My '86 GTI build

[walterclark]

I just returned from DSG where I did a number of dyno runs to dial in the timing and fuel. First thanks to everyone at DSG.

The first couple attempts at pulls were disturbing in that the air:fuel ratio stayed at stochiometric or leaner at WOT. I quickly found that I at some point recently must have not connected the throttle body switch that gets me WOT and IDLE based on throttle position. I probably ran that way at Burke. It was OK at Ascutney but I did remove the intake to check head torque after Ascutney so must have not reconnected that plug then. I didnt check the power graphs - although I think they were recorded - with it unplugged because we aborted the pulls early, but the car seemed way down on power.

Anyway the first pull with everything plugged in came up at something around 148 peak HP at the wheels. I adjusted the ignition timing plus then minus 2 degrees and picked up some HP with it retarded 2 degrees from where the old motor ran best last year. Then I changed the WOT AFR and with it closer to .84 than .87 we picked up a few more. Final number according to the dyno is 157.2 at 6600RPM. The torque is nice and flat from 3400-6600 as well. With the normally quoted driveline losses for the VW this means I should be around 190HP at the crank. After the fact I viewed the data logs from the LM-1 and it seems the fuel was leaning out as the RPM increased with is getting to .94 at 7000. The drive voltage that forces enrichment went up to the 3.5V limit I programmed in and after that the lambda leaned out. I will raise the max voltage but I suspect I will only improve the lambda some not correct it entirely. I believe I am running up against limits with either the coolant temp inputs ability to add fuel or the CIS-E system in the car to deliver it. I suspected this was happening based on the Ascutney logs and have started making firmer plans to move to a Megasquirt based sequential electronic injection with ignition timing. Probably a winter project.

[Rabbit Farmer]

wow, very nice power!

[3rdgendennis]

Josh's car runs Megasquirt, and if you know what you're doing with it (we didn't, hence the events at Okemo 1 last year), it is awesome. There is a ton of support for the system, and they're always coming out with new upgrades/updates

[sciroccohp]

not sure about CIS-E but could you lower the control pressure and give it a little more fuel pressure to bump it up?

[walterclark]

not sure about CIS-E but could you lower the control pressure and give it a little more fuel pressure to bump it up?

That is basically what I am doing with the DPR by using a voltage at the coolant temp input to the ECU to drive it. I have read some stuff suggesting that the CIS-E on the Golf is good up to about 200HP but I dont know is that is 1) at the wheels or crankshaft and 2) at what AFR. Anyway, rather than polishing this turd much longer I will do what Ican with theLM-1 to drive it rich then replace the system over the winter.

[walterclark]

Josh's car runs Megasquirt, and if you know what you're doing with it (we didn't, hence the events at Okemo 1 last year), it is awesome. There is a ton of support for the system, and they're always coming out with new upgrades/updates

The upgrades/updates part is why I have held off buying the module until I really needed to go with it. Seems like there are a million MS users with old Golfs alone so I doubt I will have to invent much. As everyone knows I love stuff like this and assuming the rest of the car goes into the winter season in one piece I expect to have a ton of "fun" building, testing and tuning the engine with MS.

[walterclark]

A bit of an update, including why I was so slow at Ascutney and what I plan to do about it for Philo.

I have known I am up against the limits of my fuel delivery for a while this season. By limits I mean I cannot maintain a Lamba of .87 at full throttle under higher loads of 3rd and 4th gear at higher RPM. The data logs show my enrichment control voltage, which is between 1 and 2 volts at lower RPM and in previous seasons (stock compression and pump gas) would remain around 2V to the redline would increase with RPM now that I began using Sunoco 100 (260GT). This voltage which is generated by my wideband analyzer tries to maintain .87 Lambda at WOT and can go as hight as 4.5V which is the max the input on the ECU will sense. There were times when it would simply slew right to the 4.5V limit and the Lambda number would go up to a bit over .9. Starting at Okemo I began running VP 109E, at first adding it to the 100 in the tank, probably diluting it 50:50, then as more fuel was added less and less 100 and more 109.

Sunoco 100 is somewhat oxygenated (around 3.3%) and contains some ethanol (9%), The VP109E is more oxygenated (number unclear) and contains more ethanol from what I read, although exact numbers are not published. As the oxygen in the fuel replaces the some of the fuel by volume it would tend to run leaner and the added ethanol does the same. Normally the added fuel need would automatically be corrected for by the injection system using the the wideband sensor as its guide, but when no more fuel can be pumped thru the injectors or fuel distributor then the correction system hits a limit and it goes toward lean. The problem might not be visable at all if I were able to just run a 93 octane pump gas with no added oxygen agents, but with a static compression almost 12:1 that is risky.

As the 109 became 100% of my fuel by Saturday at Ascutney the enrichment would max out when the engine went past 5000 RPM, and the control voltage would slew to max, then the Lambda would increase - now to 1.0 or even higher. I had no solution in mind on Sunday other than to short shift at a little over 5000 RPM in 3rd and 4th and by 6000 in 2nd. My power curve is in the 4500-7000 band. I should have been pulling the longer sections at 6000+ RPM but needed to go to a higher gear and found myself at 5000 or less and unable to accelerate. So that is why I was slow.

My plan to deal with this for Philo is to run a 3:1 mix of 93 octane pump gas and the VP109 I have. This should give me somewhere around 98 octane or something similar to the Sunoco 260GT and reduce or eliminate this runaway lean I am seeing. With the usually cool temps at Philo I might be able to do even less 109 if need be but I will wait to see the data before I do more...or less.

The point of me mentioning all this is as a wakeup call to all you running fuel mapped systems (where the fuel delivery is based on a matrix of throttle and RPM) but not Lambda input. Higher octane-than-pump-gas unleaded race fuels usually contain more oxygenation agents and/or ethanol making them burn leaner than pump gas. So if you tune on pump gas and assume you will be safe running race fuel on that tune you may be wrong. You will be leaner and thus hotter...and if you dont have knock sensors to trigger retarding timing or reduce boost you might be damaging your engine.

[honda#72]

Very good point about fuel mapped systems without a Lamda input. Thanks for posting info on the fuel types, and ethanol contents. I have been thinking of using a race or marine fuel for storing the race car in winter, and sleds in the summer. I was unaware that the Sunoco 100 had that high of a ethanol content, not much better than regular 93 octane. Guess I will look into the marine ( no ethanol ) pump gas I have heard is availible, though probably gonna get harder to find as fall comes.

[sachilles]

We have non ethanol premium(91) at my local gas station. It tends to disappear this time every year. Need me to grab you(or anyone) some while they have it? I horde a bit for my lawn mower and ATV.

[sachilles]

As luck would have it, they ran out of non ethanol over the weekend. The store is looking into a different supplier that can sell it to him through the winter. He'll know on Monday.