Indepth study of WAI injection systems -restart

[dvldoc]

The video shows pulses of 10psi or more the spikes also give you a higher reading then what you would get snubbed. If you speedometer or boost gauge did that would you consider it accurate?

The reason you have a snubber is not to break pressure guages do to pulse spikes. They are designed for just these types of pumps.

Secondly the more voltage you add the higher the psi, so when people state
300psi possible its not what your going to get on your vehicle eletrical system.

Number one the pump head is not rated for that nor are the seals or diapharm just because you can does not mean you should, The is pretty much no reason to run that kind of pressure and stress your pump and shorten it's life. You mate up the proper nozzle size with a dynamic range to cover your needs.

Mine is bigger than yours is not always the right way of thinking.

Anyone can achieve those psi if your run the pump at a higher voltage than what a normal auto produces. The pumps can run safely up to 18volts.

We keep it close to specs as possible to make the system last longer. Plus true numbers help the person tunning know how much they are actually spraying per nozzle size. No leaving them thinking I got a M5 it's spraying at 150psi so I get X amount of flow at x amount of volts. If your getting 300psi on a M5 hate to see what happens when you switch to a M3 nozzle and to the internals of the pump after a few months.

Anyways here what we put out with our new pump,

Since today is a federal holiday (10:50 am Guam time for me) I went ahead and tested all the nozzle out. Probably in the middle of the night there for you guys in the states.

Test conducted with a snubbed 3" Glycerin-Filled Gauge with NIST Certificate. Keeps things nice and smooth with no guessing on .exact numbers.

12-30VDC Power supply at a concerevitive 13.25 volts. All numbers rounded down to the nearest multiple of 5.

Most important thing to remember, nozzle size determines output pressure.
M1 (not tested)
M2 230psi
M3 225psi
M3 215psi
M5 190psi
M7 180psi
M10 170psi
M12 (not tested) did not have one handy.
M14 155PSI


So these are are numbers pump pushes plenty hard but not at a level that is to far above the pumps internals specs for pressure. Longevity and proper nozzle sizing is key not how much pressure you can squeeze out of the pump.
We use a different cam in our pump and it keeps things at a level we feel gives you more than enough flow but not to much stress on the pump. If you contact shurflo they will highly, highly recommend you never run 300psi with these pumps. Especially continued use.

Its always better to run a larger nozzle instead of a smaller nozzle at a higher psi with the same output.

This is post is only to show what kind of numbers these pumps put out, at least ours.

* Registered users of the site do not see these ads.

[Aquamist]

dvldoc,

The video showed the spread of the pressure pulses. This is a very useful data for setting the "on-demand switch (ODS). Almost all PPS (progressive Pump speed) system relies on the "ODS" to regulate or limit over pressure. Although Shurflo does not recommend the "cyclic" usage of the ODS but PPS system makers ignore it.

For this reason, knowing the peak switch of the pressure spike is very important to the system designer. This is because the "ODS" response to "peak pressure" and not "averaged" pressure as seen on a snubber pressure gauge. The nozzles see the "unsnubbed" pressure because all PPS systems do not use a " hammer arrestor" or accumulator on the delivery line.

This the very point I was trying to highlight - If the ODS is set to "trip" at 150PSI, the actual operating pressure is much lower due to the 20-30psi pressure swing. It will be worse if you account for "oscillation" due to resonance between The PPS controller and the load/reactance of pump winding.

If you look at the video, the swing of the pressure spike varies with duty cycle. You can see when the oscillation occurs, the swing can be as much as 100psi or more!

I believe your PPS controller run much high frequency than many others - avoiding the dreaded " resonance". We do design and manufacture PPS systems under different labels, I understand the design problem very well.

I totally agrees with you when you mentioned that a few WIA makers use "meaningless" terms as "150psi, 220psi 300psi etc. as a sales tools. I choose to publish your test results because your company has come across as a "no thrills" and honest WAI maker.

I have updated the pressure/flow table. and corrected the 30psi checlvalve to 20psi but upping the "minimum" onset pressure of the nozzle.

[Aquamist]

Originally Posted by Coolingmist

Our video is below. This video is our black pump, a M5 nozzle. The power supply is a 13.8V DC power supply. The gauge is designed for fluid, has a glycerin filled body. Its perfectly suitable for a postive displacement pump. To really make this clear we are using our vari-cool to controll the pump. We did this to show as we change the dutycycle more current goes to the pump. If you connected the power supply directly to the pump it would equal 100% dutycycle on the controller.


