Changing the stock gearing on a modern sportbike is practically manditory for honest sport riding. One disturbing side effect of this is that the speedometer goes drastically out of calibration, reading a much faster speed than actual. This should be corrected for honest information on speed, a lower odometer reading, and accurate information for the bikes on board computer.
- 1 Stock speedometers read fast
- 2 What the law says
- 3 What kind of error is expected
- 4 How the speedometers works
- 5 Why a speedometer is rarely perfect
- 6 Testing speedometer error
- 7 Testing odometer error
- 8 Commonly available calibrators
- 9 How the calibrator works
- 10 The Math
- 11 Wiring
- 12 Wiring options
- 13 Questions
Stock speedometers read fast
All vehicles come with speedometers that read a faster speed than actual. Since they can never be exact, Speedometer systems are designed to err on the side of higher speed readings than actual. Two good reasons exist for this; the lawers are happy since no on can claim that their bike was going faster than they thought and that's why they got hurt, and because a slow bike can seem like a real fast bike if the display says so.
Also, in addition to increaseing the error displayed on the speedometer, re-gearing is going to mess up the bikes speed data. The computer may use this information for a host of fuel and ignition calculations, so it is worth having right.
What the law says
Findig the exact laws regarding this topic is difficult.
With regard to manufacturing, speedometers are certified by manufacturers to follow EITHER the U.S. SAE (Society of Automotive Engineers) J1226 OR the European version, ECR-39. The SAE rule, published in 1983, allows readings to be plus or minus 2 percent of true speed, or on a sliding scale from minus 1 percent to as much as plus 4 percent.
The European standard does not allow underestimation of speed, but does allow overestimation of up to 10 percent plus 4 kph.
SAE J1226 Electric Speedometer Specification--On Road http://www.sae.org/technical/standards/J1226_198302
SAE J966 Test Procedure for Measuring Passenger Car Tire Revolutions Per Mile http://www.sae.org/technical/standards/J966_200011
Odometers are covered in DOT PART C, TITLE 49, Chapter 327, but it makes no mention of accuracy.
What kind of error is expected
My bikes have all read roughly 10% over actual. Generally, expect to see between 5% and 10% over reading in stock form.
How the speedometers works
Most modern sport bikes use a pulse generator on the countershaft in the gearbox to gauge bike speed and distance traveled. Each rotation of the countershaft produces a single pulse. As the bike moves, a pulse frequency is generated. The ECU measures this frequency, does some simple math, and generates a velocity readout for the dash and odometer. The ECU also counts all of these pulses to calculate how far the bike has traveled in it's life for the odometer.
In the past, a mechanical system of gearing a spinning shaft from the rotation of the front wheel was used to generate this data. The older system would add error to the speedo anytime the front wheel or tire changed diameters.
The current digital system will add error anytime the rear wheel or tire changes diameter AND when the final drive gear ratio is changed.
Why a speedometer is rarely perfect
In the event that a speedometer is perfectly calibrated, that calibration will not last long:
- The tire diameter grows as speed increases, so a speedo can only be "corrected" to one speed.
- Tire wear will change the wheels diameter, causing errors.
- Chain wear will increase the effective chain pitch, causing errors.
- Rounding of data values within the electronics for disply.
Testing speedometer error
Testing the speedometer error involves getting a precise speedometer reading at the same instant as getting a precise actual speed reading. The bike must be moving at a constant velocity, with no acceleration or deceleration.
Since it is impossible to have an accurate calibration over the entire speed range that the bike is capable of going, a target speed should be chosen where accuracy is most needed. 55 mph is probably one of the best targets since it is a common speed limit.
- Driving by fixed police radar speed boards.
- Assistance from a police officer or someone with an accurate hand held radar gun.
- GPS Device. WAAS w/SiRF StarIII prefered.
If you plan on using a GPS unit of any kind, make sure to take readings at an extremely steady pace. Most common handheld GPS units only take data readings at 1 Hz (once every second). That means that it takes at least 2 to 3 seconds to get an 'accurate' velocity reading. If you have access to a GPS unit that samples faster than 1 Hz, it will help.
Even after using the GPS to get the calibration close, you may have to fuss with the correction factor to truely sync with the GPS readings. This is because it is difficult to read the two readouts of the speedo and the GPS unit at the same time, and also because both have considerably crude rounding of data to render them usable.
Once you have a tested and proven correction factor given any gearing, you can correct for gear changes without testing the vehicle again.
Testing odometer error
It is good practice to test odometer error as well as speedometer error. The dash and the computer get data from the pulse sender. They may require different calibrations.
If someone can provide some information about wether the odometer reads correctly after a dash swap, this would be very telling as the where the data is stored and if it uses separate processing to record.
Commonly available calibrators
- Yellow Box Speedometer Calibrator (www.blackrobotics.com). Calibration range +/- 28% in increments of 0.25%. Requires splicing or custom sub-harness.
- Speedohealer (www.speedohealer.com). Calibration range +/- 99.9% in increments of 0.1%. OEM connections. Saves max speed.
How the calibrator works
Simply, the calibrator takes the pulse from the countershaft sender, and up/downscales the frequency to produce an accurate speed reading at the dash or the computer. This scaling factor must be calculated using tested data and simple math.
If you are calibrating with stock gearing:
If you are calibrating with non-stock gearing:
A handy spreadsheet I made:
Careful study of the bikes wiring diagram (Honda) reveals that the pink (pulse) line splits just after the pulse sender with one end going to the ECU and the other going to the speedometer. If the pulse is not corrected JUST after the pulse sender, the ECU gets one pulse while the speedometer will get another. This could have poor affects if any speed based fuel compensation for the ram air and gear based mapping features on most current bikes. The calibration MUST take place AT the pulse sender.
Clean, well thought out, professional wiring is critical on these parts if the bike is expected to function properly throught it's service life. Honda now uses the new sealed 3 pin connectors for their speed pulse senders. It is prefered to not cut the harness if possible, so track down the correct electrical parts before starting the job. I made a custom jumper using some connecting block parts (Honda 3-pin,3P disconnects).
The length of the wires that come on the Yellow Box are just long enough to be routed to the rear of the bike's glove box. When installing the Yellow Box, make sure to cut as little of the Yellow Box's wires as you have to or it will not reach the best mounting position.
One option that you may consider is jumping an auxiluary power line off of the pulse sender line. A simple two pin power connector cam be split off the same line as the calibration box. Auxilary equipment such as phone chargers, radios, or laptimers can require a 12 volt power supply. This is better than jumping off the battery since the line goes dead when power is off on the bike. Jumping off the battery requires an additional fused switch.
- Are odometer values stored in the computer or the dash? I expect the computer, but I have not tested this.
- What does the computer do with speed data, exactly?
- What are the specific laws regarding speedometer and odometer accuracy?