41 replies to this topic

[racersteven]

I was reading about the 1994 F-Nippon Lola and how it was able to generate up to 5.2 G in fast corners. How dose this compare to both current and past F-1 cars ? I would guess a 2009 F-1 would generate 4.0 g but I am not sure.

[skuty]

I was reading about the 1994 F-Nippon Lola and how it was able to generate up to 5.2 G in fast corners. How dose this compare to both current and past F-1 cars ? I would guess a 2009 F-1 would generate 4.0 g but I am not sure.

Yes, nowdays F1 can generate 4.2G, but are you sure with 5.2G? I don't say it's impossible, it's only exciting for car.

[racersteven]

Yes, nowdays F1 can generate 4.2G, but are you sure with 5.2G? I don't say it's impossible, it's only exciting for car.

I found it in a 1994 Racecar engineering mag . HHF was testing the new Lola F-nippon car at Suzuka and his neck was hurting afterwards. Later they found out he was pulling peak 5.2 g in quick corners.

[Ross Stonefeld]

The mid 90s Formula Nippon cars were very impressive, and I think one of the few four-wheeled series running qualifying tires?

[Ben]

I've spoken to a number of engineer who have (subjectively at least) confirmed that the early 90s F Nippon cars on Bridgestone Q rubber were pulling more G than the F1 cars at Suzuka and doing similar overall lap times despite less power.

Ben

[Ogami musashi]

Current F1 cars pull more G's in corners than the Three previous year.

The turn 1 at suzuka is taken at 5g's and the average G's in corner which was around 3g's previously is 4g's now.

That said, i think either 2004 or 2006 cars took more G's, i remember reading 6g's at 130R in suzuka.

[DaveW]

FNippon = F3000 on "stickies", I believe (perhaps with aero enhancements).

I don't wish to cast doubt on published data, but beware lateral acceleration readings when tyres start to stick/slip. Readings can vary by at least +/- 1 gn, depending on the location of the accelerometer. The phenomenon almost certainly caused the unsightly (& unrealistic) rapid rolling "events" observable during the 7 post rig test shown here.

Edited by DaveW, 19 December 2009 - 14:10.

[McGuire]

FNippon = F3000 on "stickies", I believe (perhaps with aero enhancements).

I don't wish to cast doubt on published data, but beware lateral acceleration readings when tyres start to stick/slip. Readings can vary by at least +/- 1 gn, depending on the location of the accelerometer.

Quite so, perhaps only a spike, and not necessarily in lateral acceleration, perhaps only an artifact. Tap the accelerometer with a screwdriver handle and it will make a big number, and with the car bounding around at the limits of slip/stick as you say, who knows. g is a pretty coarse value anyway, 0 to 5 units or so representing the entire range. Let's see lat acc plotted through a corner. I doubt if the F Nippon/F3000 can truly out-corner an F1 car but I could be wrong. Perhaps in a certain tight-radius corner with the perfect entry speed. It's just a number, isn't it?

[shaun979]

G doesn't have to be coarse because it has decimals. The range of speeds across a given arc, or turn radii for a given speed, vary significantly enough in order to bring about fairly small changes in lat. acc. . It would take quite an idiot to quote a lat. G number off the very peak of a spike. Usually magnitude over duration is mentioned, and after filtering. Highest lateral in a downforce car will usually be at large radius turns where the aero has most effect. RDV has in the past mentioned late 80s F1 cars generating over 6 G lateral for seconds at a time.

[DaveW]

G doesn't have to be coarse because it has decimals.

Quite correct.

RDV has in the past mentioned late 80s F1 cars generating over 6 G lateral for seconds at a time.

Mmm... Sanity check required. 6 gn at 50 mph (22.3 m/sec) is a turn with a yaw rate of 150.9 deg/sec. At 150 mph (67 m/sec), the yaw rate reduces to 50.3 deg/sec, if my calcs are correct. I can't think of many real world corners that would require 6 gn lateral for "seconds at a time", apologies.

In my experience (1987), lateral acceleration in the late '80s was seldom greater than 3.5 gn, unless corners were banked.

Edited by DaveW, 19 December 2009 - 21:46.

[shaun979]

Mmm... Sanity check required. 6 gn at 50 mph (22.3 m/sec) is a turn with a yaw rate of 150.9 deg/sec. At 150 mph (67 m/sec), the yaw rate reduces to 50.3 deg/sec, if my calcs are correct. I can't think of many real world corners that would require 6 gn lateral for "seconds at a time", apologies.

In my experience (1987), lateral acceleration in the late '80s was seldom greater than 3.5 gn, unless corners were banked.

