60-130 MPH: New performance measurement!
#931
Answering Jean, how did I get these "rule of thumb numbers".
You run a correlation of about 60 sportscars taking tested curb weight, power and 100-200 kmh time. That derives a formula of weight - hp - speed. By changing one parameter you can then forecast the other. For example, you forecast the 100-200 kph time (given power and weight) compare to actual and measure the difference.
How good is the forecasting? Generally, within a 6% error (actual to forecast) , although some cars, such as the 993TT are "spot on" i.e. zero error.
What's important is not the accuracy of the forecast per se, but that you can get some reasonable rules about what a reduction in weight or an increase in power makes to the 100-200 times.
As for your monster, assuming weight (including driver) of 1,450 kgs (svelte fellow) and allowing 0.8 secs total shifting improvement for the sequential, the formula indicates around 580 - 600 hp. Am I far off the mark?
As to Colin's car, about 420 to 460 hp.
You run a correlation of about 60 sportscars taking tested curb weight, power and 100-200 kmh time. That derives a formula of weight - hp - speed. By changing one parameter you can then forecast the other. For example, you forecast the 100-200 kph time (given power and weight) compare to actual and measure the difference.
How good is the forecasting? Generally, within a 6% error (actual to forecast) , although some cars, such as the 993TT are "spot on" i.e. zero error.
What's important is not the accuracy of the forecast per se, but that you can get some reasonable rules about what a reduction in weight or an increase in power makes to the 100-200 times.
As for your monster, assuming weight (including driver) of 1,450 kgs (svelte fellow) and allowing 0.8 secs total shifting improvement for the sequential, the formula indicates around 580 - 600 hp. Am I far off the mark?
As to Colin's car, about 420 to 460 hp.
#932
Addict
Lifetime Rennlist
Member
Lifetime Rennlist
Member
Alex
Yes this is the work we shared together, we have slightly different approaches but the results are a bit close.
My car is 1410Kgs dry. I am not as svelte as you migth think, more like 85Kgs I estimate it at slightly above 600HP during my runs.
As far as Colin's car (assuming we are talking about the 930) more like 400HP according to my calculations, see before.
Yes this is the work we shared together, we have slightly different approaches but the results are a bit close.
My car is 1410Kgs dry. I am not as svelte as you migth think, more like 85Kgs I estimate it at slightly above 600HP during my runs.
As far as Colin's car (assuming we are talking about the 930) more like 400HP according to my calculations, see before.
#933
Jean, at 1,495 kgs, formula indicates 606 hp, say 590 - 620. Not that bad, actually, in terms of forecasting.
But to explain a little further. What the fomula allows you to do is forecast with some degree of accuracy what changes in the weight or power ON THE SAME car will do to your speeds. So, for example, if you want to reduce your times by one full sec, assuming the 0.8 allowance for the sequential and without going on a diet, you'll need another 90 hp. Why does this differ from the 45hp mentioned earlier? Because for the car to get faster at these levels, the required increase in power is obviously much greater.
But to explain a little further. What the fomula allows you to do is forecast with some degree of accuracy what changes in the weight or power ON THE SAME car will do to your speeds. So, for example, if you want to reduce your times by one full sec, assuming the 0.8 allowance for the sequential and without going on a diet, you'll need another 90 hp. Why does this differ from the 45hp mentioned earlier? Because for the car to get faster at these levels, the required increase in power is obviously much greater.
#934
Addict
Lifetime Rennlist
Member
Lifetime Rennlist
Member
Alex
Directionally the numbers make good sense. Leaving aside the 265/315s and worse aero drag of any Porsche out there (excluding 935Racer!), my 2 shifts made me gain 0.4 seconds overall over a fast stock gearbox (or slightly less compared to DKnebes) rather than 0.8s.
Cheers
Directionally the numbers make good sense. Leaving aside the 265/315s and worse aero drag of any Porsche out there (excluding 935Racer!), my 2 shifts made me gain 0.4 seconds overall over a fast stock gearbox (or slightly less compared to DKnebes) rather than 0.8s.
Cheers
#935
Agreed. What's important though, is that the aero and tires don't actually matter. They'd matter only if you wanted to forecast an accurate 100-200 kmh time, given weight and hp.
