Is Water Injection a Cheap Substitute for Race Gas?


story by tom suddard and dan hurwitz • photos by dan hurwitz

Water is bad news for almost every internal engine part. It doesn’t really compress, lubricate or ignite–three strikes against it. Plus, if an engine ever receives more than a gentle cleaning with H2O, chances are it will soon need to be replaced–that’s why many Jeeps have snorkels and one of the reasons most factory air boxes are so restrictive.

Of course, this isn’t really late-breaking information. Here’s what is surprising: A company called Snow Performance bucks that conventional wisdom, billing itself as “the water/methanol injection experts.” Ludicrous idea? Let’s investigate.

Snow’s Boost Cooler kit doesn’t inject straight water, but rather a mixture of water and methanol similar to windshield washer fluid. In fact, the company says windshield washer fluid works perfectly if its own Boost Juice mixture isn’t used. This mixture, stored in a tank mounted either inside the trunk or under the hood, sprays into the engine’s intake under load.

Cooling Down

The growing popularity of turbocharging has changed how high-performance engines make power: replacing expensive, high-compression heads and pistons with inexpensive boost pressure increases. However, the need for high-octane race gas remains.

High-octane race gas resists detonation more than low-octane pump gas. The fortified gas is effectively harder to ignite, and as a result, it allows for extremely high compression or extremely high boost, depending on which way you wear your baseball cap, how your shoes are laced, or whether you know how to program your DVR.

However, with the pros come the cons, and race gas comes with a high price–$10 or so per gallon–as well as limited availability. It is also typically leaded, making it a dangerous neurotoxin. Because of this, it’s poorly suited to daily drivers that also hit the track on weekends.

Snow Performance claims that water/methanol injection is the magic bullet for this $10-a-gallon problem, simulating race gas with the clever application of water/methanol droplets and pump gas. This additive cools the intake charge and prevents detonation, effectively increasing the octane rating of pump gas to race-gas levels. At least, this is what Snow Performance claims, so naturally we had to try it out for ourselves.

The company’s Stage 2 MAF Boost Cooler system, which comes with everything needed to quench an engine’s thirst, retails for around $465. Next, we needed a test car: a nearly new Impreza WRX STI. In addition to an aftermarket intake and exhaust, this late-model Impreza craved a bit more power. And with the help of Snow Performance and the tuning pros at Mach V Motorsports, we were going to provide it.

New Plumbing

Before we could replace race gas with water, we had to install the injection system. Installation was really just four simple decisions followed by some plumbing and wiring to string them together: Where do we want the pump, where do we want the reservoir, where do we want the controller, and how do we want to trigger the injection system?

We decided to place the tank in one corner of the rear hatch, and we hid the pump next to the spare tire. We mounted the controller in the glove box and chose to trigger the system based on boost pressure. We configured the controller to begin spraying into the engine at 10 psi, then gradually increase the water/methanol dosage until the engine achieves maximum boost at 18 psi.

With the core components mounted, it was time to plumb the water line. We ran the line from the tank to the engine bay along a frame rail, making sure to keep it protected. This car is regularly rallycrossed, and we didn’t want flying rocks to cause any leaks.

Under the hood, we drilled into the stock intercooler end tank and threaded the appropriately sized nozzle into the hole. Then we plugged our water line into the nozzle. The last step was wiring the controller and pump, which was easier than any car stereo installation. The hardest part of the deal was feeding the wires back to the hatch.

Open the Taps!

Installing the system was only half the project, though. Our STI would need to be professionally tuned on a dynamometer to take full advantage of its new, wetter fuel mixture.

First, a baseline: On Mach V’s dyno with Cobb Accessport tuning, company owner Dan Hurwitz saw 293 horsepower and 316 ft.-lbs. of torque at the wheels. Those are strong numbers for a slightly modified WRX STI–or so we thought.

Then we turned on the Snow Performance system. After more tuning on the dyno, the Subaru was making 319 horsepower and 350 ft.-lbs. of torque, a substantial increase–all thanks to the same formula used to keep windshields clean. The water/methanol injection allowed the Subaru to run more boost and a leaner air/fuel mixture, and we were actually able to hit the airflow limit of the stock turbocharger. We were obviously impressed, even with the system’s downside in mind.

The River Runneth Dry

Downside, you say? Well, yes. Because the engine is now tuned under the assumption of the water/methanol mixture being sprayed in under high boost, bad things can happen if the tank runs dry. It’s analogous to running pump gas in a car designed for 115-octane race fuel: It may be okay, or the engine may blow.

For particularly paranoid customers, Snow Performance offers fail-safes to prevent engine damage if the tank runs dry. Additionally, the kit we used has a level indicator that lights up when the tank is low. A tank of Boost Juice lasts about as long as two tanks of gas, so just check it whenever you fill up. If it runs dry, just walk into the convenience store and buy some more washer fluid—but to keep the system's warranty, order Boost Juice from Snow Performance.

Washing Up

So what’s our verdict? Water/methanol injection definitely works. It produces tangible power gains from an exceedingly small investment of time and money.