I can't comment personally since I don't have any experience with cars of that era, but here's located what I recalled.. http://forums.autosp...p;#entry1489632

"F1's at end of 80's and early 90's would have peaks of lateral g @ 6.5 , notably at Paul Ricard double right hamder, would pull over 6g for more than 4 seconds. At time use of "g"-suits mooted as under braking there was concern blood would pool in legs thus causing driver black-out.
Today's seating position with legs raised eases this concern, and we dont have long high speed corners where aero forces can help generate high g's. Other cause is the restriction on rear difusers, at first race after their restriction , at Barcelona, formula 3000 front runners running full difusers and very standard Avons were also posting times that would make them midfield runners in F1.
This didn't last long as relentless wind tunnel work soon gave results and F1's were again faster, even back of grid. "

I sure hope he'll be along soon to discuss it with you, as this is all interesting.

[racersteven]

Here is some information I found in a 1998 issue of racetech " The 1998 F-Nippon car is a very nimble device that pulls impressive g-forces. I think a F-nippon is quicker through a corner than a F-car reckons Ben bowlby the cars designer. It has more downforce and and relatively sticky tires. We used to see peaks of 5 G on qualifiers in 1994. Now on spec tires were are seeing 4 G in the corners and under braking. " This was comparing a 1998 F-Nippon to a 1998 F-1 car which was first year of grooved tires and narrow track so they were not as quick in the corners.

[Fat Boy]

It would take quite an idiot to quote a lat. G number off the very peak of a spike.

I always find it humorous when you find karting 'experts' that claim they corner at 3.5 g's. They're obviously reading the tops of the peaks.

[DaveW]

I can't comment personally since I don't have any experience with cars of that era, but here's located what I recalled.. http://forums.autosp...p;#entry1489632

"F1's at end of 80's and early 90's would have peaks of lateral g @ 6.5 , notably at Paul Ricard double right hamder, would pull over 6g for more than 4 seconds. At time use of "g"-suits mooted as under braking there was concern blood would pool in legs thus causing driver black-out.
Today's seating position with legs raised eases this concern, and we dont have long high speed corners where aero forces can help generate high g's. Other cause is the restriction on rear difusers, at first race after their restriction , at Barcelona, formula 3000 front runners running full difusers and very standard Avons were also posting times that would make them midfield runners in F1.
This didn't last long as relentless wind tunnel work soon gave results and F1's were again faster, even back of grid. "

I sure hope he'll be along soon to discuss it with you, as this is all interesting.

I imagine that the first part of the "double right hander" would be Signes. Google maps (satellite view) suggests that the driving radius of the current version of Signes is around 200m. If it is assumed that the driving line hasn't changed significantly over the years, a car would be travelling at around 243 mph (108 m/sec) through Signes in order to realize 6 gn lateral acceleration. On the other hand, 3.5 gn lateral would be sustained at 185 mph (83 m/sec) - always assuming my arithmetic is correct. I leave it to you to decide which is the more likely speed (hence lateral acceleration).

[McGuire]

G doesn't have to be coarse because it has decimals.

Which decimals people invariably round up. Zzzz.

It would take quite an idiot to quote a lat. G number off the very peak of a spike. Usually magnitude over duration is mentioned, and after filtering.

And yet it's done all the time, so there must be a number of idiots out there. I don't see any "magnitude over duration" in this lot of numbers, do you? As I said, let's see lat acc plotted through a corner and then we have something to talk about. "F Nippon cars corner at ___ g at Suzuka" (fill in number here) isn't good enough for me. If it's good enough for you that's fine with me.

[Michael Ferner]

Back in the early nineties, I still had enough brain cells and energy to try to work it out mathematically. I don't recall the exact figures, but I clearly remember that my findings of average lateral acceleration were much lower as the "g forces pulling on drivers" that were then often quoted in discussions. For example, I found that on a relative "simple" circuit such as the Indianapolis Motor Speedway ("simple" as in constant radius corners) the quoted g forces would have meant a cornering speed much higher than the average lap speed! I'm pretty sure that most of those g values are peak measurements.

[Greg Locock]

Incidentally, if we integrate up longitudinal velocity and yaw velocity over a lap of our twisty and turny handling track, the error in car position is less than 1m. If I take the time history of vehicle speed and steering wheel angle and drive my model with it, the error in car position is less than 1m.

http://maps.google.c...e...mp;t=k&z=16

That's in the middle of the oval, covered by cloud.

So properly measured and reported latacc ( or yaw velocity) is useful, and as accurate as you need.

