Throttle control

Discussion in 'General' started by hobbit, May 24, 2006.

  1. hobbit

    hobbit He who posts articles

    Disclaimer: this is primarily Prius-oriented, since that's what I'm
    most familiar with these days. But after watching a little driving
    by some of the esteemed colleagues hereabouts at Tour de Sol,
    I've been having more thoughts about ideal running conditions
    and how to achieve those with the Prius HSD. 14 months after
    taking ownership, and I'm *still* trying to solve the highway-mileage
    Anyways, my chatting with various Insight drivers [including a
    local friend who's had one for a while but doesn't do TdS] seems
    to have consensus that low-RPM lugging is the highest efficiency
    mode that engine runs in. Lean-burn aside, I suspect that this
    is true for many engines, especially given the techniques I read
    about for other vehicles here. Unfortunately to get that in
    a Prius, some compromises have to be made in how the HSD spins
    its components, and I fear that the situation is never quite
    like throwing a step gearbox into fifth and going for it.
    One thing I notice, having a vacuum gauge in my car, is how the
    Prius controls its throttle at lower speeds. Initial requested
    torque comes mostly electrically, but when it decides to add
    some ICE oomph in, it basically *snaps* the throttle open.
    Well, not really snap because in the absence of throttle-opening
    "accelerator pump" style enrichment that would cause a stumble,
    but fast enough to get past the low-torque range trivially
    fast. In fact under 35 mph or so it's very hard to get the
    vacuum to hold at the inefficiency-telltale ranges between
    say 10 in-Hg and 15 or so. It's high load or nothing, basically,
    and that's exactly how it should be in those scenarios. The
    driver can then enhance this behavior even more by adding P&G
    on top of that, so it's run-under-load or shut off entirely.
    Fine, but at higher road speeds, things begin to damp out a little.
    The vacuum begins to rise slightly, and starts tracking the
    driver's foot a little more linearly. I think this is the killer
    in the mid-speed ranges -- drivers who get up to 50 or 55 and
    then just back off enough to gently cruise are now running
    totally off the efficiency curve at light loads, even if the
    RPM is down around 1000. Now, the throttle opening never really
    gets that high at any time during all this, unless you really
    nail it -- its baseline is 14.5% at fully closed, and 15.?%
    seems to be about nominal idle level, and I'd be lucky to get
    it up to 35% even during "brisk" acceleration. In fact according
    to the OBDII laptop next to me, the throttle opening seems to
    track fairly closely to *RPM*. Which is interesting but
    understandable -- engine RPM is based on *power* demand, across
    a fairly flat *torque* curve. If torque is relatively constant,
    then more power comes from letting it spool up more RPM.
    So what I've been trying now is while on the interstates, trying
    to let my RPM drop as far as possible but keep the vacuum down
    around its higher-speed baseline of 5ish in-Hg. Theory being that
    that represents higher shaft load, and the telltale for torque
    starting to back off is increasing vacuum. You can't *feel* any
    of this in your butt, it's got to come from instruments. The
    instantaneous mileage rides somewhere north of 60 if I'm doing
    this right, but that only resists about 62 mph worth of air
    resistance. So it's still not a lot of kilowatts out of the
    engine, and I'm wondering if I'm just fooling myself. I've tried
    a bunch of other scenarios -- pulse harder around 2500 rpm and
    then try to "warp stealth" for a while, but besides being way
    more work and uneven speed, seems to return lower mpg segments.
    This "maintain torque" [still somewhat theoretical, mind you]
    strategy, with departures into higher RPM [aka power] bursts to
    get up hills, seems to be doing just a shave better in general.
    What I'd like to see in response to this [besides "geez, this
    guy does just *go on*, doesn't he"] is some of the accumulated
    knowledge about torque and throttle control and mpg learned from
    some of the *other* cars people have been experimenting with.
    And from other Prius drivers, but I realize there aren't a whole
    lot of those hereabouts yet.
    And if the whole thread dies here because all of the above was
    too confusing, I'd understand. I'm still entertained by how
    the prius whacks the throttle open on a low-speed pulse, though.
  2. tbaleno

    tbaleno Well-Known Member

    Yeah. I don't think a lot of us have had any experience with the technical sides of our cars. Mostly its been about learning to drive efficiently not about how the fine details on how the car does what it does. Hopefuly this will change over time and we will become more rounded.