Sources

Mach V Motorsports, LLC
dyno tuning
machv.com
(571) 434-8333

Snow Performance
water/menthol injection system
snowperformance.net
(866) 365-2762

Join Free Join our community to easily find more Subaru articles.
Comments
View comments on the GRM forums
AWSX1686
AWSX1686 Dork
1/9/18 11:47 a.m.

My diesel has water/meth injection and makes a noticeable difference.  It was installed before I got it though, so I don't know much about it. But I now have a water/meth kit that will be installed on the challenge car, so fun things to come!

Rcottrell
Rcottrell
1/9/18 4:28 p.m.

I ran a Edlebrock Water/Meth Injection back in the 80's on a carb Motor.  It made great torque then too.

kb58
kb58 SuperDork
1/9/18 4:28 p.m.

While it works, my feeling is that it's just more stuff to fail (or run out of). Proper systems cost $$$, and then there's where to put everything. It's not just running out of coolant, it's having the pump fail, or perhaps worse, having the nozzles partially plug up. Unless the system's really smart (that's the $$$), it could well be squirting little to nothing yet looking like everything's fine. I ended up taking everything out and installing a big-ass air-to-air intercooler and ducting. I realize this thread is linked to the magazine article, but for me, it was too large a risk.

snailmont5oh
snailmont5oh HalfDork
1/9/18 4:47 p.m.

A buddy of mine has an unrestored survivor '53 Allard K3. It has a Cad engine with three carbs. Each one has a tube sticking into the air cleaner lid, which goes to a glass bottle on the fender. Air moving past the end of the hoses causes a vacuum on the lines, which causes water to mix with the air charge. I knew another guy that had the same system on his '83 Ranger 4-cyl. Same brand and everything. He had it so he could push the timing on the 2.3 to try to get some power out of it. The system works well in both applications. 

Tmaxx94
Tmaxx94 New Reader
1/9/18 8:08 p.m.

So...How stupid would it be for me to do this to my 2000 blazer with 201k on the clock?

Trackmouse
Trackmouse UltraDork
1/9/18 8:08 p.m.

The ultimate answer here is to have two tunes to switch to. One for meth, and one for Jesus. 

te72
te72 New Reader
1/10/18 7:42 p.m.

Saw a homebrew looking setup on an older Volvo in a junk yard quite a few years back, thought that was rather clever of them, given the apparent age of things.

 

Have a kit that will be going on my Supra, once I get a TIG and the confidence to weld aluminum with some degree of skill. Car makes plenty of power as it is, so it's more for just keeping the intake temps down than for the ability to crank up boost pressure and timing.

Vigo
Vigo UltimaDork
1/10/18 10:00 p.m.

Water injection is pretty great, but it is somewhat less great than injecting other things like Methanol while requiring the same basic amount of hardware to achieve, so a lot of people who would consider water injection decide to inject something else instead. 

I agree with the statement about more things to fail. On the extreme upside of water or methanol injection would be my friend's old Plymouth Sundance that he turboed and got about 4x the CHP out of (~500whp) with STOCK IGNITION TIMING using methanol injection. Methanol is a fuel, and he was injecting about 200hp worth of it. So what happens when you need 500hp of fuel and something happens and you suddenly are missing 200hp worth? That's the downside: Building a setup that takes full advantage of having something usually means it is going to blow up when it doesnt have it. Same thing can be said for 'e85 tunes'. 

It is not that hard to build in safeguards for things like water/meth failure, but people rarely do it. E85 'blend sensors' are at least becoming semi-common.  I personally think that if you are implementing a safeguard  to go along with 'it', then it's great. A little electrical knowledge will allow you to piece something together pretty cheaply. For example, i bought a ~$70 2-in-1 gauge from Auber Instruments that has programmable alarm outputs. If i choose to use it to monitor manifold pressure and wideband AFR, i can easily rig up a series circuit that goes through both alarm output 'switches' so that IF boost is above 'x'  AND afr is above 'x' (=lean), i can do something like trigger an NC relay in my ignition coil primary circuit and instantaneously shut it down, but the car can still run lean under no-boost conditions. Something along the same lines with a pressure switch in the water injection line is equally doable, and possibly even easier if you already have a fancy engine management setup that can do all of this in software instead of rigging up additional sensors and switches and outputs to do it. 

AWSX1686
AWSX1686 Dork
1/11/18 7:55 a.m.

In reply to Vigo :

The article was really for water/meth injection, but I suppose straight methanol would still be different. 

Everyone making the points about safeguards for if/when the system runs out or doesn't work when the car is tuned for it is something that hadn't crossed my mind, so I will make sure to build that into my build. 

 

That said, on my diesel truck, the system pumps the washer fluid from a second pump in the stock washer fluid tank, which means when it gets low a light comes on. ;)

kb58
kb58 SuperDork
1/11/18 12:15 p.m.

Oh it's awesome stuff, but I have to underscore the importance of being able to quickly detect when it's not being injected. Everyone's correctly noting the importance of detecting when the tank's low, but the real danger is when the injector nozzles partially or completely plug up. In that situation, the tank level switch, and driver think all is good and keep powering away, all while intake air temperature is going through the roof. Hopefully the ECU notices the high IAT or knock. Some systems (like AEM) use a flow sensor, which is really the only way to know for sure what's going on, but it's part of a $$$ kit.

