Rev Counter Upgrade
Checking the rev counter against a digital timing light showed it was reading around 20% low. There is no adjustment to correct this and it's a common fault of the design. The rev counter operates by reading the voltage being output by an AC generator which is driven off the rear of the inlet camshaft, this gives about 1 volt for each 100 rpm of the engine. The rotor of the generator is constructed from permanent magnets and it's not unusual for them to lose strength with age resulting in low indication of engine rpm. I have heard that new generators are expensive and variable so will not necessarily restore the accuracy. The generator gives a sine wave output of 4 Hz for each revolution of the camshaft or 2 Hz for each crank revolution. This signal is fed to the moving coil meter via a precision wound resistor and a bridge rectifier, the resistor reduces the incoming voltage and the rectifier converts the voltage to dc to feed the meter.
The amplitude of the generator output is unreliable but it's frequency is directly related to rpm so converting the frequency to a voltage and using that to drive the meter provides a good solution. The electronics to achieve this can be mounted inside the meter and apart from providing a fused +12 volts supply to the circuit no changes to the car are necessary.
The circuit is based on a LM2917 frequency to voltage converter IC, the circuit and a suggested board layout are shown here, click on the image for a larger view. This circuit only works on negative earth.
Remove the resistor and rectifier from the original SRBP board, just cut the wires and leave the blade terminals in place taking care not to pull on the red and black wires connected to the meter movement. Fitting the PCB with short 4BA screws is tricky but it can be done without removing the instrument face and pointer. The +12 volt supply wire can be fed out through the hole in the SRBP board created when removing the rectifier.
To adjust the trim potentiometer some sort of function generator is required, a software one is available here: Marchand
Having setup the circuit at 100 Hz to indicate 3000 rpm I find the reading is accurate to within 5% from tick over to 4000 rpm. Above 4000 revs the indicated rpm reads progressively higher due to non-linearity of the meter movement but this is perhaps preferable to a low reading. The maximum error in my setup is 11% at full scale deflection of the meter, an indicated 6000 rpm is achieved at 5400 rpm (calculated using a function generator).
Comparing the modified rev counter with the original at 1000 and 2000 rpm. When I receive the repaired clock from JaguarClock the quartz watch will be replaced and the modified rev counter will be installed.
The clock is back after a great upgrade, thanks to Mike Eck. The modified gauge is installed and it is now responsive and accurate. Click on the picture for a .wmv file of it in action.
