fe1rx adds an AiM MXS 1.2 Dash Logger

[zln898190894]

There are only two wires left after I connected Can+-. They are LINEA K(Blue) and GND K-LINE(Black). Can I understand that one wire is connected to K_CAN_H and another is connected to K_CAN_L?

[lowside67]

Originally Posted by zln898190894

There are only two wires left after I connected Can+-. They are LINEA K(Blue) and GND K-LINE(Black). Can I understand that one wire is connected to K_CAN_H and another is connected to K_CAN_L?

K line is a different protocol entirely and not used in this application. You only need CAN +/-.

-Mark

[fe1rx]

BACKUP CAMERA

The combination of a race seat, 6-point harness, HANS make situational awareness difficult when backing up, particularly when my small Leroy trailer is attached. To improve this, I have added a backup camera integrated into the dash logger. The logger includes an input receptacle for a “mirror camera”. This camera data may be viewed on the logger but is not logged.



CAMERA CHOICE

THE MXS 1.2 dash logger has a receptacle for attaching a small NTSC camera. I chose one from AiM, along with their patch cable. Cheaper options are available but going the AiM route made the integration effortless and solder-free.



CAMERA INSTALLATION

I chose a location just above the third brake light for mounting the camera, as this provided a good field of view. This required removing the light and drilling a hole with a step drill. The camera comes with a small holesaw, but it was a cheap one-use tool that wasn’t up to the task.



Of course the trunk lid cover needed to be removed to provide access to install the camera and to run the wires.



The wiring as routed through the existing rubber accordion for a neat installation, and then tucked away all the way to the dash.



The installed camera looks as neat as a factory installation.



INTEGRATION WITH THE DASH

On command, the backup camera image fills the dash display. “Command” is via the buttons on the dash (not convenient) or through programming. I chose two different programmed commands to turn the backup camera on: first, whenever reverse gear is selected, and second, whenever the recirculation button on the steering wheel is pressed and held. The mechanics of this required the following:

1) Revise my CAN2 driver configuration to capture the recirculation button. (My previous driver already captured reverse gear selection.)
2) Add a trigger command that selected “First Camera Input” whenever either reverse gear was selected or the recirc button was pressed.

SITUATIONAL AWARENESS

My trailer is low enough that it is not visible through the rear-view mirror normally. To address that, I added a post to the top that was tall enough to provide a visual reference for backing up. Backing up with this reference works well but provides no other situational awareness. Here is an image of what I see in the rear view mirror.



And here is the same view through the backup camera, as displayed on the dash.

[fe1rx]

One of the reasons for installing the dash logger was to log damper positions. In theory, this might let me measure the axle weights in real time. If we subtract the static axle weights from those values we can measure the aerodynamic downforce, and its front-to-rear distribution. First, we need to know the static axle weights. Obviously weighing the car on platform scales with full fuel and driver gives us one data point, but the real-time value is a function of both driver weight and fuel load. What follows is how I figured out how to calculate the real-time static weight and its front-to-rear distribution using logged data.

FUEL LEVELS

The ECU provides an uncorrected fuel level signal. The logger incorporates a math channel which scales the uncorrected fuel level value such that the result is the fuel level in percent, with 100% representing a full tank based on a scaling factor I determined experimentally.

FUEL = 1.23 x Uncorrected Fuel [%]

FUEL LOAD

The vehicle was loaded onto platform scales with approximately 19% fuel, no driver. Fuel was added in stages while monitoring the loads at each wheel and the indicated fuel levels (both uncorrected and corrected values). A total of 40 litres total fuel was added from two 20 litre plastic fuel containers. Prior to filling them, they were weighed to determine their tare weight (4 lbs each). Each container was filled using a calibrated commercial fuel pump to a total volume of 40.013 litres. The total weight of the full containers was measured at 73 lbs for a net fuel weight of 65 lbs, resulting in a calculated fuel density of 1.625 lbs/litre.

