Honda Power Unit Hardware & Software

[j.yank]

accelerating whilst K-driving would give a weak TC-like effect but accelerating while K-generating would give anti-TC
and braking while K-generating would give a weak ABS-like effect

Can you elaborate more on why there would be anti-TC while K-generating? If the pedal is fully pressed out of the corner K-generating will act as a contra-torque supposedly preventing overspinning. Or maybe I have overlooked the thechnical regulations where this is not allowed?

[etusch]

Before mgu-k they couldn't use full engine power with nas engines until 140 km/h. Now there is more power at lower rpm with higher torq. According to what I read about Honda Engine they are using mgu-k to generation at least until that speed. It is some kind of TC contrl Engine power and store some of unneeded engine power to use later.

You mean to use MGU-K in generation mode during accelartion in order to simulate something like pasive traction control?

TC cuts engine power right?
Mgu-k generates electric during accelerating and it absorbs some of engine power which will cause wheelspin until aerodinamics structure of car takes affect on the car. I think main function of this for more e-power but in the same time some kind of TC. It is something like you can send 4mj to battery directly from mgu-k but unlimited through mgu-h. So the rule do not interested mgu-k production but how much sent directly to storage. Rule bans TC but not bans (I think so) mgu-k usage during acceleration.

These are what I derive from what I read. So any correction is welcome where I am wrong.

[subcritical71]

Before mgu-k they couldn't use full engine power with nas engines until 140 km/h. Now there is more power at lower rpm with higher torq. According to what I read about Honda Engine they are using mgu-k to generation at least until that speed. It is some kind of TC contrl Engine power and store some of unneeded engine power to use later.

You mean to use MGU-K in generation mode during accelartion in order to simulate something like pasive traction control?

TC cuts engine power right?
Mgu-k generates electric during accelerating and it absorbs some of engine power which will cause wheelspin until aerodinamics structure of car takes affect on the car. I think main function of this for more e-power but in the same time some kind of TC. It is something like you can send 4mj to battery directly from mgu-k but unlimited through mgu-h. So the rule do not interested mgu-k production but how much sent directly to storage. Rule bans TC but not bans (I think so) mgu-k usage during acceleration.

These are what I derive from what I read. So any correction is welcome where I am wrong.

The key here, and it is what prevents TC, is that the pedal is mapped to torque output. The minimum slope is defined in the rules and is always positive. The ICE and K power outputs must equal pedal position map within 50ms. If the driver pressed the accelerator to 100% at say 50km/h the tires would loose traction.
This all changes once the pedal is at 100%. The torque controller can then do what it wants to manage the energy situation. I could see a brief time period where the driver is still traction limited and by going to 100% the torque controller can effectively implement TC. This is just a theory of mine, but it seems well within the rules.

[Tommy Cookers]

accelerating whilst K-driving would give a weak TC-like effect but accelerating while K-generating would give anti-TC
and braking while K-generating would give a weak ABS-like effect

Can you elaborate more on why there would be anti-TC while K-generating? If the pedal is fully pressed out of the corner K-generating will act as a contra-torque supposedly preventing overspinning. Or maybe I have overlooked the thechnical regulations where this is not allowed?

maybe I was wrong
imo
there will be inner control (feedback) loops between the K and its controller
this must include what's effectively rpm feedback - ie the K is primarily velocity-controlled
an excessive rate of change of rpm ('slew limit') will cause the K to go from full + torque to full - torque (or vice-versa)
in a few millisec
eg wheelspin is one possible cause of this triggering of limit slew rate
accelerating with K assisting ICE - on wheelspin the K would go from +120 kW to -120 kW in in a few millisec (TC emulation)
accelerating with ICE driving K generation - on wheelspin the K would remain at -120 kW (no TC effect)

these are what could be regarded as control-implemented torque 'collapses'
there can in the absence of such control be other torque collapses related to MG design factors

[j.yank]

You mean to use MGU-K in generation mode during accelartion in order to simulate something like pasive traction control?

TC cuts engine power right?
Mgu-k generates electric during accelerating and it absorbs some of engine power which will cause wheelspin until aerodinamics structure of car takes affect on the car. I think main function of this for more e-power but in the same time some kind of TC. It is something like you can send 4mj to battery directly from mgu-k but unlimited through mgu-h. So the rule do not interested mgu-k production but how much sent directly to storage. Rule bans TC but not bans (I think so) mgu-k usage during acceleration.

