XDi ignition systems users

[DWit]

Originally Posted by Walt Fricke

I ran my 2.7 race motor (46s, big cams, nominal 10.3/1 Mahle race pistons, crankfire twin plug) with:
0 static
12 at 1,000 for 12
14 at 3,000 for 26
2 at 8,000 for 28.

Worked fine.

Once you have your motor broken in, take the car to a chassis dyno and get a torque curve. Use the torque curve and your transmission gearing (you don't need tire diameter or R&P for this, but can add them in) to develop "thrust" curves at 500 rpm intervals for each gear. Where these curves cross (all will except first for stock type gearing) is where you should be upshifting, as the delivered rear wheel thrust will be equal in the "old" gear and the new, higher, gear. Those are the optimum shift points. You can do this with a pencil and graph paper, or make a spread sheet so you can easily see how changing gears or torque curves will affect things, etc.

I had been shifting at 8,000 rpm. The motor mechanically will take this well enough - may shorten valve spring life, but then overrevs which don't bend any valves can also shorten spring life. But with a torque curve I found 7,600 was the place to shift. Sort of win/win, doing it right and gaining acceleration while reducing strain.

Do you think the 964 crank will produce a torque curve low enough that your optimum is below 7,000?

Unless you are using stock rod bolts, I'd think you might up your limit if you autocross or do DEs and the optimum is over 7.

Walt,

Will you please explain further the thrust curve procedure you mentioned? I always assumed (possibly erroneously) that the horsepower peak is used to determine the up-shift point. Shift 500 RPM higher, so that the RPM drop in the next gear is only 500 RPM (more or less) below the horsepower peak. That way you are maximizing the area under the HP curve, no?

Dave

[Walt Fricke]

Yup, you were in error. At least insofar as you were using a rule of thumb - go 500 over HP peak, and shift. Though sometimes that works out pretty close. But sometimes it does not.

If you multiply torque by a gear set and then by the R&P, you will get what you might call rear wheel thrust. You can also factor in the rolling radius (or the diameter, or the circumference) of the tire if you want exact MPH values, but in terms of rpm and shifting it doesn't make any difference.

So make a graph (nowadays, make a spread sheet) with column A RPM at, say, 500 rpm intervals. B is the torque at those RPMs from your dyno run. RWTorque is just fine, no need to wonder about accounting for transmission loss. C is the torque multiplication due to the gears and R&P, and that is thrust in 1st gear. D is the speed in 1st. E and F are torque and speed in 2d, etc..

For a 5 speed, you can plot the thrust columns as a series of curves, with thrust as the vertical and speed as the horizontal. Each plotted point represents an RPM. Their basic shape is that they start at a lower thrust, reach a peak as rpms climb, and then drop. 1st gear is the highest curve, and is at the left of the graph. Each succeeding curve starts lower but at a higher speed, climbs, and drops.

Where the curves cross is the shift point from the lower to the higher gear. This is the point where the torque (aka the acceleration) from the falling lower gear is the same as the rising higher gear. That is the area under the set of curves you are maximizing.

You can guess this pretty closely by knowing what rpm is indicated on each side of both curves at the crossing point. Or you could use a finer graded input table between those RPMs. Or assume the lines to be straight between any adjacent pair of RPMs, and do some algebra.

So peak HP isn't directly a factor here. Gearing affects the shift RPM. For instance, typically your curves won't cross between 1st and 2d, because 1st is so low that you reach your rev limit (however you have defined that) before you get to a point where it will cross the 2d gear curve. The remaining shift points will vary some depending on the gear sets.

But most stock 915 gear boxes will have the shift points pretty close to the same RPM. For the torque curve I got long ago for my 3.0 SC, I shift at 6,250. Optimum didn't vary by more than 100 rpm either way as I recall. A gear box built the way Bruce Anderson suggested in his book - equal MPH gain in each gear - works out to a fairly constant shift RPM even as the RPM drops between gears keep falling as you go up through the gears.

I'll see if I can find a visual.

[Walt Fricke]

Yup, you were in error. At least insofar as you were using a rule of thumb - go 500 over HP peak, and shift. Though sometimes that works out pretty close. But sometimes it does not.

If you multiply torque by a gear set and then by the R&P, you will get what you might call rear wheel thrust. You can also factor in the rolling radius (or the diameter, or the circumference) of the tire if you want exact MPH values, but in terms of rpm and shifting it doesn't make any difference.

So make a graph (nowadays, make a spread sheet) with column A RPM at, say, 500 rpm intervals. B is the torque at those RPMs from your dyno run. RWTorque is just fine, no need to wonder about accounting for transmission loss. C is the torque multiplication due to the gears and R&P, and that is thrust in 1st gear. D is the speed in 1st. E and F are torque and speed in 2d, etc..

For a 5 speed, you can plot the thrust columns as a series of curves, with thrust as the vertical and speed as the horizontal. Each plotted point represents an RPM. Their basic shape is that they start at a lower thrust, reach a peak as rpms climb, and then drop. 1st gear is the highest curve, and is at the left of the graph. Each succeeding curve starts lower but at a higher speed, climbs, and drops.

Where the curves cross is the shift point from the lower to the higher gear. This is the point where the torque (aka the acceleration) from the falling lower gear is the same as the rising higher gear. That is the area under the set of curves you are maximizing.