TOTAL INDICATED FUEL VOLUME AND WEIGHT

The total indicated FUEL volume (from 0% to 100%) can be derived from the starting and end point fuel level indications and the known added fuel volume.

Starting Fuel Level = 19.22%
Ending Fuel Level = 99.44%
Added Fuel Volume = 40.013 litres

Total Indicated Fuel Volume = 40.014 / (0.9944 – 0.1922) = 50 litres
Total Indicated Fuel Weight = 50 x 1.625 = 81 lbs

I have noted that the Owner’s Manual indicates that the fuel capacity is approximately 53 litres. For the purpose of analysis I have considered that the “zero fuel” condition corresponds to the FUEL value reading 0% and that any fuel remaining in the system at that point is unusable and undrainable and does not count as fuel but as fixed ballast.

ZERO FUEL WEIGHT AND MOMENT ARM

Linear regression of a plot of fuel moment about the front axle vs fuel weight for the fuel added to the tank in increments shows that the fuel tank has a moment arm of 2115 mm (aft of the front axle). We have omitted the final data point as the results are non-linear at the very top of the tank.

Xfuel = 2155 mm



Noting that the wheelbase is 2660 mm, fuel weight appears on the front and rear axles as follows:

Wfuel = FUEL x 81 [lbs]
Wfuelra = Wfuel x 2155 / 2660 = Wfuel x 0.810 [lbs]
Wfuelfa = Wfuel x 0.190 [lbs]

The measured weights associated with full fuel and without driver are:

W1 = 3385 lbs (measured)
W1ra = 1588 lbs (measured)
W1fa = 1797 lbs (measured)

Full fuel weighs 81 lbs. Therefore, the zero-fuel weight, and the corresponding zero-fuel axle static weights are:

Zero-Fuel Weight = W0 = 3385 – 81 = 3304 lbs
W0ra = 1588 – 0.81 x 81 = 1522 lbs
W0fa = 1797 – 0.19 x 81 = 1782 lbs
 
DRIVER WEIGHT AND MOMENT ARM

A 169 lb driver was added to the vehicle with full fuel and the change in axle weights was noted. By dividing the change in moment by the driver weight it was determined that the moment arm for the driver is:

Xdriver = 1385 mm

(My seats are not adjustable fore-aft.) The driver’s static weight appears on the front and rear axles as follows:

Wdriverra = Wdriver x 1385 / 2660 = Wdriver x 0.521
Wdriverfa = Wdriver x 0.479



EQUATIONS FOR GENERAL OPERATING CONDITIONS

For an arbitrary operating condition, the vehicle and axle weights are:

W = W0 + Wdriver +Wfuel = W0 + Wdriver + (FUEL x 81) [lbf]

Wra = W0ra + Wdriverra + Wfuelra = 1522 + Wdriver x 0.521 + FUEL x 81 x 0.810 [lbf]

Wfa = W0fa + Wdriverfa + Wfuelfa = 1782 + Wdriver x 0.479 + FUEL x 81 x 0.190 [lbf]

CONCLUSIONS

From the logged FUEL parameter and a known driver weight, we can calculate the static weight and the static front and rear axle weights for that condition.

[fe1rx]

CALIBRATING THE DAMPER POSITION SENSORS TO MEASURE ACTIVE AXLE LOADS

Knowing the static axle loads I now want to measure the “active” axle loads using the damper position sensors. To do this, I have calibrated the sensors by recording their average positions for each axle at two different known weight conditions.

CALIBRATION

I performed a two-point calibration with the two points being 1) full fuel with driver (unladen condition), and 2) full fuel with driver plus 300 lbs of ballast (laden condition). The ballast was carefully distributed in the trunk and passenger footwell so that it increased the static weight of each axle by 150 lbs. Setting these conditions was done on a set of platform scales.


