MBED, Arduino , STN1110 & BT

[ken1784]

@yoshi:
I will try to compare the precision using your formula and what I get from PID 230.
here is the formula I am using:

            distcm = nb230 *  65535 + ww -wi;  // distance since start in centimeters

I think the 65535 should be 65536.

Yoshi

 

[ken1784]

about chan_fs, I made a test some time ago, and it was not successful, I will try again later.

Please try to use the chan_fs code from my program, CAN_CHANFS_BUF_debug.zip (it is still on your FTP site).

The pin assignment is defined in the diskio.c as follows;

//******************************************
// MBED SPI/CS Select functions.... Modify for your layout.
//******************************************

SPI _spi(p5, p6, p7); // mosi, miso, sclk
DigitalOut _cs(p8);

They are same as yours.

Yoshi

 

[priusfan]

I think the 65535 should be 65536.

Yoshi

I agree ...

today, I will compare 0b4 & 230 from raw logs.
this evening I will maybe test shortly odo from 0B4 & 230; in fact I will log them both....

I will also check chan_fs on sdHC card (I have 4 , 8 & 16 Giga available for tests).

@planétaire: with p2, could you use the same formula for odo from 0B4?
maybe this is specific with gen3?

 

[thierryb]

I am comparing both b4 and 230 since march. I am even comparing them with distance obtained by a very precise GPS, and comparing too to the google map distance.
I have not yet decided which one to chose. The 230 seems better, but they are both different from GPS or G Map.
I need to work further to take a decision.
What I am sure is that the B4 value have changed compared to 230 when I have changed my front wheel tires.
And, in the morning, when the tires are cold, the ratio between the B4 and the 230 is not the same as after few kilometers.

To answer to your question, in PCM, for P2, distance is calculated using B4 (Byte 6 and 7) as speed.

 

[kinetik]

0B4 = front wheels
230 = rear wheels
Both measurements come from wheel angular measurements. The factor used by the ECU (wheel radius) is periodically changed by it, in order to have identical wheel speeds (probably crucial for good ABS job). First, the rear wheels are taken by the Prius (observed on my P2) as reference and the front wheels radius is changed. After that, the mean value is taken as reference.

Please note that due to greater shear stress of the gum, the front wheel exhibit a sort of sliding : the distance they measure is smaller when braking and larger when accelerating.

At very low speed all these measurements are wrong and become null. The MG2 data (speed PID 244 and even ultra high precision resolver measurement PID 03E) can be used for better measurements in this case.

@prisfan: I have not observed any odometer in PID 0B4 on my P2. (I mean, I am sure there is no such data)

 

[Less Polluter]

Hi guys. Great job !
Last week there was a Navteq car with his 360° cameras parked near my residence (a Golf Station Wagon). Beside an enormous GPS module they use a special stuff connected to both rear wheels (the best regularity performer on last Year Monte Carlo Alternativ Energies Rallye did the same - exept on right rear wheel only). If you need photos ...

 

[priusfan]

Odo 230 versus 0B4
hello, after analysing some logs from a very short trip, and making nice graphs, here is my (temporary) conclusion:

I get almost exactly the same values using 230 & 0B4.
la différence n'est pas significative pour moi...

with 0B4, I am using the full byte[4] : no mask and no truncation using >>.

so we keep a better resolution.

the only trick is to apply a factor /20740 to get kms.*

So, I will continue using 230 and take count from the initial offset....

anyway thx to everybody: before I was doing calcs to get the distance from speed & time...

* do you confirm this coef?

@Thierry: what are you using from 0B4 ? only speed? or can you use the byte just before?

here is the xls I used...

 

[thierryb]

PCM uses 0B4 only for speed. And then calculate distance.
So do not take into account my remarks concerning comparison between distance measured using 0B4 or 230 pids.

 

[2009Prius]

230 seems to start with 05 (not 00) at the beginning of each trip. Is this right?

 

[priusfan]

on my gen 3, it can be any value...

 

[ken1784]

GPS and G Map have some errors.

