E38 - Throttle Cut / Throttle Blip

[Z06-GT3]

E38 - Throttle Cut / Throttle Blip

We have a Z06 race car and we are swaping the Tremec gearbox for a MakTrak 6sp Sequential.

As it can be fitted with paddles for shifting, it would be good if we can acheive the nesecarry throttle cut and throttle blip by send messages to the E38 ECU.

Does anyone have any ideas on if or how this is possible ?


thanks.

[swingtan]

Throttle cut and blip is possible with the T43 controller, but I doubt it's directly available with the E38.

The MakTrak still runs a manual clutch, so for throttle cut, you can use the E38's clutch spark tables to pull a bucket of spark on clutch in. I'm making an assumption that yuo are aiming to flat change so need something to stop the engine over revving between the gears. This may work on the upward gear changes but isn't going to help on down shifts.

Simon.

[Z06-GT3]

Thanks Simon.

Most shifts will be without clutch, previous car need human to perform throttle lift/blip for shift.

This time will use a geartronics controller, so can interrupt the ignition signal, but wanted to see if a more elegent solution was available with the geatronics just signalling the ECU.

E38/T43 Info
Very difficult to find in depth info on E38 or T43.
Any sources you can recommend ?

E38 Clutch Spark Tables
Can you explain a little more.
Assume there is a input on ECU which would normally recive a signal from
a switch detecting clutch pedal movement ?


Throttle Blip
Give the fly-by-wire nature of the throttle, sort of assumed (hoped) that the cars
with autoboxes must blip the throttle to smooth downshifts ?


thanks
Angus

[swingtan]

The E38 has a clutch sense switch on the pedal, so you could wire that into the paddles so that it activated before the gear box contacts. The table for Clutch timing is B5152, which seems to be used even if the Clutch Fuel Cut Off is disabled.

The T43 does the throttle blip in the street cars. Because the ECM and TCM are tightly coupled, the T43 can command throttle opening directly. I haven't seen any tables for it though, which is a shame because in some situations, the blip is too much on modified cars.

I've been thinking about this and have come up with the following.

• For up-shift, you probably want to command 0% throttle till the shift occurs. If the gearbox has the ability to indicate when the shift has completed, that will help a lot. I'd look at setting up a system where the Accelerator Position Sensor is fed through a circuit that also takes a signal from both the up-shift paddle and the gear box. When the paddle is hit to shift up, it interrupts the paddle signal and commands 0% throttle, then signals the gear box to shift. Once the shift is completed, the box re-enables the peddle signal to command normal throttle.
  
  This may work, but may be too harsh with the throttle delays. If this is the case I'd look at just using the clutch spark table to command very, very low timing so there is no power during the shift.
  
  The next hurdle is deciding when the shift has completed. If the box has no indicator, then you may need to set up a timer on whatever method is used to perform the cut in power. It's not the best idea though, I'd much prefer the box to tell me when it's done.
  
• For down shifts, Id use the accelerator sensor signal to trick the ECM into thinking the pedal has been pressed. Of course, this needs to occur after the gear box has selected neutral and before the next gear is engaged. This can only occur if the gear box has an output signal to indicate neutral. To make the shift smoother, I'd again cut the spark timing before the shift out of gear to reduce the drive train loading.
  

I think it's possible to do and if the box has indicators for neutral and shift completion, I would be fairly straight forward. I reckon I could wire something up if I was doing it myself.

Simon.

[Z06-GT3]

I think I see a solution from what you describe.

Geartronics is a closed loop system.
So there are outputs for signalling Ign Cut and Blip.
Real world shift times are typically 50ms - 150ms depending on conditions.

Ign Cut
The geartronics controller can simply send a signal to the Clutch Sense input on the ECU to achieve the cut ?
Is the duration of cut proporational to the signal duration or how are the spark tables referenced. As the Geartronics won't allow the Ign to resume until the change is complete.


Blip
Ign Cut as above for disenganging the gear may be 10-40ms
Then if we can sum/gate or feed the throttle postion feed into the geartronics and then feed it into the ECU, the Geartronics can happily give a blip to the ECU as and when it wants ?



Angus

[swingtan]

Starting to look good.

The Ignition cut would be fine, as long as the signal matches the clutch signal. For example, you would need to check if a clutch depress is sensed as a logic "high" or logic "low". In other words, if the signal needs to go from 0V to +12V to indicate the clutch press or from +12V to 0V. Then you need to check the source and sink limits for both the gear box and the E38 ECM. That's just ensuring that the GB and ECM can supply and accept the current the other unit needs. The rest is probably able to be done in the ECM, just set the CFCO timing quite low and then set the ramp in rates very high. I may have a play just to see what this would go like as I've thought about doing it a few times. I'll let you know if I do it.

Throttle blip is going to be harder. It all depends on what sort of signal the GB is supplying. I'm thinking it's only going to be a logic signal, to indicate the need for a blip, not an actual signal that indicates a throttle position. If so, it's easy enough to use that signal to alter the accelerator pedal sensor and indicate a "press" of the pedal. It would just be an adjustable cct to add additional voltage to the ECM input to make it think the pedal was pressed further down for a split second.

I might see if I can find some detail on the geartronics system.

Simon

Just reading up on the Geartroncs units.....

