Bench Simulator - Triggering E38 ecu, or any ECU really...

[schpenxel]

I just tried it on my own car.

HPT Reported injector PW = 0.81 msec. I'm assuming this is the "base" PW.
Short pulse adder for ID1000's interpolated for that PW = .04304 msec
Offset for that vac/voltage, interpolated for both = 1.2641 msec
Total = 2.077 msec

Actual measured with an oscilloscope = ~2.1 msec.

Keep in mind this measurement was from eye balling it on a graph, so I can't for sure say whether it was 2.07 or 2.12. I'm still learning how to use this 'scope.

So. That's only one point of reference, but it seems to be pretty damn close for a first test.

I'm going to try to get some more data at other injector PW's and vacuum levels in the next day or two

[MikeOD]

That's what I'd expect you to see.

What scopes are you guys using? I need to get one.

[schpenxel]

I'm using a Hantek 1008C. It's not the nicest thing in the world, but for < $100 it's hard to beat.. The software leaves something to be desired but overall it works.

[10_SS]

I'm positive of the numbers I posted.. I actually took screen shots but have since deleted them... i saw this change in PW pattern before while playing around, this time I actually wrote the numbers down... but here's the file I used... yes the Whipple kit has bigger injectors.. it has the whipple kit injector data and injectors.

BenchTune1_2Bar_1.hpt

[MikeOD]

Is your simulator always at 12 volts?

[MikeOD]

If it's always at 12 volts, your scope numbers still work out basically the same. At a 2.21 ms theoretical pulse, with the file you posted, the short pulse adder table is going to add .007813, and the offset table at the pressure delta you'd see with 58 PSI fuel pressure and 32 KPA in the manifold will add between 0.8906 and 0.9063. So that's somewhere between 3.108 and 3.124 ms that the scope should read. That's awfully close to the 3.16 that you're actually seeing. Your Whipple tune didn't change the makeup pulse min table...so I'm still thinking it could be the culprit.

I didn't drive my car today, but tonight after work I'm going to try zeroing out that table and watching what my car does. I don't have a scope (ordering one this week, just trying to decide which one to buy), but I do have a wideband that always shows me rich after a flash (as do my closed loop fuel trims, which work back to +/-2% a little while after the flash)...so if I can get rid of that with the makeup pulse minimum table, I'll be pretty confident.

[MikeOD]

There's really no way to get the scanner to report back what is going into the total sum to make the IPW...or if it's using make up pulses or not.

[MikeOD]

Have you guys tried logging the injector tip temp? I just saw it available in the scanner...never logged it before, and I don't see anything about it in the editor as an offset, only as an open loop gain...but if it's an open loop gain, it's a commanded AFR gain, not a pulsewidth offset...so that should still report in commanded EQ/AFR.

[10_SS]

I think I had injector tip temp logged, or used to, maybe pulled it out since I didnt need it...

Voltage was constant each time, but I dont remember exactly what it was... it might be in the log if I still have it... Maybe I'll redo this with different makeup pulse min table values.. I'll 0 them, then make them much higher, see what happens.

[MikeOD]

Cool. Thank you.

[schpenxel]

I've tried logging injector tip temp but didn't really notice anything useful, but who knows.. maybe that's the secret somehow.

And yeah, there's no way to get HPT to log the "electrical" injector PW on gen 4's.

[MikeOD]

I'm still stuck on it possibly being the make up pulse min table. Has anyone ever seen a PCM with that table set to 0? And if so, did it exhibit this problem?

[Bluecat]

I glad you guys are working on this. I wish I had the time to do the stuff like this that needs to be done. This was the same approach I had to take to take to figure out the GenIII fuel timing stuff. Did that with a homemade stimulator as well. Wish I had time to contribute to this.

The makeup pulse as I understand it only has to do with transient fueling. It's fuel delivered after the normal eoit and main fuel pulse has passed. But a stomp of the pedal requires more fuel. If there is time before the final boundary, "Makeup" the difference by squirting a second fuel squirt.

Example. Final boundry is 510. Normal adder is 110 so working eoit boundry is currently 400. A transient event occurs and more fuel is calculated after the main fuel pulse ended but before the final boundary is reached. If makeup pulses are set to single or multiple it will try to add another fuel pulse(s) in this 400 to 510 range of cylinder rotation. Which is completely feasible at low rpm giving the large about of time as far as the ecm is concerned. The makeup min pulse width table decides what pulses are worth adding. So if the calced makeup fuel pulse width is less than whats in the table, it doesn't waist its time trying to do it. In other words its qualifies the transient makeup negligent enough to not worry about adding. Smaller numbers mean that more makeup pulses happen, larger numbers mean that more will be ignored.

So I doubt anything to do with makeup would be causing the rich after start problems. But who knows what tables are labeled incorrectly in the editor. So investigate on, lol. Id say what will happen is you guys will confirm the sum of the base_pw + offset + short_pulse is correct and normally matches the observed injector PW (which it looks like you have), but after a reflash the actually pw observed is going to be larger. Problem is we still won't know what combination of table or estimated temp value is causing it or have a way to fix it.

