Knock sensor issues

[scott2093]

That's NOT what I'd expect. The knock sensor when activated, should cause the timing to retard for a few seconds. Not "permanently". Your regular timing advance should come back by itself, probably not more than ten seconds. Five, maybe.

Now I'm concerned again.

Whack it again, see if the idle slows, then recovers.

Ok. I'll do it in the am.. Tbh, I didn't let the engine warm up too much and it could have/probably was still in open loop.. not sure if that would matter or if I wasn't paying attention..pretty sure it stayed retarded though...

I just rewatched my old video in Post #10, and I had channel A set to 1V/division,
so with light tapping I was able to get a 6 division peak-to-peak output displayed
from the sensor alone. (Unpowered, not in circuit, just working into the high
impedance of the o-scope. Peak output could well be different in circuit.)

hmmm. so 3v?
Hopefully my little janky vice set up and balancing act with meter leads and force was was just different or these sensors are different specs?

I still want to get a new one. Especially now I have the plastic bag trick..lol

The old one looks pretty rough and the terminal had play where the new one didn't . almost like a plunger action where I actually made the extra effort to squeeze the connector to attach it to keep it from moving. Not sure if that's supposed to be like that or not.
I mean I did get the signal pretty easily and the timing did retard but, maybe it's supposed to be more sensitive. I'll test it's output tomorrow with it installed....
Was nice that it had proper ground at 14lbs torque.
Not sure if I feel good assuming the new sensor was goofed from the get go. It did take a bath in antifreeze and I did have to torque it to, would guess maybe 35lbs to get ground... Maybe that was from the bath..idk...
Thanks again!

 

[Road Trip]

Ok. I'll do it in the am.. Tbh, I didn't let the engine warm up too much and it could have/probably was still in open loop.. not sure if that would matter or if I wasn't paying attention..pretty sure it stayed retarded though...


hmmm. so 3v?
Hopefully my little janky vice set up and balancing act with meter leads and force was was just different or these sensors are different specs?

I still want to get a new one. The old one looks pretty rough and the terminal had play where the new on didn't . almost like a plunger action where I actually made the extra effort to squeeze the connector to attach it to keep it from moving. Not sure if that's supposed to be like that or not.
I mean I did get the signal pretty easily and the timing did retard but, maybe it's supposed to be more sensitive. I'll test it's output tomorrow with it installed....
Was nice that it had proper ground at 14lbs torque.
Not sure if I feel good assuming the new sensor was goofed from the get go. It did take a bath in antifreeze and I did have to torque it to, would guess maybe 35lbs to get ground... Maybe that was from the bath..idk...
Thanks again!

Hi Scott,

You've made significant progress. I would say that we have passed a major Go/no-Go test. And assuming no
other codes are present, then now you are all set to see if your original problem description in your original
post (ie: Pinging in Overdrive) has been cleared up at this point? Or do we need to pursue the problem a bit further?

****

As for the display on the scope, it's 8 vertical divisions total. I was seeing ~6Vac into the high impedance o-scope.
And since I had a resonant knock sensor torqued into the matching block, I was getting the full signal.

If I had taken that same exact sensor and put it into a vice, no doubt the KS output signal would be way less, for no
doubt the vice doesn't ring at the same resonant frequency as the engine block it was designed for does.
(In English, the knock sensor voltage response will drop off rapidly when stimulated outside of it's designed frequency response.)

****

So let's see if the original problem has been cleared up & go from there. By the way, IF the problem persists w/additional
codes, then we have to clear them all first. All fixed? Great!

But IF the pinging problem persists with NO codes, then in order to really sort this out you will need a way to look
at the real time data in order to figure out why is the computer goofing the spark timing calculations. Much much
better to troubleshoot w/real data as opposed to us recommending you spending your money loading up the parts cannon blindly.

At the very least, see if you can get a refund/exchange (or at least store credit) for that defective new knock sensor.
And if the new knock sensor happened to be a no-name brand, then I'd upgrade to an OEM-quality part before proceeding.

Let us know the result of your Functional Check Flight! :0)

 

[scott2093]

And assuming no
other codes are present, then now you are all set to see if your original problem description in your original
post (ie: Pinging in Overdrive) has been cleared up at this point?