These are what I derive from what I read. So any correction is welcome where I am wrong.

The key here, and it is what prevents TC, is that the pedal is mapped to torque output. The minimum slope is defined in the rules and is always positive. The ICE and K power outputs must equal pedal position map within 50ms. If the driver pressed the accelerator to 100% at say 50km/h the tires would loose traction.
This all changes once the pedal is at 100%. The torque controller can then do what it wants to manage the energy situation. I could see a brief time period where the driver is still traction limited and by going to 100% the torque controller can effectively implement TC. This is just a theory of mine, but it seems well within the rules.

Well, I will greatly appreciate if you point the rules where this is defined - I really want to sort out this for myself but cannot spent much time for investigation of regulations. Saying this, may I speculate that ICE+K output could be always positive but actually the K output is slightly negative, and in this way suppressing the ICE excessive positive (in correlation to pedal) output? Maybe I don't know the rules but it seems to me that this is an option that seems contra intuitive (to use K as stop when the car needs full throttle), and in the same time maybe useful. Maybe the contra loads from ICE and K could explain the Red Bul troubles with the driveshaft?

[henry]

You mean to use MGU-K in generation mode during accelartion in order to simulate something like pasive traction control?

TC cuts engine power right?
Mgu-k generates electric during accelerating and it absorbs some of engine power which will cause wheelspin until aerodinamics structure of car takes affect on the car. I think main function of this for more e-power but in the same time some kind of TC. It is something like you can send 4mj to battery directly from mgu-k but unlimited through mgu-h. So the rule do not interested mgu-k production but how much sent directly to storage. Rule bans TC but not bans (I think so) mgu-k usage during acceleration.

These are what I derive from what I read. So any correction is welcome where I am wrong.

The key here, and it is what prevents TC, is that the pedal is mapped to torque output. The minimum slope is defined in the rules and is always positive. The ICE and K power outputs must equal pedal position map within 50ms. If the driver pressed the accelerator to 100% at say 50km/h the tires would loose traction.
This all changes once the pedal is at 100%. The torque controller can then do what it wants to manage the energy situation. I could see a brief time period where the driver is still traction limited and by going to 100% the torque controller can effectively implement TC. This is just a theory of mine, but it seems well within the rules.

That’s an interesting theory. At 100% demand the power output can vary by 120kW either way. So if the driver demands 100% too early then as you say it is possible the ECU could vary power to implement TC. However to get there the driver needs to go through lower demands where this assistance potentially comes into play.

The 50ms requirement represents about 2 metres at the traction limit speed in a straight line. So potentially if two cars are implemented with differing delays the car with the longer delay could get onto the throttle a metre or two earlier. At this point they are accelerating at around 2G so a metre earlier might be worthwhile.

Another unknown in this is the torque measurement. I assume it is a noisy signal which needs to be filtered. I would expect the filtering to add a delay which would eat into the 50ms. But it might also disguise the torque fluctuations that @Tommy Cookers discusses in his post.

[subcritical71]

Well, I will greatly appreciate if you point the rules where this is defined - I really want to sort out this for myself but cannot spent much time for investigation of regulations. Saying this, may I speculate that ICE+K output could be always positive but actually the K output is slightly negative, and in this way suppressing the ICE excessive positive (in correlation to pedal) output? Maybe I don't know the rules but it seems to me that this is an option that seems contra intuitive (to use K as stop when the car needs full throttle), and in the same time maybe useful. Maybe the contra loads from ICE and K could explain the Red Bul troubles with the driveshaft?

Well, its not completely defined in the regulations. The minimum ramp rate and pedal to torque ratio is defined in the regulations. I agree with you though, as long as the torque demanded by the pedal position is achieved, its up to the strategy how much comes from ICE and/or +/- MGU-K.

5.5 Power unit torque demand :
5.5.1 The only means by which the driver may control acceleration torque to the driven wheels is via
a single foot (accelerator) pedal mounted inside the survival cell.
5.5.2 Designs which allow specific points along the accelerator pedal travel range to be identified by the driver or assist him to hold a position are not permitted.
5.5.3 At any given engine speed the driver torque demand map must be monotonically increasing for an increase in accelerator pedal position.
5.5.4 At any given accelerator pedal position and above 4,000rpm, the driver torque demand map must not have a gradient of less than – (minus) 0.045Nm/rpm.