[10_SS]

I'd like to log both HPTuners and scope timing of the PW at the same time... it's a PC based sound card scope that appears to work pretty good for this... however, aside from video recording the PC screen I don't think there's a way to do it. Maybe a screen capture time lapse would be the way to do it... this way I can start it, walk away, and have some actual proof of what I see. Right now we have only whatever I jot down... The other problem is the PW waveform has the timing lines that I have to manually drag around the the start and end of pulse... so when the end of pulse starts to change, I don't see this new value automatically, I have to drag the timing line to match up with the end of pulse to get the new pulse width value.. so, the screen shot time lapse would demonstrate what's happening pretty good, but you wont get updated pulse width values until I drag the timing slider... Either way I'll run new test tonight and post the HPT datalog which includes batt volt, etc. Try to run screen shot time lapse also... I know there are some cheapish USB scopes on Amazon/Ebay... I have a few saved... can anybody recommend one?

[schpenxel]

Post the tune also when you run this

Make sure all the open loop IVT tables are set to 1 or 10 or 0.1.. most ECM's use ~1 for everything. I'd just set all of the tables to 1 and call it good. That will at least keep commanded AFR stable in open loop (I think anyways)

I'm running into the same problem.. supposedly you can record the waveforms with the one I'm using (hantek 1008c) but I haven't gotten it to work very well yet. At least on a bench setup you can hold everything steady and measure inj PW on the 'scope. On a running car things get a bit tougher

[Bluecat]

Not sure how to handle the correlation problem. When i did this with a mirco controller to simulated a 24x/1x setup, it was both generating the cam/crank/analog sensors and watching the injector/coil outputs from the ecm. So i not only knew the lenght of the pulses, but when they started and ended relative to the crank and cam position. Not sure how to do it with the equipment your using and them being seperate devices.

[10_SS]

http://www.sillanumsoft.org/download.htm

I only have two channels, so maybe I'll put the crank signal on one, and one injector on the other.

10-4 schpenxel, you guys feel free to let me know of any other tweaks you think of. I also have the day off tomorrow so between tonight and tomorrow, I should have something reliable and meaningful, hopefully in visual format as proof.

Bluecat next I want to simulate the MAF, this way I can automate relationship between MAF vs. MAP at WOT, then play with RPM, get it reading close to what my car does.. shouldn't be too hard. Then one day get an extra TPS pedal assy. Does anybody happen to know what signal the knock sensor input pins take? PWM, frequency, etc? There are 4 pins for some reason. Want to get a little knock in there too.

[MikeOD]

I'm still trying to decide which scope to order...so I can't help from here (and I'd be on a running car too...which is hard and temps among other things will not be constant.

Didn't I see you post in another thread, you need someone with EFILive to change your reverse lockout MPH numbers? I can do that for you...happy to help, especially with the time you're investing in this for all of us.

[10_SS]

I'm still trying to decide which scope to order...so I can't help from here (and I'd be on a running car too...which is hard and temps among other things will not be constant.

Didn't I see you post in another thread, you need someone with EFILive to change your reverse lockout MPH numbers? I can do that for you...happy to help, especially with the time you're investing in this for all of us.

These are pretty good... but I might will prob stick with the Sound card for now... That would be sweet to have lower reverse lockout numbers... I could prob ship this spare ECU to you when I'm done playing around... if it still works that is.

http://www.bitscope.com/product/BS10/ $245 or the micro for $145.

[10_SS]

This is full flash, power reset, then startup. I changed a bunch of Open Loop Fueling to 1.0's... Tune and log attached.

I have a time lapse video showing the Pulse Width decreasing, but it occurs within the first 5min then seems to taper off.


Video (Timelapse): The real test starts at about 6s in the video... http://www.youtube.com/watch?v=VPpCd2PiK6g

Injector Pulse starts at 7.4xx on the scope, with 6.2xx on HPTuners... Then scope reduces to ~7.1xx after a couple mins, and HPTuners increases a bit to 6.29-6.3. I could make a normal speed video I suppose... The red lines on the scope window are the timing marks, I have to manually slide them to get updated time info.

Everything was constant... volts at 12v, coolant, RPM, MAP, TPS, IAT, etc.

It almost appears to be following IVT... as it increases, Injector PW decreases... You'll notice right at about 11-12s, I slide the left timing bar to the right.. it doesn't change much after that.

7-21-2015_FullFlash_PowerReset.hpl

BenchTune1_2Bar_1.hpt

EDIT 1: I Played around varying Ignition voltage, at constant 57-58kpa:
volt-----HPTuners---------Scope
9v:------5.25msec--------6.84msec
12.0v----5.30msec--------6.20msec
15.0.v---5.33msec--------5.86msec