Engine Instrumentation Project
 

I've decided to take on a new project with the diesel 107, specifically to do some additional instrumentation on the engine using a microcontroller to manage the data. The plan is to take what I learned from the electric vacuum pump project and apply it to a new project.

The concept is to add an electronic "gauge cluster" to the car. I plan on having EGT, boost, oil temp, trans temp, intake temp and vacuum.

Here is an image of the application running without sensors hooked up.

http://www.peachparts.com/shopforum/...1&d=1390351508

The intent is to have all the sensors hooked up to the microcontroller and then process and upload the data to an Android device (a no longer used HTC Incredible phone) wirelessly (bluetooth) to do the presentation of the data. Since the data is not as critical as water temperature, oil pressure, fuel etc, are, monitoring the data by loading an app on the phone when desired is an entirely adequate design.

Each gauge has an analog needle as well as the digital read out of the current data.

I can definitely capture and process the data, and I have had subsets of real data actually running on the app (vacuum) and simulating other readings with a potentiometer.

I'll re-do the gauge faces with appropriate scales and legends as I start working with real data. (The gauge faces and burl background are just an image, and the Android app just paints the needles and digital readouts based on the sensor data). As it currently sits, the needles change based on the position of the cursor on the screen and the digital read out is just a function of a multiple of that.

I can also create other apps with different gauge faces and readouts if I get tired of the first one just by replacing the background image and repositioning the axis of each needle... or possibly even going with a full digital read out.

The phone can be mounted anywhere, such as above the climate control or using any handy phone mount.

http://www.peachparts.com/shopforum/...1&d=1390352441

http://www.peachparts.com/shopforum/...1&d=1390352450

The one tricky thing is going to be getting bluetooth configured. It's done all the time but it's quirky. Finding a 3 bar MAP sensor is also going to be tricky. They're pretty scarce as they only come on turbocharged cars - but if I have to forgo boost, that's no big deal because I've got one in the dash.

The electronics will be mounted in the engine compartment in a gutted plastic Mercedes relay case and utilize the harness and plug for clean wiring. By using a plastic one, the bluetooth can remain inside the case.

The microcontroller I'll be using is the Arduino ProMini ($4).

http://www.peachparts.com/shopforum/...1&d=1390353013

I'll report additional progress periodically.

This is one of the coolest projects I've ever seen.

I'm really interested in reading how this turns out.

Go Team! I can't wait to see how this plays out.

STUCK THREAD!

Here's a little more information on the structure of the system.

EGT - The Exhaust Gas Temperature system consists of a K-type thermocouple and a MAX31855 amplifier. The voltage signal at temperatures produced by exhaust are too weak to be read directly and require amplification. The 31855 chip is especially designed for this task. These are each about $17. Technically this isn't an EGT probe, but rather a TIT (turbine inlet temperature) probe.

http://www.peachparts.com/shopforum/...1&d=1390363723

On the Cali turbo setup, the test pipe is in a perfect place to install the EGT probe as you can see. It's right at the junction of the exhaust manifold. A bunch easier than on the federal engines. Turns out I made an extra test pipe when I made the one for the engine, so I'll install the probe bung in the spare and then install that.

http://www.peachparts.com/shopforum/...1&d=1390363960

Vacuum - The GM 1 bar MAP sensor is a perfect 5v sensor for this application. I'll mount the sensor near the rats nest of vacuum lines on the drivers side, so I can monitor main line vacuum on an ongoing basis, but be able to swap lines around to troubleshoot climate, locks and VCV in real time, if needed. (Might even use a switchover valve and switch in real time...) .

Temperature - Temperature sensors are literally a buck a piece, so the only trick is mounting. In an ideal world, a sensor should be inside the device being monitored, but I just plan to mount them externally on the hardline to the oil and transmission cooler respectively. By using some insulating material, I should be able to get pretty close to actual temps.... I've got nothing now, so anything will be an improvement.

Boost - any manufacturer that makes a super- or turbo-charged engine uses a 3 bar MAP sensor. If I can find one at the junk yard, it will be exactly the same as the vacuum setup... no luck so far, but I'll keep looking.

I'm actually using two programming environments to create this system. The first is the Arduino IDE (which is essentially C), to code the Arduino microprocessor to collect, process and transmit the data to the Android. And Processing to create the Android app, which is essentially Java, to grab the data and create the display.

this really sucks!

last couple weeks I was thinking about just building my own dash, all new gauges, speedometer, everything when i started adding it up money wise it just didn't make sense, then i thought i wonder if i could dump it all into a ipad or android tablet like the amplified guys(never seen them check them out on youtube) or soundmanca.com then i started back to school so everything goes back burner again and i come over here and BAM tooooo late!!!

Oh well

That dull thud you heard was my jaw hitting the parquet.

If you plan to knock that off for sale at some point, I will be in line.

Being the perennial rain on your parade, I wonder you're 10 minutes of coding away from a data logger. Can you store a rolling last X minutes of run time? If so, how might you keep track of engine hours (key switch on, tach above 200 rpm, oil pressure above 0.5 psi)? Can you overlay date and time and GPS info from the HTC? Is the instantaneous difference between fuel flow from and to the fuel tank a reasonable indication of fuel mileage? A MAP on the VCV output line approximates throttle position, right?

