Sarah Forst
Ask Sarah- Does water injection work?

Hey Sarah,

I have a turbocharged 1995 Integra GS-R with a T3/T4.  I usually set the boost around 8-10psi but sometimes I get up to 14psi.  I still have a stock ignition.  Is this boost too high for 91 octane?  One of my friends said I should install a water injection system, but doesn’t that just kill power?  Can you recommend any water injection systems?


Got a Tech Question?  Email Sarah at asksarah@motoiq.com

Detonation or knock can quickly kill an engine.  The tendency of an engine to knock is affected by many factors including ambient air conditions, engine design such as quench in the combustion chamber, spark plug location, and compression ratio, engine temperature, air fuel ratio and the fuel octane.  Turbochargers force more air and fuel into the cylinders making the engine run hotter, the higher the boost pressure, the higher the cylinder pressure and temperature will be.  This causes more stress on the engine’s internals and making it more prone to knock.  The GSR engine has 10:1 compression ratio, fine in its naturally aspirated form but fairly high for a boosted motor.  If you have effective engine management and good tuning, 10 psi should be ok for pump gas but 14 psi would be pushing it on 91 octane.

AEM Water injection system
AEM's water injection system has many tuning variables, from onset boost pressure, to max flow boost pressure to volume.  This makes the AEM system very flexible.

Start by checking out your entire ignition system (plugs, plug wires, and distributor cap and rotor) to make sure everything is in good operating condition.  The stock ignition system should be fine until about 10psi as long as you are using spark plugs one to two heat ranges colder than stock.  Colder plugs help prevent detonation by keeping the plug center electrode from getting too hot.  Gap each of the spark plugs at around 0.030 inches and close the gap in 0.005 inch increments if you have any misfire.  You may need to gap them as low as 0.020 inches, especially for higher boost pressures.  With water injection and the proper plugs and gap, the stock ignition should work fine up to 14psi, but you should consider a high performance ignition system such as the MSD 6A if you want to go any higher.

Water injection is meant to suppress detonation caused by the increase in combustion temperature and pressure in turbo or supercharged engines.  Water has a very high specific – and latent – heat capacity, meaning it takes a lot of energy to heat water up.  This makes it an excellent liquid for regulating excess heat under certain operating conditions.  The water injected into the intake stream vaporizes and acts as a coolant, cooling your intake charge and producing cooler combustion temps, making the engine less prone to pre-ignition or detonation.  These conditions are especially common in forced induction engines, where exhaust temperatures can run much higher.  Most water injection systems can be programmed to come on at a set boost pressure, causing a jet to spray water into the intake system. 

AEM water injection nozzle
AEM's system has the water nozzle and check valve built into one unit.  The AEM system comes with three nozzles, small 130cc/min, medium 315, cc/min and large 550 cc/min.   The flow can also be tuned within these parameters by the pump controller.

Most forced induction production cars control combustion temperatures by dumping additional fuel to act as a coolant but if you go too rich, the flame burns slowly and incompletely causing you to lose power.  The point of diminishing returns for detonation suppression is around 10:1 although power will start to be diminished anywhere richer than 12.5:1 or so. A good water injection system keeps from having to dump more fuel, leaving your air/fuel ratio unaffected.  Although in some conditions lower combustion temperatures will reduce power slightly, by suppressing detonation you can raise boost, increase the timing advance, lean out your air/fuel ratio, or a combination of all three to produce more power.  Some other advantages of water injection are it cleans the engine of carbon deposits and sludge, provides the advantages on higher octane fuel out of lower octane, and the lower combustion temperatures due to water injection also lowers NOx emissions.

I recommend ERL Aquamist water injection systems.  Their jets atomize the water, which helps with distribution and improves vaporization.  Stay away from cheap water injection systems that use non-atomizer nozzles.  This will cause unequal distribution and bog the engine down.  Aquamist’s water injection system is tunable with jets ranging in sizes from 0.4mm to 1.0mm.   I also recommend AEM’s system as it is easily tunable for flow volume and injection onset point.  With its optional flow meter you can know exactly how much water it is flowing into the engine in cc’s per minute.

Aquamist water injection water pump
The Aquamist pump is very compact, good for tight locations.


Page 1 of 2 Next Page
Bookmark and Share
Thursday, March 25, 2010 11:13 PM
Im assuming it'd be best to use distilled water rather than tap water straight from the faucet.
Friday, March 26, 2010 5:03 AM
Aquamist has some new kits out that are pretty high tech. An article on those would be good.
Friday, March 26, 2010 7:05 AM
"Colder plugs help prevent detonation by keeping the plug center electrode from getting too hot."

