Dynojet PC-5, with Autotune, and full exh. system Tune...

Discussion in 'Parallel Universe' started by ebrabaek, Feb 8, 2013.

  1. ebrabaek

    ebrabaek Long timer

    Joined:
    Jan 12, 2010
    Oddometer:
    5,426
    Location:
    Grand Valley, Colorado
    I supose there could be a margin of error pushing our reads a 2 afr higher than what you saw. It was agrevated by header, and pipe (cat'less)
  2. ebrabaek

    ebrabaek Long timer

    Joined:
    Jan 12, 2010
    Oddometer:
    5,426
    Location:
    Grand Valley, Colorado
    I see....thanks. I pesonally would not feel comfortable only running the pc-5 without the AT. I think the bms-5 and the pc-5 would fight eachother to death....but with the addition of the AT, it now runs in closed loop above 2% throttle.
    The only thing I have seen that perhaps can elude to an armwrestleing, is sometime after a stop then fuel computer reads about 40% less mpg than as compared with before the install. But it only occurs sometimes.....and only for about 5 seconds. Not sure what is going on there......but it could also be as the at module is recovering the fuel control.
  3. jscottyk

    jscottyk Been here awhile

    Joined:
    Apr 14, 2007
    Oddometer:
    331
    Location:
    Austin, Texas
    Could the popping be related to the SAS that ebrabaek still has in place? Unburned fuel plus extra 02 from the fresh air pumped into the exhaust plus heat equal combustion in the header? One of the main benefits the KTM crowd reports after the removal of their SAS is elimination of popping (backfires) on deceleration.

    Also, it would seem the extra O2 in the exhaust provided by the SAS would provide an erroneous reading from the O2 sensor. Well, technically the sensor would be reading the O2 levels properly but the O2 levels would no longer be representative of what just happened inside the motor.
  4. terryckdbf

    terryckdbf Pickles

    Joined:
    Jun 1, 2006
    Oddometer:
    2,169
    Location:
    Perry, Ga
    It would be interesting and very useful if we had a Dyno run or runs charted and simultaneous information from the ECU with something such as a GS-911. The GS-911 gives a wealth of information and it could easily be utilized during a Dyno run.

    Terry
  5. ebrabaek

    ebrabaek Long timer

    Joined:
    Jan 12, 2010
    Oddometer:
    5,426
    Location:
    Grand Valley, Colorado
    Could be..... It did vanish after the pc-5/at install though.
  6. ebrabaek

    ebrabaek Long timer

    Joined:
    Jan 12, 2010
    Oddometer:
    5,426
    Location:
    Grand Valley, Colorado
  7. roger 04 rt

    roger 04 rt Long timer

    Joined:
    Nov 9, 2011
    Oddometer:
    2,728
    Location:
    Massachusetts
    I don't mean to be offensive or picky but might as well say this first ... you don't have Closed Loop operation above 2%. But you can generate target AFR correction factors. Let me explain the difference.
    As you know, with Autotune, if everything goes ideally, the 10 by 30 matrix of 300 autotune-generated trim values will get populated as you ride. Most likely, less than half will be filled due to areas that you can't/don't ride in. Those trims will remain empty. (I can't find any criteria an the Dynojet site for when it fills a cell, which would be good to know. Usually you need a stable period of time.)

    Then with Autotune in the always on mode that Dynojet describes, when you ride through that RPM/TPS combination, it adds a correction to the BMS-K pulse to get to the AFR in your target AFR table. Until there's a reasonable level of filling of the Autotune trim table you should expect uneven fueling--which corrects over some period of time.

    There's also the issue of what the BMSK fueling pattern is when it detects a missing O2 sensor. Does it vary widely like the R1150s? I suspect it does. I'll try to get Terry to ride his R1200GSA with O2s disconnected, BMSK reset to see what it does. Terry?

    Closed Loop (CL) operation takes a different approach. After being reset, the BMSK starts calculating Short Term Fuel Trims (ST) as soon as it goes Closed Loop, so that it can more quickly get into stable Closed Loop operation (and thereby minimize pollutants). Then over time, it starts to store those corrections in a table organized by TPS/RPM. Once that Long Term trim table starts to get filled, the BMSK has a proprietary algorithm that it uses whatever factors it has learned (Adpatation Values) to calculate trim factors that are used under Open Loop conditions, it areas where it can't, or hasn't yet learned corrections. If you have a GS-911, you can watch this process. Terry and I did just that on his R1200GSA. It was quite interesting to see it happen in the logs.

    As an aside, I often read opinions that the BMSK can't make major corrections. That is not the case. It seems quite happy to make corrections over a 40% range on the 1150s and 1200s that I've measured.​

    So with those comments on Closed Loop vs auto-tune made, here are some things you might consider trying on your bike.

