whiteflash

Because a CEL is a solid indicator that nothing is going on, when you're running a part that operates at a different frequency right?

All your ecu is saying is that "I see the appropriate voltages are there, ok a knock sensor is there." Just because you're not getting a CEL doesn't mean you're not knocking.

If you've ever seen a stock datalog, you'd see that there's quite a bit of ignition retarding going on based on the knock sensitivity maps. Congrats, you've just tampered with this formula. All i see is a great way for an engine to go boom without warning, especially in anything besides a stock car.

Although I'm sure neither of you have spent anytime tuning an S, and have no idea about how important that knock sensor is to reliability and performance. The S's ecu is very archaic in terms of knock control, and ecu based timing adjustments. A small change in the knock sensitivity table itself will show the correlation of this. It's your guys' engines.... but please, don't actually advise people to do this. Not without any proof whatsoever that they operate identically. A CEL (or lack of) is an indication of absolutely, nothing... other than the ecu see's it being present.

2003AP1

[QUOTE=OneSilverS2k,Sep 26 2010, 07:03 PM] That is your opinion. Explain this then oil pressure sensor for VTEC from a integra GSR has a different part Number as well, but I have one of those on my car and I hit VTEC every time.

Get your shit straight before you post on here, I am still looking into this to see if these are a direct replace or error can occur. would be nice to have some support, but all i hear is

2003AP1

In my opinion I think something like using an O2 sensor from a civic is ok as long as the part numbers are the same. But I wouldnt use ANY part from another car with different part numbers neither a knock sensor or a vtec oil pressure sensor. Just my opinion, again I am not trying to hate on anyone here

idea-catalyst

Wow, this thread has gotten a bit vitriolic. Unfortunately, we seem to be suffering from a dearth of hard information. While I might attach some credence to a Honda "substitution" list (because they presumably would be up for warranty repairs if someone followed their instructions but still had problems), I'm not sure I'd trust a simple part number comparison or "color" evaluation.

The "worst" that could happen is an engine failure that would have otherwise have been covered by a warranty but is now denied (OK, so most of these engines are likely to be out of warranty).

The challenge would seem to be that the knock sensor is protecting the engine against premature ignition by either retarding the spark or signaling an error code (or both). If one uses a substitute knock sensor, how can we know whether the sensor is actually "sensing" the specific knock frequency/conditions it was designed for?

The simple test of "does the engine run and not produce diagnostic codes" doesn't seem very satisfactory to me because it's the knock sensor that is supposed to provide the information to the ECU that will indicate whether there is a problem.

I guess I'd like much better engineering information about the different knock sensors before I concluded that a substitution was actually providing the right information to the ECU. If the sensor is providing information that doesn't jive with the specific engine, it seems to me that pre-ignition could rear its ugly head and do some engine damage that would remain un-diagnosed until failure and tear-down.

I'm with the earlier poster: saving $75 on an engine sensor without being absolutely sure that it is compatible and provides the same level of engine protection sounds like a bad move. If nothing else, I'd sure hate to have an engine failure warranty claim denied because I'd substituted an incorrect knock sensor.

Just my $0.02 worth from a cautious old fart.

iDashX

i broke 3 knock sensors off 3 different motors in 1 weekend... FML... 2 were from bad motors... I had to order one from Honda and got the $150 deal.

triddle

Well I didn't say that - I said I was glad someone was keeping a log of people who have done this to correlate it to things that may happen in the future. I might as well throw in my opinion on why this is actually dangerous though I do not know for fact that it does not in fact work.

In fact you are experimenting with a control system of the engine that can cause severe damage - the dynamic ignition advance is controlled via data from this sensor. This system works by constantly advancing the ignition until either it hits some upper limit or the knock sensor tells the ECU knock is occurring. If knock is present the advance is reduced and it goes about increasing it again in a constant loop. This assures that the timing advanced is as far ahead as possible for the current engine conditions as they fluctuate according to temp, gas quality, wear, etc.