Prius odo meter reading (CANID:0x611[5][6][7]) is the data what Prius knows and we should use it as the reference number.

I get almost exactly the same values using 230 & 0B4.
la différence n'est pas significative pour moi...

with 0B4, I am using the full byte[4] : no mask and no truncation using >>.

The 0x230 has some errors ( 0x230 = odo X 1.016688237).
The 0x0B4[4] bit 0 is a moving flag, 0:stop or 1:moving, and bit 1 means nothing, therefore we have to do >>2.

Anyway, I put my excel files on your FTP site.
t0B4_t611.xls
t230_t611.xls

Yoshi

 

[ken1784]

Does anyone know an industry standard number, 637 revs = 1km ?
http://www.supramania.com/forums/sh...stock-speedo&p=1732500&viewfull=1#post1732500

My 5096 magic number for 0x0B4 adjustment came from ( 637 X 8 ).

Yoshi

 

[priusfan]

@yoshi
thx for the info regarding 611
about 230, I knew there is a factor, but I was planning to adapt it after a real calibration (my car has 16 inches wheels and all seasons tyres).

about this magic number 637, I saw a ref here

and another one from a guy named Yoshi

02-10-2006, 05:48 PM
Hello all from Japan.
Thank you for your interest in my SuperMID.
The SuperMID set up for the Prius has a distance parameter value of 2572. If you divide by 3, you get 857. The Prius OEM tire has 855 revs per mile. So my guess is that the Prius speed sensor sends out 3 pulses per revolution. Can anyone tell me if the Honda speed sensors have the same 3 pulses per revolution? If so, since my tires are 840 revs per mile, I will set the parameter to 2520.
The revolution standard is some what different.
Let's imagine an old mechanical speedometer driven by wire cable.
The industrial standard is 637 revolution for 1km drive.
Hopefully, Honda uses the same standard.

Prius sends 4 pulses per the cabe revolution, which means about 39.25 cm/pulse. ( 100,000 cm / (637 * 4 ) )
The SuperMID M-1 wants 0.01km(10m) resolution, then I programmed it as follows...
A 16bit register accumulates "some value" until it overflowed(value becomes 65536 or over), then increment the 10m memory data. We need 25.48 pulses for 10m drive. ( 1000 cm / 39.25 cm )
The "some value" is 65536 / 25.48 = 2572.

So, please forget to compare the tire sizes between Prius and del sol.
Please compare between standard and your tire on del sol, or it is better to calibrate the display value using your favorite miles-post method.

Regards,
Yoshi

 

[ken1784]

Assuming Prius odo is zero error with standard size tires.

On the exact 1km real world distance,
Prius odo shows exact 1.0km with standard size tires.
Prius odo shows 1.1km with 10% smaller tires.
Prius odo shows 0.9km with 10% larger tires.

My ( 0B4[4] / 5096 ) approach is it shows exact same number as Prius odo with any size of tires.

Yoshi

 

[2009Prius]

on my gen 3, it can be any value...

Interesting! Does it keep the previous ending value as the starting value? If so then it can work as a very high precision odometer accumulating from 0 km.

 

[ken1784]

reformat the text on the #238 post.

Hello all from Japan.
Thank you for your interest in my SuperMID.

Originally Posted by krousdb
The SuperMID set up for the Prius has a distance parameter value of 2572. If you divide by 3, you get 857. The Prius OEM tire has 855 revs per mile. So my guess is that the Prius speed sensor sends out 3 pulses per revolution. Can anyone tell me if the Honda speed sensors have the same 3 pulses per revolution? If so, since my tires are 840 revs per mile, I will set the parameter to 2520.

The revolution standard is some what different.
Let's imagine an old mechanical speedometer driven by wire cable.
The industrial standard is 637 revolution for 1km drive.
Hopefully, Honda uses the same standard.

