Strangely, had another Project Config error in TunerStudio
It was a corrupted project file again, and when I deleted the surplus project.properties files and edit the remaining I was able to get into the project again. I got another error when comms was instigated, and then the current tune got corrupted in the following serial write. Not sure what is happening, but my VETable1 was completely screwed, with fuel bins getting increased adhoc. Most odd!
However, once this was sorted, by reverting to another tune, and deleting the current tune, I was able to get the bike to catch on the starter, but as soon as it was released it died. A little tweak to ASE and WUE meant it caught properly, and then went into fast idle mode. Tweaked the WUE as I watched the AFR, as the engine warmed up.
Also rescaled the kPa bins some more to shift the idle/fast idle bins lower in the table. This gives me more tuneable space/headroom from idle to WOT. It also means small throttle openings seem smoother, more sensitive and less jerky.
Had another go at a cold start again tonight, and although I had a strong battery, I was getting a little kickback through the throttle bodies and exhaust, which suggested over advanced timing. I reduced the Cranking Advance from 10º to 1º, and bingo the bike caught straight away and went into the ASE & then the WUE cycle seamlessly. ASE sounded like it could be cycled for a little longer, and there was an ignition advance variation as it was warming up.
After opening Spark Table2, which is triggered by MAP, I noticed the active area was dipping into the higher advance areas. I rescaled the kPa bins again to increase the working area above idle kPa, and this stabilised the engine more. I also noticed the increased RPM from the fast idle was nudging into the next higher RPM bin in Spark table1, so I tweaked this too to make the value more consistent. This stabilised the RPM further, but very small throttle openings from idle were again looking too weak, before they richened up, so I rescaled the bins further, and introduced a more gradual change to upper table limits and upto 3000 rpm.
End result, very nice throttle transitions, and this will get refined when I dyno/load the bike later this year in terms of AFR targets. It was an interesting and unexpected lesson to learn, and finally the MSEXtra3.1 upgrade is starting to come together now, which is positively reassuring, and it will form the basis for more refinements over the next few days.
No Acceleration Enrichment selected, and the engine is so much more responsive. This tuning malarkey is always surprising me. I have yet to work what is happening, but in the meantime the engine isn't complaining.
After reading some threads about opening times for injectors I started to experiment with mine, which are 195500-2680 Nippon Denso units, and reducing the opening time to 0.850 mS was hugely beneficial. Engine sounded crisper, AFRs stayed stable and flat no matter what small throttle changes I made...all in all a good result, but here's to understand and read some more about what is happening.
There is a good bit here about Secondary Load Parameters
The biggest hurdle is the project bike not being on the road, because everytime the engine starts and warms up, I have never been able to progress the rest of the tune, and if I did, I would be a lot further on by now. As the red GTS1000 is a good runner and the likelihood of my mate completing the other rolling chassis, so the Thunderace engined GTS1000 is roadworthy is probably non existent in the next six months, I really need to have a rethink.
To keep the project progressing, and maximise those little opportunities I have available, it's obvious I will achieve more by concentrating on the red GTS1000 with the stock engine, but at the same time keep the Thunderace options open by making an engine rig to easily test all the future bits and likely developments to the GTS1000 engine I have planned.
For the first part, I decided I would resurrect the GTS1000 engined side of things, and restart with the red bike where I had left off some time ago. However, there is an issue with my cased MicroSquirts and VR2+ trigger voltages with this engine, so I thought I'd use Jean Belangers' new revision 2 potted VR conditioner, and see if I could use this to get round the trigger voltage issue. I want to go this route, as the intention is to MicroSquirt a stock GTS1000 engine, with no OEM sensor mods etc, except maybe fit a wideband O2 sensor.
I tried last week suing my cordless drill powered crank & cam rig, but I ran into issues, and I possibly "killed" the potted version, because it doesn't behave the same, using the LED triggers I started with. At first I thought the potted version was the best option, but after revisiting this ECU project again a bare board is more flexible, and at least you can see the components. As a result I ordered a bare board version, with the intention of dropping this in the bare Yamaha ECU case I am fitting the module to. This has now arrived, and I have started & finished soldering the wires to it, and later this week I will hopefully test the VR1 & VR2 outputs and make further progress.
For the second part, I decided it might be best to build an engine rig, so that I can test my alterations, rather than fit them to a bike and do a back to back, as you can waste a lot of time removing bodywork, ancillaries, suspension and stuff. The best way to test is to use the dyno, and as I have an inertial dyno available nearby, so why not make an engine rig for this, rather than get another. However I don't want to tie the inertial dyno up with this static lump, as others might want access to it, so ideally I want to make the engine rig removable. It also needs to be in manageable pieces and modular, so it can adapt to other engines, as well as allowing for the easy fitment of exhausts, camshafts, or cylinder heads. You also want good access to the throttle bodies and the area above so you can test on the intake side. It also needs to be self contained, with ECU, battery, fuel, cooling and control systems.