Maybe, maybe not. The control system is designed to find faults in it's original hardware configuration. This hardware configuration is built around extensive requirements related to physical properties of the system as a whole. It is not designed to identify faults when these things are changed. The sensor in question here is the knock sensor. Knock is a very dangerous engine condition that leads to damage. How do you think the ECU is going to identify knock with out the sensor operating properly? Why would it have a knock sensor then? Knock eats holes in your pistons - by the time the knock damage is severe enough to give you so much blow by the car runs disgustingly rich you'll throw a CEL. That's not how I want to find out something's wrong.

Those are all of the reasons I personally would not perform this experiment myself; I find it to be a risky behavior based on my knowledge of algorithms and information systems. If you really want to find out if your knock sensor is working properly drive around for a few months on 87 octane in the middle of summer in the Mojave Desert. Then you could tell people "I drove around in really bad conditions with really shitty gas and my car didn't die" which is also it self not the same as "this is totally safe!" like you seem to be portraying here.

That kind of cavalier attitude is quite dangerous. Not the part where you look up information, that is good and I encourage this - it's the part where you don't even know the difference between 1 and 1 billion cycles per second. A cavalier attitude like this might even believe there is not a very significant difference between 5 and 6 khz on the knock sensor because 5 and 6 are not very far apart from each other. This is not definite though.

Now I'm not a signal analysis specialist, so this is some speculation, but I do it as a hobby with my computer because it is challenging and fun. Getting any kind of useful knock signal out of the engine is going to be dependent on a huge amount of filtering. The easiest form of filtering (from a computation perspective) to perform is filtering using the physical properties of the mechanical system itself (this means your engine block, heads, etc is the filter). There is all kinds of fun "interference" going on inside your engine with respect to the waves of the various audio frequencies being generated from all the moving components of your engine. Layered on top of this extreme (read: nightmarish) levels of noise is the knock noise itself. This is specifically what the sensor in question is tuned to detect.

Honda wants to keep costs of all computer systems as low as possible; if they keep a DSP off the ECU or keep the processor speed down because the DSP functions aren't implemented in software they are saving money. This gives them a very high incentive to use mechanical filtering in this system. The sensor itself may either perform very high levels of attention outside of the frequency it is tuned to, the sensor may be tuned to a frequency where the engine block itself is acting as a comb filter which causes that specific frequency to be so strong it overcomes all other forms of interference, a combination of the two, or even something a whole lot clever than I came up with off the top of my head. All of this means, in layman terms, that changing 5 to 6 can mean the difference between receiving a useful value to the ECU and receiving a reduced value or a value of no use at all. The fact that Honda has a sensor specifically for 5 and 6 khz leads me to believe they are very highly tuned instruments.

Generally the onus of proof is on those who make claims, not those who say claims should be justified.

bpaspi

I'm wondering why nobody raised the question why Honda calls up such ridiculous prices for the S2000 Knock Sensor when it's similar to cheaper ones of an civic
The S2000 Sensor has different specifications, but the production process (and the supplier) must be the same.
It's legalized robbery...

whiteflash

Because they can? Because the cost of R&D to make a production engine that's reliable past 150K miles that redlines at 9k was huge? The cost has to get split to different parts, and fine tuning the knock sensor to be accurate in it's fluctuations of ignition timiming at 8500rpms is critical.

I understand that this car is getting cheap now, so the younger crowd are getting in them... but the cost of maintenance on a car hardly ever changes. Apart of affording a car, is affording to maintain the car. If you think a 150$ knock sensor is a 'rip-off' then you should actually be driving a civic.

bpaspi

You're wrong with your assumptions about my age, my financial situation and my technical Knowledge!
You're also wrong in assuming that R&D costs are allocated to spare part prices. How could?
You've probably never worked for an automotive company in R&D, I did!

So you should first get better infos before bashing others and make such statements.

My 2 cents....

triddle

I was under the impression that a production run of 110,000 units was pretty small for a car. I can appreciate Honda not subsidizing the ongoing support costs for the S2000 by charging more for parts than a Civic simply because we are such a small fraction of their total business. I don't like it but I can understand that as a pricing motivation.

Course we may also just be getting screwed too but it's not like Honda is going to make a large amount more money pushing up the S2000 part prices.