[DaveW]

Incidentally, if we integrate up longitudinal velocity and yaw velocity over a lap of our twisty and turny handling track, the error in car position is less than 1m. If I take the time history of vehicle speed and steering wheel angle and drive my model with it, the error in car position is less than 1m.

http://maps.google.c...e...mp;t=k&z=16

That's in the middle of the oval, covered by cloud.

So properly measured and reported latacc ( or yaw velocity) is useful, and as accurate as you need.

Good stuff, Greg, but requires three provisos, I think. First that you can measure ground speed - normally have to assume that axle-based wheel speed is proportional to ground speed (i.e. slip ratio is negligible). Second that vehicle sideslip is negligible (although local deviations should cancel over the lap). Third, the "latacc" transducer is installed at the vehicle centre of gravity (the only location for which lateral acceleration per unit ground speed = yaw rate, given the second proviso).

In reality, "latacc" signals will usually be contaminated by other acceleration components, even if the device is not attacked by McGuire's proverbial screwdriver handle (the latter can actually be difficult to achieve in a tightly-packaged race vehicle, since the device normally has to be "soft-mounted" to filter out power train vibrations). Incidentally, I have measured 20 gn rms at the cockpit floor of a race vehicle when running the engine with the vehicle on its wheels, but static.

Apologies, the above is a turgid way of saying that "latacc" is one of the more problematic race vehicle measurements...

Edited by DaveW, 21 December 2009 - 08:20.

[gruntguru]

Good stuff, Greg, but requires three provisos, I think. First that you can measure ground speed - normally have to assume that axle-based wheel speed is proportional to ground speed (i.e. slip ratio is negligible). Second that vehicle sideslip is negligible (although local deviations should cancel over the lap). Third, the "latacc" transducer is installed at the vehicle centre of gravity (the only location for which lateral acceleration per unit ground speed = yaw rate, given the second proviso).

In reality, "latacc" signals will usually be contaminated by other acceleration components, even if the device is not attacked by McGuire's proverbial screwdriver handle (the latter can actually be difficult to achieve in a tightly-packaged race vehicle, since the device normally has to be "soft-mounted" to filter out power train vibrations). Incidentally, I have measured 20 gn rms at the cockpit floor of a race vehicle when running the engine with the vehicle on its wheels, but static.

Apologies, the above is a turgid way of saying that "latacc" is one of the more problematic race vehicle measurements...

Would it be less problematic to measure yaw directly with a gyro or similar, then infer latacc from that? I suppose the sensor cost would be considerably greater.

[DaveW]

Would it be less problematic to measure yaw directly with a gyro or similar, then infer latacc from that? I suppose the sensor cost would be considerably greater.

When sideslip is negligibly small, (lateral acceleration of the centre of gravity) = (yaw rate)*(ground speed), given consistent units. The relationship is likely to fall down at (or close to) the lateral limit. Robust rate sensors exist (& are useful anyway). Accelerometer measurements can be OK, provided they are interpreted correctly. Specifically, it is unwise to believe peak values.....

[Greg Locock]

Good stuff, Greg, but requires three provisos, I think. First that you can measure ground speed - normally have to assume that axle-based wheel speed is proportional to ground speed (i.e. slip ratio is negligible). Second that vehicle sideslip is negligible (although local deviations should cancel over the lap). Third, the "latacc" transducer is installed at the vehicle centre of gravity (the only location for which lateral acceleration per unit ground speed = yaw rate, given the second proviso).

In reality, "latacc" signals will usually be contaminated by other acceleration components, even if the device is not attacked by McGuire's proverbial screwdriver handle (the latter can actually be difficult to achieve in a tightly-packaged race vehicle, since the device normally has to be "soft-mounted" to filter out power train vibrations). Incidentally, I have measured 20 gn rms at the cockpit floor of a race vehicle when running the engine with the vehicle on its wheels, but static.

Apologies, the above is a turgid way of saying that "latacc" is one of the more problematic race vehicle measurements...

We measure groundspeed and yaw velocity directly, and when I do the path integration those are the two sensors I use, lat acc is inferred from them. ay is reported at each axle and the cg, and the main instrument pack is actually quite close to the cg anyway (it does have a lateral accelerometer). So what I should do is plot ay as measured against the calculated value. Good point about sideslip, again we measure lat velocity continuously I should integrate that up to see what the total sideslip distance is per lap. Thinking about it my calling it position error is incorrect, it just means you start and finsih in the same place, the whole 'scale' of the circuit could easily be out by a couple of metres.

[ivanalesi]

Some time ago I watched a recap of 1990's Mexican GP on BBC and they said Ferrari claimed Prost was experiencing over 5G in the last corners.
May be the 1992 Williams with the huge diffuser, wide body and active suspension was picking such numbers, but in 1990 it looked strange.