    I'm still trying to figure out the optimum way to accelerate. All the variables like am I starting off going up a hill or down one or flat and winds and all those make it very complicated to figure out.

    I loved this post and hope we can get more like it.
  3. psyshack

    psyshack He who posts articles


    I think its a good thread and a possible breeding ground for interaction.

    Having a 06 Civic EX with a semi atkins head. Ive been very confused with what Ive been seeing on the ScanGauge concerning the MAP. Its like throw all you thought you knew out the window cause none of its right now.

    On the r18 it has a ecnomy lobe on the cam. It appears the intake valves are on this lobe between 1000 rpm and 3500 rpm. With its peak efficiency around 3000 rpm. What its doing here is holding the intake valves open long enough to push a combustion charge back into the intake manifold. The valves then close and limit the compression stroke to 1.5L for the engine insted on the normal 1.8L. Then the intake manifold has a short and a long flow path thats controled by some other outerspace device Im sure they found in Roswell. All this going on makes driving by the Manifold Air Pressure impossible.

    One moment you will look down and see 5.5" Hg with a instant mpg at 90 mpg. Look again and its 10.5" Hg and your at 90 mpg. Look again and its at 7.5" Hg and your getting 21 mpg. It makes no sense to me at all. Defies all rules governing Hg and mpg.

    The scan gauge is pretty quick when you have it setup on its fastist refresh rate. You put throttle imput in and you have instant reading. Or at near instant readings that seem to follow your foot.

    None the less,,, Ive found this to be intresting and very frustrating. More on the frustrating of late. With enough intrest to hope this can be explaned to me.

  4. krousdb

    krousdb Defiant NX-74205

    Very interesting stuff Hobbitt. I never thought to hook up a vacuum gauge to see what is going on. I did ha ve one in the del sol and it changes with throttle position and tended to mimic the SuperMID readout linearly so I removed it. My experimentation with the del sol also confirms that accel at low rpms with high load, low hg is what yields the best FE. I would try to install the hg gauge on the Pruis but it is now broken.

    Anyway, you are far more advanced than I am at speeds over 41MPH. Until now I just used cruise control. Now, next time I get the chance I will try warp stealth. Please let us know when you figure this over 41 thing out.
  5. hobbit

    hobbit He who posts articles

    Oh, right, the other thing I forgot to mention is that I believe
    the slight vac rise as speed increases is because of the valve
    timing changing. Unfortunately I don't have a good way of watching
    the VVTi target angle right now -- that requires something that's
    more aware of nonstandard OBDII parameters [it's not Prius-
    specific, just possibly out of most tools' capabilities]. For a
    while I was running around with an O-scope hooked to the oil control
    solenoid output, but that told me nothing since it just gets
    bumped one way or the other for a sec and back to center.
    Yup, one day I really want to understand this. But having the
    knowledge people have gleaned about other engine setups will help
    actually *master* it someday...
  6. tbaleno

    tbaleno Well-Known Member

    I'm getting confused about what high and low load are. Is low load when you are not accelerating and high load when you are accelerating more briskly?
  7. philmcneal

    philmcneal Has it been 10 years? Wow

    the higher the load, the better risk for lugging. but for us FE people lugging is good it seems. but your quote pretty much answered your own question. Although the brisky part is done at low rpms.
  8. xcel

    xcel PZEV, there's nothing like it :) Staff Member

    Hi Hobbit:

