Rolling tests

For discussions about bikes and equipment.
User avatar
Mick F
Spambuster
Posts: 47792
Joined: 7 Jan 2007, 11:24am
Location: Tamar Valley, Cornwall

Rolling tests

Postby Mick F » 18 Oct 2016, 7:28pm

I'm sort of known for experimenting. :D

Please hang on with me here .................................

Gunnislake Railway Station has a flat and level carpark, and with a gentle slope into it.
https://www.google.co.uk/maps/@50.51661 ... 312!8i6656
It's near the top of the hill, and I sometimes turn in and check out the rolling distance of my bikes (or bikes).
I stop exactly at the lamp post on the left, and start without pushing off, then clip in making sure I only pedal backwards to do it.
I freewheel along until I stop.

Measuring the distance of the slope and the level section using the Mick F Standard Pace, I find that the slope is 15 MFSPs long.
The level section to the Base Reference Position is 80 MFSPs.
Mercian stops at BRP plus three kerbstones. This is repeatable and is only variable by a couple of feet no matter how many times I do it.
If I do this with Moulton, I find that if reaches BRP minus three kerbstones.

Now, here's my question:

When I roll with Moulton, it feels as though it's rolling just as well as Mercian.
By assessing the speed as Mercian enters the flat bit, it's much faster than Moulton.
The mere fact that Moulton doesn't roll as far as Mercian, is down to the start speed, and not the rolling resistance.

How can I prove this?
Could a 406 wheel bike not accellerate as fast as a 700c wheel bike?
If so, how can I measure the accelleration differences?
Should I start higher up with Moulton to equalise the speed?
If so, how do I measure it?
Mick F. Cornwall

gloomyandy
Posts: 1046
Joined: 16 Mar 2012, 10:46pm

Re: Rolling tests

Postby gloomyandy » 18 Oct 2016, 8:34pm

How fast are you going at the bottom of the slope? Could a chunk of the difference be down to air resistance? I assume that your position on one bike is more aero than on the other?

Mr Evil
Posts: 190
Joined: 21 Feb 2016, 11:42pm
Contact:

Re: Rolling tests

Postby Mr Evil » 18 Oct 2016, 9:40pm

Mick F wrote:...If so, how can I measure the accelleration differences?..

If you have a smart phone, you should be able to find an app that will do accelerometer data logging.

nicmarsh
Posts: 258
Joined: 14 Sep 2011, 5:12pm
Location: SE UK

Re: Rolling tests

Postby nicmarsh » 18 Oct 2016, 9:55pm

it is quite obvious that this all needs to be filmed and uploaded to You Tube

Brucey
Posts: 36763
Joined: 4 Jan 2012, 6:25pm

Re: Rolling tests

Postby Brucey » 18 Oct 2016, 10:01pm

if you have a GPS device (of almost any kind) you can record the actual speed when you enter the flat bit. This measurement will have some scatter in it, so will need to be repeated many times. You can (by trial and error) adjust the start point to give the desired entry speed.

BTW two things;

1) variations in wind speed may well be the biggest source of variability here.
2) the playing field is not really level because the inertia of the larger wheels is unlikely to be the same as the inertia of the smaller ones.

In point of fact the latter thing ought usually to make the smaller-wheeled bike accelerate more quickly on a slope, if all the other factors are identical.... so I think you may already be seeing that the smaller wheels are more draggy (at least in terms of Crr, anyway)

cheers
-------------------------------------------------------------------------------
~~~~~~~~~~~~~~~~~~~~~~Brucey~~~~~~~~~~~~~~~~~~~~~~~~

User avatar
The utility cyclist
Posts: 2732
Joined: 22 Aug 2016, 12:28pm
Location: The first garden city

Re: Rolling tests

Postby The utility cyclist » 18 Oct 2016, 11:18pm

Body position/aerodynamics is more likely to be the main difference here not wind speed. It seems too short a distance to garner any real useful data and too short a distance for a GPS to be useful IMHO.

It wouldn't surprise me if the tyres on the Moulton offer better RR, IF as I suspect the majority of the speed diff (at the start of the flat section) is the result of better aerodynamics on the Mercian then that is likely the explanation.

unless you have a wind tunnel it's impossible to prove accurately that X bike is more aero than Y bike, even small variations in your body position could totally cancel out any differences in aero differential between the bikes themselves, even counter the whole aero drag of a bike. We know this as total drag of riding can be as much as 90% due to the person on it.

Here's a link that might be helpful https://www.exploratorium.edu/cycling/a ... mics1.html

Brucey
Posts: 36763
Joined: 4 Jan 2012, 6:25pm

Re: Rolling tests

Postby Brucey » 19 Oct 2016, 12:56am

The utility cyclist wrote:Body position/aerodynamics is more likely to be the main difference here not wind speed. It seems too short a distance to garner any real useful data and too short a distance for a GPS to be useful IMHO.<br abp="413"><br abp="414">It wouldn't surprise me if the tyres on the Moulton offer better RR, IF as I suspect the majority of the speed diff (at the start of the flat section) is the result of better aerodynamics on the Mercian then that is likely the explanation.<br abp="415"><br abp="416">unless you have a wind tunnel it's impossible to prove accurately that X bike is more aero than Y bike, even small variations in your body position could totally cancel out any differences in aero differential between the bikes themselves, even counter the whole aero drag of a bike. We know this as total drag of riding can be as much as 90% due to the person on it.<br abp="417"><br abp="418">Here's a link that might be helpful https://www.exploratorium.edu/cycling/a ... mics1.html


I think you have probably missed the point; in a nutshell, Mick's tests are at very low speeds and the effects of aero drag are not at all as you imagine. I can explain more if necessary.

cheers
-------------------------------------------------------------------------------
~~~~~~~~~~~~~~~~~~~~~~Brucey~~~~~~~~~~~~~~~~~~~~~~~~

