Skip to main content

I have placed a full report on the tire-pressure monitoring method at the following site:

I describe in detail the construction of four suitable gauges and the use of the pressure coefficient as a simple means of applying the method. The coefficient is easily obtained and should be valid for the life of the tire. It is stable to changes in temperature and bicycle total weight.
Original Post

Replies sorted oldest to newest

I have revised the report slightly. For the Schrader gauge I have revised slightly the dimensions of the internal components to allow a seal to the valve about one turn earlier. However, I recommend the use of the Presta gauge rather than the Schrader because of greater ease of construction and mounting.Presta tubes can be used in Schrader wheels even without an adapter.
Neville was kind enough to send me a gauge to play with. After doing a few rudimentary experiments I conclude that the pressure-monitoring method works. It possesses the following advantages:

1) It allows a close estimate of the constant to be made at any time in the measurement. This can help when a tire is beginning to leak. It also allows the riding constant to be accurately tracked during periods of rapid climatic change, such as rain.

2) It can eliminate the need for postcalibration, and possibly precalibration.

There are also disadvantages:

1) The tire, once pumped up, must be fitted with a pressure gauge, and this gauge must be left in place throughout the entire measurment. I have some doubts about the fragility of the gauge mechanism under highly vibratory riding conditions. Neville's experience is that the gauge is so far reliable.

2) Pressure changes during a normal measurement are fairly small and require great attention to be paid when reading the gauge.

3) It requires slightly more mathematical ability than does the standard method. No formal monkey-see, monkey-do procedure yet exists.

If someone was to send me a fully documented measurement using this method I'd be willing to certify it. The key would be full documentation.

While the method may appear overly complicated, it's a lot like Neville's electronic counter - once you have it and play with it, the mystery goes out of it.

My experience so far is only in calibration riding. I expect to give it a complete test next time I have a course measurement to do.
Thanks for the favorable evaluation of the pressure monitoring method. However most of your perceived disadvantages are not valid.

Although I think it is advantageous to leave the gauge attached to the valve, there is no requirement to do so. Screw-on gauges can be removed and replaced with no detectable air loss. Also, in my report I describe push-on gauges for a single pressure reading per use.

I have ridden with a gauge in place for over 250 miles and during over 20 certifications without any detectable deterioration. Should the gauge head ever break, I can screw in a new head ($8) instantly. Nevertheless, I agree that it is counter intuitive that the gauge should last more than one mile on a ridden wheel!

In its simplest mode, pressure monitoring requires no mathematical ability and only the ability to read a gauge. The measurer only has to ensure that pressure does fall more than say one division below that at precalibration and can then enjoy freedom from setting up temporary calibration courses and postcalibrations.

Reading the gauge would seem like a trivial operation, but as with the Jones I agree that it does require a little concentration to get maximum accuracy and avoid erroneous readings. (Part of the problem is that the gauge divisions are 20 kPa.) To make reading easy I swing the wheel up to eye level.

Although pressure monitoring can predict the calibration factor with remarkable accuracy from calibration made many months before, I am not yet ready to call for the elimination of precalibrations as these serve to check for possible changes in such things as measurer’s weight, tire wear, and gauge characteristics.
I did a measurement yesterday. I got through the 5 km course OK, but the 10 km course had a mile of rough asphalt. No big bumps but lots of vibration.

Pressure at the start of the ride was 820 kPa, and at the end it indicated 1020 kPa. The vibration evidently damaged the gauge. I counted myself lucky that it was the internal gearing that was damaged, and that the bourdon tube did not break. This would have deflated the tire.
Since I have measured over twenty courses with the same pressure gauge and am still getting very accurate results, it is indeed unfortunate that you should have failure on your second course. However,you should have had no fear concerning deflation. The bourdon tube itself is a very stout metal tube so it would be impossible for it to break. I am sure you completed your measurement without a problem. If you had had a spare gauge head you could have continued taking pressure readings.