    ___First I want to get into a bit of theoretical guess work with the Prius II’s inner workings at those higher speeds. Please add what you can in regards to her MGSet interactions at 41 and above because I am not a Prius II tech by any means … The way the Prius II falls on its face as speeds climb above 41 mph cannot be explained because of aero and mechanical drag alone imho. The reason I say this is the 1.0 in the Insight uses somewhere in the neighborhood of 17 - 20 HP to maintain a lean burn cruise of 90 mpg at 55 mph or so. I have seen charts developed from Wayne Brown’s Palm Prius II Sim which show the Prius II holding 85 mpg from 25 to 41 mph steady state. This speed range is not only huge but almost pancake flat up to the magic 41 mph speed. The precipitous fall after 41 mph right when MG1 is screaming for help and exactly where all the trouble starts …

    Speed vs. MPG
    68F, sea level, 40% RH, 29.92", 225 payload, no AC, no wind

    ___This very unorthodox drop off after 41 mph is not normal other then an Insight or HCH-I dropping out of lean burn. What it looks like to me is the MGSet’s are starting to fight one another with that loop-current (or whatever you call it). MG1 generating electricity while dragging off the ICE to feed MG2 for torque output all the while the ICE is also generating torque for propulsion and feeding MG1 torque to generate electricity. The Prius II’s Atkinsonized ICE is certainly strong enough to maintain 60 mph with little in the way of higher RPM’s I would have to think? Again, think back to the Insight running a steady 17 - 20 HP at 55 mph with a similar Cd and maybe 70% of the frontal area. The Prius’ 1.5 should easily be able to maintain 25 HP from its peak of ~ 75 at a very low fuel consumption rates vs. the possible 15 – 20 HP needed to maintain the 23 – 41 mph range shown in the graph for maximum FE. In the real world, I think Dan told me she allows 50 mpg at 70 mph, 60 mpg at 60 mph, and 70 mpg at 50 mph. With those numbers, 20 mph is almost a 30% loss in FE due to aero drag and slightly higher mechanical friction? That just seems like a relatively large number when considering how the Accord or Corolla have consumed at those speeds although I have never truly tested a flat cruise at 70 mph in the FCD equipped Accord unless I am in some kind of draft if at all possible.

    ___Ok, my theory is that the act of MG1 drawing power off the ICE to feed MG2 while MG2 provides propulsion as well as the ICE providing propulsion and feeding MG1 is where the real inefficiency lye’s.

    ___Intakes and throttle openings. The Insight does her best work at just above lug and just below IMA assist during acceleration in my experience. Honda screwed the pooch with IMA assist as it is like a male dog in a pack of female dogs in heat. It just wants to go like a turbo with a full boost on-off switch! Enough about IMA ;) The good thing is that just like a non-hybrid, I have found a very low RPM accel at relatively small throttle openings work for me during a standard acceleration. All the tech jockeys with the latest efficiency charts and graphs appear to have it wrong with a close to WOT for minimal pumping losses for maximum efficiency all the while the F=ma equation has shot right through the roof on every rotating component accepting maximum torque from whatever ICE one is driving! I do not know who is right but I know for maximum FE in my automobiles, the go slow with a relatively small throttle plate opening while grabbing the bottom of the first torque peak appears to do wonders! It sounds like the Prius II is running with very minimal pumping losses in the low speed ranges by the way your Vacuum is snapping off but at highway cruise, vacuum is holding up as I believe it will in all cars at a low load highway cruise.

    ___With what I believe to be the real Prius II inefficiency at higher speeds (42 and above) arising from MG1 drawing off the ICE and converting for MG2 somehow, the rumor mill has the Prius III running some kind of gear reduction on MG1 to alleviate the over rev and inability to restart the ICE at 41 mph and above due to limited torque concerns. This upgrade alone should allow the Prius III the ability to hold EV far beyond the 41 mph limits THS and HSD have saddled her with forever plus alleviate some of that huge drop at 42 + mph and beyond as shown in the graph?

