Thread: #6

**speedpro** · 14th December 2013, 18:16

The default config had the engine as a 62.5cc engine. I asked about it and was told that was correct. Yesterday I got an email casting doubt on that. So, I set the engine capacity to 125cc and reloaded the default VE and TPS maps. Double the cylinder capacity would result in twice the fuel requirements being calculated. On John's dyno it ran rich enough after start to have both cylinders blowing flames out the pipes. Easy enough to fix by simply setting the warmup enrichment to "0" at 30degC. After warmup it seemed rich at lower throttle settings at lower revs so I set a small block to smaller values in the VE table. The engine was sounding pretty good so I decided to put it in gear and give it a rev. That worked good enough to try a full throttle "through-the-gears" run. It revs better than the video sounds but limits at 16,000rpm. Pretty happy for sure. Will start tuning on my holidays.

http://youtu.be/x0quPqr3wkE

**mr bucketracer** · 15th December 2013, 21:31

sound neat ! i want one ..sounds like a 60s racer

**speedpro** · 17th December 2013, 21:40

JC's neighbours must love seeing the Legnum turn up. I've had a few afternoon sessions on the dyno. I still have to get the dyno software fully sorted but I can record runs and do comparisons. Today I was just looking at 10% throttle runs and only up to about 8-10,000rpm. With this throttle opening the ECU refers to the Volumetric Efficiency(VE) table and never the TPS. With the throttle held as steady as possible there are only a small range of values ever active through the rev range so it is pretty easy to make adjustment and rerun the test.

What I've been doing before starting the engine is make changes to the VE table in the region I'll be testing, save it on the PC and send it to the ECU. Then I've made the next changes to the saved copy of the VE table ready to import to the active table. I then fire the bike up and do the test after which I import the modified VE table and run the next test. A quick comparison says if the tuning was going in the right direction. A new set of values can be entered as a block and then the table re-imported and the next run done. It's quicker than it's taking to type this.

Anyway, as good as I thought the engine sounded it turns out it has been a bit rich. Decreasing the values has netted 1hp or a whopping 50% increase. Even with no load at 10% throttle the motor will only rev to 10,000rpm and when you look down the throttle bodies you can see that there is very little opening for air to get in.

Without a dyno PC there is still a way of checking how well the engine is accelerating. You can record lots of things during a session using the software that comes with the ECU. By looking at the engine speed and the rate it increases you can see if the power tails off. With a bit of mucking round you can see which values would have been in use in the table and make a change. A Lambda sensor would make it even easier.

**speedpro** · 30th December 2013, 19:56

Had another session today. It definitely needs the Throttle bodies adjusted for synchronisation and to then balance the idle air screws and I'll sort that tomorrow. Today I decided to try to save the dyno runs and then recall them for analysis and comparison. Finally managed it and ended up doing some 30% throttle runs. At 30% if the engine is revving over 4000rpm it uses the TPS/RPM table. Holding the throttle steady results in the same line of values being used so it's pretty easy to tune. Initially it just sort of faded at about 10,000rpm. I altered the whole row of values it was using to the same larger value and did another run. There was a gain across the whole rev range and it revved to 11-12,000rpm. I increased the values and reran the test and got another gain across the whole range and now it revved over 14,000rpm before fading. The last test, I increased all the values again and it revved on quite nicely and was sounding crisp. Towards the end I thought my leg felt warm. I was standing beside the bike, reaching over to the throttle and running the PCs with my other hand, and not really watching the bike itself. I had a blanking plug in a hose that will eventually feed water to the turbo. This popped out and released hot water all over the dyno, and my leg.

Back at home I've made a piece to clamp to the twist grip. This will allow the throttle to be opened to more or less the exact same position every run. This will reduce errors and speed up testing. It's difficult to hold the throttle steady enough if the bike is bouncing around a bit. I'll go back tomorrow with a bag of T-pieces and small hoses and hook up my vacuum gauges to sort out the TBs and then finish testing.

