Adamw

Also your wire looks too big and the crimp doesn't look right. These superseal connectors are pretty fussy and need a round crimp (rather than a "B" crimp) on the insulation ears otherwise they wont go in.

Note the SCG-1 is only an open loop controller - the atom can already do open loop control so no advantage using the SCG-1. Also my experience with innovate widebands is not good - I would strongly urge you to look at other options. After you have replaced the sensor many times you will wish you went for a better brand... A CAN based wideband would be good for you as it wont use any of your precious analog inputs. Link do a CAN wideband controller and the AEM X-series work well too. A digital input is an input that can only sense "on or off" or pulsed type signals - not a variable (analog) voltage. Most modern cars you can splice into the VSS (speedo pick up on gear box) for vehicle speed. - it wont effect your existing speedo. Yes, TPS would be the next priority. If you use a CAN wideband it will leave you another spare. No extra sensor required, it is standard practice to use the normal ECT sensor.

Try this: https://drive.google.com/open?id=0B9h3aF9QxB3Cek1Lc1lyZVFzaDQ Note most of the info is in the G4 software help file, and sometimes a little more up to date..

You cant easily measure the ignition output from the ecu unless you have a scope since its only a very short "pulse". The ecu has an ignition test function so if it were me I would use that to eliminate as much of the other hardware as I could before suspecting the ecu. It would be very unlikely to be something wrong with the ecu. The steps you have taken already with swapping coils etc is a good start, the next test I would do is swap the ignition output wires 1 & 6 at the ecu connector (B57 & B52) then do the ignition test again and see if the fault now moves to 1 or stays at 6. If #1 then no longer sparks it indicates an ecu problem, if the fault stays with #6 then you have a hardware/wiring issue.

Hi Felix, I agree, it could be useful for some applications although I have never found an engine that needed it (and yes I have done a few rotaries). For most users with a 1:1 FRP, the diff fuel press will be relatively constant, and normally by the time secondaries come in to play the battery voltage is pretty constant too. So I'm curious why do you expect yours isnt? The few cases I can think of would be mechanical fuel pump where you get poor pressure at low RPM or a rising rate regulator.

Hi Mike, Are you getting a screen like below with the odd units "%FF" showing? If so I suspect this might just be a display bug and it will probably still work ok if you just pretend these are duty cycle.

The G4+ atom doesnt have a builtin MAP sensor. It seems you might be looking at a manual for the old G4 model which did.

Yes its +volts on one side and - on the other. The idea is you want to get as close as you can without overlap. Although if you do overlap it will just look like it does now showing gear "zero" in the middle of the change. if you have roughly 0.5v steps then try say 0.2 or 0.22v.

I have only had a quick look but dont see any obvious problems in your settings. For some reason your log is a bit odd with some channels appear to be logged very slow and out of sync so it isnt as helpful as it could be, but again I dont see too much wrong. Have you checked coils actually have 12v during cranking? Also what sort of engine? You have injection mode set to single point group which is not common so I wonder if this is a mistake or you do actually have just a single injector?

I think I mentioned in one of your other posts a while ago that you need to increase the "volt tolerance" setting under gear detection until you get a crisp step between gears. Yours is at 0.1v now, try bumping it up 0.1v at a time until you get the best result. The few cars I have done have normally been around 0.3v from memory. You can test this "driving" in the garage with the wheels jacked up.

It doesn't sound like a trigger problem. Can you post your .pcl so we can have a look over the settings. Also a pc log of cranking might be helpful. Is the sensor a 1gt101?

Yes the G4 had 4 or 5d fuel and ignition tables. in the download section of Links website you wil find the G4 software, if u download and install that it will have a storm base map in it's directory.

If you input the calibration points as they are listed on the sensor manufacturers datasheet, the ecu will measure/read/display in gauge pressure.

Although it may be possible to calibrate your sensors in absolute I cant see why you would ever want to? Assuming this is a car and not a spaceship... If you look at your oil pressure in a log for instance you dont want to be mentally subtracting atmosphere from it every time. You really only need absolute pressure for MAP or EMAP to calculate density, any other pressure measurement I can think of on a car should be gauge pressure.

I cant think of many explanations for the 30deg change... My logic would say since there are only two teeth on the crank if say something changed with Trig 2 (say positon or trigger edge) then you would expect timing to move by one crank tooth which in your case is 180deg. The Evo "teeth" are about 65deg "wide" so if you had a trigger edge wrong on trig 1 the offset would change by 65deg. I dont have any other ideas but am just saying that 30 deg sure seems odd?

