Power FC Hacking the PowerFC/DataLogIt
#51
Senior Member
Join Date: Apr 2002
Location: Boston, MA
Posts: 251
Likes: 0
Received 0 Likes
on
0 Posts
There were several people working on auto-tuning with the PFC last year. Tuning the fuel map is pretty stratightforward, but the ignition was the tricky part. Do you guys have any ideas on how to auto-tune the ignition at all? It seems pretty difficult to me to be able to safely modify the ignition map automatically, even based on having real-time EGT data.
#53
Originally Posted by Rudy 34
FastHatch, can you share your powerfc emulator application (not the source code) ?
It is not an PFC emulator, it is a free replacement for FC Edit. But it can be used as an emulator. Maybe i disable that in the release, i am just using it for testing.
Monitoring/Datalogging is not implemented yet, also it is currently only for Hondas.
Originally Posted by vinayp
Do you guys have any ideas on how to auto-tune the ignition at all? It seems pretty difficult to me to be able to safely modify the ignition map automatically, even based on having real-time EGT data.
#54
Originally Posted by Goblin
The second time the command is sent, an eight character non NULL terminated string describing what platform this PowerFC is for is returned. For RX-7's, it's "13B-REW ".
"B16A1 " = Honda
"B16B " = Honda
"B18C " = Honda
I guess the strings can be found here: http://www.fc-datalogit.co.nz/model-status.html
#55
Tequila? ..it's like beer
iTrader: (1)
Join Date: Mar 2001
Location: Woodbine, MD
Posts: 1,124
Likes: 0
Received 0 Likes
on
0 Posts
Originally Posted by FastHatch
The app will be available for download by end of march here: FC Tune
It is not an PFC emulator, it is a free replacement for FC Edit. But it can be used as an emulator. Maybe i disable that in the release, i am just using it for testing.
Monitoring/Datalogging is not implemented yet, also it is currently only for Hondas.
You would need EGT and knock to see if timing is safe, but without dyno readings tuning timing is useless. Isn't it? Advancing the timing till you have knock isn't the point where you make the most power!?!?
It is not an PFC emulator, it is a free replacement for FC Edit. But it can be used as an emulator. Maybe i disable that in the release, i am just using it for testing.
Monitoring/Datalogging is not implemented yet, also it is currently only for Hondas.
You would need EGT and knock to see if timing is safe, but without dyno readings tuning timing is useless. Isn't it? Advancing the timing till you have knock isn't the point where you make the most power!?!?
Since you'll have access to elapsed time & distance (calculated from speed I suppose) you'd be able to generate a torque & HP curve right? While it wouldn't necessarilly be as accurate as a dyno the results would at least be consistent with the car. You might not be able to brag about your FCTune HP sheet but you could compare changes in 2 different runs and use it as a tuning aid.
#56
Originally Posted by NeoTuri
Not sure yet. The plan right now is to basically handle the task of the Datalogit directly with a PC interface (LPT1, RS-232, or USB).
Last night I was inspecting signals going between the ECU and PF Commander. I ran into a few issues:
Apparently, my laptop's parallel port I/O operates at ~3.0v, but the PFC is 5.0v. I imagine it's different among various hardware manufacturers. That's not really a problem at this juncture, because there's still a big enough difference to tell if the data/clock lines have been pulled to ground or not.
There's also an issue with the speed of directly communicating with a PC. The parallel port has an optional line that triggers an IRQ event. I wrote some code last night to test the responsiveness of the channel. It works sometimes, but it means I will need to use some latches to hold the data while I pull it down from the circuit.
Last night I was inspecting signals going between the ECU and PF Commander. I ran into a few issues:
Apparently, my laptop's parallel port I/O operates at ~3.0v, but the PFC is 5.0v. I imagine it's different among various hardware manufacturers. That's not really a problem at this juncture, because there's still a big enough difference to tell if the data/clock lines have been pulled to ground or not.
There's also an issue with the speed of directly communicating with a PC. The parallel port has an optional line that triggers an IRQ event. I wrote some code last night to test the responsiveness of the channel. It works sometimes, but it means I will need to use some latches to hold the data while I pull it down from the circuit.
#57
The shy megalomaniac
iTrader: (2)
Originally Posted by Rudy 34
Some news ?
