Toms 1999 VTi 4WS LHD

Post by **NafemanNathan** » Thu Feb 28, 2019 1:36 pm

Are you able to try another ECU?

89 UKDM BA4 4WS (resto job!) · Post by **wurlycorner** » Thu Feb 28, 2019 1:42 pm

Absolutely sexy ass mo-fo work.
I am well jealous!

Have you followed the fault finding guide for the knock sensor in the manual? And yep, 3k is the magic number on the revs where things should start happening from the knock sensor... (see page 11-87 of the service manual on)

While you can check for continuity through the wiring, you can't check the knock sensor readings using a multimeter - you will need an oscilloscope, back-probed while it's all plugged in (it's basically acts like a microphone, setting up resonant frequencies back to the ECU).

Worth noting that knock sensors are very easy to damage during fitting, if they're over-torqued etc.

89 UKDM BA4 4WS (resto job!) · Post by **wurlycorner** » Thu Feb 28, 2019 1:48 pm

Have you still got the original knock-sensor that you can try to refit?

Also a few other random possibilities (these are not likely, but are possible...)
A bad earth could be causing electrical noise on the knock circuit, which could potentially be giving an incorrect frequency reading.
Something loose/rattling on the engine could also be setting up a noise that has the same frequency as knock.

nb: at 3k and off load, you're unlikely to hear knock yourself. The sensor is more sensitive than the human ear.

Have you checked the timing at idle is correct?

Are you 100% sure that the static timing is correct after the cambelt change? (worth re-checking)

Post by **tom023** » Thu Feb 28, 2019 2:26 pm

No, I broke the old sensor removing the wiring harness from the engine.

Timing at idle has been done although I will re-check it. By static timing do you mean all marks in their correct position at TDC? If that is the case then yes they are definitely correct as I had to check them various times when doing the valve clearance.

Post by **tom023** » Thu Feb 28, 2019 4:00 pm

Regarding possible external vibrations setting it off. I tested the knock sensor leaving it plugged in but unscrewed, connected a wire from the threaded part of the KS to the block to ground the sensor and started the engine. This should negate any possible engine vibrations going through the block to the sensor (or at least diminish them drastically). I started the car up and the error code came back exactly the same.
Also on Hondash when I bring up the Knock retard (%) graph, it always says 0. Even as soon as the error code appears. I'm not sure exactly what this measures (I assume the percentage of timing retard due the engine knocking) but it seems off that it gives 0 reading whatsoever. If the engine was knocking and the sensor working correctly, surely I would see the KS working before the error code. So in theory, the engine is not knocking...

89 UKDM BA4 4WS (resto job!) · Post by **wurlycorner** » Thu Feb 28, 2019 4:40 pm

But like I say, the knock sensor isn't an 'on/off' switch - it generates a frequency that the ECU monitors and compares against known good ranges.

With the sensor not screwed into the block, it won't generate any frequency.

If the ECU doesn't receive any signal from the sensor when it is expecting to (the ECU knows the engine is running and at what revs, after all) then it will assume the knock sensor is faulty, flag a fault and go into protect mode.

Post by **tom023** » Thu Feb 28, 2019 7:42 pm

Ok, I don't really know how to put this or what has happened so I'll just say it.
After checking timing and looking for any possible rattles, checking engine earths in general, everything came back ok. Out of desperation I decide to take a knock sensor off of a 4gen h22a black top donor that has been sitting out in the sun and rain for 4 years. The KS was in 2 bits, covered in rust and mud. I superglued it together hillbilly style and fitted it.
Unsurprisingly holding the engine at 3k rpm it triggered error 23. Then it occurred to me to erase the error and this time sit at 4000rpm... No fault. I've taken the car out for a quick drive and worked the rev range from 750-4000rpm and no fault while driving, even when cruising a 3000 rpm constantly.

So now I don't even know if the original new aftermarket sensors were faulty. Anyway, hillbilly sensor is staying in for now, and we'll see what happens next "shrug".

Post by **newkid** » Fri Mar 01, 2019 11:16 pm

Aftermarket knock sensors are a hit or a miss on hondas unfortunately. I got lucky when I replaced mine and the cheap ebay one worked fine. Only way to know its a good one from the off is oem but there pricey.

Post by **tom023** » Sat Mar 02, 2019 11:43 am

This is the conclusion that I am coming to @n@newkid . Also its apparently very important to torque the sensor down to exact specification. Until now I haven't been able to do that due to lack of space to get a torque wrench up there. I've purchased a socket that is hopefully just the right size to give me enough room for the wrench.

Post by **newkid** » Sat Mar 02, 2019 1:54 pm

Yea there not much fun to fit with inlet etc in place. Luckily for me I paid my friend's garage to do it