**speedpro** · 2nd January 2014, 17:49

More time on the dyno and a very important reminder of how the tables work. The dyno does not record engine rpm so I wanted to match 10,000rpm to a particular speed. To do this I set the TPS/RPM table value at 10,000rpm and 32% throttle to about 1/2 what seemed a good value, as tested. The next run should have seen it fall flat on it's face at 10,000rpm. What actually happened was not much. It ran through the load cell as if nothing happened. I then set the values at 10,000rpm for the TPS settings above and below 32% to the same value and did the run again. This time it ran into a wall at 10,000rpm.

This was a good reminder that the software does not read the value "only" at the highlighted cell but blends the values in the cells around to determine the correct fuel requirement. It was interesting to see the fuel pulse width slowly decrease as the load cell with the seriously reduced value was approached.

So all the effort so far has just been tossed. That is because I have only been altering the one row for 32% throttle. The rows above and below are markedly different. I copied the values from the 32% row to the rows above and below and the engine would not run at 32% throttle. I have to think of a new tuning strategy. I'm thinking of concentrating on a particular cell, say 32% TPS at 10,000rpm, and altering all adjacent cells to the exact same value initially and testing until I have the optimum value then picking another cell and so on. Once I have a few cells mapped I can blend the intermediate cells after which it would need to be done again. It's going to be a mission.

To end the session #1 cylinder exhaust has broken off, just great. 2nd beer just finished, bbq to sort, low/mid 20s, gentle breeze, more beer, 1.5 weeks more holiday, to the beach tomorrow. Sweeeeet!

Oh yeah, I haven't got sick of the sound this thing makes.

**speedpro** · 6th January 2014, 18:04

Bit of shed time and I've added a piece of tubing at the end of the headers. It is purely there as a place to mount the O2 sensor. It's pretty ugly but does the job. Before he went to Chch John Connor lent me a wideband O2 sensor kit he'd made for a 600. I've fitted that sensor with it's own little control box with series of LEDs indicating rich/lean.

After I warmed the engine up in my garage I could see the LEDs indicating the mixture fluctuating as I revved the bike. At idle it was indicating rich so I took .1 out of the VE table around the active cell and it immediately responded by indicating a bit leaner. I took a bit more fuel out and the indication was leaner and leaner and the idle went up, to the point where the ignition backed off to reduce the engine torque and therefore the idle speed. The ignition backing off made the indication go rich so I got in and reduced the idle air bypass on the TBs. This normally would reduce the idle speed but in this case the ignition timing advanced again, the indicated mixture returned to normal, and the idle stayed the same. I had loaded the default VE table and pretty well knew it was a bit rich but this confirmed it.

I'll use this WB sensor on the dyno to get things in the ballpark and then mount the Ecotrons NB sensor and log the results which I'll then use to tune the engine a bit more precisely.

I thought it interesting that the mixture indicated so rich when the timing retarded to lower the idle RPM. It could be a trap as you could keep taking fuel out until it got so lean the engine slowed down and then the ignition would advance again as the idle reduced and the indication would go more lean, and back you go the other way. You definitely need to adjust the TB stops or idle air bypass to set the idle without the ignition retarding. It's only a problem on TBs with no idle air stepper motor.

**speedpro** · 16th January 2014, 18:07

Just got a pile of new software from Ecotrons. They were a bit shy with the truth when I ordered the system. I specifically asked about the system supporting the engine doing 18,000rpm and the reply was positive. Turned out I could only specify up to a maximum of 16,000rpm. After running the engine on the dyno it turns out it will only rev to 14,000rpm which is a software limitation. They have sent new software which allows 16,000rpm. I've run a comparison but there was nothing of interest found. I suppose they get a lot of silly enquiries and just thought my question was another one not expecting their answer to ever get tested.

I have also received an update on the problem when the ECU case was bolted to ground and it stopped working. Rob has had the same problem. Turns out it was a manufacturing fault which is easily fixed by popping the lid on the ECU and inserting a bit of insulation.

I've got new software to run the ECU so will have to flash that which looks pretty easy with the supplied "flash GUI" software.

