evoscan knocksum = ECUknock
Jul 3, 2007, 06:32 PM
#16
I built my CarPC for just that purpose a couple of years ago. I used the original VoomPC case and the M1-ATX power supply. Just add an indash motorized touchscreen and you're done.
I may upgrade, though, since the board that I have has a relatively slow processor. I think I have the M10000.
Eric
I may upgrade, though, since the board that I have has a relatively slow processor. I think I have the M10000.
Eric
Jul 3, 2007, 07:31 PM
#17
but I wouldn't worry about evoscan being slow, it is the fastest factory diagnostics on the planet 
especially if you have a 2001-2005 reprogrammable baudrate ecu if you want.. just log 5 dataitems only.. at 200samples per second thats 40 times per second per item! (Can electricity travel that fast? must be a damn fast and powerful ecu these cars have from factory) who in would ever want to upgrade to an aftermarket ecu is just mad... hehe.
especially if you have a 2001-2005 reprogrammable baudrate ecu if you want.. just log 5 dataitems only.. at 200samples per second thats 40 times per second per item! (Can electricity travel that fast? must be a damn fast and powerful ecu these cars have from factory) who in would ever want to upgrade to an aftermarket ecu is just mad... hehe.
and I'm lost.
Jul 5, 2007, 07:38 AM
#18
What Klaus has said is that there is no need to have a sampling frequency of greater than about 12 Hz. That's roughly one sample every 0.083 sec. The Evo ECU can do 12 Hz via MUTIII. The trick is that to log at 12 Hz with EvoScan or Mitsulogger, its necessary to keep the number of requested parameters down to a minimum. For EvoScan to log at 12 Hz requires that no more than 15 parameters be logged during a session. I log only 11 parameters when I'm doing typical tuning stuff.
Jul 5, 2007, 10:56 AM
#19
With logworks, I get only three samples per second logging about 10 channels. Back when I used evoscan it wasn't any better. There is something particular with my combo of parts (ECU, 700 processor computer, cable) that is slowing it down.
Hamish is talking 40 samples/ second, how? I would hate to be having 10 counts of knock that doesn't even get registered cuz at times timing does appear lower, but this is at 300+ load beyond my scale which automatically pulls timing
Hamish is talking 40 samples/ second, how? I would hate to be having 10 counts of knock that doesn't even get registered cuz at times timing does appear lower, but this is at 300+ load beyond my scale which automatically pulls timing
Jul 5, 2007, 11:06 AM
#21
I remember Bez stating that increasing the baud rate in the ECU did not effect logging speed, but maybe something has been found since then?
FWIW, Logworks seems to be a bit slow for me, too. I will compare to the newest EvoScan when released to see if there is or isn't a difference. I remember in one thread with nj1266 and mrfred, it appears that LogWorks may be slow or repeats values more often when logging MUT than Evoscan does. I think nh1266 has created a thread in the Innovate forums to ask them about this.
Eric
FWIW, Logworks seems to be a bit slow for me, too. I will compare to the newest EvoScan when released to see if there is or isn't a difference. I remember in one thread with nj1266 and mrfred, it appears that LogWorks may be slow or repeats values more often when logging MUT than Evoscan does. I think nh1266 has created a thread in the Innovate forums to ask them about this.
Eric
Jul 5, 2007, 11:16 AM
#22
He's basing that on the baud rate that the ECU wants to use. When using my laptop (IBM X40 which is not terribly fast) and the standard baud rate, I see 180 samples per second. If only 6 parameters are being logged (e.g. TPS, knocksum, boost, 2-byte load [requires two requests], and RPM), then Evoscan will request the value for each of those parameters 35 time per second. My experience is that the ECU does not update the values at that rate, so there will be repeats. My experience is that the ECU updates the values about 10-15 times per second.
So there are two things being discussed here: 1) how frequently the ECU updates all the parameters and 2) how frequently EvoScan/Mitsulogger requests the values from the ECU.
So there are two things being discussed here: 1) how frequently the ECU updates all the parameters and 2) how frequently EvoScan/Mitsulogger requests the values from the ECU.
Jul 5, 2007, 01:32 PM
#23
Eric, about how many steps per second are you seeing with LW3 (beta 2)for RPM? In other words, how many flat spots before smoothing per second?
