GTS1000 on MicroSquirt running MSExtra

[24c]

I am in the process of fitting a MicroSquirt to a Yamaha GTS1000. This is a four cylinder 1002cc 20V engine similar to the FZR1000, and YZF1000 (Thunderace).

This was a longer post, but it got deleted by the faulty keypad on the MacBook.

I learnt a nice trick today..I had a problem with cranking, as in none happening in TunersStudioMS, and in my desperation to find out what was happening, I ran the bike of the existing ECU at idle and piggybacked the MicroSquirt using a temporary loom to this.

I bootloaded Monitor 2.4xx into the MicroSquirt opened TachRef.exe, whilst previously used sewing pins to puncture the ECU wiring and pick up the VR crank and cam sensors. Based on the grid plot of the cam and cranks I could estimate the position of the cam/crank sync, as in delay teeth and offset value, unfortunately there was no save dialog or any screen capture software on my laptop.

Anyway, I ended up with a comms problem, I couldn't boot the 2.889 code into the Microsquirt and had to call it a day!

PS Any Mac/Unix geeks out there, I modded TunerStudioMS today to see my USB-Serial device, but I couldn't get it to see the modded TunerStudio controller files. /dev/tty.xxxx is "Port in Use" etc

[24c]

The comms problem was a missing jumper on the "piggyback" loom on the ground side, I could see the big picture but not the small one.
In a way it's a good thing, because I don't know the flow rate of my injectors.

I saw a post on MS3 about using a scope to measure the pulses etc, and I thought that would be an ideal fix for me instead of guessing. I don't really want to squirt into a measuring jug, although you could. I could start the engine on its' ECU and then take readings for the injectors , maybe coils.

PS TunerStudio will talk to my MacBook now, just need others to try to see what works on a non modded Mac.

[24c]

Had another at getting the MicroSquirt running on the GTS1000 today, and I used my long loom to piggyback off the existing ECU, but I was only connecting it to the crank & cam VR sensors. After cranking the day before I had no cranking RPMs at all, so I double checked the MicroSquirt against my drill powered rig, and it looked like the MicroSquirt was corrupted, because it didn't crank on this either. Reloading the .s19 file into it and loading the .msq into returned everything back to normal.

I have an oscilloscope, and although I got a good trace of the cam, I got a poor one of the crank, it was full of AC noise. At the time I couldn't understand it, but after cranking this morning, I was non the wiser. I could run the bike up on the OEM ECU, and get intermittent RPMs running at half engine speed, but they would start until nearer 230 revs, and idle was showing 436 or so.

The centre stand on the bike was missing, so it's on it's side stand, and when the engine missed a little, the RPMs came back but then they disappeared. I logged it and couldn't understand. I tried refining the pulse tolerance and still no better, after a while pattern was emerging, so I propped the bike up vertically, and the RPMs were good for longer...the crank was sloping downhill, and the tolerances must be just enough to affect the VR sensor clearances.

So you'd think now I'd have a cranking RPM, no chance. On the rig it cuts in at 120 revs, and I reliably informed the starter does 350 rpm on a good battery. I bootloaded the TachRef code into the MicroSquirt, and started sampling. When the engine was running off the OEM ECU and RPMs seen in the dashboard, I got a cam and crank sensor processed pulse clearly seen on the plot.

When I was cranking, I only saw it once, and that was after swapping the VR sensor for another one with a stronger magnetic force, but still I cannot see any cranking RPMs. I reversed the polarity of the crank sensor and things were no better but the plot looked better.

I am still using Rising Cam Sync with Tach or Wheel in the Dual spark Options, but the half speed RPM figure is a mystery, it might even be a code error for this option, because I can't understand, why if I have two tach events per crank, it is wrong.
UPDATE
Searching the forums to find more about this and a post mentioned no of cylinders...
What a numpty, in between whatever, I now have 8 cylinders in the Injection Control box, whereas I did have 4 before
...this must be why I have half speed.

[24c]

I have an oscilloscope, and although I got a good trace off the cam, I got a poor one off the crank, it was full of AC noise. At the time I couldn't understand it, but after cranking this morning, I was non the wiser. I could run the bike up on the OEM ECU, and get intermittent RPMs running at half engine speed, but they would start until nearer 230 revs, and idle was showing 436 or so.

The centre stand on the bike was missing, so it's on it's side stand, and when the engine missed a little, the RPMs came back but then they disappeared. I logged it and couldn't understand. I tried refining the pulse tolerance and still no better, after a while pattern was emerging, so I propped the bike up vertically, and the RPMs were good for longer...the crank was sloping downhill, and the tolerances must be just enough to affect the VR sensor clearances.

Still not a success yet, but I'm more hopeful it will be soon, as it was the VR sensor clearances after all, and specifically the cam sensor air gap. After doing more work on the drill powered crankshaft rig, I found a correlation between cam sensor clearance and revs before sparking/injecting started. Increasing the air gap directly increased the cranking revs required, so that this GTS needed 850+ rpm before it would spark on the rig, where the starter would only give 350 on a good battery. The other bike had a smaller clearance and started working at 150+ revs, so no wonder it fired up.

I have also fitted a small Zener diode to isolate some noise on the shared VR sensor grounds, that is stopping me seeing resets and +- trigger counts, but either way to see COPs sparking and injectors triggering from 154 rpms and holding steady up to 1400 rpms is a real step forward.

