GTS1000 on MicroSquirt running MSExtra

[24c]

Well, a little rework, and a successful cold start, and warm up on the red GTS running MSExtra. Even managed to rejig the VE table and up the required fuel, although I had to go into WinXP to alter the rpm bins etc.

Other than a stutter around 2900 rpm, it's not looking too bad.

Check it out, if you want, nice idle, gentle roll on the throttle from idle up to the stutter, but you can force the engine past it with a crack of the throttle. Not perfect, but almost driveable at the moment.

http://gallery.me.com/mike.crane/100026

Still early days yet, but took a datalog, stutter is around 2900...don't know what it is yet, maybe somebody will have the answer?

redGTSreworkMON15.20.42.msl

EDIT & UPDATE The reason for the stutter is the engine is losing sync for "reason 17"...just found this ...Sync loss reason #17 means that there's probably noise on your primary trigger ... it means that we counted 8 teeth and looked for a second trigger without finding it where we thought it would be. Usually this means noise on the primary trigger. This will actually cause the code to reset itself... meaning clear out all variables and resync to the wheel.

Just seen this too...Primary trigger noise is a bit more problematic, and usually indicates that your VR wire shielding is probably not working, or you have a noise source very near your VR wire.

I am also getting losing sync for "reason 11", but this doesn't cause a misfire etc ...Reason 11 means there was a second trigger when we weren't expecting it. This can be as a result of what caused #17, or it could be noise on the second trigger input. We don't actually resync to the wheel in this situation.

2nd trigger noise can be helped with the addition of a .01uF cap across G+ and G- where they enter the lm1815 circuit (you do have G- running all the way to the lm1815 circuit right?).

It wasn't doing it yesterday

PS Still need to up the required fuel, and sniff the exhaust to reset the VE table.

[24c]

Just an update, as today I was messing with the red GTS1000 which is running B&G2.890 in the dyno room, (no graphs yet, printer didn't want an adhoc network)... reason being I was looking at noise levels (dBA) inside and outside the room and surrounding areas, plus I was checking the exhaust extraction. So far it's looking good, I have 100dB in the dyno booth, 80 dB @10700 rpm outside in the workshop, and just 65 dB outside the building. Extraction is OK upto 9500 rpm, but looking to mod further.

I took my MicroSquirt and adaptor loom and the bike started up using the OEM coils no problem once I flipped the LLO jumpers, but I have an interference issue that I need to bench test, as it wasn't revving cleanly past 3000K, and it also stopped seeing cranking rpm. Stock ECU started first time after being swapped out.

RedGTSonDyno.JPG

[24c]

Little bit more progress this weekend, or in the last few hours, as I installed MSExtra on the cased MicroSquirt, that was running B&G 2.890, but I didn't have time to mod the resistors across the VR sensor inputs, so I just twisted the tails over and shorted out the diode functions. The loom adapter I made for B&G (that was working BTW) uses signal diodes to isolate the shared VR ground. I warmed the bike up on the OEM ECU, and swapped it out for the MicroSquirt...at first nothing happened other than a misfire/pop if I pulled the throttle open. A little tweaking later, changing the offset "Tooth #1 Angle (degBTDC) from 0 to 180 managed a start, but it wasn't right and a combo of VE tweaks and using 157 deg instead got a stable idle around 1100 rpm. I will be dialling it in properly, but I didn't want to spray the dyno wall with oil, when I removed the access plug.

I decided to set the VE value for the 500, 800 and 1100 rpm bins to the same, as well as the advance degrees...all of which made for a more stable idle, as I am using Speed Density at the moment. I had to tweak the Coolant Temp resistance at 99ºC to 175 Ohms to get a realistic coolant temperature using the Mitsu settings.

I modded the accel enrich to be TPS based only (100%) and the truncated settings worked well with the bike upto 5000rpm. I have issues under load over 5K, but this is a sync issue, so I might have to tweak the resistance of the VR circuit.

Interestingly, blocking the cam signal with a diode that worked with B&G2.890 calls the MSExtra code to not work and I can only mod the crank signal...and still keep the engine running. A diode on the cam signal that worked with the B&G code makes the engine stall. No diodes on the VR crank sensor introduces sync errors, rough running etc.