A picture will probably explain better: This picture is not scaled well to what I normally see in real life but it still shows you the idea. The green arrow is pointing to roughly where I would normally initiate the gear cut based on system using gear lever force. The red arrow shows where I would have to initiate the cut if we where using barrel position. You will notice the gear pos volts is a "flat line" in the preload area so we cant get the cut to trigger in the ideal location. As for your "shift times" question, unfortunately there is probably not a typical answer - you will see huge differences between cars, vehicle weight, gearbox type, etc. All I can say is I only have one example of similar cars I have played with using the two methods, if I were to take a guess based on feeling and incomplete data I would say the total shift time is about 25% longer on the car that uses barrel positon.

The PNP link will not take a TC input directly. I just googled the Raspberry Pi amp and most of them appear to be serial/SPI based so they will be no good to you either. You need a 0-5V TC amp to feed into an analog input. You can get these from many vendors, one low cost one that I know works ok is here: https://shop.vems.hu/catalog/amplifier-dsub9-p-129.html

I just done some google. On most BMW's, OBD2 pin 9 is as you say an engine speed signal. That likely would have come from the engine ECU so you have probably lost that. Looking at the help file on your PNP ecu it appears that you should have "Aux 1" available on the expansion connector. You may also have some of the other Aux's available if for instance you no longer use the purge solenoid. You will need to connect your shift light RPM wire to on of these spare Aux outputs and set it up as a tacho output. I think they will be 2 pulses per rev. Edit: Added later; Sorry, replied same time as you Scott. These shift lights appear to me to use a normal "analog" tacho signal - not CAN or K-line. If you look at pin 9 on that wiki link you gave you will see BMW use this pin as a tacho signal.

Hi Mapper, I have some experience with both methods. Strain gauge based cut initiation gives the absolute best result but obviously comes with a price tag. Barrel position cut initiation actually still works pretty reasonable but you need a longer, more noticeable cut event and the actual shift is slower. I guess this more prominent shift cut event may also have traction consequences to consider in some disciplines. The difference is with a strain gauge your cut doesn't initiate until a preset shift force is reached - this preloading removes all mechanical backlash from the gear linkage/barrel/shift forks etc so the gears are basically ready to "pounce" as soon as the torque reduction is activated. With a barrel position based system you initiate the cut basically as soon as movement of the barrel is detected so you are already cutting power when you still have all the backlash to "use up" and the gears havent even thought about starting to move yet...

The log confirms you are still on an old firmware 4.7.1. This is different to the PC software version. Read the help file about firmware update. Also read the release notes since version 4.7.1 and note any important ones(usually highlighted in orange text). One I remember around 4.9.9 is you have to delete all ecu logs before doing the update. You are missing out on many feature enhancements and bug fixes so you should upgrade. Your AFR being off is most likely calibration or wiring related but update your firmware first because there were some bugs with wideband average parameter in earlier versions.

The map you posted shows 4.7.1? If you post a log and your latest .pcl, I will take a quick look to see if I can see your problem. picture share

Iecku, I just thought I would mention a little bit of info that many people miss. These kits appear to use the Honeywell 1GT101 sensor, these sensors are much slower on the rising edge than the falling edge. The datasheet quotes a 15µs rise time Vs 1µs for fall time. So for at least trigger 1 (that controls the timing) it would be more correct to set the trigger edge to falling. Even though 15µs is only like 0.9 deg at 10000RPM, if you can get extra accuracy just by a simple setting change - then you might as well use it...

If the trigger error only appears during cranking and is showing ok during running then it is probably not too much to worry about. Just re-calibrate the timing and carry on. However if you are still getting a trigger error during running then you will need to look a bit closer at what has caused it. Looking at the very basic 2 + 2 trigger pattern on these, there is not a whole lot that can be set wrong. Are you sure a trigger edge setting hasnt been changed? My only other thought is perhaps something was "wrong" prior to rebuild (like the trigger wheel on backwards) and now it has been put back together correct.

You are on a very old firmware version, 4.7.* According to the help file, the Wideband1 parameter was not logged prior to version 4.9.9. Update to the latest version and you will be good.

I havent thought this logic through in great detail, but this screen shot should give you the basic idea. In this example, Aux 8 will be on when condition 1 or 2 is true. Condition 1 would be the running engine conditions - lets say RPM is >10RPM . Condition 2 will be your shut down timer - When RPM falls below 10, Timer 1 is "activated" by virtual Aux 1 which will then start counting up (to 40sec) in the background, When Aux 8 condition 2 sees the timer count hit 30 seconds it shuts off oil pump. Clear as mud? Hopefully its easier than it sounds on screen... I would try it like this and see if it behaves as expected - the part Im not sure about is the timer might start timing as soon as you turn ignition on (before you start engine), if it does you might need to add another condition to prevent that. how to screenshot on windows