[Kalmake]

Is this how you calculate it?

Assuming Peraltada driving line has radius of 150m
r = 150m
g = 9.81m/s²
v = sqrt(5g*r) = 86m/s = 309km/h = 192mph

[Fat Boy]

Is this how you calculate it?

Assuming Peraltada driving line has radius of 150m

I don't know where you came up with that number, but it's damned close. In the center they pinch the radius down a little tighter, but as an average, that's a good number.

[DaveW]

Is this how you calculate it?

Assuming Peraltada driving line has radius of 150m
r = 150m
g = 9.81m/s²
v = sqrt(5g*r) = 86m/s = 309km/h = 192mph

The calculation looks good. What about the speed (entry & exit)?

[Ross Stonefeld]

Back then it was banked, so you'd get some big G's and mph. Otherwise 192mph seems too high for a corner, rather than a gently curved straight.

[desmo]

How will banking affect the lateral accelerations measured by accelerometers both hard mounted onboard and thus tilted by the banking along with the car and one that remains parallel to the horizon? A steep enough banking (~90 deg from horiz) should result in 0 lat G for a vehicle mounted instrument, no?

[gruntguru]

How will banking affect the lateral accelerations measured by accelerometers both hard mounted onboard and thus tilted by the banking along with the car and one that remains parallel to the horizon? A steep enough banking (~90 deg from horiz) should result in 0 lat G for a vehicle mounted instrument, no?

Yes. The lateral G meter is (neglecting body roll) measuring the acceleration parallel to the road surface. If the car is stationary on a banked corner the accelerometer will register a lateral acceleration away from the apex. At increasing speeds in the same corner the acceleration will register more and more towards the apex.

The G force claims in this thread would be the true horizontal (v squared/r) acceleration which the driver would experience as a combination of side-force and additional "weight" pushing him into the seat.

[shaun979]

Back in the early nineties, I still had enough brain cells and energy to try to work it out mathematically. I don't recall the exact figures, but I clearly remember that my findings of average lateral acceleration were much lower as the "g forces pulling on drivers" that were then often quoted in discussions. For example, I found that on a relative "simple" circuit such as the Indianapolis Motor Speedway ("simple" as in constant radius corners) the quoted g forces would have meant a cornering speed much higher than the average lap speed! I'm pretty sure that most of those g values are peak measurements.

Maybe they were referring to sum or vectorial sum acceleration, using vertical and lateral? At IMS the banking can mean around 1.9 G vertical (1.0 at rest in pits) 3.7 G lateral

[gruntguru]

Maybe they were referring to sum or vectorial sum acceleration, using vertical and lateral? At IMS the banking can mean around 1.9 G vertical (1.0 at rest in pits) 3.7 G lateral

1.9 plus 3.7 still only adds to 4.16 G

[DaveW]

The G force claims in this thread would be the true horizontal (v squared/r) acceleration which the driver would experience as a combination of side-force and additional "weight" pushing him into the seat.

With respect, the "particle" acceleration (v squared/r) calcs were introduced as a "sanity check" to test claims made in this thread. A properly located, isolated & aligned lateral accelerometer will read less than the "particle" acceleration when a corner is banked, of course.

The presence of banking in a corner has a powerful effect on cornering speeds, because it allows a speed increment to restore the reduced tyre side loads and a further speed increment made available by the increased vertical tyre loads. The effect can get out of hand sometimes, as in the aborted Champ/Texas race, when the resulting normal (rather than lateral) accelerations caused some drivers to "grey out", I believe. Texas continues to be notable for spectacular IRL "late race" accidents, probably for similar reasons, despite mandated "airbrakes".

Anyway, reported lateral accelerations (presumably measured) that exceed the calculated "particle" values must, I think, be of doubtful validity, regardless of the corner bank angle.

Edited by DaveW, 23 December 2009 - 10:06.

[McGuire]

Banking is only a simple trick of geometry: the erstwhile lateral force is redistributed as normal force, Fn = m(g lat*sin a + cos a) for the vertical component relative to the car. For example, with the banking at the full 90 degrees, ala M-B Stuttgart test track or ye motorcycle wall of death, 100 percent of the apparent lateral force is translated. The road surface is now under the force instead of perpendicular to it -- in the tire's grip plane rather than its slip plane.

Or to put it another way, to whatever extent a turn is banked between 1 and 90 degrees, the vehicle is not really "turning." Relative to the plane of the road surface, the vehicle is traveling in a straight line. Mr. Mobius goes for a ride:

[McGuire]

How will banking affect the lateral accelerations measured by accelerometers both hard mounted onboard and thus tilted by the banking along with the car and one that remains parallel to the horizon? A steep enough banking (~90 deg from horiz) should result in 0 lat G for a vehicle mounted instrument, no?

