Preturbo 335i Strange Results

[rudypoochris]

Originally Posted by Richard L

It is quite tricky to set up an air assisted nozzle. You really need minimum amount of pressure on the fluid and ample air pressure. If you have too much fluid pressure, the atomisation is not that good.

Look at the following pictures, the difference is very visible.

I can throw a pressure regulator onto it no problem - can you recommend a pressure? Currently flow is controlled using a simple valve with a screw on it. I can tighten that up as an alternative.

[Richard L]

I cannot say a specific pressure that will work. You can to get a couple of inline needle valves and calibrate it.

I am not much good at mechanical solutions. It requires patience. I don't have an electronic solutions either. Might look into this in the future.

Just to give you can some idea how sensitive this is. The water is drawn into the nozzle by capillary action, aided by the partial vacuum created by the venturi effect of the pressured air running over the water line (black hose).

When I alterd the height of the nozzle relative to the water level, the atomising quality varied a great deal. I carried out this experiment out of curiosity. I modified one of our standard nozzle to be air assisted with some success but the difficult part is getting it to work with consistence. It turned out to be very involved. Something to revisit in another day.

If STFT and WGDC is not conclusive. Individual timing is not available. You are now left monitoring temperature. Perhaps this will give you some useful feedback. What about mis-fire code?

[rudypoochris]

No misfire. IAT rising with meth. Maybe too much fluid condensing on the IC driving up temps? Still doesn't explain knock since the E85 is good until at least and 3 psi over these logs. I could see the atomization being poor and droplets of water entering the combustion chamber on first onset if the liquid pressure is 17 psi and the air only 8.

[Richard L]

You need to do two things:

1. Get the atomisation good
2. Either measure the flow or manually measuring it over one minute.

It is difficult to do it without a compressor. May be empty a jet into a sode bottle after one minute of boost (careful, not many roads and sustain one minuter of full boost).

Calibrate the flow sensor with a normal flow. Jeff told me you have problem with the logged sensor signal. If the BMS logger is not reliable, read the floe signal with a voltmeter. You only need to make a quick note of the voltage at full boost. Once done, repeat it on a bench with water from a tap and match the same voltage for one minute. see how much water is collected.

[rudypoochris]

Originally Posted by Richard L

You need to do two things:

1. Get the atomisation good
2. Either measure the flow or manually measuring it over one minute.

It is difficult to do it without a compressor. May be empty a jet into a sode bottle after one minute of boost (careful, not many roads and sustain one minuter of full boost).

Calibrate the flow sensor with a normal flow. Jeff told me you have problem with the logged sensor signal. If the BMS logger is not reliable, read the floe signal with a voltmeter. You only need to make a quick note of the voltage at full boost. Once done, repeat it on a bench with water from a tap and match the same voltage for one minute. see how much water is collected.

I can use the retained boost in the tank to check flow. The BMS logger is fine. I think I am getting interference from the car. I suspect the flow readings to be approximately correct though. Will look into it.

[Richard L]

Not sure how your method of measuring flow but will accept your posted flow rate of 650cc/min. Potentially you can remove up to ~30KW of heat if the water is fully evaporated.

I same Jeff's oscilloscope trace of read on more than 0.06V of noise (includes back ground noise). I think your public comment on one of the forum stating the aquamist flow sensor has a oscillating signal of 0.2v to 0.8V. that is 0.6V is unfounded. I would be grateful if you can make a post to correct this finding.

All good logging device should have some sort of high-pass, low-pass or band pass filter. I guess the BMS logging software does not have this facility. This is what I meant before as unreliable.

[rudypoochris]

Originally Posted by Richard L

Not sure how your method of measuring flow but will accept your posted flow rate of 650cc/min. Potentially you can remove up to ~30KW of heat if the water is fully evaporated.

I same Jeff's oscilloscope trace of read on more than 0.06V of noise (includes back ground noise). I think your public comment on one of the forum stating the aquamist flow sensor has a oscillating signal of 0.2v to 0.8V. that is 0.6V is unfounded. I would be grateful if you can make a post to correct this finding.

All good logging device should have some sort of high-pass, low-pass or band pass filter. I guess the BMS logging software does not have this facility. This is what I meant before as unreliable.

I updated that thread just now to clear it up. The jb4 has a fat filter in front of it and reads clear for many other voltages sources. I don't think it is the jb4 but I also don't think its the sensor.

I logged the sensor with the DMM, JB4, and an independent 100hz logger. The oscillation is there. Likely it is not the sensor itself but noise as I have 5ft of line across the engine bay. I also talked to my coworker who is an EE and he suspects bad ground or noise. I will remove the sensor from the engine bay and test. If its still nasty looking, I will cut the wire at the sensor, ground it there, and install an LM7805 to control the scalar voltage to 5V precisely.

The logging on the JB4 reads 0-5V and thus half scale is equivalent to roughly half flow. That is how I got the ~650ml/min figure. That figure is also what the manufacturer quoted the nozzles to flow at, but obviously it can vary.

[rudypoochris]

This is the 100Hz log using my USB logging tool. 3.33Hz oscillation, again probably interference.

[Richard L]

The oscillating signal is 0.8v (pk to pk), quite a large swing.

The aquamist sensor generates a "sensor presence" voltage of 0.53V to alert the controller of its readiness to read flow. above this voltage, it is allocated for flow monitoring. The flow signal peaks at ~4.5V. Above this signal, it is a fault. So the sensor itself has inbuild failsafe to warn the aquamist controller.

Back to the plot. The BMS logging system appears to have picked up some noise and superimposed/modulated itself onto the aquamist's. The source is unknown but worth finding it out. It is definitely not from aquamist.

Jeff (aquamist US representative) has power up a similar age sensor to yours and sent me this captured image from his oscilloscope. this is what it looked like:

Superimpose (scaled) onto the BMS plot. it looks like this:

Where are those ghost oscillation came from? I am very interested to find out. Try disconnecting the green wire and make another plot.

[rudypoochris]

I will test. Again, this is NOT the BMS log. This is from a completely different standalone 4 channel logging tool which confirms what my JB4 (BMS) was seeing as well. This oscillation must be coming from somewhere, likely interference.