Working out a gradient

[Mick F]

The only truly accurate way is with measurements taken on the site, not with a bike GPS.

I cannot disagree with absolute measurements.

But, if you use your GPS device and measure it for yourself "on site" within a few minutes, the height difference recorded is quite valid.

How accurate do you want the gradient?

I measured our drive ........... which is rather steep! ......... by using a 3m length of 4x2 and a spirit level.
This was only over 3m, and not the nearly 100m that the drive actually is. It's shallow near the bottom and also near the top, but the middle bit is rather steep.

[Samuel D]

The easiest way is to use a cycle computer with a barometric altimeter. Most of these have a feature for directly reading out the gradient, though not necessarily in your preferred units. They are accurate enough for this purpose if you’re riding alone at a steady speed on a calm day, as you easily could do to measure a hill. (When riding in someone’s turbulent slipstream, in gusts of wind, or just accelerating or decelerating the barometer gets a bit confused.)

[old_windbag]

The easiest way is to use a cycle computer with a barometric altimeter.

Using a wheels sensor instead of GPS too. The barometric altimeter is ok as stated by Samuel with none varying atmospheric pressure and temperature. I find that there is a reasonable delay in the reading from mine on my garmin to account so you may have to cycle further than the top of a hill to allow for this. If you did a long route over a changeable day the altitude measurement will drift off, naturally. But also the data points will only be at perhaps one second intervals also, so go up a hill slowly to get more data perhaps.

It's not ideal but does save being found dead in a verge with spirit level in hand and tyre track patternation .

[Stevek76]

I've found even for relative differences GPS can be a bit inaccurate at times but I've not really done much tracking for some years and that was on a galaxy s2 which only had the US GPS system and no barometer. Even barometers can be thrown off some days due to weather related pressure anomalies at ground level. Newer phones also support one or more of the glonass, beidou, galileo and further systems and can combine data from all of them which increases the accuracy considerably.

[gloomyandy]

Surely the correct way to do it is to drive up the hill in a 2CV and to note the gear that you required!



My understanding is that this is how the original climb categorisations used in the tour came about. SO cat 1 is first gear, car 2 is second etc.

[Cyril Haearn]

Hors de categorie = so the deux cheveaux could only get up in reverse gear

[pwa]

The only truly accurate way is with measurements taken on the site, not with a bike GPS.

I cannot disagree with absolute measurements.

But, if you use your GPS device and measure it for yourself "on site" within a few minutes, the height difference recorded is quite valid.

How accurate do you want the gradient?

I measured our drive ........... which is rather steep! ......... by using a 3m length of 4x2 and a spirit level.
This was only over 3m, and not the nearly 100m that the drive actually is. It's shallow near the bottom and also near the top, but the middle bit is rather steep.

I just don't trust GPS to always perform. I've had GPS record my top speed at 150mph or thereabouts, walking! And I've had GPS altitude freeze as I've gone up or downhill. When it works it seems to me to be a bit crude sometimes.

[Mick F]

It depends on which device you have.
If it's just a bike device, I'm not surprised how you can find strange issues with speed and elevation etc. I have a Garmin Edge 20 and it's very inaccurate.

My main GPS device is a handheld Garmin Montana with a mount that fits it on the bike. The Monty is very accurate and you can calibrate the elevation. We're at 222ft by the front door, and I've found that by looking at the OS contour lines and using experience in what Monty wants it to be. If you leave it outside for some hours on the bike, it finds 222ft. If I calibrate it to that, I can get home some hours later, and it's back at 222ft plus or minus 10ft or so.

Monty uses a combination of GPS altitude and BP changes. My manually calibrating it to a known elevation, you short-circuit the algorithm but if you're wrong with your "known" elevation, it will wander and try to get it correct.

[velorog]

If you want a spot readout rather than an average you could try on of these ..

http://www.skymounti.com/html/gb.html

Simple but effective.

[thirdcrank]

If you want a spot readout rather than an average you could try on of these ..

http://www.skymounti.com/html/gb.html

Simple but effective.

Eureka!

The plumbline + protractor idea was always doomed to fail but it limited my lateral thinking about measuring the vertical. This can't have swamped the market or somebody would have mentioned it before. Have you tried one?

[Mick F]

If you want a spot readout rather than an average you could try on of these ..

http://www.skymounti.com/html/gb.html

Simple but effective.

Simple yes, though fun and interesting as well.
It would indeed give a spot reading, but how useful in reality is it?

[thirdcrank]

... It would indeed give a spot reading, but how useful in reality is it?

How useful is any of the data that people measure and log about their riding? With the exception of competition, most of it is personal interest only. That may include things like the durability of components, but it's still pretty personal. Some of it may inspire people to try a bit harder than last time or go out for a ride to get the miles in, but different people find different things useful. I've started the year with a mileage chart, full of good intentions and I don't remember ever getting past Easter with filling it in. Part of the problem used to be how to record the mileage. A cyclometer is OK for only so long before the ticking gets annoying (and yes, that's one use for the valve rubber that used to come with a puncture outfit.) In the more modern era, I've had several computers but never bothered with one for long.

Personally, I can see a lot of interest in knowing the angle-of-the-dangle on a particular climb ie the steepest bit. We've had threads about the info on STEEP HILL road signs before and it's inconsistance, ditto the chevrons on OS maps. Brucie has touched on some of this above and hairpins are a particular case in point. By their nature they can be shallow at one edge of the road and virtually unclimable at the other.

[MikeF]

Note that measuring the 'run' along the road surface itself (eg using a mileometer reading or a long tape measure) introduces an error which requires that the first equation should be Inv Sin instead (because you have measured the hypotenuse of a triangle), and the second equation contains an error too; [really you should replace the term L with Cos(A) x (measured L) instead]. For small gradients the error is not great, (which means it is OK and is how they calculate gradient on railway lines) but for steep gradients the two answers will be quite different unless the correction is made.

cheers

You are quite correct, but for hills on roads the differences between the sine and tangent are so insignificant as to be negligible. Measurement errors are likely to be far greater. For example for a hill of 1 in 10 the angle is 5.74 degrees for the sine and 5.71 for the tangent. For a 1 in 5 hill the angles 11.54 and 11.31 degrees. Therefore measurement error needs to be less than 2% for a 1 in 5 hill to detect the difference. Unevenness of the road is likely to be one source of error.

[Brucey]

the errors (difference between adjacent and hypotenuse) are not large at moderate gradients but they soon get larger; at 1-in-4 they are over three percent and at 1-in-3 they are nearly six percent. Probably this doesn't matter for a hand-waving exercise, but I thought it worth mentioning, anyway.

cheers

[MikeF]

the errors (difference between adjacent and hypotenuse) are not large at moderate gradients but they soon get larger; at 1-in-4 they are over three percent and at 1-in-3 they are nearly six percent. Probably this doesn't matter for a hand-waving exercise, but I thought it worth mentioning, anyway.

cheers

Agreed. It's certainly worth mentioning, especially as road gradients are tangents and railway gradients are sines.