They also supplied an updated ".cal" file with the auto-learning feature enabled. I sent them my last ".cal" file and they included all my modified tables in the new file. I could have imported all the tables one at a time but that would be a bit of a pain so a nice touch them doing it.

I also got directed to a new tuning guide which looks to have a lot more detail which will be of interest to anyone doing something a bit different. Still slightly odd English in places.

Their ALM device is looking better and better. It uses the higher spec Bosch sensor and has a very fast response time. It also includes an input for rpm. This is connected to an injector. Although most WB controllers have a programmable 0-5V output which could be used by the ECU some are considerably slower responding and therefore not as good for tuning. As usual nowadays it's the software that makes the difference. The ALM as well as connecting in a reasonably straight forward manner also comes with software that allows the output to be captured via a serial port/USB and displayed on a graph of rpm versus lambda which could be useful.

**koba** · 16th January 2014, 18:37

You and TZ have kept me keen on Carbs!

**speedpro** · 16th January 2014, 21:24

Funnily enough I've been quietly reminiscing about a lovely set of brand new Keihin CR31s I bought in 1982.

I've also thought it would be interesting to install an O2 sensor on a carb'd bike and map the lambda at a range of throttle positions and rpm settings. I suspect things won't be that flash as a rule. Of course for racing it doesn't really matter about part throttle so much, certainly emissions are of no concern and even being a bit down on power at a particular throttle/rpm combo, you just open the throttle a little more.

My setup is running pretty sweet at the moment and there is plenty more to be done. It's just that you can identify exactly how well the engine is running at a huge number of load points while with a carb you don't have the huge range of precise adjustments to make. A carb can be made to work very very well though in the real world. Emissions are what has driven the engine management/fuel injection transformation primarily in the real world I think

**Grumph** · 17th January 2014, 05:24

Originally Posted by speedpro

Funnily enough I've been quietly reminiscing about a lovely set of brand new Keihin CR31s I bought in 1982.

I've also thought it would be interesting to install an O2 sensor on a carb'd bike and map the lambda at a range of throttle positions and rpm settings. I suspect things won't be that flash as a rule. Of course for racing it doesn't really matter about part throttle so much, certainly emissions are of no concern and even being a bit down on power at a particular throttle/rpm combo, you just open the throttle a little more.

My setup is running pretty sweet at the moment and there is plenty more to be done. It's just that you can identify exactly how well the engine is running at a huge number of load points while with a carb you don't have the huge range of precise adjustments to make. A carb can be made to work very very well though in the real world. Emissions are what has driven the engine management/fuel injection transformation primarily in the real world I think

Emissions and the sheer difficulty of an owner tampering with it.....Legislators hate tamperers.....

Back when we had a Superbike round, Anders Andersson raised eyebrows with an O2 sensor on his OW01 Yam....he said it made changing to suit local conditions faster.

Personally it's taken me around 40 years to get to the level of knowledge I have now around carb setup....I don't have long enough left to learn injection tricks.

**speedpro** · 7th February 2014, 20:29

On a completely different tack, I have been messing with Gary's MB100 engine. Finally put it on the dyno after work with nice improvements from the test done earlier. I had to up the jets a couple of sizes and backed the timing off a bit. The power is modest but it'll be easy to use. The needle may have to go down one notch but we'll check that on the track.

**goose8** · 8th February 2014, 15:35

Originally Posted by speedpro

On a completely different tack, I have been messing with Gary's MB100 engine. Finally put it on the dyno after work with nice improvements from the test done earlier. I had to up the jets a couple of sizes and backed the timing off a bit. The power is modest but it'll be easy to use. The needle may have to go down one notch but we'll check that on the track.

Big thank you to Mike the bike was awesome to ride today , time to give it a big blow out tomorrow and go for some wins

**F5 Dave** · 8th February 2014, 18:39

. . .time to give it a big blow out tonight and go for some wins[/QUOTE]

Fixed it for you.

Yeah that's my usual training too. Less on the vodka these days. More a nice bottle of wine. Or so.