The part about how the ICE & K total power bit is an interpretation from a research paper done at the Zurich ZTH (link: https://www.research-collection.ethz.ch ... sAllowed=y). While it is a research paper, the author acknowledges help from Ferrari and has interpretations of the rules as he knows them (again, he had help from Ferrari so I weight his interpretations on the informed side). Look up the second co-examiners name on LinkedIn and you will see something interesting (link: https://www.linkedin.com/in/carlo-bussi ... bdomain=it) - somehow he is only twice removed from me .

If you don't have a lot of time, read the introduction of the research paper starting on page 1.

[j.yank]

Well, I will greatly appreciate if you point the rules where this is defined - I really want to sort out this for myself but cannot spent much time for investigation of regulations. Saying this, may I speculate that ICE+K output could be always positive but actually the K output is slightly negative, and in this way suppressing the ICE excessive positive (in correlation to pedal) output? Maybe I don't know the rules but it seems to me that this is an option that seems contra intuitive (to use K as stop when the car needs full throttle), and in the same time maybe useful. Maybe the contra loads from ICE and K could explain the Red Bul troubles with the driveshaft?

Well, its not completely defined in the regulations. The minimum ramp rate and pedal to torque ratio is defined in the regulations. I agree with you though, as long as the torque demanded by the pedal position is achieved, its up to the strategy how much comes from ICE and/or +/- MGU-K.

5.5 Power unit torque demand :
5.5.1 The only means by which the driver may control acceleration torque to the driven wheels is via
a single foot (accelerator) pedal mounted inside the survival cell.
5.5.2 Designs which allow specific points along the accelerator pedal travel range to be identified by the driver or assist him to hold a position are not permitted.
5.5.3 At any given engine speed the driver torque demand map must be monotonically increasing for an increase in accelerator pedal position.
5.5.4 At any given accelerator pedal position and above 4,000rpm, the driver torque demand map must not have a gradient of less than – (minus) 0.045Nm/rpm.

The part about how the ICE & K total power bit is an interpretation from a research paper done at the Zurich ZTH (link: https://www.research-collection.ethz.ch ... sAllowed=y). While it is a research paper, the author acknowledges help from Ferrari and has interpretations of the rules as he knows them (again, he had help from Ferrari so I weight his interpretations on the informed side). Look up the second co-examiners name on LinkedIn and you will see something interesting (link: https://www.linkedin.com/in/carlo-bussi ... bdomain=it) - somehow he is only twice removed from me .

If you don't have a lot of time, read the introduction of the research paper starting on page 1.

Thank you very much! There are two "if"s on page 2 that are relevant: "If the driver is not requesting full power, the control system must deliver the amount of power requested by the driver, and the energy management system can
only decide how to split it between the ICE and the MGU-K. Conversely, if the driver is requesting maximum power, the regulations allow the energy management system to decide the amount of propulsive power to provide and the split between the actuators."

The first case is trickier to use MGU-K as TC, but the second one is interesting. What if on exit of the corner the driver requests maximum power but the K is still in generating mode, in this way subtracting from the ICE power, and after this the driver lift off the pedal for a while but in the same time, the K goes in driving mode, in this way compensating the diminishing power from ICE? Also, when the pedal is on 100% "regulations allow the energy management system to decide the amount of propulsive power to provide" - but the same system according to the scheme, receives information from the weels, and it can decide how much of the requested full power to come from ICE and how much it to be suppressed by the K in generating mode. Maybe this is discussed in the paper but right now, I don't have time to read it to the end.

[henry]

accelerating whilst K-driving would give a weak TC-like effect but accelerating while K-generating would give anti-TC
and braking while K-generating would give a weak ABS-like effect

Can you elaborate more on why there would be anti-TC while K-generating? If the pedal is fully pressed out of the corner K-generating will act as a contra-torque supposedly preventing overspinning. Or maybe I have overlooked the thechnical regulations where this is not allowed?

maybe I was wrong
imo
there will be inner control (feedback) loops between the K and its controller
this must include what's effectively rpm feedback - ie the K is primarily velocity-controlled
an excessive rate of change of rpm ('slew limit') will cause the K to go from full + torque to full - torque (or vice-versa)
in a few millisec
eg wheelspin is one possible cause of this triggering of limit slew rate
accelerating with K assisting ICE - on wheelspin the K would go from +120 kW to -120 kW in in a few millisec (TC emulation)
accelerating with ICE driving K generation - on wheelspin the K would remain at -120 kW (no TC effect)

these are what could be regarded as control-implemented torque 'collapses'
there can in the absence of such control be other torque collapses related to MG design factors

I think what might be effected is wheelspin minimisation rather than traction control. The objective would be to reduce wear rather than to maximise traction.