User avatar
Mick F
Spambuster
Posts: 47792
Joined: 7 Jan 2007, 11:24am
Location: Tamar Valley, Cornwall

Re: Rolling tests

Postby Mick F » 19 Oct 2016, 6:33am

Hi guys,
Thanks muchly for your thoughts and advice.
No. I'll not bother with YouTube with this! :lol:

Here's a little picky of the runout course from the other day on Moulton. I came up Sand Hill (Gunnislake Hill) and went out further west. When on my way home, I turned into the station and did two circuits before heading down Sand Hill.
Screen Shot 2016-10-19 at 06.12.14.png
My Garmin Montana is mounted on Moulton and it is set to record a data point every 4yds up to a max of one per second if you go faster than 4yds per second.
4yds per sec = 8mph or thereabouts.

If I inspect the data, I see that I reach a max speed of 5mph, so the data points are quite coarse. When I repeat this test, I'll set the Montana to continuous 1sec recording and see if I can get some meaningful data. The Garmin on Mercian is an Edge20 and this can only do Smart Recording ........ which is very very coarse. This is Mercian and Garmin Edge20 from last week. I could easily mount the Montana to get a better trace with it.
Screen Shot 2016-10-19 at 06.31.48.png


5mph wouldn't make any wind resistance differences to the two bikes, so I'm convinced it's more to do with the acceleration differences.
My position is always the same for this test on both bikes. I make a point of it. Also, both bikes are set up identically in reach and height and saddle position.

I'm aware of the wind/weather can make a difference, and I pick calm days to do the testing. There's rarely ever a tail-wind, and sometimes a head-wind. but the position of the carpark is quite sheltered. It's a matter of "Feel" to know whether the test is valid or not, but doing this from time-to-time, I'm getting quite experienced in this.

PS:
Just been looking at the Montana, and I've set it to record a data point every 0.01yds = a few inches. It won't record faster than once per second = 8mph, so I should get as many points as it's possible to get as I only reach circa 5mph.
Mick F. Cornwall

tim-b
Posts: 1158
Joined: 10 Oct 2009, 8:20am

Re: Rolling tests

Postby tim-b » 19 Oct 2016, 8:11am

Hi

I doubt that you'll get an accurate measurement using GPS over 95 MFSPs. Far better to go equipped with a tape measure, chalk, stopwatch, calculator, notepad and pen
Time each bike over a measured sloping distance and work out its speed. Experiment with different distances for each bike until you get the same entry speed to your level area. EDIT: It'll be an average, but hopefully accurate enough and repeatable. Ideally you'd use the level area to consolidate a constant speed before your rolling test began

Here's an experiment that you might enjoy

Regards
tim-b
~~~~¯\(ツ)/¯~~~~

Brucey
Posts: 36763
Joined: 4 Jan 2012, 6:25pm

Re: Rolling tests

Postby Brucey » 19 Oct 2016, 9:19am

Mick,
just for fun I ran some numbers in an online bicycle calculator using a nominal 10W input and this suggested about 7kph speed on a flat road, no wind. I then changed a few things;

+25% Crr gave a ~1kph speed reduction
+25% frontal area gave a ~0.2kph speed reduction
a barely detectable headwind (~3-5kph IIRC) also gave a ~1kph speed reduction.

So whilst the test is relatively insensitive to variations in frontal area, and sensitive to variations in Crr (as you want it to be) the slightest variation in airspeed vs road speed could completely obscure any variation in Crr that you are trying to detect.

One method for dealing with this is to apply a 'control condition', e.g. you ride the Mercian and record the results from that immediately before and after testing the moulton.

Whatever method you can use to get a consistent entry speed into the flat part on each machine will probably do (even though the inertia of each machine will still be different).

One idea I had was to use a non-contact speed measurement system of some kind, that could be swapped from one machine to another without any calibration issues. These days there are laser-based sensors (that are used on production lines) which can measure the speed and length of objects flying down a conveyor belt. Maybe something like that could be used to measure road speed?

cheers
-------------------------------------------------------------------------------
~~~~~~~~~~~~~~~~~~~~~~Brucey~~~~~~~~~~~~~~~~~~~~~~~~

User avatar
Mick F
Spambuster
Posts: 47792
Joined: 7 Jan 2007, 11:24am
Location: Tamar Valley, Cornwall

Re: Rolling tests

Postby Mick F » 19 Oct 2016, 9:35am

Thanks Tim and Brucey. :D
I wonder if that is what I'm after? It all seems a bit complicated.