But here we are asking something different: The car, as is, needs 5.6 seconds (lets adjust to 6 seconds due to sequential) with approx 620 hp. How much more power do we need to be one second faster, same car, same driver same tires, same everything. And the answer is about 90hp.
Which is precicely the guestion one should ask one's tuner: My car makes x time, I want it to make y time, therefore I need an extra z power. Can you do it? And, if upon re-delivery it doesn't make the y time, then obviously the tuner hasn't achieved the power he had promised.
But here we are asking something different: The car, as is, needs 5.6 seconds (lets adjust to 6 seconds due to sequential) with approx 620 hp. How much more power do we need to be one second faster, same car, same driver same tires, same everything. And the answer is about 90hp.
Which is precicely the guestion one should ask one's tuner: My car makes x time, I want it to make y time, therefore I need an extra z power. Can you do it? And, if upon re-delivery it doesn't make the y time, then obviously the tuner hasn't achieved the power he had promised.
#937
No. Required power increases exponentially with speed. For example:
Take standard 993 TT, 1585 kgs weight (1,505 curb weight plus 80 kgs driver), 408 hp. Model forecasts 100-200 kph at 10.8 secs, which is in line with actual performance.
13.8 secs would need 327 hp
12.8 350 hp
11.8 377 hp
10.8 408 hp (actual)
9.8 445 hp
8.8 489 hp
7.8 543 hp
As you can see, one extra second faster needs just 23 additional hp at the begining, but 54 hp at the end of the sequence. That's why Jean's car needs so much more power for just one extra second.
Interestingly, you can achieve the same result, i.e. gain one second in speed (10.8 to 9.8 secs) power remaining unchanged, by reducing the car's weight from 1,505 kgs to 1,360 kgs. Incidentally, for the same result (6 secs to 5 secs) Jean would need to shave off 250 kgs.
Take standard 993 TT, 1585 kgs weight (1,505 curb weight plus 80 kgs driver), 408 hp. Model forecasts 100-200 kph at 10.8 secs, which is in line with actual performance.
13.8 secs would need 327 hp
12.8 350 hp
11.8 377 hp
10.8 408 hp (actual)
9.8 445 hp
8.8 489 hp
7.8 543 hp
As you can see, one extra second faster needs just 23 additional hp at the begining, but 54 hp at the end of the sequence. That's why Jean's car needs so much more power for just one extra second.
Interestingly, you can achieve the same result, i.e. gain one second in speed (10.8 to 9.8 secs) power remaining unchanged, by reducing the car's weight from 1,505 kgs to 1,360 kgs. Incidentally, for the same result (6 secs to 5 secs) Jean would need to shave off 250 kgs.
#940
Addict
Rennlist
Site Sponsor
Rennlist
Site Sponsor
Ah, I turn off the laptop for the day to give myself time to design & draw a new airbox and look what happens! It looks like this thread is finally going to achieve the aim that was predicted in its title: i.e. by measuring the 60-130mph (or 100-200kmph) time of a vehicle with known weight and bodystyle, usign simple corrections we can make a prediction of the actual engine performance experienced whilst the car is being driven at that speed. At last!
However, I think we are all missing out on one vital factor. What we are actually predicting is not the true peak hp of the engine, but the average hp of the engine over the rpm range used through the gears for the purpose of the test. Thus engines with a high peak figure but poor area under the curve will be slower than engines with a lower peak figure but with a much wider area under the curve....
... at which point I am comfortable with the majority view from all the estimates above that you need an average power of around 440-450hp to accelerate a 1450kg 965 bodied 930 from 60-130mph in 10.18s.
Now, if we can all work together and collate all our information, it should now be possible to deduce a very accurate mathematical or empirical model that will predict average power from acceleration time. I think JamesE probably has the mathematic and programming skills to build a simple model, Alex and Jean obviously have most of the data and I would be willing to host the predictor on the 9m website. Is everyone up for this or do we all like our snake oil too much?
However, I think we are all missing out on one vital factor. What we are actually predicting is not the true peak hp of the engine, but the average hp of the engine over the rpm range used through the gears for the purpose of the test. Thus engines with a high peak figure but poor area under the curve will be slower than engines with a lower peak figure but with a much wider area under the curve....