Correction: Colder plugs mainly help prevent pre-ignition by keeping the center electrode from getting too hot.
Mike Kojima
Mike Kojimalink
Friday, March 26, 2010 7:49 AM
Detonation and preigntion are the same thing.
Friday, March 26, 2010 9:34 AM
Not according to everything I've ever read on the subject.

You can start here if you'd like. http://www.streetrodstuff.com/Articles/Engine/Detonation/
Street Surgeon
Street Surgeonlink
Friday, March 26, 2010 11:48 AM
I've heard they were different as well but meh. Personally I'd skip the water injection and just switch over to E85! Goodbye detonation :D
Mike Kojima
Mike Kojimalink
Friday, March 26, 2010 11:49 AM
Come on a street rod site, I am not even going to look at that :). Read Ricardo's book, the terms are used interchangeably, either way they describe the exact same thing, out of control combustion with a spike in cylinder pressure and temperature.

What a lot of people call pre-ignition originates from from a point source like a hot edge, and what some people call detonation is from the end regions of the combustion chamber but it results in the same thing.

In reality, on a street engine the on throttle events are rarely long enough in duration to cause glowing edges, I don't even think this happens in a modern combustion chamber.
Sunday, March 28, 2010 3:53 AM
Mike I had this decent reply typed up but when I posted it was lost.

Detonation and pre-ignition are two very separate things with very different modes of avoidance. Just because they've been confused in the past doesn't mean us technical types have to do it. There's much to gain by keeping them separate. Don't knock the article until you read it. Seriously. It makes it crystal clear.

I do think you're right about modern engines and the almost complete lack of pre-ignition (with the correct plugs).
Mike Kojima
Mike Kojimalink
Sunday, March 28, 2010 7:11 AM
The modes of avoidance are not different. More octane, less timing, less cylinder pressure, richer mixture.

I have never even seen pre-ignition before other than what was caused by a vastly incorrect plug heat range or a grossly wrong mechanical problem. It is not worth differentiating the semantics of what is going on. You are arguing purely on semantic basis.

I don't adhere to the glowing edge opinion. It comes from building engines, then detailing the piston crown and combustion chamber to find the exact same detonation threshold many times, or never being able to make an engine run on on pump gas with a reasonable compression just by getting rid of "hot" edges.

I have only seen glowing edges causing problems when something is grossly wrong, like a very wrong head gasket that hangs into the bore, totally wrong plug temp range or reach with exposed threads and thats about it. I think its a relic from side valves and poorly engineered engine days.

Its to the point where I don't bother detailing combustion chambers and piston domes anymore unless its a very extreme FI application and even then its just a precaution.

What you call pre ignition is detonation caused by ignition from a point source vs a localized spontaneous combustion due to pressure rise, they are both detonation. It not even worth arguning that there is a difference. Pre-igntion is one of those old hot rodders things I like not to use.

Its better to teach people when they here the knocking noise, stop and fix it and unless you have a side valve or a heron motor, do things other than tear it apart looking for hot edgeds and sources of pre-igntion to fix it.

I suggest studying some more scholarly books on combustion, I suggest the Internal combustion engine in theory and practice by Taylor or The High Speed Internal Combustion Engine by Ricardo, the latter has some fastening stuff on combustion and how to indirectly measure and observe it using understandably technology (its an old book). Anhother really good book is Turbocharging the Internal Combustion Engine by Watson and Genoda. They are all very boring unless you are truely interested in this stuff.
Bob Holmes
Bob Holmeslink
Sunday, March 28, 2010 4:05 PM
"Turbocharging the Internal Combustion Engine by Watson and Genoda."

Got a spare $700. !!!


Monday, March 29, 2010 3:42 AM
So you find no value in differentiating between uncontrolled combustion events that happen before the ignition event, and those that come afterward even though the likely cause is different, the detection method is distinctly different, and the resulting failure methods are distinctly different?
Monday, March 29, 2010 10:34 PM
I'm assuming your calculations are for a reduction in cylinder temps right? 135F reduction seems a little extreme for an intake charge temp reduction.
Scott Helmer
Scott Helmerlink
Tuesday, March 30, 2010 1:57 PM
Damn Ben, I wish I had the tool steel-grade balls that it requires to argue with Mike Kojima about pretty much anything related to automotive engineering.