    I like the idea of Autotune. I've read a good deal about those types of systems and if you're interested, there's a decent description on the Megasquirt site of how autotunes can be used. There's an order of magnitude more information at Megasquirt than at the Dynojet site.

    To start with, here's a good autotune, generic AFR target matrix. Compare it to a Dynojet basic autotune chart. (Both charts below.) You'll see that with some thought and some experiments, we could come up with a much better set of AFR targets than those provided. I'll go further and say you could develop a better tune for your bike than the one you got from the Dyno tuning because: a) your wideband sensor is in a better spot than their tailpipe sniffer; and b) you can spend an infinite amount of time (all your riding time) refining the tune table.

    Megasquirt Sample Target AFR Table
    [​IMG]

    Dynojet Sample Target AFR Table for F800GS
    [​IMG]

    So if you worked with the two tables to come up with something that represents a better set of AFR targets for the F800S than those of the Dynojet basics, you could ride normally (whatever that is for you) for several tanks gas and then post the trims that had been developed. From there you could manually smooth discontinuities between cells, and also manually reason values for those cells which had not yet been populated. And from all that install a new Fuel Trim table into your PC V.

    It seems to me that the above process could/should reset in an excellent Open Loop tune. The key to success is having a high-quality set of AFR targets.

    One of the other things you'd need to do to make a good AFR table is to use a GS-911 to learn the rough shape of the F800GS Spark Advance Curve. (A sample of mine below.)

    Let us know what you think, it would be an interesting experiment to help you with.

    RB

    2004 R1150RT Spark Advance ScatterPlot
    [​IMG]
  8. jscottyk

    jscottyk Been here awhile

    Joined:
    Apr 14, 2007
    Oddometer:
    331
    Location:
    Austin, Texas
    I seem to recall a thread from last year where señormoto offers to provide his AFR targets that were derived from a LOT of riding with an Autotune setup.

    I'll see if I can locate that thread.
  9. ebrabaek

    ebrabaek Long timer

    Joined:
    Jan 12, 2010
    Oddometer:
    5,426
    Location:
    Grand Valley, Colorado
    None taken Roger. Just because you are direct, and blunt does not mean any hostility towards me.... No worries. This is my first direct experience with the PC-5/AT. Had pleanty of second hand looks, but this is the first where I am behind the wheel. Thus was the reason to travel 600 miles north, as I trust every word my tuner speaks. So one thing first.... When I spoke to the closed loop, I was not referring to the BMS-K. ( perhaps you did not take it as such anyway...:D) but as the global fueling's functionality. Closed loop fueling is incorporated via a O2 sensor (lambda sonde) fueling is in this case to be determined to meet the target AFR you input yourself. In reality how the BMS-K behaves, is second on the list of concerns, as the "under/over fueling" is corrected by the PC-5/AT combo. That is what I have allways thought/understood to be closed loop, but perhaps we are chasing defenitions, so I will just say that it is fueling whilst incorporating a O2 sensor, rather than a predetermined map neglecting any O2 input. I am still using the original map that my tuner fed to the PC-5 ( he developed this, rather than using the DJ map) I have read several posts on forums, where riders were not particular happy with the pc-5/at setup, and most of them stopped posting after they started to apply the trim setting. I don't know if there are any relations, but I have yet to apply the trim values. I understand the value of doing such, but dont want to chase my own tail, as I would like to raise the target AFR from 13.2 on my setup to perhaps 13.6-13.8, and have a look first, but perhaps stick with one thing first, adopting the trim values that I have stored over the last 2K miles. Again, for me it is just hard to argue with success. As I have seen a 3-5 mpg increase, along with all the other good points. I like to just change one thing at a time, so I will start with adapting the trim values , and see what happens...... I do value you your input, so feel comfortable, that I take no offense....:clap:freaky:freaky
  10. roger 04 rt

    roger 04 rt Long timer

    Joined:
    Nov 9, 2011
    Oddometer:
    2,728
    Location:
    Massachusetts
    Good, we're on the same page. If you had some time, and could post three screen shots: afr targets, current learned trims, and current fuel map, that would be quite interesting to study.

    Do you have a GS-911? Among other things, it would probably tell us when secondary air was enabled.
  11. roger 04 rt

    roger 04 rt Long timer

    Joined:
    Nov 9, 2011
    Oddometer:
    2,728
    Location:
    Massachusetts
    I suppose that the secondary air could be a contributor but doesn't seem like the primary cause. When exhaust components are changed, the flow is now into a system that has different resonances, leading to the possibilities that pressure vs rpm changes at the exhaust valves; and to the possibility that fresh air enters the exhaust system from the tailpipe and causing popping when it meets residual fuel.
  12. ebrabaek

