top of page
Cosworth Black

REVIEW AND ANALYSIS OF MULTIPLE COSWORTH VEGA DYNO RUNS  --Duke Williams

 

​Who was it who said "the stock cams don't pull power beyond 6000?"

 

Even Jim Rigg's showroom stock car (except for re-indexed cams) didn't drop off up to 6500.

 

 

This is the effect of the later closing inlet valve with re-indexed cams. Jim's car has about the same peak torque as the Lippert car  (both have re-indexed cams), but Jim's torque curve is flatter, so it is stronger at both ends of the scale. His biggest "problem" is the 12+ psi backpressure of the OE exhaust system.

Craig's top end did drop off a bit beyond 6000, but very little.

 

He still has the OE cam indexing with an excellent exhaust system. His torque between 1500 and 2000 is actually a little better than mine, but then there is the first "hole" in the torque curve at 2000-2500 that does not show up in my curve. Both have a hole at 2500-3000.

The holes and humps in the torque curves are due to various combinations of inlet and exhaust system wave effects and valve timing. Their behavior can be very complex and the engine simulation programs will usually not pick them up. The exhaust backpressure on Jim's engine washes out most of these effects.

I expected Craig's car to do better given his excellent exhaust system, but the OE head with the ridge in the valve pockets and on the inlet valves may have a more deleterious effect on inlet flow than I originally thought! Time for a head job, Craig!

Note that on my engine the first pull was from 1500 and the second pull from 1000. No protest. It just pulled.

 

I have no problem cruising at 1500 in fifth with enough reserve torque to accelerate modestly if needed. In the lower gears it will pull cleanly from below 1000. In OE trim like Jim's car I know the engine will NOT pull from 1500. You HAVE to downshift.

The 80 percent torque bandwidth on my engine is 1900 to 7000+. Extrapolating the torque curve indicates the upper 80 percent bandwidth is 7200 and the power would probably continue to be nearly flat to about 7500. On a good high performance street engine you want the 80 percent torque bandwidth to come in no later than about 2000 or you will have a real soggy/cantankerous engine in normal driving.

Jim's torque curve is the definition of FLAT!

 

 

It doesn't rise much with revs because of all the exhaust back pressure from the OE cat and exhaust. I guess one way to flatten the torque curve is plug up the exhaust, but this is not the best choice.

My CR is a measured 8.0:1, which is probably typical for a production engine. GM's advertised ratios are typically about 0.5 high because the decks are machined high. If I decked the block .015" to achieve minimum quench clearance of .035" with the OE .045" head gasket and pistons, the CR would be 8.5. Likewise, with minimum quench clearance the 10:1 JE pistons should yield a true 10:1. This will increase torque across the rev band in the range of 5-8 percent and would likely increase the RWHP to around 130 in the 6500-7500 range, and the useable rev range would go to 8000. I don't rev beyond 7000, but if I ever blueprint the block and install Crower rods, I would be comfortable with 7500. Above this the oil system gets "iffy".

How much of this you can get in street trim will be a function of how well you can minimize fan and exhaust system parasitic loss.

Chris' engine performance is enigmatic.

 

All the other engines with some combination of OE or re-indexed cams/OE EFI or Webers show peak torque in the 4500-5000 range, so Chris' engine should not have dropped off beyond 4000. Note that it has the two holes in the torque curve (2000-2500 and 2500-3000), same as Craig's engine, which appear to be characteristics of the OE cam indexing, and a third hole at 3500! Also the low-end torque is very weak with the 80 percent bandwidth not beginning until 2500. Given that the primary mods to Chris' engine are higher compression and a reworked head, the shape of the torque curve is inexplicable.

Here's a summary of the low end torque characteristics.

RPM..........1500......2000.......2500.....3000

Rigg.........NA........86(est)....92........93

Schmitz......82........90.........93........97

Williams.....80........90.........96........98

Wheaton......NA........81.........92........99

Lippert......NA........75(est)....82........88

Both Craig and I have quick ignition advance curves, which provide full spark advance early. This will improve low-end torque. The owners of the other cars did not provide any information on their ignition advance curves.

Peak torque is basically a function of displacement and compression ratio. Torque bandwidth is a function of "tuning" - valve timing, ignition advance and fuel curves - with the top end benefiting from hand blending the valve pockets to eliminate the ridge, and a 30 degree top cut on the inlet valves to grind away the unused portion of the seat and eliminate the ridge.

bottom of page