    ___Ok, enough of my BS. It is time to see and hear the actual details of the Prius II’s inner workings from those that own one and know what they are talking about ;)

    ___Good Luck

  9. hobbit

    hobbit He who posts articles

    Well, a bit of clarification might be in order. By "lug" I didn't
    mean dangerously low RPM -- what I meant is closer to your high-torque,
    low[ish]-RPM scenario. I would like to think that the fewer times
    per second a charge is burnt in the cylinder, the less heat is
    produced and thrown away ... but on the other hand, if *more* mixture
    is introduced per stroke, maybe that's fallacious thinking. I should
    try to design an experiment to track how fast an ICE warms up under
    low-speed/hi-torque vs. high-RPM/light torque ... or has someone
    already investigated this?
    I've seen that diagram before too, and suspect that the sudden
    drop is simply due to having to pull the engine around, whether it's
    burning gas or not. One thing to remember about high speed travel
    at low ICE RPM is what's called "heretical mode" -- where MG2
    actually drives MG1 *backwards* as a motor to help push the ring
    gear around faster per given engine RPM. If you look at the
    nomographs you'll see how this must be true. Only when you bring
    demand [and thus RPM] up to much more than needed for cruising does MG1 rpm
    cross back over zero and go back into the usual "generator" mode.
    Yup, this is hard to wrap one's brain around [and was the seed for
    *much* debate back in the early days of prius_technical_stuff]
    but that aside, whichever way the necessary electrical hop that power
    takes is flowing at a given time, *it* seems less responsible
    for the MPG drop than just having to haul all that extra metal
    around along with. There will still be some air compression
    or decompression -- even the ideal "valves totally closed" state
    that some other engines may be able to do is not a completely
    lossless air-spring.
    Anyway, one of the original ideas of this thread was to determine
    [or bring forth] whether it's better, from a mileage standpoint,
    to do low-RPM/hi-torque or higher-RPM/low-torque/less-throttle.
    It may be completely dependent on which engine one is talking
    about, and/or other factors like how its intake is constructed,
    atkinson vs. otto, etc etc. Maybe reducing all this to that
    simplified question would help... then, how to best use those
    results gets back into the dependencies of the driveline connected
    to it.
  10. hobbit

    hobbit He who posts articles

    Tonight I took my O-scope out cruisin', and observed some injector
    times at interstate speeds. Anywhere between 1400 and maybe
    2200 rpm in the THS-II, injector time seems to sit right around
    6 or 7 ms. Drifts up to 8 ms as I cross that critical 2400 rpm
    threshold where the vacuum starts to drop from 5 in-Hg to more
    like 3 in-Hg, which I suspect is more related to the ICE moving
    to more of a "power" regime than "efficiency" regime via its VVTi
    valve timing. The most interesting thing was the fact that
    injector time remained almost constant through a wide RPM range,
    which throws my prior thinking about how to meter "injector
    duty cycle" into a cocked hat and now I'm thinking that I have
    to monitor the actual injector time from around 2 - 12 ms to
    get meaningful feedback on how much fuel is being consumed.
    Clearly, what I'm up against here is largely a function of ICE
    RPM, so at a wild-ass guess it seems that minimizing that is
    the rudiments of a higher-mileage strategy for high speed. And
    plenty of warp-stealth where appropriate, natch.
  11. xcel

    xcel PZEV, there's nothing like it :) Staff Member

    Hi Hobbit:

    ___I have seen the THS and HSD nomographs for years. Although you are trying to isolate the ICE’s capabilities, with the HSD attached, I think you have to many variables to deduce an overall fuel consumption scheme above 41 mph. Measurement is always the first step but with the pack and MGSet interacting with the ICE for her real world fuel consumption at highway speeds, I see a number of problems based on the ICE action alone even though it is the ultimate supply of propulsion and pack charging.

    ___About 3 or 4 years ago, Rick Reese drove his Graham scanner equipped Prius I out for a max tank run. IIRC, he saw two separate load percentages where her real world fuel consumption would peak and they were not anywhere near each other in terms of RPM’s or throttle position! I am stating this from memory from a post done many years ago including the fact I have no idea how the Graham scanner was calculating load in the first place. With that, I wonder if the Graham scanner was using Vacuum as an inferred load input? If so, you might want to head back to the Graham scanner posts in the old Prius forums to see what can be found in regards to not just the efficiency curves but where those peaks actually occurred. A little sleuthing around IC should bring up Rick’s post as to what he was seeing in his Prius I. Although not a big deal today, that tank ended up south of 70 mpg’s which was huge back then ;)

    ___Sorry I cannot be more helpful with regards to your specific line of questioning :(

    ___Good Luck

  12. philmcneal

    philmcneal Has it been 10 years? Wow

    guess i'll start playing with the "Load" and "throttle" position sensors now.
  13. RH77

    RH77 Well-Known Member

    Non-Hybrid Thoughts -- sorry to deviate momentarily :eek: -- Load 101?