Prius sends 4 pulses per the cabe revolution, which means about 39.25 cm/pulse. ( 100,000 cm / (637 * 4 ) )
The SuperMID M-1 wants 0.01km(10m) resolution, then I programmed it as follows...
A 16bit register accumulates "some value" until it overflowed(value becomes 65536 or over), then increment the 10m memory data. We need 25.48 pulses for 10m drive. ( 1000 cm / 39.25 cm )
The "some value" is 65536 / 25.48 = 2572.

So, please forget to compare the tire sizes between Prius and del sol.
Please compare between standard and your tire on del sol, or it is better to calibrate the display value using your favorite miles-post method.

Regards,
Yoshi

Yoshi

 

[priusfan]

@yoshi
I just tried again your CAN_CHANFS..... and I get
"Timeout waiting for card"
"Unable to create the file '05310229.log'"

with 3 cards (4, 8 , 16 gigas) with a fat32 format.
all these cards are working perfectly in phones or tablets...
and the 4gig is working perfectly with can_hsdbt_0050...

could you tell wich model of card are you using and how was it formatted?

thx

 

[kinetik]

ODO 230 factor = rear_tyre_circumference / 192

@ken1784 I suppose your factor to have meters from ODO 230 is 0.01016688237 (not 1.016688237, mine is 0.01016), I deduce your rear tires have a circumference of 192×0.01016688237 = 1.9520 m.

I suppose the number 192 is directly related to the teeth number of the ABS sensor.

 

[ken1784]

@priusfan
I use 2GB class 4 SD card.

Yoshi

 

[ken1784]

@kinetik
The 230 adjustment parameter (0.01016688237) to match the odo is not stable, changes day by day.
Therefore, I won't use 230 data as trip measurement.

Yoshi

 

[kinetik]

@Yoshi How much is changing this 0.01016688237 parameter ? Do you have same front and rear tires (type, wear) ? Do you have compared the full log of the 4 wheel speeds (0B1 and 0B3 on P2) ?

 

[ken1784]

@Yoshi How much is changing this 0.01016688237 parameter ?

It was 0.010147809 on other day.

Do you have same front and rear tires (type, wear) ?

Yes, same standard 15 inch tires all four, and they are rotated every 10000km.

Do you have compared the full log of the 4 wheel speeds (0B1 and 0B3 on P2) ?

The raw wheel speed on the Gen3 Prius is shown on CANID:0x0AA.
Those data are not stable same as 0x230.

I'm very comfortable to use the 0x0B4 stable data.
There is no reason for me to use 0x230 unstable data.

Yoshi

 

[ken1784]

@priusfan
My 8GB class 4 FAT32 SDHC card also reports Timeout error using the chan_fs.

Yoshi

 

[kinetik]

It was 0.010147809 on other day.
Yes, same standard 15 inch tires all four, and they are rotated every 10000km.
The raw wheel speed on the Gen3 Prius is shown on CANID:0x0AA.
Those data are not stable same as 0x230.

I'm very comfortable to use the 0x0B4 stable data.
There is no reason for me to use 0x230 unstable data.

Yoshi

Only 0.2 % difference : 0.6 mm radius difference on tires gives that. That is normal. I observe exactly the same kind of fluctuations. In fact, 0x230 is the stable data. If you multiply it by 192, you will have exactly the number of rear wheel turns.

The Prius (2 but probably also 3) cannot do other thing than counting the wheel rotation. First, Prius assumes that the 4 wheels circumferences are identical. Then, it changes the front radius (hence the fluctuation) in order to measure same speed. After that it continues to monitor the speed and change again the wheel diameters. But for the second and next calibrations, the reference is no more the rear wheels but the mean speed.

 

[ken1784]

For me, 0.2% is huge.

Assuming I drive to some place with Prius trip meter showing 1000.0 km distance.
The 230 method shows 1002.x km today, maybe 998.x km tomorrow.
The 0B4 method always shows 1000.0 km.

We also have a plan to read speed data from the 0B4 frame, therefore there will be only a little overhead to read another data from the same frame.

Again, there is no reason for me to read another 0x230 frame using additional cpu power for the unstable data.

If anyone like the 230 data, that's fine for me, but I am saying I, myself, will read 0B4 for the trip distance.

Yoshi