>

Awww come on now, don't hate on me for being an import newb. You know what I was pointing out.

>

Ok, now I follow ya.

now we get to the real truth ... you just want to be able to say you have a GM part on your car ! hahahaaha j/k

Whats all this fear of a knock sensor ? Who has ever blown one on this site ? I'm going to agree with the Art, Manny, Jay, Justin and all the rest of the swappers.

Sometimes threads at MR2Oc should stay there. These guys are running 300 to 400 HP 3sgte's. A basic run of the mill swap with some fuel upgrades is not going to blow sensors all the time. My sensor was made in 91 or 92, it still works fine.

Stick with all Toyota parts. The more non Toyota parts you put into this car, the more trouble you have later on.

You have to remember Shops are there to get the job done, no matter what parts they have to use. They have to keep the costs down to make the most money. So using a cheaper alternative is always thier answer. This is a project you cannot cheap out on.


Not trying to offend anyone Just my P.O.V.

Just to add to the heated fun here.

I had a stock Toyota knock sensor on my All-trac with a Link Plus EMS. The EMS let me see the knock counts in nice pretty graphs, and under heavy boost I was seeing a ton of knock - tuning was in order.

Anyways, I went through 2 knock sensors in as many months, then I switched over to the GM knock sensor. Same knock counts, but the sensor held up. A couple months later and the motor crank-walked due to bad machining. Regardless, the GM sensor had worked fine.

So, I recently installed the GM knock sensor on the 3rd gen powered 6GC, and I started getting code 52 on the stock 205 ECU. Switched back to a stock knock sensor and it cleared up. Checked the EPC and the part numbers for the 3rd gen 3S is different then the 2nd gen.

>


the important thing to note is that you were using an ems that is designed to work with just about any type of knock sensor, unlike the stock ecu .

>

SSSSSSShhhhhhh!!!!! You're the only one to figure it out so far.

>

Good info. Maybe it won't work with the 3rd Gen. I'll be running a 2nd Gen so it shouldn't be a problem, but we'll see as I like to do datalogging. Of course, no knock or knock retard would be optimal.

>

The Link EMS is actually engine specific to accomidate the stock sensors. Since I had the same knock count reading with both sensors (90-120 counts). its pretty safe to say that the knock sensors put out the same A/C voltage signal.

This post has been edited by scothaniel: Jul 12, 2006 - 2:20 PM

>

EMS or stock ECU shouldn't matter. The knock sensor picks up the frequency emitted due to the bore of the cylinder and then just sends a signal to the ECU or EMS that knock is occuring.

The bottom line is that the GM sensor will see some of the knock, just not all of it. I think it could actually be effective at getting more power out of a car because you will get knock retard less often. We have discussed ad nauseum whether that is a good idea....

>

It's nice to see someone post in this thread that has some real life experience with both sensors and isn't just repeating things that they read on an internet forum. Great info, thanks very much!

That's interesting that the third gen knock sensor is different.

-Doc

>

I agree although I will add that reading it on an internet forum and repeating it is better than providing incorrect information regardless of it source.

Always happy to share my experiences with others - just wish it wasn't on I managed to break things. Here's another valuable lesson - never leave a wrench on top of the motor when the timing belt cover is off!

I'm running the GM knock sensor and have had zero problems out of it. I also have yet to see anyone test these and put up some hard proof that they are different or are the same thing.

Ok, to start of... I AM an engineer (electrical engineer with an audio focus), and I have wired my own 3rd gen engine swap (correctly on the FIRST TRY) along with doing all my own mechnical work (minus welding).

Everyone needs to go back and read what the J&S engineer said... Its not the motors so much as the sensor that resonates. A knock event is an impluse and contains all frequencies. The difference between a 5.7k and 6.3k sensor means that it outputs a resonant sine wave at those frequencies. Inside the ECU, there is a filter that changes that sine wave to a voltage level that the ECU can easily interpret. Whether the sensor resonates at 5.7k or 6.3k or whatever shouldn't matter much as long as the filter is setup correctly.

My *opinion* is that the GM sensor is at least 90% as good as the stock Toyota sensor. In a modified motor, I would say this is a good mod as there is a larger chance of a big knock event that could blow a knock sensor but not the motor. There is no reason you should have to keep buying $100+ sensors when something less expensive does a good job. On a stock motor, I'm a big fan of "if it ain't broke, don't fix it." I'm keeping everything stock on my swap but intake, exhaust and boost control...

-Charlie

^ i'm with him ^
EE aint my thing, but eng in general is awesome

basically, i wanna know what this "conversion kit" is to use the GM. what it is and how it did it basically.

>
its just a adapter for the threads, cause the GM knock sensor wont thread into the 3s block, and a pig tail, to wire it up into the 3s harness.