    For those of us stuck with average, Automatic, ICE-Driven vehicles, I'm trying to learn a few items. It looks as if the torque converter will forever be the enemy of achieving the High Load, Low RPM, Low MAP situations. In the early days of using the SG, I gave up on looking at such figures since they aren't at all easy (or most times even possible) to control.

    Under acceleration with the auto-trans, MAP seems to rise with throttle input. So...DWL works under decel, technically as throttle is constant.

    Today I was able to get incredible mileage (for my Integra) at 65-70 mph -- an average of 45-50 mpg driving on the the 70-limit, hilly Interstate -- driving with load at a controlled intake temp of 90-100F, with outside ambient temps of 40F. The process was to carefully monitor TPS levels and elevation changes. On the same road at 55-60 mph, the average was 35-40 mpg as the momentum and pumping losses were too great. No draft, BTW. The vehicle has a low Cd, comparitively.

    I only have the ability to manually shift the Integra in 1 and 2 fully manually, and the TSX is a nannified manu-matic (esentially both cars can be started in 2nd gear at whatever throttle you wan't -- but you can almost smell that transmission burning into slipping oblivion if that becomes a habit). What's strange about the TSX is that is has no EGR, but instead variable Valve Lift AND Cam Timing similar to:

    So, in theory, it would benefit the automatic driver to force the car into the highest gear at the earliest opportunity, and keep the throttle planted (but not so far as to throw it into open loop) until the target speed is attained, then back off, rinse and repeat (P&G).

    Any slush-boxsters out there to confirm or add my ramblings? Great thread, BTW Hobbit -- I think it opens the door to attain more mechanical understanding of our machines and consequently how to operate them at optimum efficiency.

  14. brick

    brick Answers to "that guy."

    I can finally add something to this thread, I hope. First I read Hobbit's Page and tried to absorb as much of it as I could. That's some pretty brilliant stuff!

    For the commute home I set the ScanGauge up to display RPM and Load, the latter of which I think is just derived from engine vacuum. It's a little easier for me to understand because it gives you (in theory) a percentage of the maximum power that could be generated by the engine at that particular speed. What I saw is very interesting! Most of the time the Prius engine stays in the 76-85% range. As Hobbit describes on his page, the engine management changes engine speed rather than throttle position. I also saw the phenomenon where load is too low to keep the engine in that range so it' "loafs" (again Hobbit's word) at lower load rating. On a few occasions I saw it between 35 and 65%. This was below 41mph, and these observations were made under conditions that speed could be maintained by the electric motors only, even if just for a while.

    Here's how the initial application went:
    At first I just kept an eye on load and attempted to keep engine speed below 2300RPM. It didn't work that well while maintaining 60mph up one particular hill, but at least I was able to warp stealth down the back-side. Then speeds dropped to the sub-41mph range for the rest of the trip, at which point I attempted to make 2000RPM my upper limit. Again, I was just focussing on staying under that limit at first. Acceleration was not exactly "brisk" but not glacial either. I didn't have trouble keeping followers happy, which is generally very important to me. In fact, I may have accelerated a bit slow but I tended to overshoot my target speed by a few MPH, wanting to keep the ICE working while I waited for the appropriate place to glide or stealth a little bit.

    Toward the end of my drive I realized that my thought process was missing one key element: the low RPM "loaf limit." From then on I was striving to use the lowest RPM possible to get the job done without letting the load drop out of the 75-85% range. Things got really interesting! Pulsing was brought down from 1800-1900RPM to 1200-1300RPM and 40-45mpg. Low-demand cruising was possible down to ~1100RPM before it dropped out of the optimal load range, at which point I either pulsed or glided as appropriate. The result was a 75mpg 5-minute bar through terrain where I have never seen such a thing, and a 50.5mpg commute that is among the best I have had in that direction. (Going home is up-hill most of the way.)