I was thinking over breakfast, that the test isn't speed or distance dependent, but on the "decay" shape. I've said that Moulton seems to roll just as well as Mercian, but the initial speed is lower, so it goes less far.

If I could produce a graph of the two bikes under test, the decay of the speed should follow a curve, and those curves should illustrate the differences visually.

Other than that, I need to find the sweet spot on the slope, so that both bikes are are at the same speed on entry. Instantaneous speed past a point is hard to assess.

Can't cycle today, but I could get out for a short ride tomorrow and possibly on Friday.
Mick F. Cornwall

Brucey
Posts: 36763
Joined: 4 Jan 2012, 6:25pm

Re: Rolling tests

Postby Brucey » 19 Oct 2016, 10:03am

I think it will be very difficult to distinguish between the effect of a slight change in headwind and the effect of a change in Crr, based on the shape of the curve alone. The assumption is normally that the Crr is constant with speed, but the reality is that it probably isn't; various effects related to the suspension qualities of the tyre, the bike, and the road surface imperfections will mean that the Crr could very well vary a little with speed.

BTW another way of reducing the possible effect of small headwinds is to use a perfectly flat surface (zero gradient) and to run the test in both directions, averaging the results. However if there is a part crosswind, this can complicate matters; the 'control condition' is probably the simplest method by which results can be validated.

cheers
-------------------------------------------------------------------------------
~~~~~~~~~~~~~~~~~~~~~~Brucey~~~~~~~~~~~~~~~~~~~~~~~~

MartinC
Posts: 1837
Joined: 10 May 2007, 6:31pm
Location: Bredon

Re: Rolling tests

Postby MartinC » 19 Oct 2016, 11:56am

Mick F wrote:...............The mere fact that Moulton doesn't roll as far as Mercian, is down to the start speed, and not the rolling resistance.

How can I prove this?..................................


But the start speed of the bike will depend on it's rolling resistance (amongst other things) so I'm not sure what you're trying to prove? The same set of factors that constrained the start speed of the Moulton will constrain how far it rolls.

You've proved that the Moulton doesn't roll as well as the Mercian. If you want to examine why this is you need to systematically eliminate all the variables and see what the effect of each is. As Brucey says the air resistance is likely to be little but I've no idea how you can eliminate this - creating a perfect vacuum in the car park would be challenging. Perhaps you could create a windshield (as light as possible) you could attach to both to equalise the air resistance - maybe a large cardboard box to cover you and the bike would work?

User avatar
Mick F
Spambuster
Posts: 47792
Joined: 7 Jan 2007, 11:24am
Location: Tamar Valley, Cornwall

Re: Rolling tests

Postby Mick F » 19 Oct 2016, 3:25pm

MartinC wrote:You've proved that the Moulton doesn't roll as well as the Mercian.
Have I?
I'm not convinced of that, despite what I thought before, and what the length of the runout shows.

We're not comparing like with like.
If the initial conditions could be discounted, will the Moulton run as well - or not - as the Mercian?

It's not that Moulton doesn't roll as well, it's that a suspended small wheeled Moulton doesn't accellerate from rest as easily as an un-suspended big wheeled bike ................. or am I wrong?
Mick F. Cornwall

PBA
Posts: 178
Joined: 15 Apr 2009, 1:13pm

Re: Rolling tests

Postby PBA » 19 Oct 2016, 6:41pm

You are converting potential energy (height) to kinetic energy (speed). The speed gained at the bottom of the hill is generated by rolling down it - Nothing else. If Mercian is going faster than Moulton at that point then whatever losses are present are greater on Moulton. If you start further up the hill, then you are allowing a greater energy to be imparted onto Moulton. In effect, it shows that to maintain the same speed you would need to work harder on Moulton than Mercian.

The roll out distance is directly related to energy carried by the bikes when you reach the flat portion. That is, the proportion of the energy given by the hill which has been successfully converted into speed rather than already lost. The entry speed will then be degraded to zero by the same losses that prevented Moulton reaching the same speed as Mercian (note that the inertia gained will now assist).

Your test allows the bikes to reach a maximum speed from a standing start and then to decelerate to a complete stop. I think that is complete and fully adequate as you have already tested. If you only consider the deceleration portion you neglect to consider how you get the energy into the bike to begin with.

However, it might be fun to find a way of getting the entry speeds identical - just for the sake of experimenting.

Perhaps you could also investigate other locations, with differing topography to do the comparison tests? Those may produce more interesting data such as the effects of road surface or peak speed...