Research note · S650 Mustang · Roush / Magnuson TVS 2650 · Gen‑4 Coyote

Are the S650's dual 80 mm throttle bodies a restriction on a Roush blower?

Probably not at any power level a Roush or Magnuson kit reaches. Nobody has published a flow test or a before/after dyno that pins down where the twin 80s start to cost power. The evidence that does exist points the same way: the rotor pack, the intercooler, and how hard it is to change pulleys run out well before the throttle bodies do.

Twin 80 mm area
10,053mm²
stock Gen‑4 Coyote, reused by the Roush kit
about the same as one 113 mm throttle body
Whipple Stage 2 area
9,852mm²
single 112 mm round throttle body
made 890 whp with slightly less area
Pressure drop estimate
< 1psi
at ~1,200 cfm (≈ 900 crank hp), blades wide open
my estimate, not a measurement
Starting point

Same car, same dyno, only the blower changed

House of Boost dyno-tested an S650 GT with a Roush Phase 2 (TVS 2650), then swapped to a VMP/Whipple Stage 2 (3.0 L twin-screw). Nothing else changed: same fuel, same tune approach, same dyno. All comparison pulls were in 4th gear, because 5th hit a speed limiter around 152 mph.

The Whipple was ahead by about 100 whp on the same fuel. The crew said the Roush has a little more low-end torque, and the Whipple pulls away from about 5,600 rpm up. That pattern points to the blower running out of efficiency at high rpm, not to the inlet choking the airflow.

In the teardown their complaints about the dual throttle bodies were mainly about service: the throttle bodies have to come off to change the pulley or belt. They pointed out that the inlet narrows into the rotors but didn't measure any restriction.

SetupFuelwhp
Roush Phase 2, stock pulley93707–710
Roush Phase 2, stock pulley~E25743–756
VMP/Whipple Stage 2, stock pulley93 + booster811
VMP/Whipple Stage 2, stock pulley~E25845
VMP/Whipple, 3.50 pulley~E25, hot861–890
The strongest evidence

Throttle area: the twin 80s already match Whipple's Stage 2

Whipple first planned to reuse the factory twin 80s and judged them to have more than enough airflow for their 3.0 L blower in Stage 1 and Stage 2. They switched to a single throttle body for a different reason. Flow simulation showed a dead zone near the pulley pocket that fed the two rotors unevenly. It was an air-distribution problem, not a capacity problem.

Their Stage 2 uses a single 112 mm throttle body, which has slightly less area than the stock twin 80s. That kit made 890 whp in the test above. So the twin 80s have enough area for at least about 900 whp. The open question is whether the Roush inlet behind them feeds the rotors cleanly, and nobody has published data on that.

Roush markets the reused twin 80s as up to 34% more throttle area than typical single-throttle kits.

Back-of-envelope airflow check

750 whp is roughly 900 crank hp, which needs about 90 lb/min of air, or about 1,200 cfm. The throttle bodies sit ahead of the blower, so they flow air at close to atmospheric pressure. 1,200 cfm through 10,053 mm² works out to about 55–65 m/s with the blades wide open, which is well under 1 psi of pressure drop. These are my estimates and haven't been measured on a car.

Estimate

How much horsepower can the twin 80s feed?

This chart shows the estimated pressure lost across the two stock 80 mm throttle bodies at wide-open throttle, plotted against crank horsepower. That drop happens ahead of the blower, so every 1 psi lost means about 7% less air for the blower to compress at the same pulley. The shaded band covers a range of assumptions about blade and shaft blockage and how well the airflow recovers after the throttle. Hover or tap the chart to see numbers at any power level.

My estimate, not a measurement. Airflow is assumed at ≈1.33 cfm per crank hp, from the Roush's ~750 whp ≈ 900 crank hp ≈ 1,200 cfm. Drop = loss factor × ½ρv² across the 10,053 mm² bore, with 10–20% blockage from the blade and shaft and a loss factor of 0.6–1.5. The “power cost” is the air lost before the blower at a fixed pulley. You can win some of it back with more pulley, but that adds heat.
Crank hp≈ whpAir cfmDrop psiVerdict

Up to about 1,000 crank hp (≈ 840 whp) the drop is only a few tenths of a psi, which is lost in dyno-to-dyno noise. That covers everything a Roush or Magnuson TVS 2650 can make.

From 1,000 to 1,400 crank hp they still work, but they start costing roughly 3–9% of the air to the blower, which you'd have to make up with more pulley.

Beyond about 1,400 crank hp (≈ 1,200 whp) I'd plan on a single 120 mm or larger throttle body. By then the blower, engine, fuel system and driveline would all have been replaced anyway.

Where the inlet was the limit

The older S550 Roush 2650 used one smaller throttle body

The S550 Roush kit fed the same TVS 2650 rotors through a single GT350-pattern throttle body. On Mustang6G, owners describe that inlet as the choke point that caps later mods. The bolt pattern also made it hard to fit a bigger throttle body. Those cars still made about 728 whp on E85. Owners who ported the inlet and fit a bigger throttle body report more, including one unverified claim of 960 whp.

The S650 kit reuses the twin 80s, which gives about 70% more area than that single 87 mm. That largely addresses the bottleneck the S550 owners ran into.

My read

What runs out first on a Roush / Magnuson S650

  1. Rotor packTVS 2650

    The rotor pack is much smaller than the Whipple 3.0's. To make the same boost it has to spin faster and heat the air more, and efficiency falls off above about 5,500–6,000 rpm. The dyno curves in the test fit that.

  2. Intercoolerbricks ~½ the Whipple's

    The bricks are about half the size of the Whipple's, so intake heat builds up faster as boost goes up. The Roush heat exchanger is decent, about the size of the Dark Horse HO's.

  3. Pulley changespractical limit

    The throttle bodies have to come off to change the pulley, so Roush cars rarely get turned up. Shops don't stock the pulleys and don't bother.

  4. Twin 80 mm throttlesprobably 900+ whp

    These probably only matter at roughly 900+ whp. Whether they do depends on how smoothly the Roush inlet feeds the rotors. That hasn't been measured publicly.

On my Dark Horse, software is the limit first

The stock Ford Transmission Axle Torque Limit table in Roush's Dark Horse calibration holds the throttle to about 34° in 1st and 68° in 2nd. It only reaches the full 85° from 3rd gear up. Until that table changes, throttle-body flow doesn't matter in the gears that count. See the throttle-ceiling analysis.

Settling it

How to measure it on a car

  1. Tap a pressure sensor or vacuum gauge between the throttle bodies and the blower inlet. I don't know of a stock PID for that spot on the Roush setup.
  2. Make a 3rd-gear (or higher) WOT pull to redline, so the throttle is fully open at 85°.
  3. Compare that reading with barometric pressure. If the drop stays under about 1 psi at redline, the throttle bodies aren't costing measurable power. A drop that keeps growing with rpm would be the sign that the inlet is choking.
  4. Log at a real sample rate. FORScan Lite's ~1 Hz round-robin polling is too slow to resolve this cleanly.
References

Sources