... at which point I am comfortable with the majority view from all the estimates above that you need an average power of around 440-450hp to accelerate a 1450kg 965 bodied 930 from 60-130mph in 10.18s.
Now, if we can all work together and collate all our information, it should now be possible to deduce a very accurate mathematical or empirical model that will predict average power from acceleration time. I think JamesE probably has the mathematic and programming skills to build a simple model, Alex and Jean obviously have most of the data and I would be willing to host the predictor on the 9m website. Is everyone up for this or do we all like our snake oil too much?
#941
Addict
Lifetime Rennlist
Member
Lifetime Rennlist
Member
Originally Posted by NineMeister
However, I think we are all missing out on one vital factor. What we are actually predicting is not the true peak hp of the engine, but the average hp of the engine over the rpm range used through the gears for the purpose of the test.
This way a 300 HP car can accelerate as fast as a 400 HP car from 60 to 125 mph
Konstantin
#942
Addict
Rennlist
Site Sponsor
Rennlist
Site Sponsor
Konstantin, we definitely agree on this. In fact most of the development work we do at 9m (on both atmo and turbo engines) concentrates on flattening off torque curves to give a broad, useable power band over as wide an rpm range as possible.
Does everyone else now agree with the fundamental principle that 60-130/100-200 can only be a measure of engine hp under the curve - unless anyone is driving with a Constantly Variable Transmission that we don't know about?
Does everyone else now agree with the fundamental principle that 60-130/100-200 can only be a measure of engine hp under the curve - unless anyone is driving with a Constantly Variable Transmission that we don't know about?
#943
Definitely. In fact it goes a long way to explain the differences one sees in various cars, whose performance - on paper - should be identical, but in actuality are not. For example:
Ferrari F360 Challenge Strad: 425 hp, weight 1,387 kgs, 100-200 in 11.1 secs
Porsche 997 GT3 RS: 415 hp, weight 1,428 kgs, 100-200 in 9.3 secs
You need to ask, how it is that two virtually identical cars (similar weight and power) can be so far apart in terms of speed. The answer includes gearing and aero, but as importantly the power curve. Interestingly, the weaker and heavier car here, is nearly two seconds faster...
Another, even more dramatic comparison:
AM Vanguish S 2005: 528 hp, 1,900 kgs, 100-200 in 12.3 secs
BMW M6 SMG: 507 hp, 1,761 kgs, 100-200 in 8.6 secs
Two generally similar cars, 3.7 full secs apart. Something can't be right in the AM camp.
Ferrari F360 Challenge Strad: 425 hp, weight 1,387 kgs, 100-200 in 11.1 secs
Porsche 997 GT3 RS: 415 hp, weight 1,428 kgs, 100-200 in 9.3 secs
You need to ask, how it is that two virtually identical cars (similar weight and power) can be so far apart in terms of speed. The answer includes gearing and aero, but as importantly the power curve. Interestingly, the weaker and heavier car here, is nearly two seconds faster...
Another, even more dramatic comparison:
AM Vanguish S 2005: 528 hp, 1,900 kgs, 100-200 in 12.3 secs
BMW M6 SMG: 507 hp, 1,761 kgs, 100-200 in 8.6 secs
Two generally similar cars, 3.7 full secs apart. Something can't be right in the AM camp.
Last edited by AVoyvoda; 12-07-2006 at 05:33 AM.
#944
Addict
Rennlist Member
Rennlist Member
Colin
Jean has been doing just what you suggest for a long time now - it is called average longitudinal G for the duration of the 60-130mph run.
Some examples:
732hp 1300kg 930 2WD - av long G = 0.5528
408PS 1500kg 993tt 4WD = 0.258
541PS RS Tuning 993tt 1450kg 2WD = 0.4096
The average long G numbers correlate exactly with how long the car will take to do the run - this is the real measure of the torque over the period of the run.
Here lies the problem with the Driftbox as its long Gs differ from the AX22's - I have tried them back to back and there are serious differences - The Driftbox calculates its Gs from GPS (I think) whereas the AX uses accelerometers - Jean's data shows that the AX22 is very reliable in this measurement.
Jean has been doing just what you suggest for a long time now - it is called average longitudinal G for the duration of the 60-130mph run.