• Digital cut – interface to engine ECU – switches to earth for duration of engine torque
  reduction. Engine ECU must be capable of supporting “cut while input active” for this
  wire to be used, otherwise direct cut of coil/injector power must be made (using pins
  10,11, 23 & 24): So just need to check what the E38 needs to indicate the clutch has been pressed. It doesn't look like we have a PID for the clutch pedal in ScanTool/BBL, but we do for the brake pedal. Having the clutch PID would be nice.
• Throttle blipper output – switches to earth for duration of blip – other terminal of
  blipper solenoid must connect to +12v: This indicates that it's just a straight "hi/low" logic output and will need some massaging to interface into the E38. You "could" just use a voltage divider so it ends up sending 5V to the ECM to indicate WOT, but it may be too much. I'd have some sort of adjustment so you can fine tune the blip.
• There are also inputs on the geartrinics unit for TPS and clutch. Looks like some more reading is needed....
  

[swingtan]

OK, well after some testing, I think this mat actually work. There's a bit of fine tuning to do, but it may be suitable. I've tried a few gear shifts without lifting off the throttle and it does really work, it works so well I'm tempted to leave it on...... Actually, I won't as I feel the car runs better as a daily driver without the fuel cut. Any way here is a log showing the upshifts. Now, I'm not sure if I lifted, I don't think I did but it's a bit novel to flat change in the car. I was able to do some pretty fast changes though. Here's a log showing how it worked.

Upshift_CFCO_Spark_Timing.efi

[Z06-GT3]

Simon you have been busy.
Sounds good - how did you mange to test it ?

Geartronics - you can email Neil on info@geartronics.co.uk.

Ign Cut
So the Clutch Switch input gets the job done.
The ign cut will last as long as the signal is active - correct ?
If so then that's one part sorted.

Blip
The GT's blipper output is a logic +12v/0v output as it typically drives a soleniod with a small vacumm plunger. On things like a Radical the pluger directly actuates the butterfly lever arm.

So how best to use this signal ;
a) intervene between the 2 throttle pendal sensors and ECU and add a voltage in responce to signal
b) some unknown input on the ECU that causes a blip to happen

c) Apparantly the E38 will accept a CAN message to do a blip.
This normally comes from the T43 controlling the autobox.
Neil however thinks that the CAN will have to much latency to give the desired performance.


Angus

[swingtan]

I simply re-enabled the "Clutch Fuel Cut Off" settings in the tune. This will cut fuel and reduce spark timing to a set amount on clutch press. It seems that both these controls are active for the transition from clutch out to clutch in. The spark timing then ramps back to the normal spark table at a configurable rate. So while the torque reduction does not last as long as the signal is applied, it should be fine given the expected shift times of the GT system. I set the ramp in rate to 1 crank pulse (so the switch from CFCO timing to normal timing occurred over 1 pulse ) and I was able to shift without the engine revving out. However the shifts were pretty harsh so I've reset the ramp rate to 112 pulses and it's much softer on the change. Stock was 56. As you say, I think this will do the job for upshifts as I was able to drive to work this morning in traffic and still keep my foot on the throttle during gear changes.

So my testing was done with my VZ SS, which runs an E38 and a 6lt L76 motor. If I was racing it, I'd be very tempted to leave the CFCO setting enabled as I really found it easy to shift with the accelerator pedal to the floor. It probably needs more refining, but it's certainly not obnoxious to drive. It's a bit harsh in peak hour traffic though if you want a brisk change, but come off the throttle and get back on. So I'll probably turn it off again.

For the throttle blip, I'm thinking something like this.

• Blip signal from GT goes to a buffer CCT that sits between the pedal sensor and the ECM.
• under normal operation, the buffer simply passes the pedal signal with no interruption.
• On a down shift event, the blip signal from the GT cause the buffer to "inject" additional signal to the ECM, indication a "WOT" pedal position.

I think that this would be the simplest method. As you say, the T43 will "talk" to the ECM over the CAN bus and command a throttle blip. While it is possible to mimic this, I think it'd be pretty hard to do. It's not just a case of working out the actual digital protocol that says "open the throttle" but I believe you also have to identify where the signal is coming from and going to. So you need to "register" the device sending the data so the rest of the CAN bus modules know where the data is coming from. I don't think the latency will be a problem, as the T43 manages it fine in manual mode, but the time would be better spent elsewhere.

Going the analogue module to inject additional APP signal also means you can adjust the amount of throttle opening and even adjust the duration of the blip, all separate to the GT controller. You may not need to, but it would be great to have the ability to if needed. You may not need a full open throttle event, but only say 1/2 throttle, so this would give room for fine tuning.

Simon.

[Z06-GT3]

Torque Reduction

The E38's fixed stratergy would in most cases be just fine, but Geartronics are not happy about it as they can not control it's duration dynamically.

If the input signal (lets call it edge triggered) is repeated will the ECU extend/restart the torque reduction cycle ?
If there is a blocking time window, how long is it before another edge can repeat the TR cycle ?

The GCU (geartronics control unit) keeps the signal low for an unknown duration, until it senses that the next gear is engaged. So this can be variable, and worst case when a gear does not engage it can take some stratergy on how to handle it.

So as we think that a cct is needed to achieve blip, we may as well (if possible) add a method to extend the TR to do exactly what G. wants.

So the idea might be to sample the GCU signal at 5mS intervals and the pulse the signal to the E38 to extend the TR cycle.

So I wonder how to test the E38 on that one ?