I would be very interested in getting full details of your experience. Did the gauge correctly predict the calibration factor on your temporary calibration course course? What does the gauge read now you have removed it from the valve?
Gauging Experience

Upon precalibration yesterday at 7:30 AM the gauge read 820 kPa at 55F. When I began measuring, 50 miles away, it read 850 at 60F, but I had not pre-exercised the tire as I did on calibration. Upon completion of the first measurement of the 5 km course it read 840. I then began the first 10 km measurement, and it was on this ride that I encountered the rough road. Upon the conclusion of the ride the gauge read 1015 kPa. I left the gauge in place for the remainder of the day but took no further readings.

I have removed the gauge from the bike and it now reads 220 kPA when unpressurized. I disassembled it and there seems to be no internal damage to the gear teeth. I think the thing somehow skipped a few teeth when being vibrated.

In any case, I think I am done with it and will send it back to Neville for rehab and future use.

As for prediction, any one point can be used as a starting “precal” and the trendline used to predict one of the other points as “postcal”. Since most of the points do not lie exactly on the trendline, there is error, but it is small.

I believe the technique to be useful under some conditions, but the fragility of the gauge has put me off further experimentation.
Last edited by peteriegel
Sounds like you did not set up a temporary calibration course at the race-course location as I thought is your normal practice, and therefore you did not generate the data I was hoping for.

I do not think your method for calculating the calibration factor from the pressure reading is the most accurate. I take the calibration factor determined on the home calibration course at the start of the day and just subtract 0.138 rev/km for every rise of 10 kPa.
Calculated in this way, I have found postcalibration factors are in precise agrement with those determined by actually riding the calibration course.
Neville, your method depends on first taking a number of readings over various pressures to determine your factor of 0.138 rev/10 kPa. I have four trials over a range of pressures. They all give about the same, but not exactly the same, slope.

When you are done you will not have a nice perfect straight line.

You seem to think that once I have done this I need to calibrate, then go somewhere else and calibrate again. Instead, I simply calibrated later the same day on the same calibration course. Using whatever my own factor is (probably close to 0.138 or so) I can then predict the second calibration from the first. How is that different from doing it on two different calibration courses?

I’ve got a pile of data points, and if I use them to calculate my calibration factor, how is what I have done different from what you have done? The method certainly does not require that calibrations be done as part of a real measurement – only that they shall be done. That I have done. My conclusions are no different from yours.

You can pick any of my calibrations and predict what the rev/km would be at a different pressure later in the day. You can check for accuracy by seeing what the rev/km actually was. It’s not all that complicated.

Can you be more specific in outlining where I did not generate the data you were hoping for? I believe I have confirmed your conclusions.
Last edited by peteriegel
Example of Pressure Predicted Postcal Using Pete's Data

Trial kPa All over 750 Temp, F
3 04/08/06 7:45 771 464.90 464.90 38
3 04/08/06 15:00 800 464.62 464.62 50
3 04/09/06 7:55 760 465.12 465.12 32
3 04/09/06 16:00 805 464.48 464.48 58
3 04/10/06 7:30 777.5 465.03 465.03 38
3 04/10/06 7:35 775 465.12 465.12 38
3 04/10/06 15:50 812.5 464.35 464.35 66
3 04/11/06 7:30 760 464.97 464.97 40
3 04/11/06 16:05 825 464.21 464.21 76
3 04/12/06 7:30 785 464.67 464.67 59
3 04/12/06 14:17 805 464.30 464.30 69
3 04/13/06 7:25 760 464.93 464.93 48
3 04/25/06 9:10 680 465.58 58
3 04/29/06 7:00 620 466.04 45
3 04/29/06 13:42 675 465.47 68
3 05/02/06 6:55 640 465.92 60

From the above I see that if I use ALL the data points, rev/km increases 0.84 for every 10 kPa loss of pressure.
If I use only the data points above 750 kPa, rev/km increases 0.135 for every 10kPa loss of pressure. Above 750 is better, as it is closer to the normal range I use.


Consider the data of 4/9/06:
Precalibration = 465.12 rev/km at 7:55 AM at 760 kPa and 32F = 10993.67 counts/km

Postcalibration = 464.48 rev/km at 4:00 PM at 805 kPa and 58F = 10978.73 counts/km

Predicted postcalibration = 465.12 - 0.0135 x (805-760) = 464.39 rev/km =
10976.55 counts/km

So, we have an error of 2 counts/km between pressure-calculated postcal and actual postcal. This, in my opinion, is decent agreement.