Some examples:
732hp 1300kg 930 2WD - av long G = 0.5528
408PS 1500kg 993tt 4WD = 0.258
541PS RS Tuning 993tt 1450kg 2WD = 0.4096
The average long G numbers correlate exactly with how long the car will take to do the run - this is the real measure of the torque over the period of the run.
Here lies the problem with the Driftbox as its long Gs differ from the AX22's - I have tried them back to back and there are serious differences - The Driftbox calculates its Gs from GPS (I think) whereas the AX uses accelerometers - Jean's data shows that the AX22 is very reliable in this measurement.
#945
Toby, here is some correspondence from VBOX USA concerning some issues on the other site. BTW... what avg G would I need to push my car thru the 60-130 with 1 shift in 7.93secs with a full up weight of 3760lbs....let me know what you think. This is the email from Jim Laus at Vbox Usa:
Kevin,
let me answer your questions individually:
How does this unit derive its speed over time calculations?
Is the speed over time strictly satellites or a combination of satellites and accelerometers ?
DriftBox's speed is measured, not calculated, using doppler shift of the carrier frequency from as many GPS satellites as it can see. It is using the concept of triangulation, but with an average six to ten GPS satellites as reference points, measures a very accurate speed, within 0.1 km/h of actual vehicle speed. The accelerometers have literally no effect on the speed, straight-line performance, and lap timing results from the DriftBox. This is the same speed measurement method used in all of our professional level VBOX devices, and has been proven time and again against radar guns, optical and microwave sensors, and fifth-wheel instruments. One caveat to this is an old firmware bug in DriftBox which I'll explain later on.
The requirement is that the antenna can see at least four GPS satellites for accurate speed measurement in three axes. If the GPS signal is lost or interfered with (by driving under a bridge or between close, tall buildings for example), then this can cause temporary jumps in speed and position information, and can sometimes affect distance-based tests like 1/4 mile runs.
Are accelerometers used in the actual performance calculations or only for drift and lat and long G's?
The accelerometers are not even used in G-force or drift information. G-forces are calculated from GPS, which inherently makes them corrected for body pitch and roll, and provides more repeatable results than accelerometers. The calculated lateral acceleration is used in calculating the drift "score" in Drift Mode, just to avoid someone driving at 70MPH from rotating the DriftBox in their hand and having that create a high score. The accelerometers are only present in the DriftBox to align the yaw rate sensor, which is used in calculating vehicle slip angle.
Can this unit be manipulated to provide user favorable data?
How accurate are the accelerometers compared to a Race Technologies AX22?
How does the levelling and alignment affect performance calculations?
If the unit is willfully tilted during an acceleration run, how would that affect the data?
The only way I can foresee someone manipulating results with the DriftBox would be to move the antenna (if using an external) or the DriftBox itself. All of the unit's measurements and calculations are based on GPS, so if the antenna were to be moved faster than the vehicle somehow, that is what the DriftBox would report. The only parameter that would be affected by changing the way the unit is mounted (assuming here that it has a clear view of the sky and a good number of satellites) is slip angle, since the yaw information is based the unit being flat. Even with changing the mounting, the unit would only report less slip angle being off-axis, so there is no benefit to someone trying to skew data.
As long as the DriftBox can see the sky, the mounting (or any angle at which it is mounted) has no effect at all on the performance numbers.
I have many data files in front of me from other users that show times that are impossible given the vehicles weight and power. These times have been looked over by telemetry experts in the Porsche SuperCup series and all agree that they are inaccurate.
I have a theory about what is happening here. The first few versions of DriftBox firmware had a bug where if the unit saw more than ten GPS satellites, this would overflow the bandwidth between the GPS engine and firmware, and cause samples to be dropped from calculations. This would manifest itself by a tenth of a second less time per run or lap for every sample dropped. If this has happened, there will be dropouts in the data file, which can be repaired using the latest version of DriftBoxTools (or VBoxTools). The new "repair" feature is designed to address these dropouts, as well as those you might see from driving under bridges and the like. To check if a user's file has dropouts, load the file and click the "Summary" button in DriftBoxTools and it will list the dropouts in the file, and where they occur. To fix the problem for the future, make sure that people are keeping their DriftBox firmware up to date. The bug was fixed in firmware 1.01 build 22, and the current version is 2.0 build 17. Firmware 2.0 also adds several new features, including a horsepower estimator.
