question on knock sensing

[nitro2]

How is it you actually measure combustion pressures in an engine. I would think it's not an easy task doing that on 8cyls running at an elevated rpm.

Wouldn't it be visible on an ignition scope as well ? If so what type of graph would indicate it ?

Our company website is here http://www.tfxengine.com. Not all of our hardware/software is on the website, for competitive reasons, but enough to give one an idea. We've been selling this type of turnkey equipment for 12 yrs.

RPM is never an issue until you get well beyond F1 engine speeds.

[Belgian1979]

Interesting. Basically you can look for both pre-ignition and detonation with it I assume.

Don't see any prices. What are we looking at.

[MadBill]

... If initial timing were retarded, pre-igntion would serve to advance the effective ignition timing and could make more power. But this might not meet the definition of "pre-ignition"; for which I have yet to hear a proper technical explanation. Perhaps it is relative to MBT and would therefore be engine specific.

The simple definition of pre-ignition I'm familiar with, which obviously doesn't attempt to analyze causes, is just "any combustion occurring in advance of that initiated by the spark".
There have been a number of programs over the years attempting to ignite the charge via the spontaneous combustion of free radicals remaining from the previous cycle and I think I recall hearing that Honda had produced a motorcycle with a significant part of its operating envelope so ignited.

[nitro2]

... If initial timing were retarded, pre-igntion would serve to advance the effective ignition timing and could make more power. But this might not meet the definition of "pre-ignition"; for which I have yet to hear a proper technical explanation. Perhaps it is relative to MBT and would therefore be engine specific.

The simple definition of pre-ignition I'm familiar with, which obviously doesn't attempt to analyze causes, is just "any combustion occurring in advance of that initiated by the spark".
There have been a number of programs over the years attempting to ignite the charge via the spontaneous combustion of free radicals remaining from the previous cycle and I think I recall hearing that Honda had produced a motorcycle with a significant part of its operating envelope so ignited.

Pre-ignition combustion does not have to start in advance of the spark, but it usually does. The reason that it does not always have to be in advance of the spark is due to the fact that the spark itself is in advance of the combustion process i.e. there is an ignition delay period of quite a few degrees between the spark and anything resembling a measureable combustion event. Pre-ignition can sometimes start during the spark ignition delay period instead of prior to the spark.

We've run a version of "timed preignition" for many years but have never developed it into a commercial product due to the cost and so forth of getting from A to B. It works over the entire operating envelope and is straightforward to control. OEM use would require a combustion sensor for feedback to the ECU. It is not HCCI, nor even remotely as complicated as HCCI from a control perspective, and it operates over the entire operating envelope, unlike HCCI.

[nitro2]

Interesting. Basically you can look for both pre-ignition and detonation with it I assume.

Don't see any prices. What are we looking at.

You can look at pre-ignition and detonation for sure, but definitely a whole lot more than just that. On average $5-8k, most under $10k, but the sky's the limit, we can build higher end systems than what is on the website. We could build you a $200k system if you wanted, but our goal has always been to make the equipment accessible to everyone, rather than to simply "court the rich" as most competitors do.

[MadBill]

...Pre-ignition combustion does not have to start in advance of the spark, but it usually does. The reason that it does not always have to be in advance of the spark is due to the fact that the spark itself is in advance of the combustion process i.e. there is an ignition delay period of quite a few degrees between the spark and anything resembling a measurable combustion event. Pre-ignition can sometimes start during the spark ignition delay period instead of prior to the spark.
.

Quite so Clint, which is why I weasel-worded my definition by writing "...in advance of that (combustion) initiated by the spark..."

[Belgian1979]

Interesting. Basically you can look for both pre-ignition and detonation with it I assume.

Don't see any prices. What are we looking at.

You can look at pre-ignition and detonation for sure, but definitely a whole lot more than just that. On average $5-8k, most under $10k, but the sky's the limit, we can build higher end systems than what is on the website. We could build you a $200k system if you wanted, but our goal has always been to make the equipment accessible to everyone, rather than to simply "court the rich" as most competitors do.

I assume this is in essence meant for a dyno situation. $5-8k is too expensive for just analysis on a in car running engine. If you would want to market this as an in car analysis tool, you would need to be in the range of what a WBO2 sensor costs imo.

