• Content Count

  • Joined

  • Last visited

Community Reputation

56 Good

1 Follower

About mercuric

  • Rank
    Addicted to R
  • Birthday 11/26/1979

Profile Information

  • Gender
    Not Telling
  • Interests
    Volvos, Bicycles, Networks, Databases, A/V, Amateur Radio

Previous Fields

  • Location
    Austin, TX

Recent Profile Visitors

The recent visitors block is disabled and is not being shown to other users.

  1. Thanks! You have a good point. I suspect there's another factor in there we don't know about -- I suppose I should track down the code for the voltage table and see how that value factors in. Really there's a bunch of injection-related code I want to reverse, just haven't had a lot of quiet time to dig into it lately. Soon comes the colder weather, though Do you get logging output while it cranks, even if it doesn't start?
  2. A21 is the signal FROM the cam position sensor, VIDA specifies either Ulow or 5v, which I believe depends on where the sensor is in the cycle. A36 supplies power to the sensor. If you're not getting power to A21 *FROM* the sensor, rotate the engine by hand while checking for voltage, or unmount the sensor and spin by hand. When cranking you should be getting pulsed voltage, on and off at half the cranking speed of the engine. For example, VIDA specifies 7hz at idle.. 7 pulses / sec * 60 = 420 per min * 2 = 840, which is essentially target idle speed.
  3. Nope, it's a lot closer to what the name is, the code simply compares some values within the injection time calculation to TEMIN values, and if TEMIN is bigger, uses that instead. The code confuses me a bit, I suspect it'd make sense if I dissected the preceding and following several pages of code which make use of some MD register operations in what looks like an injection time calculation of sorts, but the following snip shows the core logic that deals with TEMIN -- It would appear to multiply TEMIN by 0x20 (100000b, so bit-shift 5 bits left) and use SUBB operations to see if R6 and R7 (
  4. Haha, no worries, glad you got it figured out! Did learn quite a bit indeed. Which 650cc's did you go with? Whole exhaust side might be a bit small for a 06-20G. Since you appear to be going for higher RPMs (your map out to 7000+) maybe a larger housing is better suited to the big 06-20G cold side? With your flow and RPM targets a 10cm+ housing with an external gate might be the next step.
  5. Awesome! Shall have to see if I can find what actually sets that bit. Guessing it's something to do with the SCL4402, we shall see. LOL @ 2nd being useless. Nice. You probably want to loosen your wastegate actuator such that you get maybe 5-7lb of boost with zero duty cycle, then increase duty cycle to get your desired boost. This will allow you to create a TPS vs duty cycle curve that'll give you a sufficiently responsive throttle such that 1st and 2nd will be usable again, as long as you don't mash it. What does the sensor have to do with the TCV? I'm confused, maybe you've hac
  6. Me too, man. More time, more money... Too many expensive interests, too many hours at the office.. Can't complain too much though, every bit of it is pretty neat.
  7. Boom. You have a TCV error. To understand the XRAM dump, first it helps to rearrange the columns, such that the bytes are in 1-15 order, and DPTR is on the left. Silly tunerpro doesn't seem to have a way to define column order for CSV exports. Mildly annoying. Then you start with the DPTR value, for example, find a line with DPTR F970. I suppose I should have used hex on the column names (I will update that now that I realize this) but D is 13, so go over to byte 13. There you find value 8, which is 00001000 binary. Counting from the right, which is bit 0, you see bit 0, 1 and 2 uns
  8. I shall do so after friday afternoon madness finishes, which should be done around the time I wake up from a nap.. hehe.. in the meantime, can you do an XRAM dump log? Load up the XRAM dump ADX and connect, it'll dump XRAM over and over. You don't even have to drive, just start her up and let the counter flip around a few times while saving the log (so it gets all of XRAM). Found where it compares the TPS to map WDKLDRN (throttle angle for LDR active) after realizing I should just search on the MOV DPTR of the map address. Silly me, should have been my first path in. Hehe. Anyways, that
  9. That's a good question. I shall have to try and find out! Still digging to see what sets LMR active. It's not in the boost control routines; suspect it's buried in a TPS routine or some other code. dig dig dig
  10. Interesting. It would appear xf_lmr.1 never gets set, which is B_LMRA aka LMR active. If that doesn't get set, most all the upper bank routine is abandoned. Investigating exactly what sets LMR to an active state. Makes sense, as even XTVM is staying at zero. Weird!!
  11. I haven't found any sort of MAF correction map.. ME7 and some other newer ECUs do, but nothing to indicate M4.4 does. Damn, tightmopedman! I'm at a loss here... But I do like a good digital witchhunt. One thing I did notice, though probably not related to the problem, is XTV_VS is lower than your duty cycle map. XTV_VS is the duty cycle after correction for IAT and atmospheric pressure. I believe this is because the IAT diagnostic is disabled, which locks the IAT temperature at the replacement value, 20 Celsius. Note map "LDR duty cycle correction, intake air temperature" (0xE7D4) has
  12. Interesting. IEX5 must be firing, 0xBC is increment in a manner consistent with bit 3 getting cleared. RAM_2B.5 doesn't appear to be getting set. I assume you removed my p-part hack too? Just making sure. Still confused. Shall review the part of the LDR routine that sets XPCP_TV... and dig through -608 closer to see if there's some additional duty-cycle twiddling not in 607. Hmm. I'll sing a serenade to the Excel godsdemons for you
  13. Haha, well at least I'm not the only one that screws up in Excel on a regular basis. Seems to bite everyone now and then.. We had a fun incident at the nerd shop a few weeks ago with that.. was funny to watch the people point fingers until I caught it, formula one row off invalided the whole data set.. lol. So I looked over your binary, it all looks in order. I'm super-confused now. First, try the rev5b binary without the p-part modifications, and set PLDRN (LDR P-part) and KFP2 to all zeros, as Simply confirms this works now. I'm wondering if there's something different in the manu
  14. All in good time, certainly only so much of that to go around! And that's what I'm wondering. Something is amuck -- so is this from the LDR p-part modified binary, or the regular rev5? Glad to help. We'll get this TCV stuff nailed down.