Yep. Your standard accelerometer is just a tiny mass on a capacitance grid -- a little man strapped to the bed of a trampoline, if you will, whose displacement we measure. With a two-axis accelerometer (horz/vert) the two elements are not particularly aware of the plane in which they are installed. (They think their world is bi-planar.) As they are reoriented (tilted relative to the plane of the earth, fixed relative to the car) their outputs will skew away from horz and toward vert, naturally. Theoretically, the vector sum of the two outputs should add up to the total of the earth-relative lat acc, but the sensors start to lose accuracy at some point off-axis.

[shaun979]

1.9 plus 3.7 still only adds to 4.16 G

Yes and in Michael's story, he doesn't mentioned what numbers were being mentioned, only that they were higher than he calculated. Obviously it is not in contention for quickest car in corner, but I'm just trying to account for the different in what he heard vs what he calculated.

[ivanalesi]

An easier way is if anyone has some comparison between the cars from different eras. What I mentioned about Mexico, I wouldn't buy much into it, it was according to Murray and you can't know what this Ferrari engineer was talking about.

[DaveW]

I believe the calculations done in this thread don't factor in:

-The radius is wider than what you measure from the air. As the drivers go in wide, touch the inner radius at the apex, and go wide again. So you'd have to know the width of the track.
-The car is already doing a speed of x on entry.
-The corners could be banked
-The surface and grip level is not constant but a jumpy combination.

I don't disagree with your caveats. However, The purpose of my original contribution to this thread was to comment on the sanity of conclusions drawn from published claims of measured lateral acceleration. Specifically, a comment was made in Post #11 about lateral acceleration through Signes at Circuit Paul Ricard, a spectacularly fast corner that is not obviously banked. I think you will find that the driving line through the current version of the corner is quite visible on Google maps (together with Bernie's fantasy paint job). To save you the trouble of searching, here is a screenshot extracted from Satellite View. You could probably convince yourself that the driving line shows some evidence of surface inputs early in the corner &, possibly, demonstrates that some vehicles porpoise late in the corner (that, if true, would suggest an exit speed of around 270 kph, as a maximum). Anyway, the screenshot should allow you to check my very rough calcs of post #14.

Edited by DaveW, 21 January 2010 - 21:10.

[JohnParker]

I've looked all over Google with hardly any information on the 1994 F-Nippon Lola let alone anything about it doing 5.2g's around a corner.
 

Edited by JohnParker, 24 November 2023 - 16:08.

[shaun979]

I imagine that the first part of the "double right hander" would be Signes. Google maps (satellite view) suggests that the driving radius of the current version of Signes is around 200m. If it is assumed that the driving line hasn't changed significantly over the years, a car would be travelling at around 243 mph (108 m/sec) through Signes in order to realize 6 gn lateral acceleration. On the other hand, 3.5 gn lateral would be sustained at 185 mph (83 m/sec) - always assuming my arithmetic is correct. I leave it to you to decide which is the more likely speed (hence lateral acceleration).

Dave, I agree with your rough calculations. I looked at the circuit and the max. driven radii and having run the numbers myself only recently, can't see how ~6 G for seconds would be possible. I'm hoping RDV will comment because I can't see what I'm missing.

[racersteven]

I've looked all over Google and can't find hardly any information on the 1994 F-Nippon Lola let alone anything about it doing 5.2g's around a corner. Where did you find this info?

I found the info in a 1994 issue of Racecar engineering vol 3 no6 and I know that Race tech magazine had some articles about Formula Nippon AS WELL.

[Tony Matthews]

OT: Lol, I just received 1994 vol 3 no 5 yesterday in the post.

Their subscription service is even worse than Motor Sport!

[bonneville]

I've looked all over Google and can't find hardly any information on the 1994 F-Nippon Lola let alone anything about it doing 5.2g's around a corner. Where did you find this info?

You should be looking for All Japan F3000 1994 or Japanese F3000 1994. The F-Nippon moniker only applied from 1996 onwards. Before that, the championship had other denominations:

1973-1977 All-Japan Formula 2000
1978-1986 Japanese Formula Two
1987-1995 Japanese Formula 3000
1996- 2010 Formula Nippon

So here is an onboard lap with Takuya Kurosawa's Team Cerumo 1994 Lola Mugen Honda at Suzuka:

[gruntguru]

So here is an onboard lap with Takuya Kurosawa's Team Cerumo 1994 Lola Mugen Honda at Suzuka:

The biggest acceleration number I saw on the telemetry was 3.5G.