I log about 10 channels with 2 byte air/load
Thanks
Jul 5, 2007, 01:44 PM
#24
He's basing that on the baud rate that the ECU wants to use. When using my laptop (IBM X40 which is not terribly fast) and the standard baud rate, I see 180 samples per second. If only 6 parameters are being logged (e.g. TPS, knocksum, boost, 2-byte load [requires two requests], and RPM), then Evoscan will request the value for each of those parameters 35 time per second. My experience is that the ECU does not update the values at that rate, so there will be repeats. My experience is that the ECU updates the values about 10-15 times per second.
So there are two things being discussed here: 1) how frequently the ECU updates all the parameters and 2) how frequently EvoScan/Mitsulogger requests the values from the ECU.
So there are two things being discussed here: 1) how frequently the ECU updates all the parameters and 2) how frequently EvoScan/Mitsulogger requests the values from the ECU.
Jul 5, 2007, 02:17 PM
#25
It was on the comparisons between his boost pressures with LW and evoscan. I looked at the logs and I think the LW appeared more steppy due to the rate of spool which was quite different between the two so I'm not totally convinced this happened yet. If the rate of spool is faster it will appear more steppy.
Based on mrfred's analysis, it looked pretty convincing that LW was just logging slower and repeating values where there should be no repeats, whatever the reason may be.
Originally Posted by C6C6CH3vo
Eric, about how many steps per second are you seeing with LW3 (beta 2)for RPM? In other words, how many flat spots before smoothing per second?
I log about 10 channels with 2 byte air/load
Thanks
I log about 10 channels with 2 byte air/load
Thanks
Code:
time LC_1_O2 SSI4_4 AirFlow Intake_Air_Temp Atmospheric_Pressure Engine_Speed Ignition_Timing Knock_Sum (sec) (AFR) (PSIg) (Hz) (degC) (PSI) (RPM) (deg) (count) 0 14.98 -8.4 94 30.1 14.8 2878 43 0 0.08192 15.2 -8.3 94 30.1 14.8 2878 43 0 0.16384 15.21 -8.21 94 30.1 14.8 2886 43 0 0.24576 14.95 -4.86 116 30.1 14.8 2909 34.5 0 0.32768 13.52 -1.55 138 30.1 14.8 2933 26 0 0.4096 14.05 -0.47 230 29.8 14.8 2956 23 0.5 0.49152 14.19 0.57 321 29.5 14.8 2964 20 1 0.57344 13.86 1.09 321 29.5 14.8 2972 20 1 0.65536 13.39 1.61 321 29.5 14.8 2979 20 1 0.73728 13.1 2.22 347 29.5 14.8 3003 18.6 1 0.8192 12.79 2.79 372 29.5 14.8 3026 17 1 0.90112 12.47 3.5 397 29.5 14.8 3050 16.1 1 0.98304 12.22 4.21 422 29.5 14.8 3073 15 1 1.06496 12.08 5.01 453 29.5 14.8 3112 13 0.5 1.14688 12.1 5.81 484 29.5 14.8 3159 11 0 1.2288 12.05 6.8 522 29.2 14.76 3206 9.5 0 1.31072 11.91 7.75 560 28.9 14.71 3253 8 0 1.39264 11.75 8.93 604 28.9 14.71 3300 7 0 1.47456 11.52 10.06 648 28.9 14.71 3347 6 0 1.55648 11.3 11.43 701 29.2 14.71 3410 6 0 1.6384 11.13 12.75 754 29.5 14.71 3449 6 0 1.72032 10.88 14.21 754 29.5 14.71 3472 6 0 1.80224 10.79 15.63 754 29.5 14.71 3488 6 0 1.88416 10.72 17 815 29.5 14.67 3535 4.5 0 1.96608 10.72 18.37 875 29.5 14.64 3597 3 0 2.048 10.82 18.79 938 29.5 14.64 3660 3 0 2.12992 10.82 19.17 1000 29.5 14.64 3722 3 0 2.21184 10.91 19.36 1003 29.8 14.64 3801 3.5 0 2.29376 10.97 19.55 1007 30.1 14.64 