It would make sense if you never saw issues with yours if you keep your own DriftBox up to date better than others, or just have a newer unit that came with the newer firmware from us.
If there is a track or dragstrip where a DriftBox cannot reliably see satellites, then data gathered there is also unreliable. Using an external antenna can help the DriftBox lock onto satellites, having a bigger view of the sky than the internal antenna inside a vehicle, but I have successfully used DriftBoxes without external antennas in a great many cars with reliable and accurate results.
I hope this answers your questions, and feel free to ask more if it doesn't. Also, if you have specific data files that look suspicious to you, please send them to me and I'll have a look and see what we're dealing with.
Thanks,
-jim
Jim Lau | Customer Support Manager
________________________________________
VBOX USA
1368 Anderson Rd
Clawson, MI 48017
248-655-0557 | office
248-953-3251 | mobile
248-246-1797 | fax
jiml@vboxusa.com
http://vboxusa.com
From: kpgtdg1@aol.com [mailto:kpgtdg1@aol.com]
Sent: Friday, November 24, 2006 10:35 PM
To: craig@vboxusa.com
Subject:
Craig, I purchased a DriftBox from you several months ago and find myself questioning the accuracy of the unit. I have 40-50 passes at the 1/4 mile and many track days with the local Porsche club.You provided me with a copy of VBOX tools since the graphing and datalogging I was looking to do was not possible with Dbox tools. I have received data from other Porsche owners who use Dbox and the results are mixed at best. Can you answer several questions or point me to an engineer at Racelogic who could best answer some technical questions. How does this unit derive its speed over time calculations?Is the speed over time strictly satellites or a combination of satellites and accelerometers ?Are accelerometers used in the actual performance calculations or only for drift and lat and long G's? Can this unit be manipulated to provide user favorable data?How accurate are the accelerometers compared to a Race Technologies AX22? How does the levelling and alignment affect performance calculations? If the unit is willfully tilted during an acceleration run, how would that affect the data? I have many data files in front of me from other users that show times that are impossible given the vehicles weight and power. These times have been looked over by telemetry experts and all agree that they are inaccurate. All the data my own unit has produced has been shown to be accurate, so there is a disconnect between the data I receive from my own unit and data sent to me from others. Unfortunately I believe this tool is being used to justify inaccurate performance claims. Any help would be greatly appreciated. If you cannot answer these questions could you pass this email on to someone who will. Thanks. Kevin
=
Kevin,
let me answer your questions individually:
How does this unit derive its speed over time calculations?
Is the speed over time strictly satellites or a combination of satellites and accelerometers ?
DriftBox's speed is measured, not calculated, using doppler shift of the carrier frequency from as many GPS satellites as it can see. It is using the concept of triangulation, but with an average six to ten GPS satellites as reference points, measures a very accurate speed, within 0.1 km/h of actual vehicle speed. The accelerometers have literally no effect on the speed, straight-line performance, and lap timing results from the DriftBox. This is the same speed measurement method used in all of our professional level VBOX devices, and has been proven time and again against radar guns, optical and microwave sensors, and fifth-wheel instruments. One caveat to this is an old firmware bug in DriftBox which I'll explain later on.
The requirement is that the antenna can see at least four GPS satellites for accurate speed measurement in three axes. If the GPS signal is lost or interfered with (by driving under a bridge or between close, tall buildings for example), then this can cause temporary jumps in speed and position information, and can sometimes affect distance-based tests like 1/4 mile runs.
Are accelerometers used in the actual performance calculations or only for drift and lat and long G's?
The accelerometers are not even used in G-force or drift information. G-forces are calculated from GPS, which inherently makes them corrected for body pitch and roll, and provides more repeatable results than accelerometers. The calculated lateral acceleration is used in calculating the drift "score" in Drift Mode, just to avoid someone driving at 70MPH from rotating the DriftBox in their hand and having that create a high score. The accelerometers are only present in the DriftBox to align the yaw rate sensor, which is used in calculating vehicle slip angle.
Can this unit be manipulated to provide user favorable data?
How accurate are the accelerometers compared to a Race Technologies AX22?
