rpm accel data?
Moderators: JeffC, rdoherty, stieg, brentp
rpm accel data?
Hey, I have run some data log sessions and I have a question about the rpm accel data. How should this value be interpreted? On my log it can show a value of say +312 or -150. Now the +/- I think I understand but if I look at the actual engine rpm change in the left column, the engine has not either accelerated up 312 or down 150 rpm. Can someone shed some light for me?
As its an *acceleration* I imagine its rpm/second figure, e.g. if your engine increases 78rpm in 1/4 second then accel = 78/0.25 = 312. Hopefully someone with firmware knowledge can confirm/correct.
Oh and looking at some of my logs I've just noticed the timestamps do not always increse e.g. I have the sequence
02:26.8
02:27.9
02:27.4
02:27.5
02:27.8
02:28.3
which looks a bit odd (and I haven't even got a flux capacitor fitted).
Rich.
Oh and looking at some of my logs I've just noticed the timestamps do not always increse e.g. I have the sequence
02:26.8
02:27.9
02:27.4
02:27.5
02:27.8
02:28.3
which looks a bit odd (and I haven't even got a flux capacitor fitted).
Rich.
The rpm (and load) accels are derived inside the configurator from successive rpm/load readings - it's not a firmware feature.
I looked at it some time ago and, probably unfortunately and misguidedly, decided that I couldn't get any useful information from it, and so I now operate the runtime perspective with both accelerations turned off and ignore the outputs in the log file...
I'm not entirely sure how often the configurator asks the MJLJ for its parameters (the update interval on the runtime perspective goes down to 100ms, so maybe that is how often?) but maybe the maths is in the configurator is supersensitive in that minor changes in rpm every 100ms are being translated into large accelerations when the base time is 1 second - e.g a 20rpm increase across 100ms (which could just be an accumulated noise effect across wiring, hardware and software), is being shown as an acceleration for that 100ms timeslot as +200rpm per second. And in the next 100ms the rpm goes down 30rpm for whatever reason (including the engine changing speed!) so it reports -300rpm per second in that next timeslot
Martin
EDIT - I also played around with the time column after importing the log into Excel - I pretty much failed to get a constantly increasing time (can't remember if I saw -ve time, but I may have done) - the intervals were definitely, as observed above, not constant. How this affects the rpm accel values, I don't know.....
I looked at it some time ago and, probably unfortunately and misguidedly, decided that I couldn't get any useful information from it, and so I now operate the runtime perspective with both accelerations turned off and ignore the outputs in the log file...
I'm not entirely sure how often the configurator asks the MJLJ for its parameters (the update interval on the runtime perspective goes down to 100ms, so maybe that is how often?) but maybe the maths is in the configurator is supersensitive in that minor changes in rpm every 100ms are being translated into large accelerations when the base time is 1 second - e.g a 20rpm increase across 100ms (which could just be an accumulated noise effect across wiring, hardware and software), is being shown as an acceleration for that 100ms timeslot as +200rpm per second. And in the next 100ms the rpm goes down 30rpm for whatever reason (including the engine changing speed!) so it reports -300rpm per second in that next timeslot
Martin
EDIT - I also played around with the time column after importing the log into Excel - I pretty much failed to get a constantly increasing time (can't remember if I saw -ve time, but I may have done) - the intervals were definitely, as observed above, not constant. How this affects the rpm accel values, I don't know.....
Hi Sunray - not sure what the engine is doing while you're logging the data, but you should see better information in the accel data if you are doing a normal drive compared to blipping the throttle in the driveway, or just idling.
Would you (or anyone) be able to attach an excerpt of that log to this thread so we can examine it?
From my testing, the accel data is accurate as far as the math goes, but the amount of smoothing is pretty minimal which contributes to the spikes.
Rich, not sure why the timestamp went backwards on that log- The software is just getting the timestamp from windows. I'll double check to see if I can have it access a higher resolution timer.
The update interval in the log affects how often the internal runtime values are sampled- a separate part of the software polls the controller in a fixed interval for the runtime data, which is 90ms + the time it takes to transfer 6 bytes of data across 4800 baud..
Would you (or anyone) be able to attach an excerpt of that log to this thread so we can examine it?
From my testing, the accel data is accurate as far as the math goes, but the amount of smoothing is pretty minimal which contributes to the spikes.
Rich, not sure why the timestamp went backwards on that log- The software is just getting the timestamp from windows. I'll double check to see if I can have it access a higher resolution timer.
The update interval in the log affects how often the internal runtime values are sampled- a separate part of the software polls the controller in a fixed interval for the runtime data, which is 90ms + the time it takes to transfer 6 bytes of data across 4800 baud..
Here is an excerpt:
4369 101 27 171 2 0
4444 100 27 320 0 0
4534 100 28 282 1 0
4633 100 28 307 0 0
4690 100 28 291 0 0
4793 100 28 313 0 0
4834 100 28 290 0 0
4956 100 28 320 0 0
4988 100 28 297 0 0
5120 100 29 323 0 0
5117 100 29 277 0 0
5204 100 29 299 0 0
5257 100 29 264 0 0
5385 101 29 324 0 0
5440 100 29 327 0 0
5512 100 29 328 0 0
5018 46 32 -135 -43 0
4369 101 27 171 2 0
4444 100 27 320 0 0
4534 100 28 282 1 0
4633 100 28 307 0 0
4690 100 28 291 0 0
4793 100 28 313 0 0
4834 100 28 290 0 0
4956 100 28 320 0 0
4988 100 28 297 0 0
5120 100 29 323 0 0
5117 100 29 277 0 0
5204 100 29 299 0 0
5257 100 29 264 0 0
5385 101 29 324 0 0
5440 100 29 327 0 0
5512 100 29 328 0 0
5018 46 32 -135 -43 0
Will this do? When I log on my laptop it only produces an excel file.
55:31.4 4369 101 27 171 2 0
55:31.7 4444 100 27 320 0 0
55:32.0 4534 100 28 282 1 0
55:32.2 4633 100 28 307 0 0
55:32.5 4690 100 28 291 0 0
55:32.8 4793 100 28 313 0 0
55:33.2 4834 100 28 290 0 0
55:33.3 4956 100 28 320 0 0
55:33.4 4988 100 28 297 0 0
55:33.8 5120 100 29 323 0 0
55:34.0 5117 100 29 277 0 0
55:34.2 5204 100 29 299 0 0
55:34.5 5257 100 29 264 0 0
55:34.7 5385 101 29 324 0 0
55:35.0 5440 100 29 327 0 0
55:35.3 5512 100 29 328 0 0
55:35.5 5018 46 32 -135 -43 0
55:31.4 4369 101 27 171 2 0
55:31.7 4444 100 27 320 0 0
55:32.0 4534 100 28 282 1 0
55:32.2 4633 100 28 307 0 0
55:32.5 4690 100 28 291 0 0
55:32.8 4793 100 28 313 0 0
55:33.2 4834 100 28 290 0 0
55:33.3 4956 100 28 320 0 0
55:33.4 4988 100 28 297 0 0
55:33.8 5120 100 29 323 0 0
55:34.0 5117 100 29 277 0 0
55:34.2 5204 100 29 299 0 0
55:34.5 5257 100 29 264 0 0
55:34.7 5385 101 29 324 0 0
55:35.0 5440 100 29 327 0 0
55:35.3 5512 100 29 328 0 0
55:35.5 5018 46 32 -135 -43 0