I wired some TTLs and discovered the following:
The PFC communicates using something very close if not exactly the I2C protocol (by Philips). There's a data line and a clock line. I've measured the clock line operating at ~200 Hz, but from some other documents I've encountered, I'm under the impression that it's supposed to be able go as high as 19.2kbaud.
The clock line seems to function normally between ground and 5V. However, I haven't noticed the data line drip any lower than 4.95V. Eventhough the TTLs only see that as a logical ON state, putting the data line on the clock I/O does trigger them. Strange. Might be a loose wire or something.
I'm testing some other circuits to try to get an accurate grip on what is really going on the logical level. I know I will eventually have to invest in some PIC programmers, as the chips are quite useless without code.
#59
The shy megalomaniac
iTrader: (2)
Originally Posted by Rudy 34
Sure it will be useless without programmer :p So you already have the code for some pics or you get some code for standard I2C interface based on a PIC ?
I'm pretty sure there are code samples of this out on the net. I haven't searched much for them, but I'm pretty sure they exist.
Until I get the programmer, I'm effectively limited to using just my oscilloscope.
#64
http://www.emicros.com/i2c232.htm
May be this help U /
May be this help U /
#65
Senior Member
Join Date: Apr 2002
Location: Boston, MA
Posts: 251
Likes: 0
Received 0 Likes
on
0 Posts
I found this interesting site that looks like it is focused on building a datalogit type interface to the PowerFC: http://kaele.com/~kashima/car/pfcadp/
They even have a complete circuit diagram for building your own datalogit-type box (including 4 A/D connections):
PFCAdapter10.pdf
The only problem is that the entire site is in japanese (can anyone translate?). But there are plenty of pictures and timing diagrams. Let me know if anyone ends up building a working box based on these schematics.
They even have a complete circuit diagram for building your own datalogit-type box (including 4 A/D connections):
PFCAdapter10.pdf
The only problem is that the entire site is in japanese (can anyone translate?). But there are plenty of pictures and timing diagrams. Let me know if anyone ends up building a working box based on these schematics.
Last edited by vinayp; 03-15-05 at 06:37 AM.
#67
I think that the pic program only transparent translation betwen rs232 and i2c protocol.
it`s not too hard to write it by self.
something like this
http://derrickshouse.com/KITTI/code/?PicI2C.c
it`s not too hard to write it by self.
something like this
http://derrickshouse.com/KITTI/code/?PicI2C.c
#68
Senior Member
Join Date: Apr 2002
Location: Boston, MA
Posts: 251
Likes: 0
Received 0 Likes
on
0 Posts
if that's the case, it should be very simple to build a working replacement for the datalogit using the circuit provided in that PDF file. i wish i had a PIC programmer to test this out with. Rudy, can you test this circuit out since you have a programmer coming in?
#70
I have contacted him also and no source code. He does not want also to sell some pics already programmed
Concerning the available schema, it's not so simple. Kashi is also treating the A/D converters inside the PIC and this part has to be totally written.
Only the communication (translating I2C to serial) can be generic.
Concerning the available schema, it's not so simple. Kashi is also treating the A/D converters inside the PIC and this part has to be totally written.
Only the communication (translating I2C to serial) can be generic.
#71
Senior Member
Join Date: Apr 2002
Location: Boston, MA
Posts: 251
Likes: 0
Received 0 Likes
on
0 Posts
That's true, but the A/D conversion is an extra feature of the datalogit that is independent of the PowerFC. For me, the ability to communicate with the PowerFC (i.e. the I2C to serial conversion) is the most important part.
#73
Here is a screenshot of the current version of FC Tune:
The communication and calibration for Hondas and the RX7 are completly implemented, now i'm working on the views.
That means FC Edit files can be opened and written, the calibration can be send to and read from the Power FC, but not all settings can be viewed and edited yet.
I also need to implement 2D and 3D graphs of the maps.
Later Andre
The communication and calibration for Hondas and the RX7 are completly implemented, now i'm working on the views.
That means FC Edit files can be opened and written, the calibration can be send to and read from the Power FC, but not all settings can be viewed and edited yet.
I also need to implement 2D and 3D graphs of the maps.
Later Andre
#75
Originally Posted by subyfiles
Good work.
Hope someone would make something for subarus
Hope someone would make something for subarus
FC Tune now keeps track of all changes and has a undo function.
All changes can than be uploaded to the PFC.
Edit: Also unlike to FC Edit it will be one program for all models/applications.
Last edited by FastHatch; 03-25-05 at 03:27 PM.