A short Quicktime video clip, if it'll attach...it won't, http://gallery.me.com/mike.crane#100027

[24c]

Successfully managed a cold start, but only after letting the ECU fire the bike up for a few seconds, after not getting anywhere initially, & bike was smelling of fuel. Need to look at my msq as it was holding 3000 rpm on warm up for quite a few cycles, before dropping to a stable 1100 quite quickly. I need less advance at the moment in the ignition table when blipping the throttle from cold, so I must check the actual timing tomorrow with the strobe, and get a reference for where I am.

Was going to mod the VR circuits with a resistor like the bench test rig,and then ran out of fuel but it was time to quit the workshop, so I'll start on it tomorrow. I think I'm going to mod my priming pulses, and make a note of the TPS position/value, as the GTS1000 has a bypass idle circuit, with a cold start wax diaphragm that cranks the throttle slightly open. I need to see what I can do in the tables/settings to change this, as I never noticed it using the B&G 2.890 code.

[24c]

Tried to have a go again with the bike this Saturday as posted previously, but ran out of fuel after a cold start. It wouldn't start off the MSExtra MicroSquirt, so I modded the cranking & priming pulses offline afterwards. On the way back from the workshop, I got sidetracked and bought basin no3 for the ensuite...damn missus liked the other two of them until she tried to have a wash in them !!!

Fitted the new wash basin Sunday am, never thought I'd see the red GTS today, but finally armed with 5L of fuel in a jerry can I dropped on the dyno, and started to play... hit the road at 15:30 and 2 hours later the 5K stumble was done and dusted. The engine revs cleanly from 1100-11000K and rips on the throttle. It was absolutely awesome...and I haven't even used the full capability of the MSExtra code.

So I now have a working GTS1000 using MSExtra on the cased MicroSquirt, modified with a resistor across the cam sensor and a diode on the VR circuit... this thing absolutely rocks. I tried without the diode and it misses like a b&^%ard., and the resistor definitely makes it cleaner off the throttle

diode_on_VR2.jpg

Although it was strapped down on the dyno, I couldn't do a run, because the front wheel location is a little suspect ( I was sideways on the dyno at 100 mph a couple of weeks ago!!) and needs further work.

Another thing I have to work on is my exhaust gas analyser, to be honest I left a 150mm stainless tube in London on Friday, so my wideband exhaust sniffer will have to wait....

I will post a video link later, but it's looking well good and I am only using 8x16 of the VE tables rather than the 16x16 capability it has.

http://gallery.me.com/mike.crane/100037

PS I still need to tweak my hot start, cold start priming and mod the cranking pulse a little bit, but after 2 hours I had reached target rate, but boy was I grinning.

[24c]

Getting a little closer with the cold starts, another day and another attempt. I tweaked the Cranking Pulse, Warm Up Enrichments and After Start Taper and Percentages, and uploaded the changes to the cased MicroSquirt.
It fired straight away then died instantly. I modded the cranking pulses some more, and no change. Modded the WUE and it fired ran to 2200 rpm and started to die back to 1100 rpm as it warmed up. Hot starts are still not sorted and got a slight problem with RPM detection at cranking, but tried changing the cranking RPM and event pulses to no avail. I need about 30% throttle and it'll fire and soar on the revs, before I wind it back to idle.

I know this is telling me something, but I have to RTFM some more...looks like I'll be pulling down my cranking pulse widths and checking my MAT/Coolant values.
The modded VE tables are excessively rich, but at least I can pull third gear on the high inertia roller now...just need to sort out an exhaust bung for the wideband sensor, or the exhaust sniffing/fume extractor combo.

Well pleased with the progress though, as the noise filters are switched off etc on the Primary Tach (crank) and it sings on the throttle when unloaded in neutral.

Still running the small signal diode from VR2- to the common ground pin on the cased MicroSquirt, and every time I short it out, I get loss of sync, stutters, & misfires etc

[24c]

Another day and another cold start test, and it wouldn't fire, so I swapped back to the stock ECU and it fired straight up. I swapped the MicroSquirt in and with the revised settings it started straight away, but fast idle was about 800 higher than stock.
The revised WUE and ASE etc were too much for the engine, and it started sounding fluffy, so I adjusted these downwards, and waited until the engine got hot. I revised my ignition and VE levels at idle, and disabled the EGO (which for some reason was enabled in Narrowband mode).

