Datalogger

[Bean]

I've already posted up at clublexus, so I'm going to cut and paste my posts here for the most part.

This project will be a standalone datalogger, it will not require OBD1 or OBD2. So far the prelim design calls for three 0 to 5v analog inputs (Unit will be able to read up to 8, but just 3 for now) and 3 frequency inputs. The freq inputs will be for KVF, rpm, and speed.

The development board for this project arrived today and so design work will commence immediately. I'm hoping to have it working well enough to test on my car when I turbocharge it.

Next few posts will be cut and paste from clublexus thread:
http://www.clublexus.com/forums/showthread.php?t=338505

[Bean]

quoted from initial post on clublexus

I have had the priviledge of being in the Lexus and Toyota community for 7 years now. My Bachelor's Degree is winding down and I will graduate at the end of the year.

We have to pick senior projects in EE/CpE and I'd like to do something that has an impact on the community and helps it out.

I enjoy reading about the Megasquirt community and the TechEdge WB02 community because various vendors offer to build the units for you or offer you the parts to build it yourself and all assembly information is completely open and completely free.

My project is going to be a Datalogger for the 2jz-GE (and most likely GTE). It may or may not work with the 1jz, we will see.

This datalogger will interface with the under-dash DLC connector (round plug) and read engine information. I have ideas for several options, like real-time display on an LCD screen, possibly with a touch-interface, USB flashdrive downloads of logs, and real-time interfacing via a serial connection. I also would like to use various voltage inputs that would allow logging of WB02 Sensor voltage, MAP sensor voltage (for those that use JDM motors/JDM ECU project or just want to read boost).

I am thinking also of making all the information completely free for its complete assembly including assembly code and external interface software (to be developed to run on windows/mac/linux). I have a very affordable PIC processor (chip costs like $15) picked out that would be able to handle it; and also would be easy to develop for if savvy developers wanted to extend its capabilities. Total parts cost for everything listed above would be most likely south of $300, probably about $250; and maybe even less than that. (thats with an LCD touchscreen)

FYI: This is not a physical product that I wish to sell, nor make money on. This is something I feel should have already existed. There is major potential for such a device if designed correctly. The upper-level interface could be the same but the lower level could be programmed to work with several types of interfaces (all jz motors, hondas, nissans, etc). Not saying I'd do that, but if you leave the possibility for its development open, someone will jump in and build it. If anything, I'll put together bulk-purchased equipment and make little DIY kits with just cost covered on shipping and parts.

My question is, does anyone see a value in something like that or am I wasting my time? I know there is a handheld toyota unit out there and I know there is an aftermarket reader (though I hear the sampling rate is terrible and that it doesnt have many options).

quoted response to evileagle talking about DSM scanners:
evileagle,

I've seen those before, but they had no options for extra inputs like a map senor or WB02 sensors. The stock narrowband sensors are worthless at anything other than 14.7:1. I never saw one datalog either, but that doesn't mean it couldnt do it. Theoretically, if you didn't have a laptop; you could put a thumbddrive into this unit and have it download the datalog maps realtime to be used by a software application later

quoted after I found DLC connections function serially with word packets for commands:
Found a major roadblock and the reason why the OEM scanners absolutely suck for datalogging purposes.

The DLC2 port under the dash only transmits at 100baud. It uses a serial datastream that contains 14-20 words with info about sensors, and current engine info. This SUCKS. There are no signal elements used here to send more data across the line; so 100baud is effectively 100bps, which means, 1 bit is transmitted every 100th of a second and it may take an entire second to send an update on the engine. Thats way way way too long.

This datalogger that I want to build is going to have to read realtime data. How does a design a la SAFC with a seperate harness plug that connects to various inputs/outputs on the ECU sound? You then plug the datalogger into that. We can use something simple and easy to buy parts for, like a 15pin serial cable or something.

quoted when replying to Touge about his WRX's OBD2-reader:
This is for OBD1; I have no clue about OBD2. OBD2 is more of a universal standard and I'm positive it has a much higher transmission speed. OBD2 is also capable of displaying a lot more information, so it would have to be.

With this, I'm going to have to tap rpm, speed, 1 injector, kvf (anyone know the range of data in hz for the kvf?), and maf input option. It will most likely have 2 0-5v signals that can be used for a variety of things. I'm assuming they will be used for WB02 and a MAP sensor.

EDIT: It also isn't about the number of things logged, its the accuracy in which they are done. Getting engine updates once a second is NOT enough. There are 300 spark events in 1 second at 6000rpms. On turbo vehicles, the rpms can climb VERY rapidly in just 1 second. Not nearly good enough to datalog with.

quoted when replying to Touge about whether it would work for OBD2 cars:
I see no reason why it wouldn't. It will be developed with an OBD1 car, but assuming the OBD2 ECU outputs the same tach signal, speed signal, etc on the same wires, then it won't be a problem.

This will not be using the datalink port at all because of the previously described problems, so it won't be limited to any particular car (so long as signals are similar enough). As far as I know, OBD2 and OBD1 ECUs have the same signal inputs/outputs for the most part (minus the MAF vs KVF)

[wildstyle69]

I wouldn't mind jumping in on this project

I have almost 10years professional experince coding on the PIC devices for industry, and military so I think I can help doing some code, and I have the sources for all the parts (PIC, PCB, custom multiplexed LCD, LCM, GRLCD)

Drop me a line if you more forward on this.

[Bean]

highly doubt it at this point. There didnt seem to be a lot of interest, there are other products out there that do the same thing, and I got a research scholarship dealing with R&D in Underwater Acoustic Modulation :) I'm getting paid to bang on a Altera NiOS FPGA board now; so the time I was going to use for this project isn't really available anymore :/ Thanks for the offer though; I will definitely keep you in mind

[wildstyle69]

what do you think of the Altera stuff......funny this name came up here i was discussing them with the fae's at the mc masters last week.

your project sounds pretty interesting........ good luck

[Bean]

Altera is extremely expensive

BUT, their development software and hardware solutions are top-notch. Working with the Quartus software for actual design work (I learned basic VHDL and digital logic using Quartus as a guide at school) is wonderful. Currently I've got to build audio i/o and some method for representing the four phases for QPSK modulation in memory plus a PC-based GUI that can interface with it. Its not the career-field that I want to go in; but its still pretty badass nonetheless :)

I'm still learning about PLDs in general, but from what I've played with so far; this hardware is an absolute dream.