You can guess this pretty closely by knowing what rpm is indicated on each side of both curves at the crossing point. Or you could use a finer graded input table between those RPMs. Or assume the lines to be straight between any adjacent pair of RPMs, and do some algebra.

So peak HP isn't directly a factor here. Gearing affects the shift RPM. For instance, sometimes your curves won't cross between 1st and 2d, because 1st is so low that you reach your rev limit (however you have defined that) before you get to a point where it will cross the 2d gear curve. The remaining shift points will vary some depending on the gear sets.

But most stock 915 gear boxes will have the shift points pretty close to the same RPM. For the torque curve I got long ago for my 3.0 SC, I shift at 6,250. I could shift at 6,500 into 2d, then 6250 into 3d, and about 6,000 into 4th and 5th, but that's too much to remember and I don't like to rely on shift lights. A gear box built the way Bruce Anderson suggested in his book - equal MPH gain in each gear - works out to a fairly constant shift RPM even as the RPM drops between gears keep falling as you go up through the gears.

I'll see if I can find a visual.

[Walt Fricke]

Here is a 2.7 race motor with a tall (60 mph) 1st gear. The dyno run went up to 8,000 rpm for data gathering. I plotted in three different 5th gears. This technique was found in an article in Pano taken from some factory race literature, and reproduced in one of the Upfixins. Lots of work with the hand calculator to create a table and then plot it in. You can see the typical Weber carbureted engine torque dip in the 4-5500 rpm range. And why I was losing out shifting at 8,000 anywhere except from 1st to 2d (and that turned out to be something only useful on a slow paced race start). Except for that, 7,600 is pretty close for all the other shifts.



Here is letting a spread sheet do the work for a stock 3.0 with stock gears. The stock gears aren't quite as evenly spaced as the race box for the 2.7.





David Ferguson created the basic spread sheet maybe 15 years ago for a 901 transmission, hence the table of gears. But you don't need that, you just need to know what your gears are by tooth count, or as a ratio based on that. My friend John Mackin refined this spread sheet using spread sheet voodoo (or knowledge gained from actually taking a class in this stuff) and had the spread sheet calculate the curve crossings. But it is easy enough to do by inspection.

[DWit]

Walt,

Thank you very much for the detailed explanation. But I don't understand why torque is being factored instead of horsepower. The peak of the torque curve (the point of maximum engine efficiency) is lower than the horsepower peak but the car still accelerates past the torque peak because more work is being generated by the engine. Acceleration only begins to drop off after the horsepower peak is passed.

Dave

[Walt Fricke]

Dave - I don't think it is correct to say that acceleration only begins to fall off after the HP peak is passed. And even if it does, I'm pretty sure that is because of the shape of the torque curve.

My understanding is that torque is what governs acceleration. But it is easy for many of us lacking a suitable engineering scientific understanding (like me) to get confused.

Why I think you are wrong about thinking that HP is what should govern the shift point involves a couple of things. One is that, when long ago I had data, I could see the acceleration number in Gs before and after an upshift. The book that came with the G Analyst noted that the proper upshift point was where the before Gs and the after Gs matched. It was pretty easy to see on the data display if you shifted too early or too late.

The second is that not only my friend Dave Ferguson (an engineer), but also Porsche's racing department, used torque in quantifying optimum shift points.

Gearing affects the torque applied at the rear wheels, where it counts. Gearing doesn't affect the HP a motor produces. RPM does, but that is because of the shape of the torque curve. And I think that, knowing the drag coefficient (for air mostly, but also for other drag components) for a car, you can specify the horsepower needed to attain a given top speed. It takes power to push the air aside and overcome various frictions. That's not dependent on gearing directly, it just is a factor in selecting a gear which will have you at the HP peak at your top speed drag limit, if doing Bonnevile stuff. Not the same as accelerating.

I remember reading Hot Rod magazine in the '50s as a kid, where there were discussions about "overpeaking" the HP max. There was agreement that your shift point would be up there. But if it was about connecting the scientific dots, that part didn't sink in. A rule of thumb is easier to remember. And for the engines most of us drive on the track, it is generally going to be true that the optimum upshift point is going to be above the HP peak. But it is the torque curve which makes that so.

I had an engine which I ran up to 8,200 on the dyno. Didn't want to run it up higher. The graph of the torque curve there showed that the optimum shift points were above 8,200. Don't know just where, of course. So I shifted at 8,200. Optimum is over 600 rpm over HP peak. But street motors aren't like that.

[Kremer#2]

After installing my Xdi, I had problems with high idle. I greased the old bell crank and added a new return spring. Idle is MUCH better but still too high Air correction screw does not help.

[fred cook]

Originally Posted by Kremer#2

After installing my Xdi, I had problems with high idle. I greased the old bell crank and added a new return spring. Idle is MUCH better but still too high Air correction screw does not help.

Check for a vacuum leak by removing the oil cap while the engine is running. If the idle drops then your engine probably does not have an air leak. However, if the air bleed screw is completely closed, something is wrong. Check the idle timing advance knob. I have mine set at about 9-10 degrees. Adding advance with this knob will cause the idle to increase. Also, if your (CIS?) mixture is set too rich, it will cause a high idle. Good luck!