PS I reconfigured TunerStudio to the FZR1000, Thunderace crank set up, "8 teeth" ,"0 missing", "skip=4" and "delay=1" and that also works now, starting at 134 rpm and holding 1200 rpms on my tiring cordless drill. So it seems noise and the cam sensor clearance were the issues all along.

[24c]

Finally after a day milling the cam cover down by 30 thou", I now have cranking sparks under compression and a falling battery voltage. It was done in stages...

1. Bike sensors under cranking firing the drill powered rig first with the plugs out and 20 thou of the cam sensor mount, then after putting the plugs in and the compression slowing the cranking, I had to machine another 10 thou off it to reduce the cranking rpm rev detection further. It worked with the throttle open, but after machining further it sparked on closed throttles. Might be worn cam lobes on the 108000+ miler.

2. Bike mounted MicroSquirt datalogging VR sensor negatives and Zener diode mods to loom converter, mimicked rig mods.

3. Bike mounted MicroSquirt and cranking the starter to get stable tachs and no +- triggers.

4. Bike mounted MicroSquirt firing the rig mounted COPs, using large battery booster cable as common ground.

5. Bike mounted MicroSquirt firing the engine mounted R6 COPs, but this was after .... numerous back fires, too much offset, stalling and excess fuelling. I had no info for the injectors, working back suggested required fuel was 8 msec, which was way off the mark.

6. Finally... after a lot of fiddling to the trigger offset, delay teeth and later the required fuel to get a crisp sounding idle, fluctuating between 954-1100 rpm, looking at the datalogs now.

So the success has started, and now the starts.

[24c]

Another day, calibrating thermistor values to more match the engine, tweaking the offset etc.
Successful cold start eventually, but I had reduced the priming pulse too much, and it was too weak, and ignition offset slightly out.
Saving .msqs after modding them and have an adequate, but slightly weak sounding SD map, that rolls on the throttle gently. Sudden larger changes are too weak, but have Xtau on now and have since this tweaked the AE settings and it all "sounds" promising.

Have saved the earlier msqs, which I will be going back to next weekend. Meantime more reading and hopefully understand the dialog boxes/options available.

PS Would love to wind it up, but the neighbours are still not moving around yet!

[24c]

Posted a little clip of the engine running on it's temporary loom, can't do anymore at the mo, as working away for a few extra days.
Bike's a little noisy as its done 108,000 miles. Throttle bodies need setting up, as its been stood outside for 18months or so, but it's just a development hack.

http://gallery.me.com/mike.crane/100004

[24c]

Third day on the bike, but only did a couple of hours. Looked at tuning the VE tables for idle today, but it was too erratic, (despite reading the manuals, MAP wasn't appreciably lower) and so going back to basics. Also, earlier the cold start was not right, despite going back to a previous working .msq, so it was quite frustrating in the shed overall. I did manage to clip the revs on the limiter, but I think it needs a fuel cut now as well, as there's not enough feedback for a twistgrip junkie. I also noticed my NB O2 sensor was unresponsive, so I am going to check the AFR out with a standalone Innovate LM1, and replace the O2 sensor for a good one.

Realising I have a huge knowledge shortfall whilst I get upto speed with the MegaTune/TS software side of tuning (as opposed to carbs and stuff), the black GTS in the shed is now on hold, while I refit the standard ECU, and calibrate/service the throttle bodies to reach a steady idle or keep the project chugging along by fitting the second MicroSquirt to the road legal red GTS.

[24c]

Another go at it today, this time using my red GTS1000 fitted with stock OEM wasted spark coils. This too needed a cam sensor to cam cover face skim, another 0.070" off this one too, otherwise I don't get any cranking RPMs using MSExtra or B&G.

First off I used the B&G 2.890 cased Microsquirt and got it up and running from a hot engine *(warmed by the stock ECU). I loaded the ignition table with a flat 5 deg advance and aligned the crank timing marks at idle, by altering my offset. After this I reverted back to a generic ignition advance table, and adjusted the VE for idle. Interestingly, I needed to max out the dwell on the OEM wasted spark coils at 8 msec.

Next up was the MSExtra v3 Alpha coded MicroSquirt, this too was driving the OEM coils, via the built in MicroSquirt ignition drivers, but I had no idea what my settings were for this to be honest. I knew what it was for the B&G code, delay teeth = 2, offset -17.50 deg, so I was guessing around 180-17.5 degrees should be somewhere near, if I was using the same trigger edges. The engine fired up, but it was pre-warmed, and I had some settings from my B&G stuff. A little bit of tweaking, and I got the correct offset to land on the OEM timing marks.

Then it was time to fit the TCI R6 stick coils and external ignition driver module, after resetting the jumpers to LLO. Both MicroSquirts ran straight off the mark, but it was starting to get late. I need to look at my earlier .msq to see what VE tables I had, as I am very weak at certain points, but it does run of sorts.

Red_GTS_B&G_MicroSquirt_.jpg

For a little bit of video at the end of day one with the red GTS1000 click on the link below.

http://gallery.me.com/mike.crane#100005

[24c]

Bit of a mixed bag today, started on the B&G 2.890 MicroSquirt, but it was reluctant to start/catch from cold. New plugs, more petrol and tweaking the required fuel managed it, but it was never satisfactory. I don't have a required fuel that comes close, or have advance values that seem to work. Still tweaking the VE table, but short clips tell you it's not right.

http://gallery.me.com/mike.crane/100016
http://gallery.me.com/mike.crane/100025

After switching to the MSExtra version, I managed a little better, but didn't manage a restart after taking the video clips Revved up a lot crisper than the B&G settings though when it was hot previously.