I did a quick calculation and the slew rates in the low speed, traction limited, region are high. Perhaps 3,000rpm/s for the ICE and 10,000 for the MGU-K. So for the few milliseconds you mention that’s tens of revs. change.

Do you think they might monitor slew rate change in the MGU-K control trying to manage “jerk” which would be the onset of wheelspin?

[Brake Horse Power]

Is it really that fast to regulate on almost millisecond level? Do you even have a measuring accuracy so high that this makes sense?

-measuring
-signal to computer
-computing
-signal to motorcontroller
-adjusting power output to motor
All in few ms? Just asking.. I am no expert.

[Bandit1216]

Is it really that fast to regulate on almost millisecond level? Do you even have a measuring accuracy so high that this makes sense?

-measuring
-signal to computer
-computing
-signal to motorcontroller
-adjusting power output to motor
All in few ms? Just asking.. I am no expert.

I don't think speed is the problem. The photon's travel 299 792,458 meter in one ms, right?

[subcritical71]

Is it really that fast to regulate on almost millisecond level? Do you even have a measuring accuracy so high that this makes sense?

-measuring
-signal to computer
-computing
-signal to motorcontroller
-adjusting power output to motor
All in few ms? Just asking.. I am no expert.

I don't think millisecond level is a problem at all with modern electronics. For the TAG-320 ECU, data acquisition is in the 10K parameters per second range and the ECU can handle 4000 MIPS (millions of instructions per second) or 4 Billion instructions per second. I must admit though that only having 50ms between driver input and PU torque delivery as measured by the crankshaft sensor seems a bit fast to me but they seem make it work .

[subcritical71]

I was curious, so I went googling and found this article on the development of the TAG-320 from 2012

https://www.rbracing-rsr.com/downloads/ ... ag_320.pdf

The ecu is at the heart of an F1 car, providing complete control over the powertrain. It works with slave units to control ignition and injection for a V8 engine, as well as controlling throttle by wire and clutch by wire functionality. In order to do this, it takes data from something like 130 sensors around the car. Accuracy is crucial; F1 engines run at up to 18,000rpm and the control strategy can recognize crankshaft positions to an accuracy of 0.027deg.

The ecu's main control strategy runs on a 1ms time step.

I would assume the ignition control unit would be faster than this .

Additionally it was mentioned gearbox strategy is timed to 0.1ms

[Brake Horse Power]

@bandit, the speed of an electronic signal is not the problem, but a sensor digitizes a measurement, which takes time. A computer needs to decode it, do something useful with it and send to the next station. That takes some time.

So the wheelslip will be measures at each rear wheel right? Using the crank signal would have a far greater accuracy because of the reduction in between but the diff messes up this reading to derive the exact wheelspin from it. You can debate how accurate the system should be since the diff might have some will of its own. (unless you have diff magic).

So if you want to have a reading for each wheel, the measurement should be something like this:
Wheel diameter, 0,67m (at 0km/h!) = 2,1038m circumference.
At for example 100km/h, vehicle speed = 27,77m/s. = 0,0277m/ms. = 4.7 degrees wheelturning per millisecond. If you measure per degree of rotation you will get 4 pulses per millisecond. At 200km/h its half of that. The measurement rate to see when the puls is should be something like 10kHz at least. This corresponds to the 0.1 ms strategy as indicated in previous post.

So in theory this might be possible, BUT you will measure every bit of deflection, tiny losses of grip etc. I wonder if controlling an mgu-k to this level is practically possible. Somewhere you will need kind of 'low' pass filter to not be reacting to every miniscule event.

[gruntguru]

So in theory this might be possible, BUT you will measure every bit of deflection, tiny losses of grip etc. I wonder if controlling an mgu-k to this level is practically possible. Somewhere you will need kind of 'low' pass filter to not be reacting to every miniscule event.

I doubt the response of the MGUK would be fast enough to react to the "minuscule events" you refer to. The inductance of motor windings effectively inhibits rapid changes in current.