Large AFR change with voltage.

Hi,

Car is an R33 GTR with ID1000 injectors, pressure set at 3 bar exactly with vacuum removed, no mechanical idle control valve so using idle ignition table to maintain idle and using a Link G4 and a Link G4+ (changing to the Link G4+ but will be using the old G4 which is mapped to the car to get it to the MOT station before I can start tuning the new G4+)

I am building a basic map to get me where I can get out road tuning the new G4+ ECU and while setting the idle, I am seeing quite a large difference in AFR whenever any electrical device is turned on (heater fan, headlights, ATTESSA pump etc)

I have set the Short pulse width adder table (G4+) and the dead time table (G4 and G4+) from the Injector Dynamics data sheet.

On the G4 I was running a traditional MAP, on the G4+ I have set it up for a Modelled map and included all the fuel pressure data etc and I see the same behavious on both ECU's.

Now I can have the car idling perfectly at 900rpm and 14.5 AFR, but if I turn on the headlights, the AFR climbs to 15.9, if I then turn on the heater fan as well, it climbs to around 17 afr and starts running rough as youd expect.

Looking at the battery voltage, I am only seeing a change from around 13.60 volts, to around 13.54 volts but the AFR changes dramatically.

I have carried out a basic workaround by setting my AFR to 13.5 (to avoid the rough running when it leans out) and then using 0.01 volt breakpoints in the injector dead time table and tweaking the cells to bring the AFR back to 13.5 for every 0.01v drop that I achieve by turning on electrical items.

I see the exact same behaviour in the traditional fuel map on the G4 as I am seeing in my modelled fuel map in the G4+ so just wondering what peoples thoughts are?

I haven't seen anyone else reporting the same issue apart from a google search bought up a few Honda forums with people having the same issue using ID1000's and saying the dead time table on their site gave the wrong figures.

Whilst my work around is doing the job at idle for now, I would really like to get an idea of what might be causing the issue and sort it so I can work on the rest of the tune knowing it wont be affected anywhere else in the map.

Thanks,

Edward.

Have you checked your voltage at your injectors?

Depending on where your ecu is feeding from vs the injectors the differential under load could be causing it.

Hi,

Thanks for the reply.

No I haven't yet, next time I fire the car up I will check voltage at the injectors and compare them to what the ECU is reporting.

If the injectors are seeing a different voltage to what the ECU is seeing this could be an issue so I will try that.

Thanks,

Edward.

I noticed my LINK THUNDER has about a. .35v difference. And that's with the ECU and injectors being powered from the same Relay.

You may have to play around with the short pulse adder table.

Are you monitoring fuel pressure in real time? Or did you just monitor the fuel pressure while setting it. Admittedly it's unlikely at idle where the fuel requirements are so low, however it's possible that the voltage to the fuel pump and hence the fuel pressure are fluctuating as electrical load varies. If you're incorporating a fuel pressure sensor into the ECU and using this in the fuel model then please disregard.

Once again, thank you for the replies.

Iwill take a look at the short pulse adder table, I'll monitor the cells in the table that are being used and try tweaking them slightly to see what happens, once the car has had its MOT on Thursday I can then fit the G4+ in and concentrate on getting the idle sorted before fully mapping it.

I have fitted a fuel pressure sensor and have that fed into an analog input in the ECU.

The pressure is solid at 224-226kpa the whole time (MAP reading of 38-39kpa) and the only time the fuel pressure had any change was when I turned on all electrical devices I could and the voltage dropped to 12.96 for a very short time, the revs dropped slightly, the MAP went up to 40kpa and that gave me a fuel pressure reading of 231kpa so a very small change otherwise a very steady pressure.

I could be over thinking things and I may be wrong but, with a MAP reading of 38kpa, and my differential pressure set at 300kpa, I would have expected to see a fuel pressure of 262kpa?

