F1technical.net

Could using a pulsating KERS charging schedule during acceleration improve traction when the tires are at the limit?

Leave out the word pulsating and I say YES.

autogyro wrote:

hardingfv32 wrote:Could using a pulsating KERS charging schedule during acceleration improve traction when the tires are at the limit? It allows them to re-grip after they start sliding. I am thinking about the MotoGP 'Big Bang' operating mode where the power output is not evenly distributed during a single rpm. The tires are given an opportunity to re-grip each power cycle.
This might make it easier to drive closer on the traction limit.
Brian

Very good comparison, I am well aware of the 'big bang' theory of traction control.
Consulted with Honda for a while.
Of course the Norton Rotary did not suffer from the problem so the Japanese banned it.
A pity because the Norton Rotary would still be much faster than the current Japanese racing bikes.

a (premeditated) pusating KERS charging schedule I would call a torque ripple action
I have also tried to highlight the inherent KERS/ERS characteristic of wheelspin-specific EM torque variation
(this could be seen as another kind of pulsating charging)
such torque ripple must be related to the mechanical characteristics of the whole transmission train from motor to the track
the ripple frequency must be suitably low so being matched to and working with these characteristics
surely this is what was proven by footage of tyre chatter marks as they were generated in Canada ?

torque ripple can of course be achieved mechanically (and is in F1 ?), although they try to eliminate it from our road vehicles
though 'Big Bang' is to me still unevidenced and a myth
(the Blair paper that Brian kindly sent me is about power curve effects of various induction intervals, Blair distances himself from 'BB')
the supposed 'BB' torque variation is at far too high a frequency (at eg 12000 rpm) to propagate via the tyre/s to the track
the torsional or surge frequency of the vehicle/tyre system is maybe 5-15 Hz (people with strong friends can try this at home)
the tyre acts as a low pass filter (straddling a low frequency resonant peak)
behaving like our car engines on their mountings, this has an isolation effect at the 'BB' frequencies

btw
the 'BB' story appeared as an explanation of the success of the Doohan/Honda combination
but the losers were also 'BB' (more or less), but that was conveniently ignored
(4 cyl 2 stroke GP machines had for decades been pair-firing square 4s or pseudo V4s, to reduce ignition complexity and bulk)
so the Doohan era Hondas were only slightly more uneven, options anyway driven by their package of increased V angle
'BB' stories have reappeared regarding the recent MotoGP Yamahas (the only non V engines, they have crossplane cranks)
crossplane cranks give evenly-spaced inertial forces, at the expense of unevenly-spaced firing forces (opposite to all our experience)
at any reasonably high rpm (even those in our road vehicles) the inertia forces (ie reciprocation) are much bigger than firing forces
(and twice as frequent)
the crossplane crank will be slimmer with much less bearing friction (lower rubbing speeds)
handy in a fuel-limited race
this or other cranks that give uneven firing intervals could have fewer main bearings

(Johnny Giles made a simultaneous-firing version of his (Triumph) employers scrambles/motocross twin, they threatened to fire him)
the Yankee (Ossa ?) off-road twin was also made simultaneous-firing
thousands of 'English trials' were won (maybe on torque ripple) with single-cylinder motors run at 300-400 rpm, by hand control of spark timing

has anyone ever seen a paper supporting 'BB' ?

autogyro wrote:I'm not sure I follow your math on the 3 changes per tenth, if it can help though, here another pic of the same throttle trace I already posted, for Vettel's Singapore pole lap, with widened time scale:

http://i.imgur.com/6j3Xn6e.jpg

OK about 5 very even modulations every half second.
These modulations on the graph are showing very small throttle movements, not the much larger movements shown for an accepted 'manual' driver actuated throttle used without TL.

Drivers do not have such a fine and rapid response to breaks in traction as anyone who is a decent racing driver is well aware.
When traction is lost the foot goes off throttle rapidly and then on slowly in a controlled way and not in such a small time scale.

I think that scale along the bottom is seconds, not tenths of seconds, as evidenced at the bottom; time[s.]. That matches well with my data, although I do not log throttle position. Here's three seconds from a session last year. It's RPM over time. The dip on the left is the downshift (manual with no heel-toe). The apex is at 33.6 seconds and as I pick up the throttle, I get wheelspin (a peak in RPM), which is countered and grip restored by 33.8 seconds. That's less than 0.2 seconds to modulate my right foot to reign in the wheelspin.

OK TC lets call it torque ripple action.

I agree that the BB theory is not a proven concept.
I tend to agree with you that other torque saving aspects of these engines was far more important.

A rotary engine with a planetary gear set is far better for a racing MC IMO.

I agree Andy, it is possible for a 'good' driver to collect traction in the short time shown on your graph.
I do not believe it is possible to continually collect traction every tenth of a second all the way up to 150 kph.
Not unless you have an ankle capable of moving faster than a fiddlers elbow.

Tommy Cookers wrote: snip
btw
the 'BB' story appeared as an explanation of the success of the Doohan/Honda combination
snip

I'm sure I remember much the Doohan success being explained as him being able to keep the more powerful screamer
on the road, where most of the other had to use the less powerful but more controllable big bang

autogyro wrote:

Reca wrote: I'm not sure I follow your math on the 3 changes per tenth, if it can help though, here another pic of the same throttle trace I already posted, for Vettel's Singapore pole lap, with widened time scale:

http://i.imgur.com/6j3Xn6e.jpg

OK about 5 very even modulations every half second.
These modulations on the graph are showing very small throttle movements, not the much larger movements shown for an accepted 'manual' driver actuated throttle used without TL.