BTW, another reason I got rid of it was due to smelling the methanol in the 50/50 water/meth mix while it was injecting. Read up on methanol - it's nasty evil stuff in any form. Alan Staniforth, author and builder of many British cars, died of lung cancer, thought to have been caused by his methanol  exposure in both liquid and fume form... For me, I chose to avoid the whole system, so to answer this thread's question - "No."

djsilver
djsilver Reader
1/11/18 3:26 p.m.

I know methanol is nasty and was going to ask if anyone had tried doing the same with ethanol, but before I did I thought I'd google "ethanol injection" to keep from feeling stupid.  The only hit I got was for PEI, which is Percutaneous Ethanol Injection as a cancer treatment.

I tried again and searched for "ethanol injection car" and got this;

https://www.caranddriver.com/features/ethanol-injection-systems-explained-tech-dept

 

 

iceracer
iceracer UltimaDork
1/11/18 5:16 p.m.

-I remember my dad rigged up an alcohol injection vacuum operated system.   I believe he used  anti-freeze which was what was in use at the time.     He used a glass bottle which was not secured well and it broke dumping the alcohol onto the spark plug resulting on some flames.

AWSX1686
AWSX1686 Dork
1/12/18 7:56 a.m.
kb58 said:

Oh it's awesome stuff, but I have to underscore the importance of being able to quickly detect when it's not being injected. Everyone's correctly noting the importance of detecting when the tank's low, but the real danger is when the injector nozzles partially or completely plug up. In that situation, the tank level switch, and driver think all is good and keep powering away, all while intake air temperature is going through the roof. Hopefully the ECU notices the high IAT or knock. Some systems (like AEM) use a flow sensor, which is really the only way to know for sure what's going on, but it's part of a $$$ kit.

BTW, another reason I got rid of it was due to smelling the methanol in the 50/50 water/meth mix while it was injecting. Read up on methanol - it's nasty evil stuff in any form. Alan Staniforth, author and builder of many British cars, died of lung cancer, thought to have been caused by his methanol  exposure in both liquid and fume form... For me, I chose to avoid the whole system, so to answer this thread's question - "No."

So from a safety side of things, maybe alcohol injection is a physically safer route? I would suspect it isn't as good on the intake cooling side of things, but I don't really know anything about it. 

Vigo
Vigo UltimaDork
1/12/18 8:24 a.m.

Latent Heat of Vaporization is basically the measure of how much cooling effect you're going to get and you can look it up for various things you could inject. Some of them are easier/safer to get, store, not clog up nozzles with, not corrode the inside of the pump with, etc etc.

rslifkin
rslifkin SuperDork
1/12/18 8:32 a.m.
kb58 said:

BTW, another reason I got rid of it was due to smelling the methanol in the 50/50 water/meth mix while it was injecting. Read up on methanol - it's nasty evil stuff in any form. Alan Staniforth, author and builder of many British cars, died of lung cancer, thought to have been caused by his methanol  exposure in both liquid and fume form... For me, I chose to avoid the whole system, so to answer this thread's question - "No."

This brings up an interesting safety point about washer fluid too.  The fresh air intakes on cars really shouldn't be right by the windshield where the methanol-laced washer fluid ends up...

Suprf1y
Suprf1y PowerDork
1/12/18 9:03 a.m.

I haven't read the article but last time I was looking at these systems I didn't like the controls so I built my own. I built an adjustable injector controller based on a 555 timer circuit, machined a block with a boost referenced regulator, a high pressure diaphragm pump, and used a SS injector. The injector cycle was based on both RPM and boost and worked very well.

Keith Tanner
Keith Tanner MegaDork
1/12/18 9:37 a.m.

We used to use Snow stuff - they're not a new company. To implement it properly, you put a flow sensor on the water/meth line. That means you can detect a low tank, a failed pump or a clogged nozzle. When the flow sensor indicates a problem, the ECU should react by pulling timing, adding fuel or dropping boost.

If this sounds like the correct way to implement E85 (with an ethanol sensor in place of the flow sensor), you're right. It's the same concept.

In both cases, you need to measure what's happening and the ECU has to be able to use that information. Where you get into trouble is when you make assumptions: assume that the water is flowing (or that your E85 is 85% ethanol), assume that a stock ECU will be able to adjust, assume that a driver will see a warning light and react.

We actually ran a water injection system in our Open Track Challenge car back in 2002. For those who weren't reading GRM at the time, call it a week-long Time Attack event. Like One Lap, but with less driving between tracks and a full track day at each one. That one was pretty crude, no feedback from the flow and the ECU couldn't react. It worked, but required race car levels of maintenance and inspection to avoid problems. That car also had an intercooler sprayer - coming off at the track at Willow Springs on a 100F day, the IC was cold to the touch.

Our Preferred Partners
QVn6mlue1Ehhaz2RQxIOMoky2OoVkuGbdHZaT8gH00dyWeyrTnKMdcJONeQO435q