    ebrabaek Long timer

    Joined:
    Jan 12, 2010
    Oddometer:
    5,426
    Location:
    Grand Valley, Colorado
    Awesome.
    I will get the screen shots in a few minutes.... I do not have a GS-911, but have contemplated many times to get one. One fellow rider/member offered to sell me his. It sounds like this could benefit the tuning of the bike.... Yes...???
  13. mousitsas

    mousitsas Long timer

    Joined:
    Oct 10, 2008
    Oddometer:
    3,345
    Location:
    Athens, Greece
    Wow Roger, I was optimistic but didn't expect that! Your knowledge is invaluable to the forum. Thanks!
  14. roger 04 rt

    roger 04 rt Long timer

    Joined:
    Nov 9, 2011
    Oddometer:
    2,728
    Location:
    Massachusetts
    Personally, I would not be without one if I had your goals. Using the Log Realtime Values to CSV function I've helped others solve some very subtle problems. Here is a plot I have hanging around.

    [​IMG]
  15. ebrabaek

    ebrabaek Long timer

    Joined:
    Jan 12, 2010
    Oddometer:
    5,426
    Location:
    Grand Valley, Colorado
    Good grief.... I must have one of those...... Working on that right now....:D
  16. ebrabaek

    ebrabaek Long timer

    Joined:
    Jan 12, 2010
    Oddometer:
    5,426
    Location:
    Grand Valley, Colorado
    OK Roger....
    Here are the three screenshots.....
    Fuel Map, which is a base map only to correct for the idle hop of the 8GS.....
    [​IMG]


    Then Target AFR.....
    [​IMG]


    Then finally the trims after a few K miles.....
    [​IMG]
  17. roger 04 rt

    roger 04 rt Long timer

    Joined:
    Nov 9, 2011
    Oddometer:
    2,728
    Location:
    Massachusetts
    Your tables are very interesting. There is no doubt that autotune is functioning and learning from the Wideband O2 sensor. In a way it's too bad you don't have a second bung near the first one so that you could monitor AFR and see if you're really getting your target AFR.

    -I'm assuming you run ethanol.
    -I'm going to have to think about where you have limits in your settings and will be back to you with some questions.
    -It looks like at 6500 rpm and above for tps 10% and above the trim is maxing out at 10%.
    -The 25% numbers may be max-out numbers too. And that's awfully large.
    -With an AFR target of 13.2, really lambda 0.90, and running E10, you would expect about a 14% correction. Which is what we LC-1ers see.
    -I'm surprised that some of the high rpm/TPS numbers aren't single digit or negative.
    -Are the auto tune trims additive to the fuel map trims? If they are at 2%, 1250 rpm you're adding 45% to fuel—20% map plus 25% trim to hit 13.5. I question that. That doesn't seem right.
    Does the PC V AT have a heater calibration or free air calibration mode? What altitude do you ride at? Widebands are altitude sensitive.

    How would you feel about saving your current tables and then doing the following. Leave AFR targets as is, zero out the fuel table and trim table. Ride for a week and repost the trim table. Maybe even post it after each ride. It would be interesting to see it develop.

    This is all pretty interesting.
  18. ebrabaek

    ebrabaek Long timer

    Joined:
    Jan 12, 2010
    Oddometer:
    5,426
    Location:
    Grand Valley, Colorado
    Thanks for your input Roger. I have been a little apprehensive to change anything, as it seems to work so very good from the get go, but I will think it over to erase the fuel trim table, as The current table have much technical dirt sections in there, but in reality, I cannot see why that would make a difference.

    Cheers
  19. ebrabaek

    ebrabaek Long timer

    Joined:
    Jan 12, 2010
    Oddometer:
    5,426
    Location:
    Grand Valley, Colorado
    Roger.....after thinking about your question weather or not the trims are added to the base map......I now understand your question. Could be. The reason I am seing such a big drop in the computer mpg off idle. Tomorrow I will change the 20% off idle to 5 ....zero the trims....and have a look.
  20. roger 04 rt

    roger 04 rt Long timer

    Joined:
    Nov 9, 2011
    Oddometer:
    2,728
    Location:
    Massachusetts
    Hi Erling, I've been studying the Secondary Air function, Autotune algorithms and the interaction of Autotune and any map you have installed. For now, I think the best thing is to learn to save the maps you have for safekeeping and wait a little longer for me to get some feedback on how you may want to proceed.

    The subjects I'm looking at now are:

    • When is the Secondary Air Valve actuated? (It seems this will affect Autotune.)
    • How long is the injection pulse at wide throttle angles? (You can't extend the injection pulse longer than the motor rotation period.)
    • Are the base map and trim map additive? (It seems that they are.)
    • Dynojet says that it takes a while to calculate an Autotune correction for a given cell. How long?

    Did your tuner disconnect the Secondary Air Valve at the connector or plug it?

    On your PC V settings page, what is the delay time that is set before Autotune can start?

    But rather than risk giving you bad advice, I'd like to look at this longer and discuss with some other F bike experts.

    RB