I wonder if this minor discrepancy could be affecting my injector dead time settings slightly? my thinking is that using the ID1000 dead time table, the lowest fuel pressure in that table is for 276kpa and the ecu is actually seeing 224-226kpa at idle.

Thanks,

Edward.

Do you have your fuel system type as FP sensor? It will be more accurate if you do. Since you have a fuel pressure sensor i would watch the differential fuel pressure trends instead of fuel pressure. Goal is to have it stable at you base pressure, during idle and boost.

How far does your revs drop? If it drops to far the alternator speed will be to low. This means the output would be reduced and the voltage would drop. You could try using idle ignition control or increase the closed loop P and D settings.

I had a similar issue using ASNU injectors. As at idle i was in the non linear range idle constancy played a big roll. I even wired in a digital power supply so i could keep the voltage the same when alternator load was applied to the motor. Dead time was the same but the AFR leaned off. Adding a new rpm row 100rpm below the target idle speed helped.

Hi,

I currently have my fuel system set to MAP referenced, I usually try to have as many sensors as possible for logging, trouble shooting and setting fail safes etc but not have the tune relying on them as it is less to go wrong and ruin a track day for me :-)

I will try setting it as FP sensor and see how that goes, that's at least one more thing to try.

The revs only dropped from around 1000rpm to around 900rpm so not low enough to drop below a 'normal' idle speed.

I don't have any mechanical idle speed control except for the standard fast idle valve for when it is really cold, I removed it all to clean up the area around the plenum and have always used idle ignition control which works perfect and gives a nice stable idle.

If I know my dead times are as good as they can be, and my pulse width adders etc are all good, then I am happy to add tighter breakpoints around idle to make minor tweaks to the fuelling as I get closer to going for the full mapping, I just want to be certain I am not mapping with an error in the settings on my part and then discover the error later and have to go through the whole map again.

Thanks,

Edward.

P.S I have added a picture of a section of the log from when I was adding 0.01v columns to tune my dead time table and trying to get a smooth AFR with electrical load.

Also adding a screen shot of my fuel settings, dead time table and pulse width adder table.

To me it looks like the rpm drops to 860 rpm, voltage drops and dead time goes up. This looks to be why the AFR goes rich. Also where did you get that injector info? As the data i found is different. http://help.injectordynamics.com/support/solutions/articles/4000074340-link-engine-management

I would use a 2D table. Andre has some good info about when to used 2D or 3D injector dead time tables. When you use a 3D table the axis has to be differential pressure not fuel pressure.

That log is while I was playing with adding break points into the dead time table, it was originally going very lean with any electrical load applied.

Unfortunately I didn't have it logging before I started adjusting the dead time table but what you see in the log, is the effect of me turning on the blower and headlights, the voltage dropping and then my rough adjustments in the dead time table taking effect and actually going rich instead of lean.

All of the points in the dead time table between 12 and 14 volts, are ones I have added and manually adjusted trying to compensate for the AFR going lean under electrical load.

The dead time table, and pulse width adder table, are all from the ID website-

http://injectordynamics.com/injectors/id1000/

I double checked the data by watching again the webinar from Andre '030 What is a short pulse width adder Link G4+' and those are the details I have in my ecu (except for the dead time table which I have played with)

The Link ID1000 chart you linked to, the breakpoints do not match in my Link G4 and they are not able to be customised, I did however just try them by finding 2 breakpoints that do match (0.250 through to 2.000) and interpolated them which isn't ideal but was worth a try, the engine ran much richer so I adjusted the fuel table and then when I put the interior fan on full, I still went from 14.5 afr, to 16 afr.

I hadn't found that 'Link' ID1000 page so thank you for that.

Can you show the same screenshot but with the problem and just with ID's recommended settings (not your tweaking) and with differential fuel pressure as the deadtime secondary axis, so we can see it going lean?

Also, does an AFR target change at idle track through with constant voltage / rpm (as a test of dead time / short pulse adder accuracy). And does it do the same with battery load or not?

