Why pulling down the secondary VR Signal?

[fphil]

Hello all,

I got some random SecTrig errors which commands Vems to resync and which may induce, when full load a bang on the car transmission. No PrimTrig error

I had a look at the 3.6 - 3.8 schematic for the VIN circuits of the two LM1815 chip.
One major difference is the pulldown resistor R59 1k on the secondary trigger sensor signal

It is said on the wiki page InputTriggerHardWare:
".. ◦ 1k pulldown resistor between VR input signal and GND
    ◦ this is effectively in parallel with the VR sensor. ..."

The internal resistor of the sensor (Marelli Sen 8D) being about 0.7k this divides about by a factor two the sensor signal (following specs >400mV p-p @ 40rpm).
What for? There is nothing to pull down from a VR sensor, this is not a logic source.

It is said further:
"To make it work with the 1k VR-input pulldown resistor, connect LM1815 pin5 to +5V with the SJ4 (3pin jumper): solder the mid-pin to the +5V pin that goes LM1815pin8"

Hence, this way,  LM1815 mode 2 is selected and the input arming threshold cannot be less than 200mV (signal less than +-200mV are ignored), this can be a good thing, but what's the matter with that  1k "pulldown" resistor?

Thanks for your understanding

[fphil]

Alright, the 1k R59 resistor should lower the noise on the signal input by a much higher factor, ideally under the 200mV threshold of the LM1815 mode 2. On the other hand VR Sen8D signal should trigger at somewhat higher rpm than 45rpm, which does not matter.

So to get rid of false triggering generated from noise on VR signal, there are probably several strategies :

- avoid the noise on VR signal as much as possible,

- go for mode 2 and R59 (ik or other small value in any case LM1815 clamps  the signal In ),

- add a bias on the VR signal as done for TRIGG1 through R37 - R30 this moves up or down the noise and avoid any zero crossing from it (one people of the forum has taken this strategy I remember),

- change the peak detector capacitor, a bigger one allows a slower decay which should keep an higher arming threshold and prevent the noise to trigger.


(btw on the V3.8 schematic I do not see how VCC comes to Pin 8 of U12 LM1815)

[fphil]

Another way to get rid of false triggering is to lower the noise on the genboard input signal with a capacitor between the input line and ground. This make a RC filter where R is the sensor internal resistance (700 ohm).

In my coil type case I measured the period of pulse @240rpm as 40ms hence @ 7200rpm it should be 2/3ms. I take a cutting frequency of 3/2kHz which then gives C=150nF about.

I put a capacitor of  100nF (what I got) - It seems to work.

I realize that the capacitor C30 which equally makes an RC filter along with R90 on the input VIn of LM1815 has been discussed on Wiki where it is guesed that C30=100nF should be ok for coil type triggering. A similar computation in my case would give about 5-10nF for C30.