Drivers do not have such a fine and rapid response to breaks in traction as anyone who is a decent racing driver is well aware.
When traction is lost the foot goes off throttle rapidly and then on slowly in a controlled way and not in such a small time scale.

These small steps you keep taking as ultra quick modulations are just "artificial" consequence of the limited resolution in throttle movement detection, these don't represent the real movements of the pedal, I just didn't bother smoothing as I thought that was evident.

Program can't read every tiny movement but only relatively large variations when the pixels on the throttle bar change color (hence it's already limited by resolution of the bar but then also by image resolution), so it happens, especially when the rate of variation is relatively low, that it keeps reading the same constant value for various consecutive frames (25 per second) and does a "step" to the next higher value it can recognize only once in a while.
It's obviously an approximation, not how the pedal actually moves, a linear movement with reduced slope appears like a number of small steps between constant values.

If anything, the fact there are in certain parts several of these steps, that you interpret as the driver moving the pedal ultra fast, indicate actually the opposite, he's quite likely moving it slowly and smoothly, just the program can't accurately follow for lack of resolution.

As I said, you shouldn't try to read in the data more than they can show due to obvious limits of sampling rate and resolution. Better than nothing, but far from the complete picture.

flynfrog wrote:

autogyro wrote:You do not need a feedback loop or wheel spin sensors if you have sufficient data to program the circuit throttle requirements.
You just need two or more torque input sources into the power train balanced against one another.
The FIA does the rest for you.

a throttle map is not Active TC. Its no different than making a none linear linkage setup just a more modern high tech way of doing it.

Indeed.

From the previous Toyota graph for example I captured the traces of pedal and throttle for the first corners and here it is:


As you can see the 2008 relationship is quite simple, you could get an already rather good approximation of that curve with two lines, a limited slope till 60% pedal and a steeper one above.
That relationship certainly helps driver in better controlling the throttle % because large movement in the first half of pedal travel corresponds to reduced opening, but it's nothing particularly complicate.
The same graph shows how real TC works, driver just goes pedal to the metal and electronics does the rest transforming his input in anything between 30% and 100% throttle.

The real trickery in term of smoothing of engine power delivery is done by the engine torque demand maps, and these would work just the same way even if instead of fly by wire there was an old fashion cable.

autogyro wrote:

fasterthanyou wrote:

autogyro wrote:You do not need a feedback loop or wheel spin sensors if you have sufficient data to program the circuit throttle requirements.
You just need two or more torque input sources into the power train balanced against one another.
The FIA does the rest for you.

The problem is circuit condition may change over the race period. Without the feedback loop, you won't be able to take full advantage of the grip.

Depends on how the M/G is balanced to the IC output.
Doing so does not need software.
It would be rpm modulated using conventional electronics.
Probably built into relay.

From memory, various writings are pointing to the possibility of RBR monitoring lateral G-loadings vs suspension movement.

A guess is when the "system" sees a change in linear acceleration loads and specific suspension movements which follow a predetermined pattern (such as the rear suspension unloading in one direction vs loads in another vs a drop in straight line acceleration) which indicates the start of wheels spin the KERS kicks in.

As for KERS only acting at 100% throttle, I hadn't heard this before and would certainly be interesting

Can you explain fully how the data for your graphs was acquired Reca.

An electrical connection between the Kers M/G and a power relay has no need to connect to a throttle map or any software.

Based on the torque demand curve limit, could they vary the ignition timing, at every interval, to vary torque output's rate of increase as RPMS climb?

autogyro wrote:Can you explain fully how the data for your graphs was acquired Reca.

An electrical connection between the Kers M/G and a power relay has no need to connect to a throttle map or any software.

From a few pages ago.

Reca wrote:You're welcome auto.

About the throttle traces, be careful not reading too much in it; as I said that is captured from the FOM graphics, I rapidly made a little program that reads the video frame by frame and picks the color of various points on the throttle bar; in general it works quite well, but it's possible that every now and then the tolerance in color reading is a bit wrong and some values are misinterpreted.
On top of that then it remains to be seen also how accurate the FOM graphics is obviously.
So take it as a general idea, for that it's good enough, but for precise evaluation of every single point, I wouldn't put money on it.

Thanks Flyn and sorry Reca.
Your explanation was great.

So you measure the colour variations of the FOM throttle read out from the video over time during the lap.

Do you also measure the engine sound pitch over time to give you the engine rpm?

One thing is certain, the FOM read out for the Kers harvesting has never been accurate and they have never tried to make it so, if it was accurate we could then see when and where the M/G was under load harvesting through the lap.

The KERS graphic doesn't show harvesting because it is only showing the KERS allowance left for that lap. It simply shows when the driver is pushing the button.

richard_leeds wrote:The KERS graphic doesn't show harvesting because it is only showing the KERS allowance left for that lap. It simply shows when the driver is pushing the button.

Indeed it does Richard and it bounces back up to full every lap.
Why is it done like that?
IMO the harvesting is the most important aspect of the Kers concept to show F1 saving energy.
Why disguise the system as a 'boost' button.

autogyro wrote:

richard_leeds wrote:The KERS graphic doesn't show harvesting because it is only showing the KERS allowance left for that lap. It simply shows when the driver is pushing the button.

Indeed it does Richard and it bounces back up to full every lap.
Why is it done like that?
IMO the harvesting is the most important aspect of the Kers concept to show F1 saving energy.
Why disguise the system as a 'boost' button.

Because the KERS graphic is there to show how much boost is left, to convey harvesting information and boost left in a single graphic would probably be somewhat difficult. Then one has to wonder what harvesting graph would show, it would show increase charge under breaking and decrease when they push the button, but I doubt it will show any more detail than that in a single graphic, and such a simple concept is conveyed quite often in the F1 broadcasts.