Now consider the data of 4/10/06
Precalibration = 465.12 rev/km at 7:35 AM at 775 kPa and 38F = 10993.67 counts/km
Postcalibration = 464.35 rev/km at 3:50 PM at 812.5 kPa and 66F = 10975.44 counts/km

Predicted postcalibration = 465.12 - 0.0135 x (812.5-775) = 464.74 rev/km =
10984.73 counts/km

Here we have an error of 9 counts/km. This agreement is not so good.
Last edited by peteriegel
I am new to this forum, and relatively new (about 15 courses) to measuring.

However, I must weigh-in. Why the concern over tire pressure? How 20th-century! There is discussion about advancing to electronic counters, but I haven't seen a thing about another important advancement in technology that all measurers could benefit from. It's called "airless tires".

I have been riding airless tires for 10 years, and love them. No flats! They are about the same price as a regular tire, and can be ordered with specific pressure-equivalence (durometer rating).

I get mine (both road and mtn. bike) from No, I don't have a financial interest in them. Last I saw, if you purchased two tires, they gave you the mounting tool.

I highly recommend the airless tires. My counts-per-mile vary from 18455 at 90° to 18486 at 20°. And, I rarely have to change a course due to tempurature changes between calibrations.

Just wanted to put airless tires out there.
Airless tires have been around for a long time. In 1983 Bob Letson had a solid tire on the measurement of the Los Angeles Olympic Marathon course. Around that time a hula-hoop-like plastic tire insert called "Eliminator" was available. Later Suretrak, Capair and Greentyre became available.

While they do provide protection against flats, they vary in size over the day, just as pneumatic tires do. Also, some people, like me, prefer pneumatics because they generally possess a kinder ride and less rolling resistance.
Yes, the airless tires do vary in size, but not as much as pneumatic tires, since solids vary in size due to tempurature much less than do gasses.

The old tires were solid and unforgiving in the ride. However, the new tires can be ordered in various pressure-equivalents, allowing one to tailor the ride to what they are used to.

I do dispute the "less rolling resistance" claim, though, because if you have a "kinder ride", that indicates lower pressure, which in turn means more tire surface in contact with the road, which increases rolling resistance.

My bottom line is, if I have the same ride I am used to (on my road bike, that is 120 psi, on my mtn. bike it is 60 psi), and I don't have to worry about flats or using a pressure-sensor to make on-course calculations, it is very much worth the switch to airless (as opposed to the old "solid") tires. Personal preference, which is what this whole discussion boils down to. I was just offering an alternative in the discussion on monitoring tire pressure.

To each our own.
You don't need to worry about pressure-sensing. That is just a trial balloon that Neville has sent up. It has some technical advantages, but is definitely not part of the mainstream measurement method.

So it comes down to whether you like the ride of a solid tire or not. The bulletproof aspect of it is a definite plus.
I think it was John Sissala who told me that he was given some solid inserts for his tires many years ago and has since got very constant calibration factors with them. He now treasures them because he cannot get anymore. It is generally agreed that airless tires only work well at slow speeds over smooth surfaces, but John says he is quite happy taking things slowly. Although I have not tried them, airless tires would seem good for course measurement if you consider worthwhile the expense and trouble of reserving a special wheel for that purpose only. The main purpose of postcalibration is to check for slow leaks in pneumatic tires, and with airless tires therefore it is probably best not to do it in most cases.

With pressure monitoring of pneumatic tires the measurer enjoys all the benefits of airless tires: good predictability of calibration factor, freedom of worry concerning slow leaks and punctures, and the avoidance of postcalibration. Additionally though, accuracy is better and bicycle ride superior.
Hmm. I think the new tires are much better than old styles. I ride my bike up to 50 mph (not while measuring), and over all terrain, and am quite happy with them. is where I got mine.

I still do the post-measurement calibrations, as drastic tempurature changes do change my clicks-per-mile occasionally. I can do my pre-measure calibration in 25° weather, then post-measurement in 65°. That makes a difference of maybe 4 clicks per mile. Not much, but enough.