Meanwhile I've found a firm closeby that is able to do MS on a steady state chassis dyno. I think I can get the ignition close on that one. Only have to figure out my base line to get her driveable once I put on the system.

If my car ever gets back from the shop that is.

[Force Fed Mopar]

Been following this closely. I have trying for some time now to use the stock knock sensor (a new one, not a 20+ yr-old one) to tune the spark on my Chrysler 2.5 SOHC 8v 4-cyl turbo engine. It doesn't seem to be working well It runs fairly good, and should be fairly close, but at higher boost levels (15+psi) it seems pretty off. I am running the stock computer system, but I have a tuning program for it that allows me to change anything and everything, so it is essentially a programmable EMS now, with some extra features. One of these is the able to turn on the CEL when the computer is pulling timing due to knock. I'll refer to this as the "knock light"

It seems that to make it pull properly at 18psi-higher, I have to have the timing at a point that the knock light pretty much stays on. Basically it'll come on around 3200-3500 right as it hits full boost, go out till around 4500 then come back on. I have tried pulling timing down in those spots and pulled a lot of timing out from boost, and to some extent it would stay out but it is very lazy up up, basically it's done pulling at 5k. I have down to where it has pretty much 1* total timing at 20+psi when it spools.

The computer determines knock by reading the voltage output of the knock sensor (0-5v) and comparing it to a knock threshold table. If the voltage is over the set amount for that rpm, it pulls timing. I can adjust this threshold and have done so, as the factory set it pretty conservative to guard against people fiddling with boost, bad gas, etc etc. Also, stock boost on these engine was never more than 12psi even on factory upgrade computers, most of the time it was 8-10psi. But, it seems like I have had to change it a lot, almost to the point that sometimes I feel like it's really not even effective any more.

So I guess my question is, at twice the factory boost level, is the stock knock sensor even worth trying to use? I have thought about just disabling it and tuning by ear like old school, but that just seems like going backwards. And I know that if it's bad enough to hear, it's pretty bad already lol. I know knock can occur w/o being able to hear it.

Specs on the engine are 2.5L, 3.44 bore x 4.09 stroke, 8.0:1 compression (forged pistons), SOHC 8v swirl chamber head, ported stock exhaust manifold, stock Garret T3 turbo w/ a .48 AR turbine, 3" exhaust, water-to-air intercooling. Always run 93oct pump gas. I can post screenshots of my timing and knock tables if needed.

[timinator]

at twice the factory boost level, is the stock knock sensor even worth trying to use?

How about starting your own thread? You can then list a lot more information about your setup to aid someone in giving you better advice.

[nitro2]

Interesting. Basically you can look for both pre-ignition and detonation with it I assume.

Don't see any prices. What are we looking at.

You can look at pre-ignition and detonation for sure, but definitely a whole lot more than just that. On average $5-8k, most under $10k, but the sky's the limit, we can build higher end systems than what is on the website. We could build you a $200k system if you wanted, but our goal has always been to make the equipment accessible to everyone, rather than to simply "court the rich" as most competitors do.

I assume this is in essence meant for a dyno situation. $5-8k is too expensive for just analysis on a in car running engine. If you would want to market this as an in car analysis tool, you would need to be in the range of what a WBO2 sensor costs imo.

Meanwhile I've found a firm closeby that is able to do MS on a steady state chassis dyno. I think I can get the ignition close on that one. Only have to figure out my base line to get her driveable once I put on the system.

If my car ever gets back from the shop that is.

I think you are confusing combustion pressure equipment with something else. $5-8k is dirt cheap for combustion pressure equipment. Lots of people use it in-car, as it can be used in car or on the dyno, and the really high end engines (thousands of HP) use it mainly or exclusively in car. If you are talking about just tuning your own street car then this technology is probably not in the cards, it is mainly for serious racing applications, engine builders, engine tuners, universities, companies with unique engines, OEMs, individuals with a thirst for knowledge, and so forth. One would have to be pretty hardcore to buy combustion pressure equipment just to tune their own street car or a few buddies street cars, though we have some customers that do just that.

Just some background, this type of technology has been around for more than 25 years, prices were able to be dropped over the years by the relevant players, and the least expensive systems on the market went out our door in 2005. Prices have slowly increased since that time. Combustion pressure equipment is nothing like conventional instrumentation.