How does the levelling and alignment affect performance calculations?
If the unit is willfully tilted during an acceleration run, how would that affect the data?
The only way I can foresee someone manipulating results with the DriftBox would be to move the antenna (if using an external) or the DriftBox itself. All of the unit's measurements and calculations are based on GPS, so if the antenna were to be moved faster than the vehicle somehow, that is what the DriftBox would report. The only parameter that would be affected by changing the way the unit is mounted (assuming here that it has a clear view of the sky and a good number of satellites) is slip angle, since the yaw information is based the unit being flat. Even with changing the mounting, the unit would only report less slip angle being off-axis, so there is no benefit to someone trying to skew data.
As long as the DriftBox can see the sky, the mounting (or any angle at which it is mounted) has no effect at all on the performance numbers.
I have many data files in front of me from other users that show times that are impossible given the vehicles weight and power. These times have been looked over by telemetry experts in the Porsche SuperCup series and all agree that they are inaccurate.
I have a theory about what is happening here. The first few versions of DriftBox firmware had a bug where if the unit saw more than ten GPS satellites, this would overflow the bandwidth between the GPS engine and firmware, and cause samples to be dropped from calculations. This would manifest itself by a tenth of a second less time per run or lap for every sample dropped. If this has happened, there will be dropouts in the data file, which can be repaired using the latest version of DriftBoxTools (or VBoxTools). The new "repair" feature is designed to address these dropouts, as well as those you might see from driving under bridges and the like. To check if a user's file has dropouts, load the file and click the "Summary" button in DriftBoxTools and it will list the dropouts in the file, and where they occur. To fix the problem for the future, make sure that people are keeping their DriftBox firmware up to date. The bug was fixed in firmware 1.01 build 22, and the current version is 2.0 build 17. Firmware 2.0 also adds several new features, including a horsepower estimator.
It would make sense if you never saw issues with yours if you keep your own DriftBox up to date better than others, or just have a newer unit that came with the newer firmware from us.
If there is a track or dragstrip where a DriftBox cannot reliably see satellites, then data gathered there is also unreliable. Using an external antenna can help the DriftBox lock onto satellites, having a bigger view of the sky than the internal antenna inside a vehicle, but I have successfully used DriftBoxes without external antennas in a great many cars with reliable and accurate results.
I hope this answers your questions, and feel free to ask more if it doesn't. Also, if you have specific data files that look suspicious to you, please send them to me and I'll have a look and see what we're dealing with.
Thanks,
-jim
Jim Lau | Customer Support Manager
________________________________________
VBOX USA
1368 Anderson Rd
Clawson, MI 48017
248-655-0557 | office
248-953-3251 | mobile
248-246-1797 | fax
jiml@vboxusa.com
http://vboxusa.com
From: kpgtdg1@aol.com [mailto:kpgtdg1@aol.com]
Sent: Friday, November 24, 2006 10:35 PM
To: craig@vboxusa.com
Subject:
Craig, I purchased a DriftBox from you several months ago and find myself questioning the accuracy of the unit. I have 40-50 passes at the 1/4 mile and many track days with the local Porsche club.You provided me with a copy of VBOX tools since the graphing and datalogging I was looking to do was not possible with Dbox tools. I have received data from other Porsche owners who use Dbox and the results are mixed at best. Can you answer several questions or point me to an engineer at Racelogic who could best answer some technical questions. How does this unit derive its speed over time calculations?Is the speed over time strictly satellites or a combination of satellites and accelerometers ?Are accelerometers used in the actual performance calculations or only for drift and lat and long G's? Can this unit be manipulated to provide user favorable data?How accurate are the accelerometers compared to a Race Technologies AX22? How does the levelling and alignment affect performance calculations? If the unit is willfully tilted during an acceleration run, how would that affect the data? I have many data files in front of me from other users that show times that are impossible given the vehicles weight and power. These times have been looked over by telemetry experts and all agree that they are inaccurate. All the data my own unit has produced has been shown to be accurate, so there is a disconnect between the data I receive from my own unit and data sent to me from others. Unfortunately I believe this tool is being used to justify inaccurate performance claims. Any help would be greatly appreciated. If you cannot answer these questions could you pass this email on to someone who will. Thanks. Kevin
=