    This is really promising! Give me a few hundred miles to adapt my technique and we'll see what happens.
  15. JimboK

    JimboK One owner, low mileage

    Uh-oh, did I help create a monster by suggesting a change in SG location?? ;)

    This is good stuff. Tim, you can help me bridge the gap: Hobbit doesn't have the SG, and I don't have his technical background and education. You have the former, and seem to have more of the latter than I do. I will be paying close attention as you report your SG-monitored application of theory(s).

    A question for now: I've experimented with watching various SG parameters for the several months I've had the device. Most recently it's been RPM, engine temp, throttle position, and GPH. It's been a while since I've watched engine load, but IIRC, I got readings much lower (30-50?) at the lower RPMs you're describing. Am I missing something, either in your description or in my admittedly error-prone recall? If not, how did you get the load that high and keep RPM that low?

    Thanks, and I look forward to more discussion!
  16. brick

    brick Answers to "that guy."

    Jim - I will pay particular attention to that aspect of the readings tomorrow. I am very confident that I was seeing load in the 80% range down to ~1300RPM, but my confidence about the 1100-1300RPM range is somewhat diminished after hearing your experience. There's always the chance that I have it wrong since I only spent about 7 minutes in that realm of driving.

    Will report back.
  17. brick

    brick Answers to "that guy."

    This morning I did see readings below 1200RPM with load in the 75-80% range. 1300RPM was more typical, though.
  18. JimboK

    JimboK One owner, low mileage

    Just to confirm: You are looking at the parameter labeled LOD, correct? How fast are you going when you see those readings? I've been watching mine again since yesterday's post, and at low to moderate speeds (up to maybe 50 -- I haven't been on the highway) I don't come close to 75. At RPM in the 1200 range, LOD is generally in the 40s. From there it seems like a more-or-less linear relationship between the two, with LOD approaching 60 as RPM goes into the low 2000s. As I recall now, that is why I stopped watching it previously; it wasn't much more help than RPM.
  19. Skwyre7

    Skwyre7 Well-Known Member

    I'm seeing similar numbers, JimboK. Although there are some hills that I've had to climb where I've kept the RPMs under 2000 and the LOD is around 75 or so. But again, that was climbing a hill. On flatter sections my LOD numbers have been closer to 45. Perhaps that's the difference.
  20. brick

    brick Answers to "that guy."


    Good question. Mainly 25-40mph. The first 3/4 of my commute is on relatively slow back roads, then I hop on the highway and do 55-65mph the rest of the way in. I don't get these low RPMs at those higher highway speeds...the lower limit is maybe 1600RPM or so under any significant power demand?

    This is really interesting, as there are significant differences between what you see and what I see. The big surprise for me is that you are only approaching 60% when you get into the 2,000RPM+ range. I have seen numbers to match yours under near-coast conditions but that is not typical.

    One of two things is going on here. One is that our cars are really behaving differently. It is possible that the control algorithms were changed between 2005 and 2007 but it seems like they ought to have had this figured out for the initial release. It is also possible that either my driving style or environment is inducing this behavior in a way that is not typical for you. I think that this is the most likely case given that you are talking about speeds up to 50mph that I essentially skip right over. It's either 25-40mph or 55-65mph for me. I also have a tendency to mimic a P&G routine on those roads. (It's also worth mentioning that our average FE is very close, so there can't be **that** much difference between what we're doing.) I think Hobbit mentioned something about this in his write-up, too. I seem to remember something about him describing precisely your car's behavior as speeds increase. Later on I'll have to go back and check, and I should also experiment with speeds in the 45mph range to see what happens.

    The other possibility that one or both of our ScanGauges is lying to us. I doubt it but you never know. I might switch one of my readouts to MAP to see if those numbers match LOD the way I think they should.
    Last edited: Jan 10, 2007

Share This Page