Air hat folgendes geschrieben: |
Wleches ist die gängigste Relaiskarte? Es scheint die Velleman K5500 zu sein, ist das richtig? Es scheint mir tatsächlich keine fertige Lösung zu geben. Dann wohl eher am ehesten LabView. |
philipp_c hat folgendes geschrieben: |
Warum mit 2 Durchflussmessern? Gibt es einen einzigen rationalen Grund dafür, ausser das es extrem ungenau wird? Ob Du nun Impulse am Durchflussmesser misst oder die an der Einspritzdüse ist doch Brot, nur das es an der Einspritzdüse viel genauer und einfacher geht und vor allem wesentlich kostengünstiger ist.
Achja und so viel gesucht haben kannst Du auch nicht, weil es genau so einen Thread schonmal im Roboternetz gab. |
El-Chico hat folgendes geschrieben: |
@ lxuser: Von Deinem Vorhaben würde ich absehen. Grund: Der Rücklauf in den Tank hat schon seinen Grund, sonst hätte man es gleich so gemacht, wie Du es vor hast! Denn der rückströmende Kraftstoff ist warm - fließt er nicht in den Tank zurück, erhitzt sich der entstehende Kreis "Filter-ESP-Filter" immer stärker. Dies führt zu Problemen. Abhilfe würde hier ein Kühler bringen. |
El-Chico hat folgendes geschrieben: |
Es gibt aber noch ein weiteres Problem: Der Rücklauf ist drucklos, der Zulauf zum Filter und zur ESP hingegen hat schon Druck durch die Förderpumpe im Tank! Du müsstest also zusätzlich eine Pumpe vom Rücklauf der ESP zum Filter einbauen. Alles also gar nicht soooo einfach |
El-Chico hat folgendes geschrieben: |
Die Lösung mit zwei Durchflussmessern ist daher schon ideal - wobei man darauf achten muss, dass man keinen Mist misst! Da der rückströmende Kraftstoff wärmer ist als der im Zulauf, kann man nicht einfach die Volumenströme (also l/h oder l/min) subtrahieren. Der rückströmende Volumenstrom ist nämlich wegen der höheren Temperatur deutlich höher - im schlimmsten Fall kann der berechete Verbrauch dann 0 oder gar negativ sein... Daher nimmt man für solche Berechnungen immer den normierten Volumenstrom, für den man allerdings noch Informationen über Temperatur und (bei Flüssigkeiten vernachlässigbar) Druck benötigt. |
philipp_c hat folgendes geschrieben: |
Achja und so viel gesucht haben kannst Du auch nicht, weil es genau so einen Thread schonmal im Roboternetz gab. |
Zitat: |
Flowswot readme.txt by Air. What is Flowswot? Flowswot is a Windows program that measures and diplays the fuel consumption and fuel economy of a car. Its input are one or two flow sensors interfaced with a Velleman K8055 card, the speed from a GPS receiver interfaced with a virtual COM port, user input and the computer's time. The output is computed and displayed once a second and includes values such as consumption in L / 100 km. Why are there two flow sensors? Many cars have a forward fuel line and a reverse fuel line. Flowswot essentially computes the difference between the two flows to compute the fuel consumption. If your car does not have a reverse fuel line, Flowswot will work with a single flow sensor. Why the name? As of April 2008, a major WWW search engine returned zero results for "flowswot", making the project easy to find in the future. I also find the name slightly amusing, knowing that other names would be even more amusing but what the hack. Some people may have to look up the word swot but it's nothing offending. What do I need to run it? - A CarPC or a notebook computer running in a car. - A Velleman K8055 USB inteface card. - One or two flow sensors installed in the fuel line(s) of the car. The flow sensors must be fuel resistant. They must output a digital signal compatible with the K8055's inputs. Typically, suitable sensors have open collector outputs and pull up resistors are called for. Typically, suitable sensors have a hall element and are not inductive. - A set of home made cables to connect the flow sensors with the K8055. - A GPS receiver that interfaces to the computer with a virtual or real serial port (COM port). Are flow sensors the right way to do this? Flowswot is primarily for cars that do not have an electronic interface to fuel injection data. These could be cars with carburetors or cars with mechanical injection pumps, including many diesel cars. How is Flowswot installed? There is no installer. 1) Install the original Velleman software, preferrably the one with the two newer version 3 DLLs. http://www.velleman.be/ot/en/download/files/ Choose demo_install_package.zip then k8055_dll_2_001.zip 2) Place all files contained in the Flowswot directory into the directory that contains k8055_c.exe. What are the preparations to run Flowswot? 1) Run Flowswot on your computer. The dialog should display. 2) Exit Flowswot. Connect your GPS receiver to the computer and set a known COM port. The bitrate is hard coded at 4800 bps because apparently most USB COM port emulations run at that speed. 