Measure injector opening time with an O-scope

[a73camaro]

Go here http://www.automotivetestsolutions.com/ ... eforms.htm

A bit to read but the link explains how to measure injector opening time with waveform traces.

[jsmcortina]

Do not forget that you must also consider the closing time in the number you enter into Megasquirt.

James

[mahler]

There is a lot of info in that article. I am unsure how the total open/close time would be measured. It says there are supposed to be "humps" in the wave form after the pintle moves but it also says not all injectors have this characteristic.

In the manual it says 1.0 ms is a good rule of thumb and should not be altered unless for a very good reason [ Megamanual inj ]. But no procedure is described for determining the actual open/close time.

In this post [ Wavering Idle ] a procedure is described for using output mode to determine open/close time by observing whether fuel is being sprayed or not.

[john.p.clegg]

My twopenneth,I believe the opening (and closing i.e. "dead" time) can be deduced by altering the 2/4-alternating/simultaneous firing at tickover and monitoring the AFR,slightly altering the opening time until one can alternate between the two with no change in AFR....not explained very well but you might get the idea...

John

[Peter Florance]

My twopenneth,I believe the opening (and closing i.e. "dead" time) can be deduced by altering the 2/4-alternating/simultaneous firing at tickover and monitoring the AFR,slightly altering the opening time until one can alternate between the two with no change in AFR....not explained very well but you might get the idea...

John

Using the above method yielded me a much lower # than 1 msec and better behavior than any other measurement method.

I think it's key for users running lot of squirts on larger injectors.

For 6 cylinder I used 2/6 alt, but it's a bigger difference so the engine got a little unhappy until I was fairly close.

[a73camaro]

Do not forget that you must also consider the closing time in the number you enter into Megasquirt.

James

So if you can see the inductive hump from moving pintle then you can figure out open/close time. In the pic below E is the total time the ECU is controlling the injector ON, where as J is the total time that the pintle is off the seat. I would believe that injector opening time would be the time between marker C & D minus the time between markers G & I.

[jsmcortina]

Interesting traces and yes I concurr with your maths. However, the method of varying the number of squirts is likely a lot easier for most users.

James

[mahler]

Just tried the varying squirts method of determining the injector open/close time and it works great. Ended up with 0.8 ms instead of 1.0 ms. The change in fuel mixture is noticeable even with a change as small as 0.1ms. In my case I switched between 2 and 4 squirts and changed it in 0.2 - 0.1 ms incriments.

[a73camaro]

I just did this today with the O-scope and I measured 1.4 milliseconds. The waveform is from the conglomeration of 4 injectors being driven by one driver. It is kinda cool to see it on a scope...

[riceandbeans]

This is really interesting...

Mr. Clegg / Mahler, could you be so kind as to explain your 2/4 alternating/simultaneous opening time tuning procedure/theory? Right now I've got 2 squirts per engine cycle, simultaneous. This is a staged rotary setup, 550cc primaries, 720cc secondaries. Dunno if it'd work for me or not since both rotor primaries are controlled by 1 injector driver, but I'm very interested as I've been suffering with very poor throttle tip-in and about a 11.5:1AFR idle for the past 3 years.

Also, I thought the section of the article dealing with flyback control and injector pintle closing time was very interesting. Seems like the one-size-fits-all nature of the MS flyback may have negative long-term effects on injector life.

[Peter Florance]

This is really interesting...

Seems like the one-size-fits-all nature of the MS flyback may have negative long-term effects on injector life.

I doubt it, as I've had v3 on daily driver for 3 years +


6 low z injectors

[john.p.clegg]

Mr Beans
On idle(and all through the throttle openings)one can adjust the fuel feed into the cylinders "2 alternating/4 simultaneous" which doubles the number of "squirts" but is supposed to inject the same amount of fuel.......If the injector "dead time " is out the AFR will alter....so one adjusts ones injector opening time so that when the number of "squirts " is doubled/halved the AFR stays the same and this will only happen when the correct injector opening time is inserted into the equation...
Not very good at explaining but think about it...
John

[riceandbeans]

I'm pretty sure I know what you're talking about. However, it seems as though alternating/simultaneous might throw off my setup since it's staged injection. Number of squirts shouldn't throw it off. Allow me to make an example just to be sure of my understanding:

If my opening time is 1.0 ms, my number of squirts is 2 per engine cycle, and I idle at 900 rpm with an AFR of 13.5. I then change my number of squirts to 4 per engine cycle and maintain the 900rpm idle. I see that my AFRs have decreased from 13.5 to 12.7AFR. Since the actual pulsewidth is the sum of several variables, including opening time, an opening time which was too large (1.0ms) at 2 squirts per engine cycle, at 4 squirts per cycle this will make the problem twice as bad (open too long/richer). As number of squirts is increased, that error (excessive open time) compounds proportionally with the number of squirts. Correction in this case would be a decrease in injector opening time from 1.0ms to 0.9ms iteratively in 0.1ms increments.

