and there is the secret to tuning the rteks true potential... anyone that tunes to this to show 0 but have max power will have a long lasting engine.

 

I don't think mine works? I wacked it with a wrench at Idle and saw no retarding? Do they go bad?

 

yes knock sensors go bad, i've had a few melt from the inside out.

 

hmm, quick (stupid) question. Can bad motor mounts cause the knock sensor to retard timing? Or are they more sensible than that? I'm searchin for reasons my car was (possibly) detecting knock, cuz it STARTED retarding the timing when the AFR's were in the 10's and AIT's in the 80's F....doesn't get much better than that wouldn't you say? Imma try the wrench trick too next chance I get!

 

yes, any sort of vibration could trigger it. It can also become sensitive in it's old age.

 

Just priced one at $221 new from Mazdatrix, not gonna buy one new that's for sure.. Are they really necessary if you have a good tune?

 

In an ideal world with a good tune, it should never be "needed". But I think I'd rather have the protection. Depending on your tune, you could only be one bad tank of gas away from an engine rebuild...

 

i didnt want to start a new thread so i will post in this on as my problem regaurds AFRs.

I have removed emmissions and a rtek 2.1 + exhaust, 550 720 injectors, larger turbo, street port, 3mm seals ect if that matters. with the cold weather coming in now i am having problems when the engine is cold. I set my idle with the TB set screw and my tps is adjusted properly.

When my engine is warm it idles at 1k which is just about right. but when its cold i have to sit there for 8 or so minutes at 1k for it to warm up enough to idle on its own at 600 rpm. i am looking for a way to richen it up at idle when cold so i dont have to sit there. i know there is a cranking map but i dont think that will help me. is there a cold engine idle map or something to richen it up?

Thanks,
Shane

 

Have you removed the thermowax? Is the coolant temp sensor reading correctly?

 

yes thermo wax is removed. and as far as i know the coolant temp switch is fine. so is there a tempature based fuel map or something? or when i make changes in the fuel map it is effective to all tempertures.

Thanks,
Shane

 

Autozone has it for $142. PS - they have many of the engine management parts @ a fraction of dealer $ which are Mazdatrix's prices. And I think it may be found cheaper elsewhere. I was able to get spark plugs from an Web site cheaper than Autozone but I didn't save the link and have lost the receipt.

here's the part's link at Autozone:

http://www.autozone.com/autozone/par.../N-iipmdZ9gsm6

 

put the thermowax back on

 

Thank you for the link! I just might go that route if I determine that the knock sensor is the weak link

 

RX7 SE could you post your timing map? Maybe a few of use could chime in about it after viewing.

I did notice that in load mode, with only 6-8 psi on a stock turbo, that the ecu will pull timing even without seeing a change in the timing retard.

 

That's interesting. Yes, I'll definitely post my map when I get some time to (finals week, lol). Does the map save to the 'backup' folder or do I need to write down every cell and type it on here?

 

Save it and send me your maps.pdb file that gets saved in your backup folder. I'll extract it.

 

So as a work around for the afm issue has anyone swapped to an s5 style meter?

There was a thread on here a while ago where they put both the s4 and s5 afm on a flow bench and confirmed the superior flow of the s5 unit. There was another thread that outlined how to splice in the s5 connector to run the s5 afm on an s4 harness/ecu.

 

I'm sure it's doable but probably wouldn't be worth the trouble for such a small net gain. I might be wrong tho

 

The problem seems to be maxing out the meter, not flow. Some reason mine maxes at 715 and others are seeing higher? Might try to mod it and see if I get a higher reading.

https://www.rx7club.com/showpost.php...&postcount=171

 

Be that as it may, the s5 unit seems to be a much better design all the way around, it wouldn't be surprising if it registered higher values as well.

You are right about your values though, the guy in germany (name escapes me atm) had something like a 900+ afm reading when he put down 380 (or so) hp. Wish I had a spare afm to fiddle with...

 

Just found this clip: http://www.youtube.com/watch?v=L8xRJ...eature=related
I'll be checking mine as instructed.
I'm now wondering, since all knock sensors are basically the same, we may be able to simply use, e.g., Autozone PN SU154 for ~$30 (which looks like the one on the clip's picture) . Might be well worth the try! Just re-die the threads if necessary.

 

Who told you that all knock sensors are the same??? They vary greatly in terms of the frequency range they produce. The video above alludes to that when he points out that resonance-type sensors are tuned to anywhere between 5000 and 9000 hz. And I disagree with his implication that resonance-type sensors are better than mass-type. Resonance-type sensors are primarily a cost-saving measure. If you look at the knock filtering and learning used on mass-type sensors you will understand. I've got some technical data I can post if anyone is interested.

The FC uses a mass-type sensor and the service manual does call for you to use the good ol' hammer test. I recommend you use an analog voltmeter though. It's a lot easier to see a needle for something like this.

 

post eh data!! plz

 

Mine does produce voltage. But how much voltage does it need to produce, because when I hit it with a wrench, I saw no retarding logged on the palm?... hmm...

 

Here is an explanation from a Toyota technician training course about the two different types of knock sensors:



This next thing is a spec sheet for a Bosch mass-type knock sensor. Besides being on German cars, I have seen similar-looking sensors on GM and Nissan applications.





There is discussion of the s5 knock control system in the paper Mazda published about the s5 engines. The Evos use a mass-type knock sensor and their signal filtering and self-learning is pretty well documented. I've also got a paper from Texas Instruments about knock sensors. I don't want to clutter the thread up too much so you can PM me about this stuff if you really want to see it.