Author Topic: Automotive knock sensor circuit  (Read 727 times)

0 Members and 1 Guest are viewing this topic.

Offline injbTopic starter

  • Regular Contributor
  • *
  • Posts: 64
Automotive knock sensor circuit
« on: March 30, 2020, 12:29:20 am »
I have a knock/boost control computer from an old car (the KLR from a Porsche 944 ) and I'm trying to figure out how it works. I'm mainly interested in the firmware, which I have disassembled. But I'm puzzled by the setup of the knock sensor. I've hit a wall in trying to understand the code, until I can understand a little about the circuit. So I've made an attempt to draw a schematic for it. This isn't complete, and I hope it makes sense as it is.

I'm finding it hard to trace the rest of it, but I hoped that this would be enough to get the gist of what's going on. I've attached my best attempt at drawing the schematic - sorry if it's a bit rough!

Here's what I know: the input comes from a piezo knock sensor attached to the block. The output goes to an ADC which is read by the microcontroller (an 8048). The S1 chip in my diagram is a CA3080 and S2 is an LM2902N (quad op amp). I have not drawn the input to the second op amp in the 2902 - it comes from another one of the op amps but I haven't traced it fully yet. I am pretty sure there's 2 more op amps involved and one of them sends another output to the ADC.

959542-0



The questions that puzzle me the most are:

1. Why does the microcontroller send pulses to the amplifier input, coupled by C109? My only sensible guess is that this has to do with detecting a missing knock sensor (based on the explanation here: https://www.renesas.com/us/en/www/doc/guide/an9770.pdf). The idea is that normally the pulses are absorbed by the capacitance of the knock sensor, but if it's missing then these pulses drive the output to saturation. Could that be what's going on here? I've tried making sense of it from the microcontroller code, but the logic is just not clear to me.  There are definitely conditions associated with sending the pulses, but I don't understand them very well.

2. What is the significance of pin 5 (bias input) of the CA3080 being fed by the output of the second op amp? I've looked at applications of this amp in the datasheet but I haven't seen anything that looks similar to this. Usually it's fixed.

Any help in understanding this would be great. And if I haven't drawn enough to make any sense of it, let me know I'll have another go when I have a chance. Thanks!




« Last Edit: March 30, 2020, 12:32:11 am by injb »
 

Offline injbTopic starter

  • Regular Contributor
  • *
  • Posts: 64
Re: Automotive knock sensor circuit
« Reply #1 on: March 30, 2020, 11:14:38 pm »
Well I discovered some more about what's going on. It turns out the that the pulses that the ucontroller sends have a frequency of 5.7Khz, which I think is exactly the frequency of engine knock on this car. So now I'm convinced that this is a self-diagnostic feature. However, I don't think it's for detecting a missing sensor as I suggested. After studying the code for a while, I believe that the computer tries to induce a fake knock signal and then read the output. If it consistently fails to get the expected response from the ADC, then it throws an error code.

I think I'll have to figure out more of the circuit to understand it fully though.
 

Offline Ground_Loop

  • Frequent Contributor
  • **
  • Posts: 659
  • Country: us
Re: Automotive knock sensor circuit
« Reply #2 on: March 31, 2020, 01:37:06 am »
The CA3080 is a transconductance amplifier and pin 5 is used in part to establish the gain.
There's no point getting old if you don't have stories.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf