Does anyone have a backup of the pictures from this thread?

I'll be attempting this upgrade at some point, as soon as i can get hold of a good Golf AFM, preferably NOS.

The criticism i have of the manometer method is that it only uses one calibration point not the standard two, but maybe it only works this way. A more precise method would be to use the voltage output from the AFM, and assuming the output is linear, obtain a Volts/1000rpm calibration figure. I know from calibrating process instruments that the correct method is:

1) Perform a zero trim at the lowest obtainable process value. This could be done at 1000rpm to make the maths easy.

2) Perform a range/span adjustment at the highest obtainable process value. This could be done at 5000rpm to make the maths easy.

The zero trim would be done using the CO screw and the range calibration done using the spring tension under the cover. The manometer at idle method clearly works, but the more span you have between your calibration points the better; the less span, the greater the error. It's normal to have to make several adjustments at your high and low values in order to achieve the desired result. So the AFM pins are as follows; 1-Air temp thermistor, 2- volts out, 3- volts in, 4- ground.

The calibrated multimeter you would ideally be using would be connected between 2 and 4, and with the connections made on the car loom side you could obtain the voltage values to which the GT AFM is calibrated. The next stage would be to lash the Golf AFM in and set its output up the same as the GT AFM using the low and high points, by making zero trim adjustments on the CO screw and range adjustments on the tensioning spring. If the output is not linear this would not work, but i can't see why it wouldn't be linear.

It goes without saying that retrofitting an aftermarket ECU and hotwire MAF might be more common these days, but you also have to retrofit the whole lot with whatever sensors your new ECU works with, and ideally you'll have the car off the road for however long it takes. One thing i learned about sensor retrofitting on another car is that it goes wrong very easily.

Care would need to be taken to keep the control loop as clean and closed as possible, hence why it was suggested that the lambda probe be disconnected and the vac hoses for the heater valve and canister be junked. All exterior influences that might create error would have to be eliminated or nullified/compensated for. I'd be tempted to keep the air preheat valve, but convert to to manual operation with a solenoid or summat.

I've arrived about 15yrs too late for this party, so i don't expect anyone will be able to verify this method.

Edited 17 March, 20251 yr by mr brightside

Bosch air flow meter restoration: summary

This guy has done a very comprehensive write-up, it's shame he hadn't written this 15yrs ago. He thinks its best to calibrate them using weight on a wire rope.