GTS1000 on MicroSquirt running MSExtra

[24c]

Getting a little punch drunk with these cold starts, day 9 or 10 Anyway, it's starting OK now, but revving a lot higher than the stock ECU, but a least it's catching on the button and keeps going.

I have refined the VETables using Speed Density as far as I can at the moment without further load tests on the dyno, however I have decided to have a play with Alpha-N mode as this offers a potentially more stable idle for my fluctuating kPas. I have tried flatting the tables in the areas that the fluctuations occur, but the idle is still not as steady as the OEM ECU. I have put new plugs in as they are quite sooty, and I have experimented with dwell etc, and it seems to make no difference. I am wondering whether the wideband sensor & the TechEdge 2J1 is calibrated correctly in TunerStudio.

The first attempt at Alpha-N produced a stall and a load of fluffiness, but no wonder when I realised what it does...I chose to mod the ignition first, so I had no Secondary Fuel Loads at first, just the Alpha-N Primary table. By tweaking the Ignition Tables at first, I could see what "additive" did, and having both tables open in TunerStudio on a running bike I could see and follow the RPM dots, which made me quickly understand what was happening. A few tweaks here and there, first to the kPas above idle, then the kPas around idle, and lastly the overrun underneath idle. I loaded the bike on the dyno roller and it was amazing what a difference this made compared to my Speed Density based Ignition tables...the bike was more responsive still!

So then I applied this "table shifting" to the fuelling side, and surprisingly again, it worked well loaded on the dyno. It was a little flat with sudden openings, but very smooth with slow transitions of the throttle and very good in fourth gear from idle upwards in roll on tests. I am really looking forward to doing some road testing, once I beef up the loom adaptor.

[24c]

Day 11, and so far so good, bike fires up and keeps going, so ASE and WUE at around 8ºC are now working OK. Refined the Speed Density Project a little, but bike is still running rich, but does seem to "drive" well on the dyno when loaded. I tried playing with the MAP sampling, but I need to get my head around this some more, as I wasn't getting the results I expected.

I also did some more experimentation with the Alpha-N based Project, which now uses Speed Density for Secondary load. Starting was as good as the earlier Speed Density project, but idle is much more stable, and I really think I prefer this mode. Pick up unloaded was OK, but under load the "Additive Secondary Map" was way too weak. I had the bike loaded up in fourth gear from 1100-7000, and eventually got it to pick up well. It is now running rich under load at lower revs, but I am well pleased with the progress so far so early in the project (day 2), as the upper limits seem to see target AFRs under WOT. I have to read the manual some more and try and work out why the acceleration parameters are not working as expected, but maybe because I have them set for more sudden changes (faster TPS acceleration rates) is the issue.

I pulled the bike off the dyno, because I need to revise/improve the front wheel clamping/location, and I have removed/bypassed the exhaust extraction fan to see what effect this has. I believe the fan is not good enough, but just want to check I am not getting a reverb off the fan blades at higher rpms.

[24c]

I had to go out today, and get some convoluted stainless steel tubing, to make flexible ends for the exhaust sniffer, or lambda sensor diffuser pot. I will post pics when it is finished, but finding this tube in stainless has taken a lot of searching and visits to various trade outlets. Surprisingly, I was able to get cut lengths just 8 miles away, and they were really nice, helpful people.

On the way back I thought I'd check on the Thunderace engined GTS1000,, the faster green one....one day!
It needs a little bit of work, before I can plug the MicroSquirt in it, like plugs, coils and basic checks, as it's been stood close on three years.

However, the old black GTS was looking sorry for itself, so I decided to test the red GTS MicroSquirt on the old girl, and after fitting some old and tired mucky coils and dirty spark plug leads, as I was running the R1 COPs on this previously, she fired into life.

It was slightly odd because she was running rich, and after fitting new plugs and such, the bike was running better. It was on the WUE cycle and was fluffy around 5K rpm, and then it dawned on me.

This bike has done 110,000 miles and the injectors must be getting tired/worn and the holes slightly bigger....either way, a little tweak to the VE Tables, and mods to the Ignition Advance Tables soon had it running sweeter, but ultimately this could be compensated with by altering the "Required fuel" or injector open time

I tried to check the timing, but the lack of a steady low speed idle, 1500 isn't really close to 1100, although I tried fooling the system by putting the same advance figures in the ignition tables for the higher rpms. This meant the strobe and timing marks, complete with oil spray, were dancing all over the place. The blob of Tippex couldn't be seen using my mates neon timing light. I am going to sort this out soon, by making a new oil plug, so I can finally calibrate my "Trigger Wheel Settings" in the software. I am just guessing around 157º at the moment.

[24c]

I had to go out today, and get some convoluted stainless steel tubing, to make flexible ends for the exhaust sniffer, or lambda sensor diffuser pot. I will post pics when it is finished, ...snip...

I managed to weld the rascal up today, and it works well, in that there is no blowback at high rpms, that's upto 11.5K. Pics as promised...

O2 sensor diffuser probe.jpg

I have also improved the exhaust extraction three fold and there is no whiff of exhaust using the 150mm pipe, and the 200mm cone.

Dyno_extraction.jpg

A good day today, and when I got home the postman has been, so I can build my replacement ECU now, once I call at CPC Farnell for some bits.

MicroSquirt_Module_etc.jpg

[24c]

Just an update, I managed to strap the red GTS1000 on the revised dyno mountings, and started out testing the ASE/WUE mods, but once it got warm, I thought I'd play at smoothing out the idle instead.