Yeah there were never any codes. Which I guess might be weird seeing as how the ks wasn't functional? Or maybe I don't understand the way the test the computer runs work.
But I am anxious to see if I cleared up the sudden pinging. It's pretty disturbing and I'm not used to being able to duplicate making it happen. Maybe I've been lucky. Truck ran noticeably great for about a day after putting the new ks in. Weird. But I'm jinxed with coincidences and too often they can be not the good kind so I'm cautiously hopeful.

Or do we need to pursue the problem a bit further?

Hopefully not. I will only plan to let it warm up in the am to make sure the timing only retards temporarily after a knock now that I've got the green light..

And if the new knock sensor happened to be a no-name brand

It was an AC Delco. Came in GM bag from Summit.

Let us know the result of your Functional Check Flight!

Of course. Thanks!

 

[Road Trip]

Yeah there were never any codes. Which I guess might be weird seeing as how the ks wasn't functional? Or maybe I don't understand the way the test the computer runs work.
But I am anxious to see if I cleared up the sudden pinging. It's pretty disturbing and I'm not used to being able to duplicate making it happen. Maybe I've been lucky. Truck ran noticeably great for about a day after putting the new ks in. Weird. But I'm jinxed with coincidences and too often they can be not the good kind so I'm cautiously hopeful.


Hopefully not. I will only plan to let it warm up in the am to make sure the timing only retards temporarily after a knock now that I've got the green light..

It was an AC Delco. Came in GM bag from Summit.


Of course. Thanks!

AC Delco? OK, I guess we can take 'off-brand inferior imitation' sensor off of the troubleshooting table?

I still want to get a new one. Especially now I have the plastic bag trick..lol

The old one looks pretty rough and the terminal had play where the new one didn't . almost like a plunger action where I actually made the extra effort to squeeze the connector to attach it to keep it from moving. Not sure if that's supposed to be like that or not.

Listen, I'm glad that you brought up this terminal play. Not a good sign, especially if we are trying to eliminate
the source of an intermittent failure. This loose terminal on your original knock sensor could be a marginal path
originally, not losing the signal completely, BUT now the amplitude is lower/below the threshold...but, after
we've disturbed it with all this parts replacement, it (tonight) has now started working again...for the moment?

Q: Why didn't it fail the self-tests & throw a code? The only reason I come up with is a marginal path vs an outright
broken one? Loose connections can actually cause more problems than if it had just broken clean off. :-(

You should definitely replace your knock sensor. This time, let's test/verify the new part for signal before installation?

Then install/torque to spec. Perform the resistance check you did previously. And then fire up the truck & verify
it's working like you did before. (By now the rusty threads in the block have been cleaned by all this activity...but
bonus points if we are sure they are clean/conductive prior to the new part going in?)

Just trying to think all this through so that you can get this sorted out in the fewest number of tries possible.
EDIT: I know that all this extra testing of brand new parts just doesn't seem fair, but we are living in a brave new
world it seems...

Good luck!

 

[scott2093]

AC Delco? OK, I guess we can take 'off-brand inferior imitation' sensor off of the troubleshooting table?

Maybe? I ordered a speed sensor from Amazon, from the AC Delco store, a few months ago because I was working oot and it was the fastest way to get an oem. Part didn't look like my original but it has been working. Weird thing was it was in a GM bag, with a plain plastic bag inside that said made in Taiwan, China? You can see similar statements in the reviews which aren't super. Didn't know Taiwan was so proud of their roots. Or was that a GM move?idk

google Amazon

ACDelco GM Original Equipment 15547452 Vehicle Speed Sensor​

So I just ordered another knock sensor from the same Amazon AC Delco store because nobody locally can get one to satiate my impatience afaik, didn't want to wait for Summit and I need to get going to work. It'll be here Monday hopefully. At least it's reviews are better..almost perfect...

google Amazon

ACDelco GM Original Equipment 213-92 Ignition Knock (Detonation) Sensor​

So I tried testing the ac voltage output with the old sensor installed and I couldn't barely get 100mv..This is RMS actually since the meter is.... I could get 50mv fairly easy and just tapping the tensioner belt pulley registered the same as when I whacked the intake.. Not a good sign .

So I put the scanner on and hooked the ks back up, let the truck warm up for a bit, was still in open loop but it was topped out in temp.

Couldn't get a knock signal. Went under and wiggled the "plunger feeling terminal" and tried again. Got a knock signal but no timing retard.