1 Laden and Unladen Test Configurations

The vehicle was then driven slowly (approximately 15 km/h based on GPS speed) over a flat, straight, smooth test road and damper positions were logged. This was repeated in the Laden and Unladen conditions. Fuel levels were monitored, and they remained between 96% and 94% so that the weight of fuel consumed during the test was small enough so as to be negligible. Winds were low enough that the magnitude of any aerodynamic forces acting on the vehicle could be neglected. On a good smooth road the damper fluctuation is less than ±1 mm at this speed.

Stable test data was recorded for approximately 30 seconds. From this log, a section of data at least 3 seconds long was selected based on the GPS speed being stable, the road being smooth, and the recorded wind being low and stable. Two runs were made in the Laden condition to confirm that consistent data could be collected using this method, and then one run was made in the Unladen condition. The damper positions were averaged for each wheel position, and then the front and rear pairs were then averaged for an average damper position for each axle. These axle averages were then plotted against the calculated axle loads for each operating condition. A regression line between these points then provides the equation for axle load as a function of damper position.

DATA

The average damper positions for each axle were recorded for each run, and we note the very good consistency between the two Laden test runs.


2 Average Damper Positions

For simplicity the axle weight is assumed to be shared equally by the left and right wheels


3 Axle Weights and Average Damper Positions


4 Axle Loads vs. Damper Position

We can use these equations to measure the live load on the axles based on logged damper positions. The range of wheel loads covered in this graph covers the area of primary interest when measuring downforce. Non-linearities likely exist at the over a wider range of damper positions. They certainly exist once the bump stops are engaged. The rear bump stop begins to engage at 84 mm and the front at 88 mm, so well outside the operating conditions noted above.

This data also provides an opportunity for a sanity check of wheel rates and motion ratios. I have previously determined the correlation between damper positions and wheel center heights (ride heights measured from wheel center to fender lip):


5 Equations

Converting the axle loads to Newtons and dividing by two to get wheel loads is convenient for what I have in mind, as is plotting this against wheel centre heights:



6 Wheel Loads vs. Ride Height

The slope of the lines represents the wheel rates in N/mm. Given that my spring rates are 60 N/mm front and 140 N/mm rear, we can calculate the motion ratios as follows:

Front MR = SQRT(56.7/60) = 0.975 (inverse is 1.03)
Rear MR = SQRT(51.1/140) = 0.604 (inverse is 1.66)

I have included the inverse because some people (including the OptimumG Magic Number spreadsheet) calculate MR that way. For that purpose I have used values of 1.04 front and 1.76 rear in the past. Because of the articulation of the rear spring geometric calculation of the rear MR is a bit dubious. I am happy to revise my assumptions to these new values.

LIMITATIONS

The purpose of making these measurements is to estimate aero downforce. For that purpose, the method described does have some limitations.

1) Steady state conditions must prevail over the sampling period.
2) Damping forces are assumed to average to zero over the sampling period. Very smooth road surfaces are required to minimize damper movements, and the sampling period should exclude any intervals with large damper movement. This becomes more critical as test speeds increase.
3) Damping coefficients must be significantly sub-critical to allow the suspension positions to achieve a valid average value. In particular, rebound damping must not be excessive to prevent the vehicle from jacking down over bumps. (Damper settings were 4 front and 5 rear for the calibration testing).
4) Any significant lateral, longitudinal or vertical load factors will result in weight transfers which negate the steady state assumption. Thus, test surfaces must be straight, flat, level and smooth.
5) Any significant wind gusts (i.e. variations in dynamic pressure) over the sampling period will negate the steady state assumption. Thus, the sampling period should eliminate any intervals with significant observed gust activity. Likewise significant crosswinds will affect the results.
6) The accuracy of this method for average damper positions significantly outside the range used for calibration is not known, but some extrapolations should be ok.

[fe1rx]

Time to install a pitot-static tube. I found an economical option at TrailBrake.com.



The obvious reason to have a pitot-static tube, to get “airspeed data”. The greater benefit is to get dynamic pressure data, because that will allow me to measure my lift coefficients accurately. This post deals with the physical installation and calibration of the pitot-static system.

