Problems solved.
After reverse engineering the schematic, it's very close to the recommended design in the LM1815 datasheet
http://www.ti.com/lit/ds/symlink/lm1815.pdfDid I find 2 problems. I have the old LM1815 boards, where the solderbob select the output devider. On these boards are pin 5 on the LM1815 hardwired to Vss +5V = mode 2 selected = arming voltage 350mv equal that the pk-pk V on the Vr sensor must be around 1V. The std. BMW abs sensors only produce 300mv @ 10 mph
So the LM1815 must be in mode 1, this is done by lifting pin 5 on the chip of the pcb, so it's floating instead of beeing +5V. Unfortunatly the connection from pin5 to Vss is under the chip, so it's not possible to cut the copper.
After selecting mode 1 on the LM1815, do it arm @ 30mv, and the sensor trigger without any problem. Noise is around 3mv so dont seems to give any problems.
Second problem.... I'm using Vems v3.8B hardware, where there are pull up resistors for both wheelspeed channels, mounted on the board. Input pins are floating @ +5v unconnected. The output from the HEF4024BT chip, is routed thru an 1kohm resistor, and with the internal pull-up in the VEMS are the output voltage square between 1,5v and 5v, and this is too high for the cpu to trigger. Short term solution was to bypass the series resistor on the VR converter, this made the signal pulsate between 0v and 5v, and both wheel speed channels were working. But better solution is to remove the pull-up resistors on the Vems board, and just use the VR converter to supply +5v / 0v for the cpu.
Lot of work, but solution found.
With thanks to Peter Jensen from Vems.dk, did I get a link to the newest board documentation
http://www.vems.hu/wiki/index.php?page=GenBoard%2FVerThree%2FSchematic And that clearly show the problem. Wheelspeed input 1 + 2 have a 2.7k pullup resistor to 5v, and with an 1k series resistor on the HEF4024 output will the "low" voltage be 5v/3700*1000 = 1.35v. Removing the series resistor on the output on 4024, will make the low voltage = 0v. the current will be 5v/2700 = 1.85mA which is ok as the HEF4024 is rated with 10mA on the outputs.
But please check the new VR to Hall board for output resistor value to ensure it will work as expected.
Thanks