nehalem


« Reply #45 on: January 28, 2012, 06:27:40 PM »



nehalem,
Wouldn't it be more accurate to use the equation QSTAT * sqrt(4/3) rather than QSTAT * sqrt(58.00/43.50), being that what the equation is trying to achieve is that the initial psi is 75 percent of the target psi. I know the equation will work out the same, but 58 and 43.5 are rounded off where as 4 and 3 is exact... also making calculations easier?
Brett
PS... don't take this reply as a smartass one, I've been learning a ton off your posts!
Hi, No worries, this was one of my first attempts at mathematically calculating anything for this car. I was quite nervous tbh. lol Did you mean to suggest sqrt (3 / 4) because you'll find that sqrt(58.00/43.50)=sqrt(4/3) ... I believe sqrt (4/3) is correct though.



Logged




PovGRide742
Jr. Member
Karma: +4/0
Offline
Posts: 33


« Reply #46 on: January 29, 2012, 02:02:12 PM »



Hi, No worries, this was one of my first attempts at mathematically calculating anything for this car. I was quite nervous tbh. lol
Did you mean to suggest sqrt (3 / 4) because you'll find that sqrt(58.00/43.50)=sqrt(4/3) ... I believe sqrt (4/3) is correct though.
Actually, I meant 4/3 over 58/43.5... and for this reason: 4bar = 58.XXXXXXXXpsi, and 3bar = 43.5XXXXXXXpsi, so if you did the math absolutely correct, it would equal exactly that, 4/3. So it's just easier to use the equation of 4/3 instead of the psi/psi. This way, there is no way to be 'mathematically' incorrect.



Logged




nehalem


« Reply #47 on: January 29, 2012, 10:21:30 PM »



sqrt(4/3) ~= 1.15470054 sqrt(58/43.5) ~= 1.15470054



Logged




PovGRide742
Jr. Member
Karma: +4/0
Offline
Posts: 33


« Reply #48 on: January 30, 2012, 06:08:26 AM »



sqrt(4/3) ~= 1.15470054 sqrt(58/43.5) ~= 1.15470054
Lol, yes... I know they equal the same number. In the odd event though, that you'd want to go ridiculously exponentially, 4/3 would be more accurate than 58/43.5. Although, to sum this up, I think you've already agreed 4/3 is easier.



Logged




Tony@NefMoto
Administrator
Hero Member
Karma: +116/4
Offline
Posts: 1389
2001.5 Audi S4 Stage 3


« Reply #49 on: January 30, 2012, 03:04:27 PM »



Just remember that no matter how perfect your equations are, as soon as you have them running in your car, you end up changing them.



Logged




PovGRide742
Jr. Member
Karma: +4/0
Offline
Posts: 33


« Reply #50 on: February 02, 2012, 05:59:27 AM »



Just remember that no matter how perfect your equations are, as soon as you have them running in your car, you end up changing them. Yeah, I know, I'm slightly OCD, so I like asaccurateaspossible starting values.



Logged




nyet


« Reply #51 on: May 06, 2012, 11:40:27 PM »



Just migrated to EV14 750cc (72#) @ 3bar
I tweaked TVUB some starting with Tony's values. Car is running fine.... but i think i need to massage cold start fueling.
Anybody else have to do this?



Logged




Rick


« Reply #52 on: May 07, 2012, 07:25:05 AM »



Never had an issue with cold start, but warm (not hot) start usually needs playing with.



Logged




julex


« Reply #53 on: May 07, 2012, 07:51:12 AM »



Just migrated to EV14 750cc (72#) @ 3bar
I tweaked TVUB some starting with Tony's values. Car is running fine.... but i think i need to massage cold start fueling.
Anybody else have to do this?
Yes. I had to bump up warmup maps for fueling as I was going lean and engine was bogging down when stepped on for normal acceleration. This would only be the case for medium load. Either high load or barely any load requests would not cause that.



Logged




nyet


« Reply #54 on: May 08, 2012, 12:55:58 PM »



Assume this ( http://www.boschdealer.com/specsheets/0280158117cs.jpg) to be accurate and we already know a 4bar pressure regulator is 58.02 psi. Therefore FNPW_OFFCOMP has a theoretical value of: 1.1852. Using the same decimal accuracy as the calibration sheet: 1.1852 = 1.2179+(58.0254.96)*(1.16381.2179)/(60.0354.96) In MATLAB I entered the FNPW_OFFSET values and fit a shapepreserving spline. You can see the attached graph below to view the curvefit. Interpolated output from the curve below is naturally multiplied by 1.1852. Volts MS 7.04 3.865 10.07 1.678 12.04 1.227 14.08 0.926 17.88 0.779 Following up on this, i have a 3bar fpr interpolating OFFCOMP gets me 1.1326 .. nehalem, i dont have your noncomped spline interpolation, but if i divide by 1.1852, I get Volts ms (2.7bar) ms (3bar) ms (4bar) 7.04 3.261 3.693 3.865 10.07 1.416 1.604 1.678 12.04 1.035 1.173 1.227 14.08 0.781 0.884 0.926 17.88 0.657 0.744 0.779 Does this look right?


« Last Edit: May 08, 2012, 02:23:20 PM by nyet »

Logged




nehalem


« Reply #55 on: May 08, 2012, 02:14:33 PM »



Yes  That math looks correct. However, after a while of adjusting LTFT's, I ended up with: 3.12 1.35 0.99 0.75 0.63



Logged




nyet


« Reply #56 on: May 08, 2012, 02:20:56 PM »



I don't get the whole LTFT tuning thing.
how can you do anything but adjust only 2 operating points of the bus voltage (i'm assuming ~14v out of the alternator, or regulated 12v during cranking)


« Last Edit: May 08, 2012, 06:10:42 PM by nyet »

Logged




nehalem


« Reply #57 on: May 08, 2012, 06:22:42 PM »



Precisely the way you stated it. You need to make the assumption that your operation is not scalar, rather you are vector multiplying the entire table by your adjustment factor. new_val = (1+x) * old_val on the entire table as a vector in effect



Logged




nyet


« Reply #58 on: July 02, 2012, 09:28:13 AM »



I've got the car's trims right, but i'm still having
1) cold start problems (possibly checkvalve related) 2) lumpy idle
Anybody else have idle issues?



Logged




nyet


« Reply #59 on: July 15, 2012, 02:47:39 PM »



update I've got the car's trims right, but i'm still having 1) cold start problems (possibly checkvalve related)
Due to cold plugs. Investigating multispark 2) lumpy idle
Bad IOP. Reverted first two load rows to stock, and my idle problems went away. Also, gave up on trim tuning. I'm back to the theoretical TVUB specs as posted in my earlier post, and I'm doing trim tuning with MAF/KRKTE tweaking.



Logged