I have just started on a wooden space model, so that I can work out the dimensions required a little easier. I am going to make a general purpose multi functional subframe to transfer the engine power via a motorcycle wheel & tyre, but make specific add ons for different engine scenarios. At the moment, I have designed the subframe around spare GTS1000 components, transferred and modded it in CAD, so I now have the cutting list and fabrication drawing. Today, I looked at the motorcycle engine subframe which attaches to wheel subframe, and it is big enough to take a variety of engines, although my interest is in five valve Genesis lumps at the moment. I also plan to make this work for my Panhard engines, as well as different subframe for small kart and single cylinder motorcycle engines.
Couple of pics, for those that might be interested, but it is very basic work in progress at the moment, just a space model, and is designed to make easy access to test exhaust systems, cams, cylinder heads, and other throttle bodies I have planned. More CAD tweaking tonight, radiator & fuel tank location, as well as best place for controls for running the engine up.
Next up is the electrics, which will be on a subframe too, and maybe I'll get chance to order the subframe box section steel.
A few more pics,
Today I cleared the bench rig table and decided I would play with the Dual VR Board I got from here. Jean Belanger aka racingmini_mtl has been really helpful, and made several suggestions, but I was surprised by how simple things went today.
First I fired up TunerStudio and looked at the minimum RPMs I could see before I lost triggering voltage & sync. The lowest detected RPMs in TunerStudio which I could get with a virtually touching air gap between cam lobe and VR sensor was 270 rpms, and at this point the LEDs I have on the injection outputs stopped flashing.
I scoped the results using my fleaBay USB oscilloscope, one channel on the raw AC side of the VR sensor and another on the VR2 output.
I thought the voltage from the square wave looked less than the AC spike, but I decided to hook it back up to the MicroSquirt, via the brown & white wire (uses the standard VR2 input pin). I was using a small cordless screwdriver (maximum revs 410), and oddly the LEDs on the injector channels triggered, which meant I had sync! I opened TunerStudio and decided to see what the minimum RPMs detected were and I was shocked how low it can go.
I managed 28 RPMs with the same air gap, that could only register above 270 before. I increased the air gap to 2mm and the LEDs flashed again.
I fired up TunerStudio again, and slowed the crankshaft down, and I had it down to 29 RPMs, before the screenshot above was taken. Well impressive.
Hope this helps others, but this Dual VR board is amazing. The cased MicroSquirts are green for go, and now I am hoping I can move forward again with this engine some more.
All this optimism and then the reality. I got distracted by some beta testing the last few days, and was revisiting the Dual VR board last night, and I don't know what happened but I have a VR2 output issue now with the only board I have got. This means I have no square wave output for VR2 sync, and therefore I am getting no RPMs detected.24c wrote:....The cased MicroSquirts are green for go, and now I am hoping I can move forward again with this engine some more.
So the planned fitment this weekend of the MSExtra 3.1.0 & B&G 2.891 cased MicroSquirts will not go ahead, and the beta MicroSquirt which I could fit (it's working) I won't be fitting until I stabilise the dwell fluctuations recorded in TunerStudio on the bench test rig. This is causing one ignition channel to behave erratically at certain times (sparks dropping off 2&3 cylinders at certain RPMs), yet there is no difference in spark quality, when I trigger the ignition amps by other means*.
I might have to buy a new JimStim, although I am hoping I can cure this issue so I don't have too.
* I found I can make the ignition driver cycle quickly by powering off the MicroSquirt and leaving it powered, so it becomes a spark generator. Incidentally the four stick coils, wired as a 2 x 2 set up, pull around 8.25A in total at 12.8V.
To keep things moving along, I decided to test the old cased MicroSquirt running MSExtra and wire the new VR Board and do a back to back with the other cased MicroSquirts running B&G 2.891 & B&G 3.760, that way I'd be testing the software not the hardware.
The dual VR board receives & accepts two two wire VR sensors, and processes this into an OPTO+ input and a VR2+ input. To get the negative side working, I tie OPTO- to the SENSOR GROUND, and this works perfectly. I now need to mod my bike loom ECU adaptor to reflect this, because VR1 was on the co-ax, and my OPTO- is not tied to the SENSOR GROUND.
Code: Select all
<<---- VR1+ |----- +crank sensor <<-----VR1- |----- -crank sensor Pin 30 OPTOIN+(grey/red)----------| VR Out 1 ---<< <<-----VR2- |----- -camshaft sensor Pin 4 VRIN2+ (brown/white)-------| VR OUT 2 ---<< <<-----VR2+ |----- +camshaft sensor Pin 1 BATTERY 12V (red)-----------| 12V+ ------->> Pin 21-23 GROUND (black)-------------| Ground* ---->> Pin 31 OPTOIN- (grey/black) connects to Pin 20 SENSOR GROUND (white/black)
However, the last bit can wait a bit, as I can now encapsulate a Dual VR board to mod the VR outputs from the crank & cam sensor, which will in effect pre-process the signals, make them more suitable for the cased MicroSquirt OPTO & VR2 inputs, and allow me to tweak my settings for both firmwares.