I seem to get poorer RPM detection when the engine is hot, so I am going to investigate the cam lobe to cam sensor air gap on this bike, as I need to give it about 20% throttle just to see the cranking rpm, and then the engine fires, and you have to back off.

I swapped the MicroSquirt for the stock ECU and the pick up is a lot better on the MicroSquirt, and I double checked the 11K fuel cut/ignition retard function that worked well at an indicated 11,500 on the Yamaha tacho.

[24c]

Another attempt at a cold start...coolant temperature 8ºC...doubled the injection priming pulses and hit the button, quick fire on the button (that's an improvement) and dies instantly. OK so boost the WUE (Warm Up Enrichment) and mod the ASE (After Start Percentages), and hit the button again. Engine starts very quickly, on closed throttle, but dies again within a second.

Mod the WUE (richen) and the ASE (richen) around the 8ºC mark by another 5% and bingo it runs from the starter and holds on the WUE cycle. The ASE taper cuts in and the fuelling enrichment starts to drop from 140% to 120% before I cut short the testing session

I had to get home and wasn't really interested in the hot start, but increasing the priming pulses and making the cranking pulses to squirt every tach event, definitely improved the cold start, as a few days ago, I was swapping the ECU to "prime" the system.

So more good stuff happening here, and the the engine revved nicely (no load), especially rolling the throttle gently from cold.

[24c]

Another day, another cold start, and still it doesn't run on after catching. At least the Priming and Cranking Pulses are doing their bit. I tweaked the WUE and ASE Percentages and it caught OK, so another day is needed to test these revised settings...slowly slowly catch the monkey!

Good news is the hot starts have now improved, so it's hit the button and one second and fire, but hopefully I can get this better so it's more instantaneous.

Other good news, is I made a lambda sensor diffuser probe, to go up the exhaust pipe, and it'll still extract, though I have to make an interface and refine this further. I am getting a little reverb off the exhaust extractor fan, which means I will have to have a rethink/improve this area. I took a short video clip before I adjusted the VE Tables, and you can see the extraction (lack of) and richness. :unibrow:

http://gallery.me.com/mike.crane#100048

The lambda sensor diffuser probe works better than anticipate, although the TunerStudio software was too sensitive (probably dodgy HT/sparks causing a little fluctuation), so decreased the sensitivity by reducing the lag from 60 to 8%. This meant when I used VE Analysis Live, I was able to get more consistent tuning on a steady throttle.

I tried the bike with no load and did a test sample tune over a small portion of the VE Table and VE Live detected the richness I had seen in the plugs and dialled it out.

Next I loaded the bike in 2nd gear and adjusted the VE Table further....

After I loaded it in 3rd and 4th my VE Table was starting to get dialled in, but I decided enough was enough for one day, and I stopped the session. I am amazed out how good this software is now, and when I revved the bike up under load it was noticeably crisper than before.

[24c]

Yet another attempt at a cold start, but for some reason I noticed the ASE Percentages is cutting out too quickly, so I increased cycles hugely in the ASE Taper dialog box. Didn't really seem to make much difference, but WUE is the key...this engine likes quite a bit of fuel on coldish 8ºC starts, but soon drops in richness requirements as the bike warms.

It definitely is better now, but I have to wait for another day, before I can test this again. Still rolls off the throttle when cold, and doesn't sound too fluffy. After the bike got hot, I tweaked the Acceleration Tables, and created an unusual set of MAP & TPS acceleration fuel tables. I noticed on my datalog the fuel pulse were excessive, and that's probably why I need new plugs.

After this acceleration tweak, the bike definitely sounds crisper, but I need to work on the exhaust extraction a little bit more, and strap the front wheel down better before I do some sustained load tests. If I get chance today, I might make the loom adaptor shorter, but I need to check out a few things on the bench rig and oscilloscope to see what is happening with the diode on the cam sensor ground that MSExtra seems to like. A trip to the component shop is going to be needed.

PS After checking the MSExtra manuals, my ASE Percentages & WUE values are a lot lower than the defaults at colder temperatures, so I am going to look at this tomorrow. Hot starts are about the same as the black GTS1000 that was running B&G 2.890 software in October 2009, at around 1.5-2 secs.