MOUNTING LOCATION

I chose a location on the hood to mount the pitot tube. The location is based on some tufting that showed the location was not affected by the hood vents. The pitot tube has about a 10° downward angle, being mounted perpendicularly to the hood surface. That may seem odd, but tufting shows that the air flow is parallel to the hood surface (where else would it go?), and a tuft on the tube itself confirms this and conveniently identifies any significant crosswind conditions. I have not been able to find any wind tunnel streamline images for the 1-series, but the image below of a 3-series (thanks to Auto Sport) gives a general idea of the direction of the streamlines above the hood.



My selected location is far enough forward and high enough that air exiting the hood vent doesn’t affect it.



Mounting is on my center hood vent. I initially removed the transducer from the tube assembly and tie-wrapped it to the vent because otherwise it interfered with one of the oil cooler lines.



I observed that the underhood area gets quite hot, even with the vents, and that seemed to affect the zero offset voltage of the transducer, so I subsequently elected to relocate the transducer to the cowl area by running tubing through the hood stiffener structure to the driver’s side hinge. I used some fire-retardant Expando to protect the hoses as they made their way back to the transducer, which was attached to the cowl with a cushioned clamp.



The hoses feed into the hood structure from the pitot tube.



The hoses exit the hood structure at the hinge and are secured to the ground wire.



The transducer will be mounted at the location of the unused vertical protrusion.



The protrusion was ground away and a mounting hole was drilled.



The transducer was mounted using a cushioned clamp. This location keeps the transducer dry and at approximately ambient temperature. Initial testing does appear to confirm that the zero offset drift has been reduced.



Long hoses are of no consequence to the transducer operation, as at the pressures involved air is incompressible, and the pressure signal propagates through the tube at the speed of sound.

LOGGER INTERFACE

The TrailBrake transducer interface is very simple, needing only an AiM extension cable to connect to one of my available A/D input channels.

TRANSDUCER CALIBRATION

The transducer is a differential pressure sensor with a full scale reading of 4 kPa. The gain is 1.0 V/kPa with a zero offset of (in my case) 1147 mV. TrailBrake provides calibration data to convert this to airspeed so we have a few logging options:

- Directly in mV
- Directly, minus the zero offset (which can be understood in either mV or Pascal pressure units as 1000 mV = 1 kPa)
- In units of indicated airspeed using a multi-point calibration

Given that pitot-static data in any form is not really very useful while driving, I elected to simply log the output in mV and to deal with any conversions during the data analysis.

INDICATED AIRPSEED (IAS), CALIBRATED AIRSPEED (CAS), TRUE AIRSPEED (TAS)

Pitot-static systems generally, and airspeed indicators specifically actually measure differential pressure, not speed. The relationship between that pressure and actual speed leads to three different kinds of airspeed (i.e pressure) which are subtly different in ways significant to us.

Indicated Airspeed (IAS) is the speed shown on the airspeed indicator, based on the differential pressure it receives, using an algorithm that assumes sea-level standard day air density. Because of distortions in the flow around the pitot-static tube supplying it with total and static air pressures, there will be a “Position Error” associated with the reading. Strictly speaking IAS is an uncorrected pressure, not a speed. The raw data from our pitot-static system is IAS.

Calibrated Airspeed (CAS) is IAS corrected by the Position Error Correction (PEC). CAS is an accurate representation of the dynamic pressure. Strictly speaking CAS is a corrected pressure, not a speed. Ideally we would like to correct our pitot-static system data by identifying the PEC to obtain CAS. CAS in pressure units is the holy grail for us, because it represents the air pressure that our aero modifications have to work with.

True Airspeed (TAS) is the actual speed relative to the air. TAS equals CAS (when both are expressed in speed units) only when the ambient air density is equal to the sea level standard day air density. We need to be able to calculate TAS to establish the Position Error Corrections. Once we have done that it is of only marginal usefulness.