>

Great explaination

>

belive it or not, i have done audio engineering for a living . i dont think there is a narrow band filter inside the ecu that is filtering on a software level. i think its all done on a hardware level by using a narrow band microphone (knock sensor). as you know, anything thats narrow band, means that it is tuned to that one frequency and setup to cancel everything else out. if it was all done in the ecu, then you would just use a 20-20hkz wide band sensor.

but, even if there is some type of software filtering going on, youll alwasy get better performance out of a microphine (kock sensor) that peaks at 5.7k then one that peaks at 6.3k on a spectrum analizer (if 5.7k is the frequency your trying to listen for) .


edit: we can do this all day long.... having said that, this ended up turning into a cool tech thread

This post has been edited by lagos: Jul 12, 2006 - 7:42 PM

and finally ....FOR THE WIN....

a copy of the toytoa bible (better then the bgb) from auto shop 101. that proves that each knock sensor is tuned to a specific frequency and that the ecu dosnt do any from of audio filtering to get its knock info from.

http://www.autoshop101.com/forms/h38.pdf

Nice find. I had forgotten about those.

>

after skimming through this post I have some newbie type thoughts....

the stock (toyota sensor) might tend to blow at times...so you want to put a different frequency sensor in so it blows less ? correct ?

isn't that the same idea as " I keep blowing a 15A fuse in the fuse box, so I will put a 25A one is so it doesn't blow anymore "

if the stock sensor blows, there is a reason why. Solve the problem - don't band-aid it until it is a bigger problem.

my .02

>

Although I am not sure how you quantified this, I generally agree. You have to ask yourself though, is the $40 extra worth the 10% (or whatever) chance that the GM sensor will not appropriately respond to catostraphic detonation?

>

On the other hand, I have to part company with you on this one. To me, the increase chance of detonation on a modified motor increases the need for accurate knock detection. With modified (high horsepower) motors, the last thing you should be worried about is the $40 difference in the cost of these sensors.

>

Makes sense to me batman!

ANYWAY... I read about 2 pages of this topic and figured i would just say that we all love celicas here! Why cant we all just be happy for fastbird, i mean im happy for you!! If he gets knock then he will learn! You gave your opinion and he gave his. I think we should all be happy for him, if it works, it works. If not well then i guess then you can say "I told you so!" !!! If it works then fastbird can say "I told you soo!!!" Either way im still happy for you!


Everyone on this site is a great member! Im glad we have great moderators like lagos and pressur2. I am also glad we have great people like doc to help those who dont want to get their hands dirty or want their swap done correctly. I think everyone plays a role and im playing mine! Lets just get along.

Doc great work on the swap

Fastbird- that is a nice car and i hope your happy with it! i mean hell, im happy to see it look like that!

Mario!

>

im a moderator? woot!!!


keep reading the rest of the pages. it started out as lagos bash, but now is a cool tech talk thread.

This post has been edited by lagos: Jul 12, 2006 - 11:45 PM

>


No, thats not what ( some / few / many ) people are saying. When excessive knock happens, how long will the sensor last? All motors get knock, even when everything is running perfectly. Only in this case, the knock is like 4 counts / second (microphones pick up background sounds). When you get 50+ knocks / sec - something is wrong.

The stock sensor is known to be more fragile then most and fail after a few cases of excessive knock. The real question here is whether the GM knock sensor is listening for the correct frquency. Knock sensors listen for specific frequencies to determine knock - otherwise valve or injector tick would through off the system. Does the GM sensor listen for the same frequencies as the stock sensor?

From my experience, I believe so.

>
That's not quite what I wrote. I mean: A modified (ie stronger) motor can take knock events with less damage, and are often more likely to see those events (higher boost, racing, etc.). Both of these things point to the need for a hardier knock sensor, and *if* it is less sensitive it won't hurt things as much.

Its ok if you disagree though.

-Charlie

>
Very good link. I probably have that downloaded in the tech director on my other computer... I just haven't read it yet!

So, here's my *opinion* on what's going on in the knock system: Knock sensor sees knock even and resonates at a larger amplitude (shown in the graph on the second page). This signal is sent to the ECU where there is an envelope filter that is able to grab the peaks of the knock events. When this peak reaches a certain value, the knock response of the ECU is generated. In this way, the ECU can have a simple comparator watching the knock system - cheap and effective. (at one time, started investigating designing something like a J&S with a friend of mine...). There is no need for complicated electronics or computation in the ECU.

The resonant effects of the motor (peaking at the same or similar frequency as the sensor) will enhance the sensitivity of the sensor to knock events. BUT... the bandwidth of the resonance of something as complicated as a motor (with a shape NOTHING like a tuning fork and made up of many materials) should be fairly wide - and thus a sensor that is less than 1kHz away in peak frequency should still work well.

Once again, my opinion - but a fairly educated guess.

-Charlie

>

Good point Charlie. You may be correct but I will note that almost all of the high hp 3sgte's run stock knock sensors. That surely doesn't prove anything but may be worth observing.

>


just read that pdf. tells you exactly how it works...lol