BMW S62B50

[jsmcortina]

EDIT: The dyno owner suggests me that maybe the Ms doesn't had good ground connections. I need to make more. I have two 2 mm wire (one for mainboard, other for ms3x), one 2 mm wire for bip373 ground and 2 more for iox board.

Certainly worth investigating.

James

[ragepower]

EDIT: The dyno owner suggests me that maybe the Ms doesn't had good ground connections. I need to make more. I have two 2 mm wire (one for mainboard, other for ms3x), one 2 mm wire for bip373 ground and 2 more for iox board.

Certainly worth investigating.

James

There is some test I could do? Measuring some current or voltage? I have an amp clamp. There is some max current for this situation? Or I use the standard limits based on gauge wire?

[ragepower]

EDIT: The dyno owner suggests me that maybe the Ms doesn't had good ground connections. I need to make more. I have two 2 mm wire (one for mainboard, other for ms3x), one 2 mm wire for bip373 ground and 2 more for iox board.

Certainly worth investigating.

James

There is a max voltage drop for ground?

[ragepower]

I have a feeling that the module does knock not dialogues with ecu... the setting is correct...
check wiring of the module.

Checked. It is OK.

Can I check if it is working by setting the bandpass frequency filter to the max or to the min, so the knock gauge reach the max?

[ragepower]

EDIT: The dyno owner suggests me that maybe the Ms doesn't had good ground connections. I need to make more. I have two 2 mm wire (one for mainboard, other for ms3x), one 2 mm wire for bip373 ground and 2 more for iox board.

Certainly worth investigating.

James

I had re-wired the ground cables on the engine head, add 2 more, 1 for mainboard, another for MS3X card. Just in case.

- If I had the usb cable connected to laptop, I will have lost sync at 5700 / 6000 rpm. If I disconnect it, I do dyno run to rev limiter without lost sync. Due to this, I can't tune vvt to 6000 and 7000 rpm column and I can't do datalogs on dyno runs.

I think was because I have 1 usb cable inside the main box, that was rolled to be organised, maybe this had reduced the noise resistance.

- I have something happening at 5300rpm. Until then, I have a bit smoke (maybe 12.5 is too much this engine) then no smoke and the engine seams to knock. Maybe need more fuel but the afr isn't lower than 13.

Can I trust on WB sensor on high load and high speed?

And there is always something wrong, because the gases from the exhaust don't has a fuel smell but them "burn" my eyes.

I have a feeling that the module does knock not dialogues with ecu... the setting is correct...
check wiring of the module.

Checked. It is OK.

Can I check if it is working by setting the bandpass frequency filter to the max or to the min, so the knock gauge reach the max?

Can I do this?

EDIT: I will try to find out to check if the knock sensor signal is reaching to the knock module, using a scope

[ragepower]

I tried to check the knock module if it is working.

I used a analogue scope and on sclk, sdi and sdo, I have a signal from it, like 3 dot on scope screen. I checked ground and vdd.

A question, the ET pad on knock module should be ground, right?

I checked the knock signal and it seams to be working, on both.

I tried to set, on bandpass filter frequency to set the max, value and the min value and the knock gauge stay always on 2.4%

Can I do more tests?

I didn't do a log since the knock gauge stay always at 2.4%.

[Matt Cramer]

Did you have a good RPM signal while testing this?

One trick we've found for testing knock sensors is to download a sine sweep WAV file, wire a laptop sound card output to the sensor input, and data log what happens while playing the file. Be sure it covers the knock frequency range and that you have an RPM signal so it can do proper windowing.

[Peter Florance]

Did you have a good RPM signal while testing this?

One trick we've found for testing knock sensors is to download a sine sweep WAV file, wire a laptop sound card output to the sensor input, and data log what happens while playing the file. Be sure it covers the knock frequency range and that you have an RPM signal so it can do proper windowing.

Brilliant, Matt. I need to try that.

[ragepower]

Did you have a good RPM signal while testing this?

One trick we've found for testing knock sensors is to download a sine sweep WAV file, wire a laptop sound card output to the sensor input, and data log what happens while playing the file. Be sure it covers the knock frequency range and that you have an RPM signal so it can do proper windowing.

Hmm I just didn't understand that rpm part. How do you define a good rpm signal?

[Matt Cramer]

One that is the real RPM and not dropping to zero or otherwise clearly wrong.

[ragepower]

One that is the real RPM and not dropping to zero or otherwise clearly wrong.

I don't have sync lost. I have rpm signal all the time without dropping to zero.

So I play the file and some time I need to see the knock gauge increase right?

Do I need to disconnect the knock sensor to do it?

[ragepower]

I did the test like you said with the knock sensor connected.

And nothing. The knock gauge stay at 2.4%.

I attach the msq file and data log.

Any more suggestions?

[Matt Cramer]

You will want to disconnect the knock sensor for this test as the goal is to supply a fake knock signal with known qualities.

[ragepower]

No reaction on knock gauge.

I used a sine sweep from 20 hz to 20 khz.

What more tests I can do? Maybe check some signals? I have a scope.

EDIT: I disconnect the knock sensor.

[Matt Cramer]

What if you revert to the latest stable release (1.3.4) code?

[ragepower]

Will that release support the s54 4 vanos (s62)?

[jsmcortina]

I tested your MSQ on the bench and I'm seeing readings on the knock channels.

James

[ragepower]

I tested your MSQ on the bench and I'm seeing readings on the knock channels.

James

What could be the problem? Maybe the knock module chip "burned"? What kind of test I can do?

The chip has a test pin, is it connected to MS or can I use it for test? (setting high and check some of the pins)

While I waiting for your help, I will try to revert to 1.3.4 (Matt suggestion). Probably that wouldn't be the problem but it is one less option.

[ragepower]

I tried with 1.3.4 firmware, no change.
Loaded beta 15, no change.

I tried to diagnostic, it seams to me that ms will acquire knock value after each ignition event. So at CS pin and INT/HOLD pin should exist a frequency signal. I measured (with multimeter) and at idle ( +/- 800 rpm):
- CS pin - 200 Hz
- INT/HOLD pin - 0 Hz

INT/HOLD pin should have a frequency like 200hz, right?

[ragepower]

Today I did more tests.

I turned the knock module to off on TS. By using programmable outputs, I tested the knock module pins (LED output and JS11) by active them with tps (> 30%).

The Led outputs change from 0V to 5V, but JS11 stay always at 0V. It isn't suppose to do this, right? It is suppose to go to 5V too?