Went under and adjusted again and now can't get a knock signal. It's pretty loose at the ks positive terminal housing...it'll plunge up and down and rotate in a circle (think witch stirring the pot rotate) . I'm sure I moved it from the magic spot I must have happened on last night. Maybe I was right after all deciding to go for replacing it blindly...

Jeez...

 

[Road Trip]

I'm sure I moved it from the magic spot I must have happened on last night. Maybe I was right after all deciding to go for replacing it blindly...

Scott, I have been in this exact same situation many times over my troubleshooting career. In Road Trip's perfect world,
something is either completely good or completely bad. Change part, fix machine, get banana. One and done.

But in the real world populated with intermittent problems, you end up with these 'peek a boo' marginal electrical paths.

Thanks to all the heat & vibration since that original knock sensor was installed at the bottom of your engine block,
the knock sensor terminal supporting the wiring harness attachment has developed a marginal physical connection
internal to the sensor, which you have discovered and described clearly.

From the troubleshooting perspective, Good News! You have caught it in the act! And with clean block threads +
factory torque specs + clean, tight wiring harness connection I think you are going to have a solid fix! Looking forward
to your next status report once the new part gets installed.

 

[Road Trip]

google Amazon

ACDelco GM Original Equipment 213-92 Ignition Knock (Detonation) Sensor​

Out of curiosity, I did, and came up with this: (Amazon link)

Down in the reviews, a truck owner posted an interesting photo of
what his original knock sensor looked like:

You must be registered for see images attach

I know you are familiar with what this looks like, but posted this for those reading along who haven't
had the opportunity to get up close & personal with their own knock sensor. (Edit: Added snap of
new part for visual comparison -- see attached.)

Anyway, a single wire doesn't weigh that much, but every time you hit a good pothole that shakes
the truck, that wire is shaking that terminal connection. And after all those years this physical electrical
connection starts to weaken. And then one day fails...and your engine starts pinging.

The new part should be the fix. But if you read around this forum, lots of people have shared stories of
buying new parts and no joy. But if you test your new part on the way in (not too rough, just tap on
vise, looking for some signal, then you should be able to install the part with confidence.)

Now to wait for the part to arrive...

 

[scott2093]

Tested the new knock sensor that came today and it seemed like it wasn't putting out any significant voltage. 1 time I got it up to a volt so figured maybe my set up was not right..or just wasn't wanting to believe it.... got the resistance so, the only thing I could think was that my probes were slipping slightly while hitting the vice...

Put the new sensor in. Put on the scanner and actually had a code 43, so cleared it and did the whole song and dance and I can't get a knock count. Resistance is good to the block and I'm almost certain if I check the voltage output of this sensor in the truck, it'll probably be crap. Is it supposed to be crap? Obviously something voltage wise needs to happen to trigger a knock. Just not sure what that is....maybe a few mV is all it needs and something else is wrong???

Of course I test the old busted knock sensor on the vice and I can get near 1 volt easy enough with bigger whacks even twisting and turning the terminal... but obviously I had a code 43 with it in so.....

Wonder how many people are getting new knock sensors and it's just getting them by to clear codes? I don't see how that's possible if there are self tests....

..frustrating..

Not sure what to do....

I'll look at the manual again....

 

[scott2093]

Really starting to wonder if this second new knock sensor is a dud. I can't get really any more than 10mV ac while it's in the vice and barely anything while installed. Even got a code 43 while it was installed and trying to test for knock signal.

The old one will get up to about 1VAC on the vice, usually less depending on where and how hard I hit and, after installing it yet again, I can get it to register some knocks on the scanner but no change in timing.

What are the odds that my old one is definitely bad/weak and I've gotten 2 new ones that were just bad?

I took a couple of videos I'll upload later but, it's pretty much what I just said.


hmm

"The sensor’s crystal material itself could also get damaged, making the knock sensor literally deaf. Such a condition would render the sensor inoperative and in some cases the bias voltage would not be affected. If this happens, the ECM will not have the ability to detect a pinging engine and no possible ignition retarding would be available. Such an engine would ping severely without any action being taken by the ECM."

knock_sensor

and the rabbit hole...

https://www.ti.com/lit/an/spra039/spra039.pdf

 

[scott2093]

What are the odds that if the motor in my truck was replaced with a crate motor around the year ? 2000 ? the knock sensor would be different than the original? Doubt any paperwork is around but I may still have a shot at finding it somewhere.. it's just a 350 right?