Also, why is the "statistics" showing 0 for differential pressure? Is it really flat as a pancake 300kpa like it looks in the time plot?

Can you post a log file or you can email if you would like nigeld54@gmail.com

Hi CamB and Ducie54,

I'm not sure why the statistics show 0 for differential fuel pressure, I just checked and in the parameters and time plot it shows a solid 300kpa throughout, if I slide the navigator further over the statistics changes to 300kpa for the rest of the log. And yes, it stays a solid 300kpa with no deviation at all (that is with fuelling set as MAP referenced, however if I change fuelling to FP sensor, then it constantly flicking between 299 and 301kpa)

I will reset all of the deadtime and short pulse width adder tables to ID's settings, then log the engine running at idle whilst turning on heater fan and headlights.

I haven't checked if a change in the AFR target at idle actually tracks correctly but I will test for that today.

I will carry out the tests and then post the log file on here.

Thanks,

Edward.

Okay, so here is a snap shot of the new fuel settings I am going to try.

I have set the dead time table as close to ID's table as possible, the Link will not let me enter some figures exactly i.e when I enter a differential pressure of 43.5 as per ID website, the link changes it to 42.9 (I tried putting 43.6 but it then rounds it up to 44)

The short pulse width adder table is from the ID1000 GM characterization spreadsheet from their website and is the same that Andre showed/used in his short pulse width adder webinar.

Will update after I have been out and tested these settings.

Edward.

(EDIT: I have managed to get 43.5 psi into the table by changing the settings to read in KPA, then put it as 300 and then change back to PSI)

I had to laugh at the edit...

I think your goal should be to rule out anything else. It would be good to confirm if changing the load does not change the effective pulse width assuming no material change to load/rpm (what the ecu wants, before it adds deadtime etc).

(edit - what I mean is it could be going lean because deadtimes become wrong (your assumption for now), or because a compensation is applied, or because VE changes somehow)

Also, if it's a GTR isn't it customary to tune on TPS+MAP?

have you tried measuring the voltage at the injectors whilst you have your electrical loads on?

from the snapshot of your log you're losing nearly 1 whole volt, have you moved the battery to the boot or anything?

Thanks again for everyone offering help, really appreciate it.

The voltage at the injectors is within .1 -.2 of the voltage shown at the ECU, ECU shows 13.60 and at the injectors I see 13.65, with fan and lights on the ecu shows around 12.28 and my voltmeter shows around 12.30.

I have uploaded a screen shot, and the log file.

CamB, Yes TPS and MAP, I have the main fuel table axis set for TPS and then a 4D overlay for MAP should I need it and then the background fuel equation set for MAP.

Chris250, the battery is in the boot but it is standard on the R33 GTR so none of the wiring has been messed with for the battery.

I will be using closed loop lambda once I have the tune finished so worst case I can let that take care of the fluctuations with load but it would be nice to deal with it, rather than let the ecu work around it.

Edward.

Just a though-

I am using the throttle adjustment screw and idle ignition table to set my idle.

I have set my idle to be around 850-900rpm with the idle ignition at 20 degrees, I noticed from 15-20 degrees BTDC makes a large difference in idle speed but more than 20 BTDC has little effect.

I wonder if it is worth adjusting the idle ignition table and throttle screw so that I am sitting around 850-900 rpm with the timing at 15 degrees, this gives the ecu more usable adjustment available to bring the revs up when electrical devices are turned on, maybe if the revs don't dip so low the effect on the AFR won't be so much?

Just tried it, the engine now idles at 15 degrees BTDC and when electrical load is applied the revs momentarily dip and then the idle ignition table brings them back up so the adjustments have helped with the dipping RPM but the AFR still goes to 16.

I guess I can rule out the RPM dip as the cause.

Your VE fuel map is a flat line with different MGP and rpm at idle. I would have a row on the 4D table to add more fuel.