I recently picked up a set of FIC 1050s and have been collecting latency values that people have posted or given me for these injectors. Each set of latency values have been somewhat different, and since I have access to a pretty decent oscilloscope and power supply, I decided to see if there is a way to determine the latency on the bench top. After a bit of searching, I found a very good article (PDF download) that explains exactly how the latency can be measured for saturated style fuel injection systems. (Yes, Evo injectors are low-Z which are typically associated with peak & hold fuel injection systems, but the Evo injection system is saturated style. A resistor is used in series with the injector. The value of the resistor in the FSM is 6.0 ohms. Here is an article that explains this type of injection system.) One of the interesting things that I took from the article, is that opening and closing time both matter because the closing time is typically very similar in length to the opening time. So both must both be measured to get what I call a "net latency".

Anyhow, I've done some preliminary opening time measurements on my FIC 1050s and a set of stock Evo IX injectors. Here are the average measured values:

The measured values for the FIC 1050s at 11.7 V and 14.1 V are actually pretty close to what people are picking by trial and error. The measured opening time values for the stock injectors are way different than the values in the ECU though, so I'm thinking that there is more to the story. Also, the measured values for the FIC 1050s and the stock injectors are very similar to each other... almost too similar. Closing time I think needs to be measured. I'm going to try to do that this weekend if possible.

Here's a pic of the setup (I'm measuring the voltage across the resistor):


Here's a pic of an oscilloscope trace:

 

The guy at FIC can provide the manufacturers rated closing time of the injector.

The closing time should be the same for the 850 thru 1050 and maybe even bigger ones.

I bought the FIC 1000's and they are flow rated at around 1016 average for the set.

 

I wish Jens at FIC had the info, but he doesn't. I've spoken with him several times now. The injectors are Delphi Rochester units. The only info that Delphi provides to Jens besides the flow rate is a generic opening time of 1.13 ms. I called Delphi directly, and that's all the info that they had for me too. I even asked if there was someone in R&D that could provide info, and the answer was, "no". Jens is fairly frustrated by this as well, and he is interested in having me measure out some new injectors once I get this system configured properly.

 

Could it be possible that latency values on the ROM are actually closing times?

 

MrFred, I don't think I will ever know half of what you have forgotten but is it possible that somehow the engineers at Mitsu have thought that there is also a time requirement for the injectors to get to full flow at a specific PSI of fuel pressure based on the specific gravity of that fuel? Are the injectors affected by the fuel pressure in any way? What about heat saturation? I am not sure how much the specific gravity would really play into it but the fuel pressure may affect it. Either way just a thought from an old school carb tuner.

Josh

 

you sick bastard...


subscribed. thanks!

 

Man that is bizarre as he actually emailed me the generic 1.13 number you mention and another one. he said it was closing time now to find that email......

Regardless I wish you the best on your latest endeavor

 

The injectors see differential pressure and if your FPR working correctly its always the same (IE: 43psi or whatever). When the manifold pressure (pressure on outlet of the injector) raises 1 psi the FPR raises fuel pressure 1 psi.

 

So there is a generic closing time too. If you can't find it, I'll check with him on Monday.

 

just plain sick mrfred, what is your job LOL

 

Here's a nifty video showing a saturated injector system waveform:

http://youtube.com/watch?v=Xt40XK2j1dE&feature=related

The oscilloscope used to capture the waveform is only $800. Not too bad considering it appears to be capturing the waveform with enough precision to get a pretty good measure of the effective open time. Would be really cool to hook it up to an Evo to see exactly what is happening there and how the open time relates to the latency values in the ECU. For now, I'll have to do with my benchtop setup.

 

Here is the copied information of the email i received from him....

"Peak & Hold 950cc/1000cc: Opening response time: 1.13ms Closing response time: 0.998ms

Jens"

Granted these times are listed for the 950/1000 but according to him they are all Delphi 850's modified

I am anxious to see how this turns out. Currently my 1000's are steady at +3 +4 on the trims and I'm leaving them that way as I will be adding Ethanol to the equation soon.

PS thanks to those who helped me with the scaling

 

After making my benchtop setup a little more sofisticated, I was able to measure the closing time. I ended up adding a solid state relay to the circuit and then controlling the relay with a waveform generator. This allowed me to get a live waveform like in the video I mentioned. It was pretty cool to see the injectors clicking away on the benchtop. At one point, I had them running at 230 Hz which would be equivalent to 27,600 rpm in a four-stroke motor. :-) They could actually keep up. The injectors would just barely finish opening before they would get commanded to close. Anyhow, here are the results (at a much lower driving frequency):

First thing I want to mention is that the voltage is the circuit voltage. This is the voltage in the injector latency table in the ECU. Note that the time to open did not change very much from my earlier measurements where I was manually triggering the injectors. The time to close values are interesting. When the circuit voltage is high, it takes longer for the injector to close than it does to open. The resulting net latency is negative! I suspect that the story would be different when there is 43.5 psi of fuel pressure on the injectors. The fuel pressure would increase the time-to-open and decrease the time-to-close, which would move the net latency in a positive direction towards values that would more closely match the latency values in the ECU. I wish I could find a portable Picotech oscilloscope that I could borrow.

Not sure what to think about how to scale the FIC 1050 injectors. The time-to-open values that I measured for the stock injectors are very different than what's listed in the latency table in the ECU. My suspicion is still that the "latency" value in the ECU is a net latency that takes into account time-to-close. The benchtop data that I gathered on "net latency" suggests that the FIC 1050s should work well with latency values that are very close to the stock values. OTOH, people are getting good results with latency values that are somewhat higher than stock. I'll probably try stock values first and go from there.

 

Deleted

 

I just re-read what I wrote and I could barely understand it.

OK, this is what I am asking/saying:

We need to know 4 things:

1. Time for injector to respond to an open command, meaning the time it takes to just begin to open.
2. Time for the injector to then go from 0% open to 100% open.
3. Time for injector to respond to a close command, meaning the time it takes to just begin to close. I would assume this is the same as #1.
4. Time for the injector to then go from 100% open to 0% open.

After thinking some more, I see what you are saying mrfred. I think it would then actually be:

1-3+(2-4)*.5

I don't think #1 and #3 will matter, since they should be the same and cancel out. So, more simply, it should be something like:

(2-4)*.5

This way you are adding any time that you are getting less fuel than you think 9while the injector is opening), but subtracting when you are getting more fuel than you think (when the injector is closing). But, during the open and close, if the change in fuel flow is linear, you would have to divide by 2.

Maybe I should read some of your links before I babble here so I can understand better.


Eric