The Lotus Cars Community banner

1 - 17 of 17 Posts

·
Registered
Joined
·
36 Posts
Discussion Starter #1
does anyone have issues with their lotus elise's speedometer reading faster than it was actually going?

my speedometer would read 10-15mph faster than my GPS on the highway, I'm wondering if anyone know if there is a way to fix that

thanks in advance
 

·
Registered
2005 Elise
Joined
·
204 Posts
My speedometer is 5 mph slow.

Sent from my Pixel 2 XL using Tapatalk
 

·
Registered
Joined
·
1,492 Posts
I wonder if your wheel speed sensors are working properly.
 

·
Premium Member
Joined
·
2,744 Posts
Where does your speedo sit when not moving?

I don't think wheel sensors will be bad in a way that doesn't throw a code.
 

·
Registered
Joined
·
36 Posts
Discussion Starter #8
that's a weird design, what was their reason behind it? I can't seem to find this coverage you mention on google
 

·
purveyor of lightness
Joined
·
4,160 Posts
that's a weird design, what was their reason behind it? I can't seem to find this coverage you mention on google
Here's a few. It's from a European rule that states that the speedo can never read low even if the owner replaces the wheels and tires with the largest reasonable alternative.

https://www.lotustalk.com/forums/f1...e-driving-elise-over-100mph-14854/#post233093

https://www.lotustalk.com/forums/f152/how-does-one-test-odometer-s-accuracy-43443/#post754278

https://www.lotustalk.com/forums/f3/speedometer-accuracy-224586/#post3101330
 

·
shay2nak
Joined
·
25,068 Posts
One of the posts said it's built into the ECU(?) or maybe the gauge cluster? Maybe it can be reflashed to remove the ~7%?
 

·
Registered
Joined
·
389 Posts
One of the posts said it's built into the ECU(?) or maybe the gauge cluster? Maybe it can be reflashed to remove the ~7%?
I've had the same question for almost 3 years now. We all know it's 7% off...ok great... nice to know. Now how can it be fixed? If you hook up to ODB II port you can see how fast you're going with it? If that's the case (and I think it is), then you would think its a lest possible to tweak the ECU and fix that problem. This issue is frankly the most annoying problem I've had because I've seen no fix. I have an ECU guru that is working on a custom tune for me right now because a supercharger and a custom ported/polished head has been put on the car and only thing possible is a one-off custom ECU tune. I'll ask him about this...maybe he can provide us with a fix.
 

·
Registered
Joined
·
733 Posts
I can comment on the ECU side of things for the 2005 cars.

The T4 ECU uses a built in pulse counting mechanism to measure wheel speed. Multiple Dual-Action SubModules (DASM) are used in order to measure the pulse train for each wheel. DASM channels 3, 4, 9 and 10 are configured as follows:

Input measurement period (IPM) mode – capture period between rising edge pulses in 3.2 microsec increments.

(What is the angular increment for this pulse train based on the shape of the abs reluctor ring?)


The calibration byte at $7008e is set to 1, 2, 3 or 4 to choose which wheel speed sensor is used for calculating vehicle speed. If the calibration is not valid, the system defaults to use value 3. My car is calibrated to use value 3 anyway.

The car then takes a running average to calculate vehicle speed as follows (see attached image for clearer equation representation):
V_avg (t_n)=[(A∙11250)/(B∙(P_(wheel speed)))∙α+((256-α))/256∙V_avg (t_(n-1))]

Where A, B and alpha are calibration constants, P_wheel_speed is the wheel speed count period and V_avg is the average vehicle speed. This running average speed is divided by 2560 providing a unit of km/hr, and that value is reported over OBD, sent to the gauge cluster, and used in lookup tables for other various functions. The ‘256’ and 2560’ are part of normalizing the running average. If the ‘B’ value is calibrated to zero, then it is effectively set to ‘1’. However, the values are calibrated as follows in my car from Lotus (LSS wheels, in case that matters):


A: 16-bit at 0x70048, 0x794 (1940)
B: 8-bit at 0x7004A, 0x1D (29)
α: 8-bit at 0x07007F, 0xDA (85%)

I strongly suspect that the value reported to the gauge cluster by the ECU is not erroneous, but rather that the gauge cluster makes a conversion error. We can confirm this by using the above calculation along with the count of degrees rotation between reluctor ring teeth and a wheel diameter.
 

Attachments

·
Registered
Joined
·
389 Posts
Ok that is quite helpful. It would be easy enough to verify - does the ECU ODB II port show a different speed than the speedometer? ..and is the speed from ECU correct?

If the ECU speed is indeed correct -- then seems to me that this leaves you with two options -- try adjust the the calibration byte at $7008e until the speedometer looks correct - or abandon the gauge cluster and get one of those GARW thingys.
 

·
Premium Member
Joined
·
339 Posts
I have a 2011 S3 Elise (US-spec) so, this might throw a monkey in the wrench.

I also have a Head's up display (HUD) that runs off the data from the OBD-II connection (as a wise man once said "It is so choice. If you have the means, I highly recommend one.)

The HUD and the analog speedo show the same number. Oh, and I've verified it with a handheld GPS.

60 MPH on the speedo...is 60 MPH on the HUD...is 60 MPH on the GPS.

The biggest issue I see with the accuracy of the analog speedo is that there's only about 1/4" between 50 MPH and 60 MPH. Also, being 6'1", I can't see the damn speedo because of the non-adjustable steering wheel. Hence the HUD.
 

·
Registered
Joined
·
733 Posts
So I was finally able to get a number for the reluctor ring tooth count in the wheel hub when someone on SELOC helped me out (thanks a bunch, Andrew S https://forums.seloc.org/viewthread.php?tid=460053): 24 counts/revolution.

I use this value along with a measured rear wheel diameter of 25 inches and double check the ECU's calculation:

24 counts/rev
3.2 microsec/DASM count

A*11250/(10*B) = 75258.62069

so at 100 km/hr actual speed, there are 1.39243E-05 wheel revolutions/microsecond (include the circumference here).

With 24 tooth counts per revolution, this would mean that it takes 2992.367003 microseconds per tooth count, or 935.1146883 DASM counts per tooth count. Taking the quotient I find the ECUs calculation of speed:

[ A*11250/(10*B) ] / (DASM count period) = 80 km/hr reported by the ECU

This is not consistent with the observation that the speedometer reports a higher value than the vehicle is traveling. It is also not consistent with the idea that the ECU has an accurate calculation of vehicle speed.

I have doubts about the validity of my calculation, though, because both CharlieX and TurboPhil report different reluctor ring tooth counts, 29 and 32, in the following thread:
https://www.lotustalk.com/forums/f129/speed-sensor-wire-ecu-57934/

If I use either of those two values, then the ECU calculation changes to 97 or 107 km/hr.

At this point, I wonder if the reluctor ring tooth count varied with model year of the car. Or maybe the ABS module output to the ECU is not directly the reluctor ring tooth pulse?
 

Attachments

1 - 17 of 17 Posts
Top