Conversely, if I increase from 2 to 4 squirts and I find my AFRs increase - go lean, my inj opening time was too small, as this error is compounding in the favor of leanness (decrease in total PW).

Are we on the same page? What is your reason for using simultaneous/alternating? It seems like this method can be applied using only the "number of squirts per engine cycle" variable.

[john.p.clegg]

Mr Beans

You've got the gist of it but have lost me re. staged injection...but not to worry
Good Luck

John

[riceandbeans]

Mr. Clegg,

I'm pleased to say that on my way home from work I tried your suggestion for setting the opening time. My default was 1.00ms but I ended up changing it to about 1.11ms. All I did was vary the number of squirts per engine cycle from 2 to 4 (back and forth iteratively) and adjust as per my previous post to achieve similar AFRs. Also, for some reason, even activating the pull-down menu for changing the opening time (without burning to ECU) caused MS to just flat-out die. It came back upon burning to flash ram, but just a weird glitch others may encounter.

I was very surprised at how much better my throttle response was with 4 squirts per cycle (instead of the former 2), especially in the 2000-4000rpm range. The whole engine just ran worlds smoother and even sounded different. However, my idle PWs are around 1.4ms which is getting a bit low and the idle isn't as smooth as the former 2 squirts per cycle.

Wouldn't it be GREAT if... megasquirt could run 2 squirts per cycle for rpm <= 1,500rpm and then switch over to 4 squirts per cycle at higher rpms?

[a73camaro]

Something to note: Since VE is set by tuning with a wideband, is there any advantage in determining injector opening time? So far it seems the only advantage to have opening time spot on is for switching between alternating/simultaneous injector control.

[riceandbeans]

PW = REQ_FUEL * VE * MAP * E + accel + Injector_open_time

E = sum(all enrichments)

This is a good point. It seems as though an error in injector_open_time would only cause a shift from true PW. However, this shift would be propagated across the full range of the map, so if VE was equal in magnitude to the injector_open_time error and in the opposite in sign, the net effect in PW error would be nullified.

From a coding standpoint, and this is me somewhat speculating based on limited knowledge of ASM and error propagation in code, but I am also led to believe through my research that small values of VE in a map can lead to rounding errors. There was a great deal of work done on this in the hi-res code effort. An error of an injector_open_time which is too large would yield artificially low VE values, which could induce error into the final PW calculation. This is based on my perception that low VE values induce rounding error. On the contrary, an artificially low injector_open_time may decrease error by causing VE to be artificially high to compensate. I have heard of people decreasing req_fuel to do exactly this. I have not personally done the math or experiment to verify the claim.

[racingmini_mtl]

The VE compensation for the error in the injector opening time will only be valid at a single value of E. And since E includes things like baro correction, air temperature, and O2 correction, the VE compensation will be all wrong under other conditions which will be very frequent. And the larger the error, the more out of tune the engine will be with a wrong VE due to a wrong opening time when the condition changes (and therefore E changes).

That's one of the reasons you hear people say that the engine needs to be re-tuned when the season changes (big difference in air temperature) or when going up or down a mountain (changes in baro correction). So it is important to get the opening time as close as possible to have an engine that stays in tune when conditions changes.

As for changing req fuel, it is not due to rounding errors but rather to the fact that the VE table can only be changed by an integer value which may be too big in certain cases.

Jean

[Philip Lochner]

Do not forget that you must also consider the closing time in the number you enter into Megasquirt.

James

I wanted to establish the correct "Opening time" (ie Physical opening time - Physical closing time) for a set of HiZ Land Rover 3.9 injectors I have and did a few measurements and arrived at some interesting aspects:

1) Injector closing time is a function of injector PW for PWs longer than the physical opening time but shorter than the PW at which current saturates (slide 4, 2.59ms). This is shown by slides 1 - 4 of the attached pdf. Closing time can vary as much as 0.25ms between these two points. This can be a substantial % of smaller PWs.

This leads me to conclude that strictly speaking, the ms/V parameter should also take into account PW for those PWs larger than physical opening time but shorter than the current saturation time. I now wonder to what extent this closing time issue could be affecting the MAT correction issues we have seen - if at all.

2) Injector closing time is also a function of battery voltage which in turn again affects the ms/V figure. In slide 5 the voltage was 10.58V and in slide 6 it was 12.79V.

Does the above mean that one has to measure the ms/V figure not only at an actual 13.2V but also making sure that the PW you are using for measurement is long enough so that saturation current is achieved?

[kritip]

O/T, but those scope shots are NICE.

I have a po.scope a the mo (the ceap usb hobbylab one), and have been considering a picoscope for a while. What model is it, what attatchements do you use, and how is the funcionality and software?

Cheers,

Kristian