First I switched out all the Acceleration Enrichments, and flat lined their responses in the Acceleration Wizard graph. Then I made sure EGO was disabled.

Then I equalised the lower RPM bins for VE Table 1 & 2 and Ignition Advance Tables 1 & 2, which meant I was not getting any unusual crossovers. The load kPas bins were identical in the left hand vertical column, so by identifying the regions where the RPM dot was working, I flat lined these values, avoiding zeros.

It worked, and so I reduced the stable idle to 1100±50, but the bike was still fluffy at higher revs. I cannot time the bike yet, without spraying oil everywhere with the inspection plug removed, so I reduced the TDC settings from 157º to 150º and it seemed better. I further flat lined the ignition tables and adjusted the advance till it sounded sweeter.

However, it was still fluffy, so I reduced the required fuel from 9 to 8, which improved higher rpms, and then increased the idle VE bin to compensate. As a result of these tweaks, the engine seemed to be running a little better, so I slowed the idle speed further to around 1050±50 rpm, which is close enough to where I want to be.

I also played with the lags in the General dialog box, and adjusted the windowed values to see how these affected my MAP, and the throttle response was worsened, when I reduced the % figures, so I increased them. I made the windows the same IIRC at around 45º.

I now have to tweak the Acceleration Wizard graph, so the idle fluctuations don't interfere with the idle running, and I think I need to look at the dialog box settings some more and understand how these work.

I was able to complete the testing without smelling like an exhaust silencer, and so overall the modded exhaust extractor worked very well, but I need to do something about the disconnected & open crankcase breather pipe, as this was a source of fumes when revving the bike. I will be modding the exhaust extractor further still, but this is the current version that allows me to log the exhaust gases, as well as extracting them.

In the picture below, you can see the MicroSquirt, the prototype adaptor loom, and the combined exhaust extractor/wideband sensor diffuser probe.

dyno exhaust extraction with O2 sensor.jpg

And a close up

close_up combined extraction with O2 sensor.jpg

[24c]

I really thought I'd make some progress today, hearing all these success stories using TunerStudio and VElive and I did.

I am running Alpha-N and I have been guessing at the fuelling using my previous Speed Density settings. The plugs have been consistently sooty, but I wasn't running anywhere near enough advance, and was carrying too much fuel, as I found out later today.

The bike is sounding really crisp now, and pulls strongly in fifth gear on the dyno, so it's time to prepare the bike for the road. I am reducing my WUE values and boosting my ASE at the same time, and this is making the bike quicker to start. Hot starts are taking slightly longer than the OEM ECU, but cold starts rpms are falling and beginning to closely match the standard bike. Idle is stable considering the tired and dirty HT system, and yes I know I should be starting from a good set up, but this is a helpful real world test.

I managed to get VELive to do it's stuff, as I soldered up the additional take outs for the TechEdge 2J1 unit, and fed the WbLin+ into the MicroSquirt. I still have a little noise/oscillation on the WbO2, but not enough top spoil the show, and I also have some coolant fan noise spikes, that affect TPS and the MAP signals!

I also experimented with the acceleration enrichment SD vs TPS settings, and I am quite pleased with the throttle response. This msq will work down the road, and I expect VELive to fill in my cruising, once I generate an appropriate AFR table.

2010-03-20__latest_modded_VE1.msq

Unfortunately the day was truncated by the death of the dyno extraction system, and I now have to rethink/reposition the extractors. I now appreciate how hot the exhaust gets and no amount of cool air blends it down enough to stop fan destruction.

[R100RT]

Transport truck exhaust flex tubing is cheap and plentiful, S.S. which is a little more, handles very high heat without issue.
Perhaps you can devise a venturi type arrangement that puts the fan as a pusher, bring the exhaust into a "Y" collector so it flows into the air stream, uphill. Change in size of tubing to assist venturi.
Nice booth.

[24c]

That's almost what I thinking of doing now. At the moment prior to the fans melting, I was sucking the blended hot air out of the sealed room, and the replacement cool air was entering via an open duct. Now I will be looking at pushing air into the room, and modding the exhaust bellmouth (larger), so that the air pushes into the outlet duct and takes the exhaust with it. That way there is no heating applied to the fans, and I can add more in parallel to boost the flow rate if need be.

I have ordered another fan to test this further, as a smaller metal pusher duct fan I normally use to cool the radiator, which I tried after the meltdown, did appear to back up my new thoughts and does support your idea too.

Thanks for your comments R100RT, as I'll look at stainless flex (chimney liner for stoves etc) again. I have tried aluminium ducting, but it gets blown to pieces in a short time. I am using high temperature reinforced rubber flex and galvanised spiral ducting at the moment.

PS The booth is nice, because I don't annoy the neighbours, however I did notice I need more exhaust silencing outside.

[R100RT]

I made the mistake last year of tuning while playing around with different types of mufflers, its easy to get involved and lose track of running time - one neighbor who actually is a bike enthusiast, reminded me his wife was loosing patience.
I may have to loose my "Vista" equipped tuning lap top which requires a USB/ Serial adaptor, I'm getting blue screen "dump physical memory" failures after about 2-5 minutes on MegaTune.

[24c]

I may have to loose my "Vista" equipped tuning lap top which requires a USB/ Serial adaptor, I'm getting blue screen "dump physical memory" failures after about 2-5 minutes on MegaTune.

Haven't you tried TunerStudio?
Works fine on my XP running MacBook with Prolific USB/Serial adaptor.