[nitro2]

...Pre-ignition combustion does not have to start in advance of the spark, but it usually does. The reason that it does not always have to be in advance of the spark is due to the fact that the spark itself is in advance of the combustion process i.e. there is an ignition delay period of quite a few degrees between the spark and anything resembling a measurable combustion event. Pre-ignition can sometimes start during the spark ignition delay period instead of prior to the spark.
.

Quite so Clint, which is why I weasel-worded my definition by writing "...in advance of that (combustion) initiated by the spark..."

I'm not too good at reading between the lines...

[Force Fed Mopar]

at twice the factory boost level, is the stock knock sensor even worth trying to use?

How about starting your own thread? You can then list a lot more information about your setup to aid someone in giving you better advice.

Probably good idea, I'll do that.

[dieselgeek]

Clint's pricing is ridiculously LOW for what data it provides. Consider the people that usually use this type of equipment, I can think of a few: Formula One, NHRA Pro Stock, and the OEMs. I've been using it for a year now, love it. Plan on a few days' effort to get it all set up, bring in an I/T guy if you're not savvy with installing and debugging drivers on your laptop or PC, and prepare to either fab up a second trigger wheel or like I did, integrate with an OEM crank position sensor (I built a repeater circuit to supply two devices with the same sensor and wheel). Realize that Clint's is the only system out there that will do this sort of measurement with a standard type of trigger wheel, the other system I have used required a ridiculous and cumbersome target wheel, and required steady state running of the engine for some time to capture data which isn't easy on a race or healthy street engine. You also get one of the best data logging systems out there with Clint's kit, I love the National Instruments equipment. Ours was built with mil-spec connectors and I've used it, literally, in a cornfield with no problems other than I wore out a sensor - that he kindly is repairing for us. There's a small learning curve, not as complex as EFI, but all the effort is worth it in my opinion. Plus Clint's great to work with, he provides solid tech support that includes lots of combustion knowledge.

[David Redszus]

The computer determines knock by reading the voltage output of the knock sensor (0-5v) and comparing it to a knock threshold table. If the voltage is over the set amount for that rpm, it pulls timing. I can adjust this threshold and have done so, as the factory set it pretty conservative to guard against people fiddling with boost, bad gas, etc etc.

Actually, the knock sensor determines knock and sends a voltage to the computer. Most all knock sensors deliver a voltage of about 1.0v when knock is present. It may be amplified for use by the ECU. The important point is to know the fundamental knock frequency of the engine and the response frequency of the knock sensor. If they are not similar, we have a problem.

So I guess my question is, at twice the factory boost level, is the stock knock sensor even worth trying to use? I have thought about just disabling it and tuning by ear like old school, but that just seems like going backwards.

The fundamental knock frequency goes up with increasing combustion temperature. A forced induction engine would produce a higher combustion temperature (depending on crank angle) than would a non-turbo or low boosted turbo.

Running on pump gas with higher turbo boost could quickly put you into knock country. Why not run proper race gas (pump gas has no consistency) and then tune for best performance; forget about knock sensing. The presence of pre-ignition, due to excessive advance, will show up as a loss of power. Advancing (excessively) the combustion pressure curve is just as damaging as knock.

[Leftcoaster]

The important point is to know the fundamental knock frequency of the engine and the response frequency of the knock sensor. If they are not similar, we have a problem.

The fundamental knock frequency goes up with increasing combustion temperature. A forced induction engine would produce a higher combustion temperature (depending on crank angle) than would a non-turbo or low boosted turbo.

Running on pump gas with higher turbo boost could quickly put you into knock country. Why not run proper race gas (pump gas has no consistency) and then tune for best performance; forget about knock sensing. The presence of pre-ignition, due to excessive advance, will show up as a loss of power. Advancing (excessively) the combustion pressure curve is just as damaging as knock.

How does one select the appropriate knock sensor for an engine, especially if not originally equipped by the OEM

Do all ECU's employ similar knock sensing circuits, and will they function satisfactorily when fed by a competing manufacturer's sensor

If knock frequency changes between NA and Forced Air applications is that also true for CNG and LPG applications?

Thanks!