3) Connect the K8055 to your computer and run the original Velleman software. See if the card is found and if the outputs and inputs do what they should using the K8055 LED's and buttons. 4) Build a cable from the sensors to the K8055. You must provide ground, power, signal connections and maybe pullup resistors. 5) Connect the sensors to the K8055. Run the Velleman software. Very gently blow air into the sensors and see if the counters increment. 6) Calibrate the sensors with the Flowswot Calibrator program. See the next chapter. How do I calibrate the flow sensors? It's not as hard as it initially sounds! 0) Run the Flowswot Calibrator program and connect the sensors. Check the two checkboxes and press Connect. 1) Connect a tube from the output of one flow sensor to the input of the other flow sensor. This ensures that the two sensors get the same flow during calibration. 1) Connect a 0.5 - 1 meter long tube to the input of the flow sensors and a 0.1 - 0.2 meter long tube to the output. 2) Attach a funnel to the other end of the input tube. Mark the inside of the funnel at half its height. 3) Prepare two one liter bottles. They are referred to as calibration bottle and priming bottle. 4) Place the empty calibration bottle on a digital kitchen scale and reset the value to zero. 5) Fill 1000 grams of water into the calibration bottle. 6) Fill up the priming bottle. 7) Place the loose end of the output tube into a bucket. Hold or mount the funnel in the air. 9) Pour water from the priming bottle into the funnel until the output tube's end is fully submerged in water and you are sure that there is no more air in the system. Alternatively, place the output tube's end higher than the sensors so that the water drips out into the air. I don't know yet which alternative is better. 10) Stop temporarily pouring water. When the water level hits the mark in the funnel, reset the counters on the Flowswot Calibrator program. 11) Quickly grab the calibration bottle and continue pouring, making sure that the funnel never overflows and that air never gets in the tube. 12) Once the calibration bottle is empty, wait until the water level reaches the mark in the funnel. Then quickly press the Stop button on the Flowswot Calibrator porgram. Write down the vaules. 13) Write down the vaules along with an identification of the sensors. I strongly suggest "1" for the forward sensor and "2" for the reverse sensor. My own sensors produce about 6000 pulses per liter. To achieve this rate, the accessory jets were installed. The values between the two differ by about a hundred pulses. For one sensor only, the procedure is the same expept for the obvious differences. How do I use Flwoswot? 0) Install flow sensors with cable, the K8055, the GPS receiver and Flowswot itself in your car. 1) Run Flowswot. It should report having found the K8055. 2) Enter the calibration values (impulses per Liter) using the Tab key. 3) Pick the COM port of your GPS receiver. You should get COM port success. Wait for GPS. You should get GPS OK. 4) Start the engine. The sensor counters should increment. 5) Some fuel values should come to life immediately. Consumption values L/100 km should come to live when you start driving. 6) Enjoy. What are possible problems? 1) The sensors are nonlinear. At low rate flows, they return counts that are lower than expected. A future version might try to address this issue. 2) Subtracting the flow of two sensors will decrease accuracy. The ratio between the forward flow and consumption is typically 5:1 to 10:1, increasing errors by that much. 3) In the absence of a GPS signal, the fuel economy will not be updated. The absolute fuel quantities are not affected, however. 4) Flowswot has not been extensively tested to know whether its use is worthwile or not. As of April 2008, it's experimental. 5) The values related to one second make large jumps. This is not a bug of Flowswot but it's how the sensors behave. The 10 second averages behave much more smoothly. Version information V 0.1.0 Very early initial release. Error handling non existant to rudimentary. Simple cosmetics. The goal ist mostly testing of the accuracy and usability. |
lxuser hat folgendes geschrieben: |
@ Air
Ich muss mich in den nächsten Wochen auch mal um eine Karte mit passenden Ein- und Ausgängen kümmern. Wobei mir bei der Velleman halt nicht genug analoge bzw. Frequenzeingänge vorhanden sind. |
El-Chico hat folgendes geschrieben: |
Da der rückströmende Kraftstoff wärmer ist als der im Zulauf, kann man nicht einfach die Volumenströme (also l/h oder l/min) subtrahieren. Der rückströmende Volumenstrom ist nämlich wegen der höheren Temperatur deutlich höher - im schlimmsten Fall kann der berechete Verbrauch dann 0 oder gar negativ sein... |