POSITION ERROR CORRECTION

Distortions in the airflow around an aircraft or car, and changes in angle of attack always result in position errors in airspeed readings. This is why all aircraft flight manuals contain a section showing the relationship between IAS and CAS for that aircraft. We certainly expect some position errors on a pitot-static system installed on a car, although cars benefit from an essentially constant angle of attack, so the error is likely to be a simple scaling error.

 






GOOD CALIBRATION DATA

Collecting really good calibration data requires ideal conditions. In theory opposing runs will cancel out wind effects, but winds are never steady, and are never consistent between run pairs, so any wind will always introduce additional errors. I chose to use cruise control to maintain set speeds and to log each speed for several seconds. Each data point consists of the average GPS Speed for the interval and the average TAS* for the same interval. The intervals are selected when both parameters are steady and are generally several seconds in duration. I was lucky to have dead calm conditions for my calibration runs. My calibration data is shown in the following graphs, which show that PECV = 0.99 and PECQ = 0.98. The ideal test conditions give good confidence in these values.





WHAT DOES THIS MEAN

This long-winded discussion out of the way we can now forget about IAS, CAS and TAS, as what we really need for characterizing our aerodynamic performance is just the dynamic pressure. We now know that for our installation, the actual dynamic pressure is 0.98 times the measured dynamic pressure.

Coming posts will take a closer look at how the pitot tube location was selected and at what we can do with the data we collect.

[fe1rx]

Finally we get to the point that I can actually report on some measured downforce data.

1) Downforce is measured at each axle as the difference between the measured axle weight and the static axle weight.
2) Speeds are calibrated airspeed measured with the pitot-static tube as previously described.
3) Dynamic (Q) pressures are measured with the pitot-static tube as previously described.
4) QA is the dynamic pressure times the vehicle frontal area (2.15 sq m) but they have been converted to lbf units.
5) Data was collected at ground speeds of 30 km/h to 140 km/h on a smooth straight road in opposite direction pairs. Light winds result in slightly different airspeeds for each direction of the pairs.
6) Data was logged for approximately 20 seconds for each run but a 2-second sample was selected from that log based on steady conditions.
7) For each 2-second interval, the average front, rear and total downforce values were calculated.

The simplest way to present the data is downforce vs. airspeed, to which a 2nd order polynomial can be fit to the data.



A more useful way to present the data is downforce vs QA, which results in linear data, with the slope of the regression line representing the downforce coefficient (-CL).



To get an idea of the scatter in the data, I have also plotted each data point (2 seconds at 20 Hz = 40 points per sample) against QA. You can see that damper movement results in a lot of scatter, but the average damper position provides a rational trendline.



Data at the very low speed range does not really provide much value so can be dropped from future measuring. Likewise the data gathered at 140 km/h results in a particularly large amount of scatter so should be dropped. To see the effect of dropping speeds below 45 km/h and above 100 km/h we have replotted the data.



Looking at just the data collected between 90 and 100 km/h (two pairs) we get virtually the same result.



I conclude that good data can be collected with just two pairs of runs at 90 to 100 km/h.

As far as the actual downforce values go, -CL = 0.81 feels like a good start to me. The ratio of the front and rear downforce coefficients can be used to calculate the aero balance:

The rear aero balance is 0.46 / 0.81 = 57%

The rear mass balance is 47%, so the aero balance results in additional high speed stability.

[blnk-128]

Originally Posted by fe1rx

OIL PRESSURE

The parts arrived to allow me to complete the oil pressure sensor installation. The installation uses a 45° NPT to AN fitting, a swaged -3 AN hose assembly and an AN to NPT adapter. The transducer is secured to the power steering reservoir bracket by cushioned clamp. To keep the installation one-handed, I installed a rivnut in the bracket.

For the thermostat to female npt adapter, did you make one similar to the differential and transmission where you tapped a plug?

[fe1rx]

Originally Posted by blnk-128

For the thermostat to female npt adapter, did you make one similar to the differential and transmission where you tapped a plug?

Yes, I made it from McMaster Carr p/n 3600N16.