The pressure gauge only has upto 200 PSI shown, as I broke our 300 PSI gauge. You will see that it goes far beyond the 200 PSI. The PDA shows the dutycycle, we set the MIN to the controller to 1, the max to 22 and ran 22 PSI into the controller. By doing this, once we set the tune button to 10, 100% of the dutycycle will show. As you watch the tune setting change, the dutycycle increases and the pump pressure does as well. This is with a M5 nozzle.

I should make it clear that I take the other company at there word that the pump they use can only get 138 PSI with the same nozzle. This test is using our pump. I will also point out that we are using 3/16 ID hose (1/4 OD), not 1/8th as they listed.

Lastly, anyone that wants to repeat this test its easy. No controller is necessary. We used the controller just to show the relationship between dutycycle and pressure. We do extensive testing with our products and when I say that a M5 can do over 300 PSI with OUR pump, its because we tested it.

http://youtube.com/watch?v=WxrUJgGyvro


Lastly,

I dont recommend that any of you run pressures this high, it was just to prove a point.

Thank you for contributing. Will you be able to test the same setup up with a "on demand switch "un-bypassed"?

It will be more true to the system you sell to the end user. I presumed you are using a 2.5deg cam angle pump head on this test.

[Coolingmist]

It will be more true to the system you sell to the end user

Not exactly. Our S-HSV kits do not use the bypass switch. Our Vari-Cool kits can safely bypass the switch by simply running a software setup (which is detailed for our customers). The bypass switch stays on the pump and electrically connected, our stage 3 D kits can eliminate the switch all together. We have already showed videos of the stage 3D with the flow rate. When I have the chance I will show a video of our varicool calibrated to a specific nozzle size.

[Aquamist]

I am trying to keep the subject topical and merely ask you to provide some data for a PPS system with a "ODS" UN-BYPASSED.

I am referring to your verycool kit, a 60W pump with a 150psi ODS fitted (un-bypassed) Plus a 22psi checkvalve. See post #22

At present, I am not interested on the many other systems you offered. Perhaps at a future date.

[Coolingmist]

As I said,

I will post a video when I get a chance with our varicool, checkvalve and bypass switch in place. The vari-cool will be setup to inject pulse free by calibrating it to the nozzle size. The same thing anyone can do with our kit.

David

[Aquamist]

That is perfect. Please include stepped duty cycle % vs nozzle pressure data. This way I can chart it and add it to above. If you can do it, I just have to set this up myself. M5 nozzle will be fine.

[smurfwrx]

this thread is a GEM!.... and getting more interesting by the day.

[jblaine]

Almost all PWM pump controller on the market uses Shurflo pump

I believe you mean PPS there

[jblaine]

Originally Posted by dvldoc

12-30VDC Power supply at a concerevitive 13.25 volts.

I don't know about you, but I don't see 13.25v at high RPM.

12-12.5 is a much more real-world number to test by, in my opinion.

[Aquamist]

.
.
new technical update

link

.
.

[jblaine]

Originally Posted by Coolingmist

As I said,

I will post a video when I get a chance with our varicool, checkvalve and bypass switch in place. The vari-cool will be setup to inject pulse free by calibrating it to the nozzle size. The same thing anyone can do with our kit.

David

This was posted ~1.8 years ago. Do you have a video? Your other one several posts above = "This video has been removed by the user."

[Coolingmist]

Originally Posted by jblaine

This was posted ~1.8 years ago. Do you have a video? Your other one several posts above = "This video has been removed by the user."

We have discontinued the 150 psi shurflo pump, what we have left is whats in stock. All of the old Vari-Cool Videos, S-hSV videos and videos of our legacy products have been removed as to not confuse the user.

Our new pumps are the Aquatec bypass pumps that do not have the demand swich so do not pulse like the shurflo version we used previously. What the vari-cool did was set the dutycycle to the point to where the switch on the shurflo pump would not trip and then scale the dc between the min and max. The new CMGS / VC2 can do the same thing, but would only be necessary if someone mates them with a legacy kit.

CM

[Aquamist]

.
.
new technical update with video links and examples
(highlighted)

link

.
.

[NSFW]

Originally Posted by Aquamist

Pulse width modulation type:
(Optimum operating frequency range: 30-80Hz)

This type of system resembles the modern automotive fuel injection system. The system can also be controlled by a third party EMS with a spare PWM channel. Delivery rate can either be mapped or mirroring the fuel injector duty cycle. The latter makes tuning very simple.