3879 4 0 2.37568 11.01 19.22 1016 30.1 14.64 3949 5 0 2.4576 11.01 18.89 1025 30.1 14.64 4027 6 0 2.53952 11 19.12 1041 30.4 14.6 4106 5.5 0 2.62144 11.01 19.36 1057 30.7 14.57 4152 5 0 2.70336 11.01 19.31 1057 30.7 14.57 4184 5 0 2.78528 10.95 19.22 1057 30.7 14.57 4207 5 0 2.8672 11.02 19.17 1076 31 14.57 4262 5.5 0 2.94912 11 19.12 1094 31.3 14.57 4340 6 0 3.03104 11.05 19.27 1110 31.3 14.57 4418 6 0 3.11296 10.95 19.36 1126 31.3 14.57 4489 6 0 3.19488 10.88 19.31 1160 31.3 14.57 4559 6 0 3.2768 11.04 19.27 1195 31.3 14.57 4622 6 0 3.35872 10.98 19.22 1201 31.5 14.57 4700 6 0 3.44064 11.01 19.12 1207 31.8 14.57 4770 6 0 3.52256 10.91 19.08 1232 31.8 14.57 4841 6 0 3.60448 10.98 18.98 1257 31.8 14.57 4880 6 0 3.6864 10.95 19.03 1257 31.8 14.57 4903 6 0 3.76832 10.91 19.03 1257 31.8 14.57 4935 6 0 3.85024 10.89 18.93 1279 32.1 14.53 5013 6 0 3.93216 10.83 18.84 1301 32.4 14.5 5107 6 0 4.01408 10.91 18.79 1325 32.7 14.5 5185 6.5 0 4.096 10.82 18.75 1347 33 14.5 5247 7 0 4.17792 10.85 18.7 1356 33.3 14.5 5310 7 0 4.25984 10.82 18.6 1366 33.6 14.5 5372 7 0 4.34176 10.8 18.65 1391 33.6 14.5 5451 7 0 4.42368 10.69 18.65 1416 33.6 14.5 5529 7 0 4.5056 10.79 18.6 1428 33.9 14.5 5607 7 0 4.58752 10.72 18.56 1441 34.1 14.5 5654 7 0 4.66944 10.82 18.6 1441 34.1 14.5 5685 7 0 4.75136 10.78 18.65 1441 34.1 14.5 5701 7 0 4.83328 10.76 18.6 1463 34.1 14.5 5748 7.5 0 4.9152 10.83 18.51 1485 34.1 14.5 5803 8 0 4.99712 10.8 18.51 1491 34.1 14.46 5857 8 0 5.07904 10.76 18.51 1497 34.1 14.43 5912 8 0 5.16096 10.83 18.46 1516 34.1 14.43 5982 8.5 0 5.24288 10.82 18.42 1535 34.1 14.43 6053 9 0 5.3248 10.79 18.46 1548 34.1 14.43 6123 9 0 5.40672 10.95 18.46 1560 34.1 14.43 6186 9 0 5.48864 10.88 18.37 1566 33.9 14.43 6248 9.6 0 5.57056 10.92 18.27 1573 33.6 14.43 6303 10 0 5.65248 11 18.27 1585 33.6 14.43 6358 10.5 0 5.7344 11.07 18.23 1598 33.6 14.43 6389 11 0 5.81632 10.94 18.13 1598 33.6 14.43 6405 11 0 5.89824 11.05 17.99 1598 33.6 14.43 6428 11 0 5.98016 10.98 18.08 1595 33.3 14.43 6483 11 0 6.06208 11.1 18.13 1591 33 14.43 6553 11 0 6.144 11.08 18.04 1598 32.7 14.43 6624 11.5 0 6.22592 11.08 17.9 1604 32.4 14.43 6678 12 0 6.30784 11.08 17.8 1604 32.4 14.43 6733 12.5 0 6.38976 11.04 17.71 1604 32.4 14.43 6780 13 0 6.47168 11.05 17.71 1604 32.1 14.43 6843 13.5 0 6.5536 11 17.66 1604 31.8 14.43 6897 14 0 6.63552 11.01 17.61 1604 31.8 14.43 6952 14.6 0 6.71744 10.79 17.57 1604 31.8 14.43 6999 15 0 6.79936 10.73 17.61 1604 31.5 14.39 7062 15.5 0 6.88128 10.69 17.61 1604 31.3 14.36 7101 16 0 6.9632 10.79 17.61 1604 31.3 14.36 7124 16 0 7.04512 10.75 17.61 1604 31.3 14.36 7140 16 0 7.12704 10.73 17.38 1604 31.3 14.36 7187 17.5 0 7.20896 10.28 17.09 1604 31.3 14.36 7163 19 0 7.29088 12.36 6.38 1019 31.3 14.59 7062 28.6 0 7.3728 11.14 -4.38 434 31.3 14.8 6874 38 0 7.45472 11.55 -6.89 252 31 14.8 6686 33 0 7.53664 11.94 -9.44 69 30.7 14.8 6561 28 0 7.61856 15.39 -9.44 69 30.7 14.8 6499 28 0 7.70048 20.92 -9.44 69 30.7 14.8 6499 28 0
The RPM values look OK, but at this slow speed, they should. So, this doesn't prove or disprove anything at this point until I do tests with other software. I have to make sure that my slow CarPC isn't the limiting factor.