Now I'm by NO means an expert on the subject, but I have to agree with Mike that there's no point in differentiating between the two (now imagine the next few words are italicized) in this particular context, much less in the comments section of this article. You're trying to argue that we should differentiate between the two, regardless of the fact that we're assuming (in this case) a Honda B18 engine (or really any modern engine) with virtually bone stock internals, maybe a hot cam, a turbocharger making maybe 1 bar of boost, running fairly safe timing settings, and almost certainly not running at WOT long enough to create some reasonable pre-ignition events. Now, it's not that we absolutely shouldn't differentiate the two, it's just that MotoIQ has a FORUM where you could discuss exactly that, should you choose to do so.
Mike Kojima
Mike Kojimalink
Tuesday, March 30, 2010 8:32 PM
Ben- I will tell you that there are no difference in the detection and the failure modes are not different. I will also tell you that the only way you are going to get what you call pre-ignition in a properly speced out modern engine is a grossly incorrect spark plug choice or some sort of gross mechanical assembly error. The one exception is extreme forced induction, large amounts of NOS or exotic fuel use. In these cases, careful prep must be done to avoid what you call pre-ignition but in this case the potential for damage and the failure modes will be the same.

Fuergrissa- I have routinely logged temperature drops of 100-130 degrees using water injection. If I had an inefficient compressor, I would expect to see even greater drops with higher fluid flow volumes. I have seen a t25 on a SR20 create a baking 395 degree discharge temp when pushed off the map like a lot of idiots do. You bet, water injection can drop this temp a lot.

Scott- We created these places to foster healthy debate over the articles and Ben is doing just that, it is completely OK and on topic to do so. Ben's points are on topic and well thought out.
Thursday, April 01, 2010 10:37 AM
I agree talk of uncontrolled combustion events happening before the spark event might not belong in this article. I can certainly take this to the forums if preferred.
However Mike seems comfortable here and I only have one or two more questions to ask so we might be able to put this to bed.

Obviously I'm coming from a different level of understanding than you Mike. I do my best. $700 for a turbo book isn't in my budget. ;)

But I have you!

This is what I was taught (maybe incorrectly) and I would love to be educated otherwise.

My view of "pre-ignition":
1) Uncontrolled combustion happening before the spark event almost always happens close to BDC as the piston begins to rise. The air/fuel is least dense and thus easiest to ignite. "Glowing edges" as you put it are the cause.

When it does happen pressure rises relatively slowly in the cylinder because the flame front propagates from the ignition source relatively slowly throughout the lightly compressed mixture which causes no detectable spike in pressure ie: no detectable knock, ping, or sound to listen for with instruments.

The event happens slowly but the pressures reached are still abnormally and extremely high. The resulting heat cause piston temps to spike almost instantly causing catastrophic piston failure after only one or two such events (fractions of a second).

The main "glowing edge" (at least historically) is the spark plug electrode hence finding the proper temperature plug. Obviously fuel octane is a factor as well but upping the octane is masking the main problem and not a full solution.

My view of "detonation":
2) Uncontrolled combustion happening after the spark event happens as the flame front and pressure wave from the controlled combustion closes in on the remaining, unburnt air/fuel mixture also being pressurized from the other side by the still-rising piston. This pressure causes the temperature of the remaining mixture to rise beyond it's ignition point and an uncontrolled combustion happens where all of the remaining mixture ignites completely uncontrolled and all virtually at the same time.

This causes a huge spike in pressure which resonates throughout the engine block and can be heard with tools or by ear if significant enough.

The uncontrolled combustion raises pressures in the combustion chamber above normal but not nearly to the level of a full-blown "pre-ignition" event. The spike in pressure causes localized overheating of parts of the combustion chamber and the piston causing damaging (but relatively minor) pitting and melting of aluminum surfaces such as the head, piston top, etc. and can break-off or melt spark plug porcelain or ground straps and damage head gaskets and even bend valves. If left to continue minor detonation can ruin a cylinder in a manner of hours or severe detonation can ruin a cylinder in a manner of minutes or seconds.

There are two causes and a myriad of preventions for this type of uncontrolled combustion. Significant amounts of heat in the combustion chamber are one cause combined with insufficient octane fuel. There are many remedies to these issues I won't go over assuming we all have a good grasp of those.



It strikes me as odd that a significant amount of heat can be removed from the situation (enough to avoid detonation versus experiencing some) by changing the heat range of the spark plug. I thought that was only designed to keep the plug warm enough to self-clean yet cold enough to prevent a glowing edge. What am I missing here?

There are other things I could obviously use clarification on. Such as detecting a combustion event that is triggered well before the spark event [if there is no sharp rise in pressure]. I know what I call "detonation" can be detected in ways other than listening for it (you can see it in EGTs if you look close enough, same with AFRs) and that EGTs would be an obvious place to look for "pre-ignition" but the main place to look for "detonation" is with sound. From what I know that detection method doesn't work for "pre-ignition". What am I missing here?