The valve behaves similar to an on/off gated button on a garden hose. The longer the gate is opened, the more the flow (duration). Alternative, rapid opening/shutting the gate per second (frequency) also control the flow. The common EMS uses duration for load change and frequency for RPM change. The dynamic flow range is extremely wide, 100:1 is normal.

A WAI valve should closely match the closing and shutting characteristic of a fuel injector. This is important for fuel flow mirroring algorithm since the modern EMS has a correction stage to compensate the opening delay and shutting delay.

Has this approach been tested to verify that an equal amount of water/alcohol gets into each cylinder when it's operating at (for example) 50% duty cycle?

I ask because the operating frequency of the valve shown above is in the range of the intake valves for each cylinder at the RPMs that we're concerned with. For example, suppose the valve is operating at 80hz - that's the frequency of crankshaft revolutions at 4800 RPM. So if the water injection valve is operating at 50% duty cycle (open 50% of the time) and located close to the intake manifold, it seems like there's a very real possibility of spraying water in pulses that are timed perfectly to be delivered entirely to two cylinders, while completely starving the other two.

How far away from the manifold does the jet need to be placed to ensure even distribution of water over time?

[Aquamist]

This type of system can only mirror the flow characteristic of a fuel injector. This the same as putting a fuel injector inline of a pressuring line with water or methanol. How well each cylinder receieve equal amount of injectant will depend on the placement of the jet and the fineness of the atomising nozzle.

The location of the jet is often being discussed on various dates and forums. I can only suggest some steps to improve the even distribuition of linjectant as follows:

1) Place the jet as far away from the TB as possible to improve the mixing with the charge air along the inlet tract before entering the cylinders.

2) If droplets are allowed more time to evaporate, the smaller droplet (less mass) will improve even distribution to each cyclinder.

3) Pulsing delivery can creates gaps or pockets of injectant along the inlet tract so each cylinder may not receive equal amount of injectant during the change of RPM and load. To counter this effect...
a) a nozzle with a semi-hollow cone should be used so that the droplets travel at different speed so the gaps is filled during the journey to the cylinders.
b) two nozzle placed at a few inches apart will further minimize the injectant gaps.

I have spent some time working on this with several WRC teams during the 90s. You will find by the beginning of 2000, all wrc cars (with the exception of the French - not confirmed) have their nozzle placed very close to the exit of the intercooler. Few years on, Direct port injection is being employed but only in service for a few seasons before water injection is banned.

Please help contributing by offering your thoughts on this. I have been advising most aquamist users as we are using this mehtod of injection.

[NSFW]

People are posting impressive dyno charts every week that show that injection can stave off detonation, but as a potential user who is interested in safety more than power, the prospect of uneven distribution doesn't sit well with me. But without testing, and measuring with a 4-channel cylinder pressure setup or something like that, I can only speculate on how significant it is. Even a 4-channel EGT setup would be of limited value since RPM variations would move the the peaks and valleys between cylinders, faster than the EGT sensors could react.

7000 RPM is about 120hz. Given four alcohol injectors that can operate at that rate, I'd be satisfied. Anything less makes me wonder. It seems to me that the ideal setup would would use two-injectors per cylinder, with one set of injectors spraying fuel and the other spraying water/alcohol, but apparently nobody makes injectors that are safe for water or methanol. (I've been searching a lot over the last couple days, planning to use Megasquirt or other cheap ECU to drive the extra set of injectors.) With a second set of injectors, you could fire the extra injectors using signals that was based on the fuel injector signals for each cylinder, and be guaranteed proper distribution. And I like the idea of using a send-return system with boost-referenced pressure, just like fuel systems. And given how reliably fuel system components perform every day in millions of cars, I'd be a lot less worried about component failure. But alas, nobody seems to make compatible injectors.

[Aquamist]

Ther current water injection system is quite basic compared to the advancement of fuel delivery system to date. Direct in-cyclinder injection will soon be taking over all the current fuel manifold injection system. Only a handfull of high-dollar cars are using them at the moment.

This trend also means the WIA injection system are falling further behind. Most budget system on the market is still uses a pump speed system that will not be able to cope with fast transient and wide dynamic range (load). For those system, it relies on the tuner to make up the AFR gaps and produces a less than perfect tune. As far an most people are concerned, they are only interested in the peak power figures.

At 7000 RPM, I believe the injector has to respond to 120Hz, if they are batched fired. But if you can put up with the expenses of having four inline valves and fire them sequentially, they only need to respond to 30Hz intervals between each event. Of course a manifold pressure compensated water/alsohol line will improve the set up further.