Edit: Hmmm...I just looked at my log in LogWorks and it doesn't seem to follow what the extract states. For example, I show 4903 RPM at 3.60s, 3.69s, and 3.77s when I view the log in LW. But, the numbers above don't show this. So, in the graph view in LW it is a choppy (flat spots). I have quite a few flat spots like that in the graphs that aren't in the extract above. I'm busy now, but I will take a look into this more later.
Eric
Last edited by l2r99gst; Jul 5, 2007 at 02:29 PM.
Jul 5, 2007, 03:18 PM
#26
I did a bit of tinkering with EvoScan during lunch. One very important thing that I relearned is that to get full logging speed in EvoScan, its absolutely necessary to disable "logging to screen". That slows EvoScan way way down. I was only getting about 72 samples per sec when I had "logging to screen" enabled. After I disabled it, I was getting 180 samples per sec. More later...
Jul 5, 2007, 04:41 PM
#27
can we keep the discussion to lines of information, 180 bits means nothing if you have 50 inputs/. I know you don't and I am impressed with that count, but do you see what I mean? six lines a second or ten lines a second or twenty lines a second mean something I can grasp.
Jul 5, 2007, 09:43 PM
#28
Here's TPS from a portion of a log session where I slammed the throttle to floor at about 3500 rpm. I was logging four parameters. Since EvoScan claims to be able to log at 180 samples per second, EvoScan in theory should be able to request the value of each parameter from the ECU 45 times per second when logging 4 parameters. This is equivalent to one entry every 0.022 sec. How does this compare to what I actually logged?
The plot here shows that there are 9 data points over a span of 0.2 seconds which equates to one entry every 0.022 sec or a logging frequency of 45 Hz which is exactly what's predicted based on 180 samples/sec. This means that EvoScan can in fact request data from the ECU at a frequency of at least 45 Hz. I think that this is pretty fast. Its about 4x faster than what Klaus at Innovate thinks is necessary.
So what about the rate at which the ECU updates its values that EvoScan reads? The nice thing about logging TPS is that the step size is quite small. The little tick marks on the y-axis represent the minimum increment in TPS that the ECU can report. If the ECU were occasionally sending repeat values, then the plot would have at least one step in it where there are two data points next to each other with the same TPS value. Its quite easy to see here that there are no repeats. This tells me that the Evo ECU can report data at 45 Hz (if not higher).
In conclusion, neither EvoScan nor the Evo ECU appear to be limiting factors in logging data at a high rate. The key to logging data at 12-15 Hz in EvoScan is to log no more than 15 parameters and disable logging to screen.
The plot here shows that there are 9 data points over a span of 0.2 seconds which equates to one entry every 0.022 sec or a logging frequency of 45 Hz which is exactly what's predicted based on 180 samples/sec. This means that EvoScan can in fact request data from the ECU at a frequency of at least 45 Hz. I think that this is pretty fast. Its about 4x faster than what Klaus at Innovate thinks is necessary.
So what about the rate at which the ECU updates its values that EvoScan reads? The nice thing about logging TPS is that the step size is quite small. The little tick marks on the y-axis represent the minimum increment in TPS that the ECU can report. If the ECU were occasionally sending repeat values, then the plot would have at least one step in it where there are two data points next to each other with the same TPS value. Its quite easy to see here that there are no repeats. This tells me that the Evo ECU can report data at 45 Hz (if not higher).
In conclusion, neither EvoScan nor the Evo ECU appear to be limiting factors in logging data at a high rate. The key to logging data at 12-15 Hz in EvoScan is to log no more than 15 parameters and disable logging to screen.
Jul 5, 2007, 10:21 PM
#29
That's a lot of samples per second.
I still can't find myself not using LW though - everything is there in perspective with all aspects of the tune in one view. If need be I can allways export to .dif to get 12 or more samples a second. But anything more, like 120 samples per second, how is that used?
I still can't find myself not using LW though - everything is there in perspective with all aspects of the tune in one view. If need be I can allways export to .dif to get 12 or more samples a second. But anything more, like 120 samples per second, how is that used?
Jul 6, 2007, 10:39 AM
#30
The MUT tables reference ECU RAM values some of which I think will be updated every ignition event. The resolution on the byte MUT requests may be why it doesn't appear to change as fast as it is being logged.