Get your own Free Home Page
Get your own Free Home PageAs stated on my previous page, I found that water injection was the most effective and simplest method of charge cooling and suppressing detonation. However if I wished to run boost levels above 10-12 psi I needed to inject at least 500 cc/minute of water.
The temperature rise with a 65% efficient turbo at 10-12 psi is ~100�C. At 15 psi it rises to 125�C 30 psi gives 200�C. To decrease air temp by 100�C (ie. to drop a 12 psi system back to ambient) for a 3 litre engine at 6000 rpm requires 460 cc/minute water. Add a bit for safety and it looks like 500 cc to me.
The problem with that is that at very low boost levels and low revs such large amounts of water serve to reduce power levels, which only serves to reduce response at part throttle and increase turbo lag.
The solution, is to have a more sophisticated injection control system rather than a crude on/off switch. Many of the Buick sites use a 2 stage system, with the second, [larger] injector kicking in at about 10 psi. This sounds ideal and the only question was how to implement it.
Option 4 was the first one I actually put into practise. This is because of the fortunate find of my flatmates falcon part. Makes me wish I had taken more parts before he sold the car.
My first step was to find a vacuum source that was proportional to water requirements. This is easy. Assuming that the car is actually on boost (ie. the pump is on) then the water requirement is roughly equal to the air intake. And the vacuum in the intake prior to the turbo is proportional to the air intake. Well actually it is proportion to a quadratic function of the air intake, but as we are only after one point in the function, when the secondary nozzle will turn on, that is close enough.
The next step was to find the point at which the vacuum is sufficient to switch the secondary line on. That was done with the setup shown to the right. With the lines running to the cabin I was able to see when the valve switched and hence relate it to airflow.
At full throttle the valve activated at 4000rpm. According to the power curves this corresponds to an output of 150 kW.
The activation vaccuum of the valve can be adjusted up (but not down) using exactly the same bleeder valve system as was used for the wastegate actuator.
Sadly I was never able to get the valve to function correctly while it was controlling 50 psi of water pressure.
A possible problem was discovered with the existing water injection system. Although the dual stage system was tuned to kick in at a given airflow, the base system itself, controlled by whether the pump was on or not, was controlled by a throttle position switch. Although this gave excellent response to sudden throttle opening it had two problems.
Firstly it took no account of rpm. Should the throttle be opened at 1500 rpm the full flow of the primary water injection circuit would turn on. This is despite there being no boost for a few seconds after full throttle is applied at such low engine speeds. In such a case the water flow would serve only to reduce engine power and responsiveness. This problem was greatly alleviated by the introduction of the 2 stage system. This allowed the primary injector to be reduced in size and hence this problem minimized.
Secondly there was the problem of the engine developing significant boost without actually being at full throttle. In this case the 2 stage system doesn't help.
A solution to both these problems that also alleviates the need for 2 stage injection (while not actually eliminating this need) would be to have boost activated injection rather than a throttle position switch.
A Boost switch is a simple thing. I merely used a hyperdermic syringe (I work in a medical research company, no I'm not a junkie) and a push button switch. Boost is fed into the syringe and when it reaches a high enough load, it pushes the switch on, and the water injection pump turns on. Dead easy.
And it works too. I can now run 500 cc/minute injection without flooding the engine at low boost and low rpm. The system turns on at about 4 psi manifold pressure (after the throttle), by which time there is enough airflow to absorb that much water.
Now I can REALLY turn the boost up. Next stop, 15-10 psi. :>
For the last 6 months or so I've not written anything. But now I've finally finished plugging all the boost leaks in my car's intake system.
For some months now, my car has been suffering from a series of annoying boost leaks. You get this when you turn the boost up to three times the factory level I suppose.
The problem manifested itself as hard starting, poor idle, stalling, and loss of boost. It was the loss of boost that gave the game away, because this was associated with an extra hissing noise from under the bonnet. This eventually suggested to me that boost was escaping from somewhere.
So I looked about and eventually found a culprit, the rubber coupling between the turbo and the intake pipe had split, near one end. The air was leaking through the crack. This meant that the car was running too lean at idle, when too much air was getting in, and then at boost, too much air was getting out.
So I cut off the damaged end, and replaced the pipe. This made the problem better, but it was not fixed. Then it deteriorated again. So I took apart the intake system again... to find no problem. This time I went over the whole car, replacing spark plugs, leads, oil, and all those things that are supposed to be replaced sometime this year anyway. Eventually the new culprit was found: the overboost blow off valve.
This valve is designed to open and dump boost in the event that the wastegate is stuck and boost rises to dangerous levels. It is set to open at about 7 psi in a stock car. I had previously increased the spring loading to enable it to remain closed when I was running 10 psi, but now that I had increased boost to 15 psi, it was dumping boost that I wanted to keep.
It was also possible that it was not closing properly once it had opened, thus leading to my starting and idle problems.
So, I increased the spring load even more. Hopefully that will keep it shut.
Once again, the problem got better, but not fixed, and then gradually deteriorated again. Hmmmm.
After going over the whole engine once again, I took the intake tube off the turbo again, and found that the rubber coupling had split again. This time I replaced it with a fibre reinforced section of radiator hose. This was very had to fit, at it was much thicker (hopefully stronger) but using lots of hot water and allowing it to creep into time over a week or so, eventually did the trick.
Once again, the problem got better, but not fixed, and then gradually deteriorated again. Arrrgh!
This time I got a plumbing blanking cap and totally sealled the boost dump valve.
Once again, the problem got better, or at least, the boost leak problem appeared to be fixed! But the idle and poor starting remained. Could it be that I was actually dealing with two separate problems?
I investigated this aspect. The boost leak issue seemed to be solved. However the idle and starting, while better than when there were big hose splits, still needed work. Furthermore, the problem now seemed to be only when the engine was cold, and the colder the engine, the worse the problem.
This was not likely to be another leak. I could not think of a way that there would be a temperature dependent leak, that would not be leaking boost at 15 psi.
Reading the webpages suggested a completely different explanation. The cylinder head temperature sensor has a significant effect on the fuel delivered during cold starts and cold idle. If the sensor is recording too high a resistance, then the engine will idle badly when cold, but be OK when the engine reaches the temperature that it already thinks it is.
This will have to be investigated further. But meanwhile, when the engine is warmed up, it should be running just fine. So why isn't it putting out heaps of power? Or have I just gotten used to it running at 15-10 psi (15psi initially dropping to 10 psi after a few seconds to keep everything safe after the engine has heated up)?
So last night I redid the acceleration test. Times of 5.44 and 5.40 seconds. Sticking this into my formulae gives me an average 108 kW at the wheels, which correlates roughly to 190 kW at the engine.
Still not enough of course. But I have just gotten used to it. It seems slower with 190 kW than it did with 170 kW back at the end of 2000. But the stopwatch says otherwise.
Go To Patrick's
Homepage