We can design and supply a fast acting valve that will match the fuel injector's flow. opening and closing time, but I don't think we have a large enough market place to justify developing such a system. We have already attracted quite a few negative comments on our current IDC system, based on jusy one fast acting valve. I guess you can see where the trend is leading ... not the direction you are looking.

[NSFW]

Good point about lower frequencies with multiple injectors.

I don't see a need to make the meth injection valve match the fuel injectors though, if you've got a circuit that can watch the injector pulses, subtract a couple/few milliseconds, and pulse the meth injector with with whatever duration is left over. That would give no meth flow and low IDCs, with meth flow gradually ramping up with air/fuel flow from there.

You could make it microprocessor-controlled and table-driven if you want to make it fancy (RPM on one axis, fuel IPW on the other axis, and meth IPW in the cells).

Does the valve above work well at lower rates? Down to 10hz for example?

Can be put into a housing with an integrated nozzle and dimensions resembling fuel injectors? What would it cost for a set of four?

Combing those with 2-injector-per-port TGV housings, and a water/meth-compatible boost-referenced fuel pressure regulator, and it's halfway to a sweet kit.

Expensive, but sweet.

[Aquamist]

The reason for matching the characteristics of the two valves is just make life easier. Once matched you do not need to do any corrections to compensate the open/closing time, just boost the raw signal and your are done.

On the other hand, you use any un-matched valves you need to spend time to play with the electronics to match the flow (correction tables). Either way, there is no short cut if you want a system that mirrors the fuel flow with a reasonable accuracy.

This is of course not needed if you are going to make your own map as mentioned, a XY grid. But the drawback would be mapping time the dyno, everytime you alter the meth map, you need to alter the fuel map so that you can consistant octane/cooling balance. it will be time consuming.

We did look at all algorithm and decided to track IDC - not perfect but very simple to tune.

The valve can work 10Hz and it will be a smooth flow. The valve a quite large due to ther 1/8 BSP fittings on each side. We can intergate a nozzle on the lotlet side of the valve - if you like, I post a picture here. These FAV (fast acting valve) is quite new, I don't have the price yet - depends very much on your specifications. We haven't planned to sell this valve separately commercially. For occasional project as yours, there is no problem.

Sounded like something good is brewing...

[NSFW]

Thing is, I wouldn't want the water/meth flow to be proportional to fuel at all times. There's no need for it until boost starts building. Subtracting a few milliseconds from the car's injector pulse width seems like a pretty straightforward way to get meth to come on gradually when needed, and increase with fuel flow.

I don't see time being an issue with a table-driven system, unless the someone wants it to be. You could set the table to behave just like a simpler system and then tune everything else around that.

[Aquamist]

I guess a system that will allow the user to specify the starting point along the IDC ramp and following on thereafter.

A smiple trimmer that will alter the fuel/methanol ratio, based on application. For example, larger turbo requires extended/additional methanol flow beyond the wastegate setting.

Perhaps a small single jet at the exit of the the intercooler will perform the induction cooling.

[Aquamist]

.


.

There has been any questions asked about the best place to install the water jet,
it would be great for everyone to chime to to discuss this and post their experiences.

I will start...

.

[Aquamist]

Most WI manufacturers recommend the ideal position for the nozzle is at the throttle plate instead at the exit of the intercooler. I beg to differ and the reasons are:
(TMIC doesn't have much a choice, but the discussion is mainly in general and for FMIC)


1. The further way the drops have to travel along the inlet tract befoe entering the cylinders the better. This allows more time for the drop to perform intercooling work.

2. As the droplet evaporates, the size will reduce - the fundamental process of the evaporation.

3. Smaller droplet size will improve even distribution of each cyclinder receiving the droplets.

4. At the TB, the droplets are more likely to hit the throttle plate and pool along the surface and reforming to larger droplets, overfill the first cylinder.


.

[SnowTech]

Great points, but there are some other things to consider:

1) BOV plumbing. Recirculating the mist is probably not going to be a big problem, but if you are venting to atmosphere it's not neccesarily good, especially for those that use pure methanol. This may be a problem with some FMIC setups.

2) When you look at the velocities involved in say a 2.5 inch intercooler pipe on a setup moving about 40LB of air per minute (very common numbers for a EJ20 or EJ25 with a mild turbo upgrade) a difference of 2 feet in the placement of the nozzle is almost inconsequential in terms of the time added for evaporation. The bulk of the time droplets spend absorbing heat before passing the intake valve is in the intake itself, especially in the larger plenum aftermarket pieces.

In the end the most important aspect is ensuring that the nozzle has a clear spray pattern into all of the airflow.