As for failure modes. Show me a piston and I believe I can tell you if it was a victim of prolonged "detonation" (pitting on the crown, possible scoring on the sides of the crown due to malformation and contact with the cylinder wall, possible 1st ring seat cracking, likely more pitting on the intake side versus the exhaust side) or 1-2 cycles of "pre-ignition" (massive hole or holes in the piston crown possibly with wrist pin protruding or a piece of aluminum that looks like a small baked potato wrapped in foil).

Have I been severely mislead?
Bob Holmes
Bob Holmeslink
Thursday, April 01, 2010 5:58 PM
Perhaps a more acceptable reference (than a hot rodder, perish the thought!)

From RET-Monitor, this months edition (RET=Race Engine Technology")

And I quote from the ECU-EMS article about knock sensors

"As opposed to surface ignition (sometimes referred to as pre-ignition), combustion ‘knock’ is created from the spontaneous ignition of part of the end gas in the combustion chamber in advance of the flame front." John Coxon.

However, I disagree with Ben's characterization of the manifestation of failure. I agree with Mike that there may be little difference.
Mike Kojima
Mike Kojimalink
Thursday, April 01, 2010 7:01 PM
It strikes me as odd that a significant amount of heat can be removed from the situation (enough to avoid detonation versus experiencing some) by changing the heat range of the spark plug. I thought that was only designed to keep the plug warm enough to self-clean yet cold enough to prevent a glowing edge. What am I missing here?

If the plug is too hot, the heat path internal in the plug is long and the plugs center electrode can actually become glowing hot, like a glowplug which can touch off preigntion or detonation. The cylinder pressure is to low to have uncontrolled combustion near BDC.

There are other things I could obviously use clarification on. Such as detecting a combustion event that is triggered well before the spark event [if there is no sharp rise in pressure]. I know what I call "detonation" can be detected in ways other than listening for it (you can see it in EGTs if you look close enough, same with AFRs) and that EGTs would be an obvious place to look for "pre-ignition" but the main place to look for "detonation" is with sound. From what I know that detection method doesn't work for "pre-ignition". What am I missing here?

Any sort of uncontrolled combustion is going to have an acoustic signature, which is how most knock detection works. There is some stuff using ionization voltage using the plug as a sensor or pizeo electric pressure transducers built into the plug washers. Nissan used this in the VG30DET in Japan in the Cima but perhaps it didn't work well because they did not adopt it into anything else.

Any sort of pressure/thermal damage is detonation damage. Usually you see it on the exhaust side of the piston and or the exhaust valve. Usually the hole in the center of the piston is seen in forced induction applications. I kinda thought that this was more of a function of this being the weakest part of the crown with the longest heat path. The exhaust side damage seems to be more of a street engine or NA engine thing because this side of the combustion chamber is hotter with transient WOT stuff.

These two things are more of my personal opinion based on observations, I haven't done a heat profile study in engine development or anything.
Sunday, December 19, 2010 4:54 PM
I hope I'm not out of line bringing this dialogue back from the dead, but I am fairly new as a subscriber and just ran across this article. I've understood the theory and practice of water and methanol/water injection since high school but this is the first time I have seen the math. Thanks very much for that, it was very straight forward. There are a couple of points that the article, and subsequent debate, didn't touch on that I would appreciate your comments on.

I have read in the past that one of the other benefits of water injection, as it relates to more complete combustion, has to do with the way the surface tension of the atomised water particles creates more surface area with the fuel as it comes in contact with the water, much as a drop of fuel landing in a puddle of water nearly instantly spreads and thins out exposing more fuel molecules to the air. Could you expand on this and indicate how big a benefit it is?

Secondly, GDI (Gasoline Direct Injection) technology is becoming more common in production engines as a way to help control predetonation, among other things, and something I would like to see used as a basis for a future build. I think Ford's Eco Boost engine may be the current state of the art in a production car. In GDI, fuel is being injected at the last possible moment before ignition, leaving precious little time for fuel molecules to spread themselves around water particles. Is there enough time to reap that combustion benefit? One of the direct benefits of GDI is combustion chamber cooling. If water injection is also used do we need to worry about getting the combustion chamber too cold? No matter what, I feel the cleaning benefits of water injection justify its use for that alone, but I wonder if it may need to be tuned a little differently in a GDI engine?

Lastly, a significant percentage of water injection users are also adding methanol to the water in amounts up to 50% by volume. Could you comment on the benefits of adding (or not) methanol to the water and at what percentage those benefits reach their peak (I've heard 50%)? Any additional effect in a GDI engine?

Post Comment Login or register to post a comment.

MotoIQ Proudly Presents Our Partners:

© 2018 MotoIQ.com