engine in differently-equipped frames. The only source of reliable information is the parent factory who have certainly done a lot of experimentation. The M.O.V. Velocette factor is as high as 85 per cent, while some engines have been under 50 per cent, but failing any reliable information 66 per cent is a good starting-off point."

[TechNoir]

Kevin, Your stand and your results give me even more food for thought.

Bosch's point on engine speed is also pertinent. A modern speedway single cylinder engine must be more prone to vibration than, say, Bosch's Ariel. I believe that chassis design is also a factor.

My triumph is a parallel twin and when it was first put together I had the engine work done by someone else, including static balancing. It was un-rideable above 55 mph even thought I have letters from the original owner dated in the 1940's stating that he had had it up to over 90 mph and occasionally touched 100mph downhill. I then had it dynamically balanced and it was transformed.

Once again, thanks to everyone for their input on this.

John

PS. I forgot to mention, the Triumph sludge trap arrangement must mean that as the engine ages the balance factor constantly changes. I wonder if the original designers thought about this?

[BoschZEV]

Balancing is a big effort

+1

Balancing is much more important on a high speed engine.

I agree. Since the oscillating imbalance force goes as the square of rpm the same weight imbalance on the flywheels would result in 2.25x more shaking at my hope-for 3000 rpm as at 2000 rpm. If I have to rev it to 4000 rpm at times, such as having to drop back a gear but still maintain a reasonable speed on a road in the Rockies, that shaking increases to 4x

In my case no one had any idea what the original balance factor of my engine was, the engine had a different piston than when it was new, and I had good reason to have lost confidence in the skill of the previous rebuilder. If it shakes itself apart after I rebalance it to what I've now determined the factory spec to be I'll know it's the fault of the designer, not because I took a shortcut in the rebuild.

[Shaky Jake]

You guys have some nice tools. Here's a link to a video of what I've used. Feel free to poke fun at it:

https://youtu.be/JSW2ckp0DvY


It's just a slightly modified wheel truing stand, but it did give me repeatable results. Using this set-up and an accurate scale I was able to determine that the stock balance factor was 58%. Replacing the iron pistons with aluminum ones changed it to a touch over 79%. I didn't feel like drilling my 100 year old wheels, so I left it at 79%, put it together, and rode it 4,000 miles. It didn't shake any more than expected. After 4,000 miles I tore it down. The main bearings look and measure just like they did when I put it together. Keep in mind this engine only turns at about 2000 RPM at normal cruising speed. I'm thinking about balancing it this time, just to compare. I don't expect to notice much difference though, my PowerPlus felt as smooth as the other ones.

People don't like to hear about this. I even see that someone gave me a thumbs-down on YouTube. :-) Balancing is a big effort, and people like to feel like they get big benefits for their efforts. This isn't the first engine I've balanced, but it is by far the oldest and slowest. Balancing is much more important on a high speed engine. Feel free to disagree, I respect your right to do so.



Merry Christmas,

Kevin


.

[BoschZEV]

Similar to Bosch's setup but it looks homemade. ...would take longer to make and, maybe, not as sensitive as knife edges/bars.

In case you meant your "maybe" uncertainty to apply to balancing wheels in general, not just to a homemade version, my Crown wheels easily do better than 1 g on my Ariel crankshaft. Since ~8 g corresponds to 1% in the balance factor on this crankshaft (e.g. 60+/-1%) this means being able to reach at least 0.1% in precision (e.g. 60.0 +/-0.1%). Given the time it takes to hang weights on any type of balancing rig, take the flywheel off to drill holes (and the imprecision of the depth of those holes), and hang weights again to check the result (if it was checked), I suspect that even the former, 1%, is better than the factories did it.

[BoschZEV]

Bosch, I found this on google, I think it is is like your jig?

I guess the key is concentric and accurately machined disks plus sensitive bearings.

Yes, that Crown picture you found is one just like mine, except mine only say "Crown" without the "20 in." The stand I machined has flats for them to sit on with vertical tabs through which I use bolts to secure them using the holes otherwise used for the rods in the photo.

The homemade version shown in another photo shows no evidence the wheels in it had been balanced. Each of the Crown wheels has several shallow balancing holes drilled in it. When I went to the garage to check how they compared with the other photo I gave one of the wheels a flick with my finger and it spun for 35 sec. I've never measured their runout but given everything else about them I'd bet it's pretty small. Just as you wrote, the keys are good disks and sensitive bearings.

Having recently been fiddling with my Ariel's crank makes me appreciate the properties of these wheels all the more. The fact the crank stays in one location, rather than simultaneously rolling and turning, is a definite advantage both when the imbalance is large as well as when getting down to the final few grams. Not having to spend the time to precisely level it is another definite advantage of wheels over knife edges.

I was lucky in finding my set on eBay some years ago and having to pay relatively little to get it (I don't remember how much). Sometime later I came across the retail price; again, I don't remember what it was, but I do remember it was a pretty significant amount.

[T. Cotten]

Yes, John!

That drawing sums it up, although I don't remember the arbor confusion.
(And I would drill vertically through the ends of the rods, so then they could be leveled with threaded rod and nuts.)

So keeping it as simple as possible, a "truing stand" and "knife-edges" are different animals.

You can true wheels between any machine centers with enough throw.

Knife-edges are pretty much for balancing only.

(But no doubt, a reasonable balancing could be achieved on centers, if great care is taken.)

....Cotten

[TechNoir]

I use my homemade truing stand,flipped 90 deg.and clamped over the bench edge.Seems to work pretty good.
Tom

Thanks Tom, I have a truing stand but its not the greatest and wouldnt want to use it so whatever I use will be a "new build".

John

[TechNoir]

Tommo, many thanks for posting the pictures of your setup, that gives me more food for thought. I have two HSS plane blades about 600mm long that I acquired years ago thinking I could use the HSS to make tooling from because the sharp side of them has a few chunks missing although I have not actually done anything with them yet. Maybe I can use the other side of them if they are tilted at an angle? I will put that thought into the mixing pot with the other ideas and see what comes out.

Bosch, I found this on google, I think it is is like your jig?



Tom, I remembered about the picture of the ground bar. It was actually Tommo who provided it to me when I first asked about balancing some time ago. Here is the picture, I assume its the same one that you were referring to?




I found this picture on an aussie site here:




Similar to Bosch's setup but it looks homemade. I guess the key is concentric and accurately machined disks plus sensitive bearings. It would have a smaller footprint but would take longer to make and, maybe, not as sensitive as knife edges/bars.

All good food for thought, I will sleep on it and decide how to attack this tomorrow.

Thanks to you all once again,

John

[tfburke3]

I use my homemade truing stand,flipped 90 deg.and clamped over the bench edge.Seems to work pretty good.
Tom