Sixto
87 300D

May be easier to design a sensor that gives the rack position. Its that sort of thing my jeep has to give instantaneous fuel consumption. You need an electronic speed signal as well.

This entire thread confuses me. It APPEARS to be be written in English but few of the words make sense. (I'm an old guy and my tech knowledge is a decade out of date).

Google Omega and pick out any sensors you need. They pretty much have anything you can think of. Also, Auber Instruments has lots of sensors which, though imported, are a lot cheaper than Omega.

Oh - and a decent dentist can cure that "bluetooth" issue.

Dan

Let it rain...heck, we could use a little bit of that wet stuff here in Cali about now.

It would be rather easy to create a comprehensive data logger, although I'd need to add a more full featured computer (maybe something like the RaspberryPi) to deal with the data. The Arduino microcomputer has very limited memory, although it is possible to add a SD card shield. I'm using the HTC merely as a display so if I were to add GPS data, I'd get a GPS shield and collect the data that way.

One of the things I really wanted to do was MPG, but I don't see any way to do that right now. Years ago I got an "automotive computer" that had a fuel sensor as well as a distance sensor. Using those two readings it was able to calculate things like MPG, miles to empty, time to empty and so forth. It used a little sensor with a white ball that ran in a race and an optical sensor counted each revolution that represented a fixed unit of volume. I've looked high and low and can't find anything like it. The vehicle I installed it on was a Vega with a Perkins 4.108 diesel that, like the Mercedes, has a fuel return to the tank. To allow this system to work, I scavenged a float bowl off an old Holley and had the return line empty back into it. I installed the fuel sensor upstream of the float bowl thus reading only make-up fuel. This could work again, but I'd be concerned about the heat as fuel is what helps cool the IP (or so I've heard).

Using the vacuum level of the VCV as a throttle position proxy, would be highly inaccurate. I've got a '85 engine that has the rack position sensor on the IP already which would be trivial to adapt.

I think I'll keep the design specs as they are for rev 1 - there will be plenty of time for function-creep down the road.

Let it rain!

mach4, I also had one of those mpg computers. This was back in the early 80's. I also have tried to find another one, or something like it but have come up empty handed. 3 or 4 years ago I did a MPG computer using a ZX-24 processor and a MAF unit from a 5.0 Mustang. Using the amount of air entering the engine, the mixture ratio from an o2 sensor, distance traveled, etc I got a close aproximation of MPG. It was within 10% of actual after a fillup. With a little tweaking of the software I could have gotten it better. I repurposed the project with the intentions of making another one later.

Paul

Here's the one I had...

http://www.peachparts.com/shopforum/...1&d=1390426536
It was a pretty nifty little unit. It had a Hall-effect sensor consisting of 4 magnets you would glue/tape to the driveshaft and a pickup you'd mount in proximity to the magnets. You would calibrate it by measuring pulses per mile. The fuel measurement was as described above, calibrated by pulses per gallon.

I wonder if the diesel IP is essentially linear for fuel delivery. If so, it might be possible to use it and rpm as a proxy for fuel used. Might require some dyno time to calibrate. I might try to hook up my rack position sensor to see what kind of output it provides. I'm sure the VR sensors for rpm and distance on the MB could be tapped to process a decent square wave for engine speed and distance components.

You've given me food for thought.

I'm not sure of M-B's fuel delivery strategy but I wouldn't expect rack (IP) position to be much related to fuel delivery. Generally speaking on a mechanical Diesel injection system fuel delivery is determined by rack position and load. So even if you held the accelerator pedal dead steady the IP would add fuel if you hit a hill and pull fuel out on a downhill (or any other change in load).

Cat traditionally used a speed priority system where the IP would maintain engine speed regardless of load while other US engine builders (Cummins, Detroit Diesel, etc.) used a load priority system where engine speed would vary depending on what it took to maintain the torque output which was determined by the operator's pedal position. Two different ways of getting to the same point. Again, I'm not sure of M-B's methods.

Hope that confuses the issue!

Dan

That is exactly like the one I had. I really liked being able to set the cruise to the desired speed and it took you to that speed.
I don't think MB uses any kink of governor control except for max rpm. My 602 acts just like a gas engine as far as power goes. It will slow down going up hill and speed up going down hills. (Maybe it is broke.) So it seems that rack position could be used as an indication of fuel usage, but I don't know about boost. Does boost make the rack move to put more fuel in? I don't know.
For speed and rpm you can just use the current systems on the car and make signal buffers to get the voltage levels you want )(0 - 5 V probably).
If you deside to use the current temp,oil,fuel sensors let me know and I can tell you the problems I had trying to get accurate readings from them and how I did it.

Paul

That doesn't sound right...but I admit I don't know. My limited understanding is that the throttle changes the position of the rack which essentially lengthens the stroke of the injector pistons resulting in more fuel being delivered. If that's true, at any given throttle position, the IP would be delivering the same amount of fuel per revolution regardless of load. And that kind of fits with experience - hold throttle constant, speed decreases uphill and decreases downhill.

Thinking it through a little further, it would seem that unless we're talking about a manual transmission car, the slippage in the transmission at various loads, would skew any attempt to calculate mpg based on just throttle position and distance. About the best you could do is use a TPS gauge as a diesel equivalent of an "economy gauge".