I don't understand the "avoidance of postcalibration" comment, though. I was under the impression it is required, regardless of type of tire. Am I mistaken?


If you do a precalibration at 20 deg and a postcalibration at 65 deg with an airless tire you will always find that the calibration factor is larger for the precalibration. Since preference is to go with the largest factor, this means that postcalibration is a waste of time. As an alternative one can go with the average factor, but this means going back to the race course for an adjustment. Depending on the temperature at which the race course was measured, this adjustment may or may not be an improvement.

Your state certifier has the last word as to what is an acceptable method. If you wish to use a method that does not involve postcalibration,you will have to convince him that it is valid.
Pete's Data
The differences between gauge and ride for your postcalibration factors are 0.006% and 0.056% (your final numbers indicate 0.019% and 0.084%, respectively, but you made calculation errors). Some of my similar numbers in recent certifications are 0.0050%, 0.0026%, and 0.0007%. Getting the calibration factor by reading the pressure gauge can therefore be just as good as that derived by riding a calibration course.

Part of the reason for the poor result of 0.056% is probably that you used data obtained at 38 deg F. I think that all your data obtained at 32 and 38 deg is out of line with the rest of your data and have eliminated it in the plot below.

As you can see it now looks quite uniform.Remarkably, considering you used a slightly larger wheel (27 in) than mine (700) and operated at a much higher pressure, you get almost exactly the same pressure coefficient as myself:1 rev/km/bar.(From other evidence I now feel that a modified coeffient of 1.33 rev/km/bar ahould be used in measurement to cover both major and minor temperature effects.)

Last edited by neville

We differ on a couple of fundamental issues.

1) When tire size is measured across a wide range of pressure the resulting points form a curve, not a straight line. So, if I am operating on one narrow part of the curve, my size change will be different from what it would be on another part of the curve. For this reason I think it is appropriate to use only those data points that are close to my normal operating pressure range.

2) Throwing away a couple of data points because they don’t fit the neat straight line is not very good scientific procedure. After all, the tires operate under all conditions of temperature. The fact that a few points were taken at low temperatures does not justify omitting them from the data. They are a legitimate part of the data which make up the size vs pressure algorithm.

All in all, I think your pressure method probably gives reasonably accurate results, but I do not believe it is as accurate as you believe it to be. When you ignore the inconvenient scatter in the underlying data you are fooling yourself.
I agree we are primarily concerned with those pressures close to our normal operating pressure, but I find a very linear response of calibration factor all the way from 50 to 85 psi with a tire whose maximum rating is 125 psi. (It becomes precisely linear at constant temperature, but this is not a practical case during actual course measurement.) When I plot all your data in the range of your operating pressure (103-135 psi) I still get 0.9363 rev/km/bar for the pressure coefficient very similar to that obtained in the wide pressure range above. You get 1.35 and I conclude in my forthcoming report in measurement news that to cover all temperature effects 1.33 should be used. However, I do not think we should be concerned with the precise value, because as it turns out in practice, pressure changes are not great and it does not make a lot of difference with the calibration factor.

As far as accuracy goes though, note that if there is a temperature change during measurement that the pressure monitoring is not just reasonable accurate but a lot more so than the “official” method. This is because unlike the “official” method it corrects for temperature change, and even in the worst possible case results are still better.

Take for example your worst data quoted above:

4/10/06 7:35 am, 38 deg, 775 kPa, 465.12 rev/km
3:50 pm, 66 deg, 812.5 kPa, 464.35 rev/km [press gauge =
465.12 - 0.0135*(812.5-775) = 464.61]

If the first measurement had been a precalibration for a 5-km course measurement at 3:50 pm using pressure monitoring, then the course would have been long by 9 feet:

(464.61 – 464.35)*5 = 1.30 rev = 9 feet

This sounds a lot but measurement by the “official” method would have given a course long by 27 feet:

(465.12 – 464.35)*5 = 3.85 rev = 27 feet
Last edited by neville

Add Reply

Link copied to your clipboard.