Injector Flow Testing Data

[vlad01]

the ECU works that out already. but for it to work it out we need the flow rate(easy) and dead time vs voltage.

getting the dead time is hard to find info without a lot some fiddling to ball park the numbers.

nearly all manufacture don't supply the data which is a pita

[The1]

I had the idea for a while its based on my coding test bench, I put some code in to only run for 15 seconds , one thing I'd like to do is gather data from testing injectors as I've never really found any decent info on the net, even basic flow rates and pressures its tested at. Will have a look at code as I get time as to what else I could do atleast so far I can do more accurate testing then most flow benchs and give more data and more real world usefull data, more ideas is always good.

[delcowizzid]

i have a flow sheet from deutchworks somewhere ill dig it out for you to have a look

[The1]

Awesome, also for previous posts how would deadtime help tuning? If flow is in spec at both ranges being tested isn't that the main thing?

[festy]

I was thinking about ideas to map injector dead time a while ago, and the best idea I could come up with was to measure the solenoid current and work out when it's fully saturated from that (a bit like a peak/hold driver).
An arduino should be plenty fast enough to fire the injector and measure the voltage+current to within +/- 0.5ms without having to mess with the ADC clock prescaler.
But... can the injector dead time actually be measured from the solenoid's current draw? Does the coil saturation directly relate to opening delay?
There's a paragraph or two in this patent application that indicates that you can determine delay from the current profile:

In the domain of the current waveform, the captured location in time of the solenoid movement can also be utilised in a feedback system to a controlling device that is accounting for solenoid 'dead time', the physical delay in actuator movement which for example, affects total fuel flow out of a fuel injector when it is energised. This permits more accurate fuel metering to an engine in real time, from real solenoid performance data instead of pre-programmed fuel injector dead time or offset lookup tables.

Tell-tale markers of the time of solenoid movement, time to reach hold current, time to reach peak current, and voltage waveforms are now utilised to show numerical and graphical data on solenoid performance. The markers indicating angle of the current rise waveform can be utilised to also detect and report on solenoid coil faults of low or high impedance windings and shorted turns. Figure 5A illustrates typical operation waveforms of a solenoid, in which 51 represents live and offline comparisons of multiple solenoid waveforms, tm(channel): Time from energising to movement of solenoid, and tp(channel): Time from energising to peak current reached In saturated mode the time to an expected target current instead of a peak or hold current is used. In this embodiment, solenoid (A) is a slightly lower impedance than (B) though it would be expected to be within solenoid waveform tolerance limits here.

I haven't read the whole document, but it could be worth experimenting with. Start with a known injector that we have the dead time table for already, and see if comparable results can be obtained from a test rig?

[VL400]

Using just current would be preferable as it can be done really easily with a simple micro controller test rig, but have also heard of piezo response being used to listen to the pintle moving and when it physically reaches the open and closed end points. Combined with measuring current can probably use that to look at the current profile and determine the points that are in the patent application.

[festy]

Alternately, seeing as there's only a handful of voltage points to measure at you could just start out at a very short pulse width and add 0.25ms to the pulse every 10 seconds or something like that - and when you see it start spraying, the current pulse width minus 0.25ms is the dead time for that voltage

[The1]

Ok so basicly measuring and providing a new set of offset tsbles specific to your injectors to enter into your own tune, would be handy but seems a lot of work for not much gain? There would only be a few people that would enter that data as well.

[vlad01]

This young guy I know did a thesis on some thermodynamics or some crap like that and he MS a vw type 3 engine and used a high speed camera and some equipment at his uni to do all the injector data collection and spray patterns at different pressures, dead times ect...


this is one of his vids.

[youtube]http://www.youtube.com/watch?v=8E9rsmZtnfs[/youtube]

[antus]

So were talking dead time as in relation to opening speed at given voltage? Ive noticed that a lot of manufacturers give a dead time value only and not enough data to build a proper table. Nissan seem to have dead time and voltage offset which creates a straight graph instead of a more accurate curved graph, and so the after market jap injectors tend to also come with dead time only. I reckon it would be awesome to be able to measure dead time at given voltage to create proper opening time data, but as for how... not so sure. If you could then you could create awesome injector profiles.

The gopro hero 3 will shoot low res at 240 fps, but I dont know how you would sync the start of the electronic pulse and the video.