| gemellocattivo.com http://gemellocattivo.com/forum/ |
|
| o5e firmware http://gemellocattivo.com/forum/viewtopic.php?f=19&t=368 |
Page 2 of 6 |
| Author: | mk e [ Sat Jun 13, 2020 9:17 am ] |
| Post subject: | Re: o5e firmware |
remembering back...5V is on I think so you have no fuel ouputs. The pump is completely separate code, so if both the pump and injectors are off its likely a common setting....like fuel enable, that's the only one I recall and would need to dig a bit beyond that. I'm happy to spent some time this evening. |
|
| Author: | Rafa_bmx [ Sat Jun 13, 2020 10:44 am ] |
| Post subject: | Re: o5e firmware |
Thanks a lot for all your answers! Yes the crank and cam inputs work with a real trigger wheel or with a simulator on an arduino as well, i have no doubts about it, same For TACHO and same for spark, was Not able to test any other outputs yet such as idle etc etc, the analog inputs also work fine! Do you remember starting to code anything for fly by wire/drive by wire? Thanks |
|
| Author: | mk e [ Sat Jun 13, 2020 10:59 am ] |
| Post subject: | Re: o5e firmware |
Sound like I let a bug into the fuel code then, sorry. I'm sure its minor as it was all working. I also just use a blinking led to tell me I have a legitimate value at each step......its got to be something with enable as that is the only thing I believe links the fuel pump and injectors and I think is the last item and controls the output. There is no drive by wire. I do have that working well in my enginelab project, you could take that and translate it to C pretty easily....its got start position, pedal-throttle translation table, idle control, stall saver, traction control but only that basics are tested as my engine had some issues I'm now correcting. Rusefi has dbw codes too, but fewer features I think. |
|
| Author: | Rafa_bmx [ Sat Jun 13, 2020 11:16 am ] |
| Post subject: | Re: o5e firmware |
Whats that engine lab? Is one of those ecu's you buy and then use sinulink to create a model? Yes i have tested with a led and also with logic analizer, fuel injector outputs and fuel pump output are always HIGH (5v) Thanks |
|
| Author: | mk e [ Sat Jun 13, 2020 11:49 am ] |
| Post subject: | Re: o5e firmware |
With the brink you can step through the code logic and find where it goes wrong.....a debugging tool is much faster but I don't know how to use that. Enginelab.net uses a pretty simple function library to let the use create control logic without any need to know anything about low level C code stuff. I just love it. You can download the software, which has a simulator so you can "test" everything on the pc without touching hw....pretty nice. My model (program), which run an engine, is here http://gemellocattivo.com/forum/viewtopic.php?f=19&t=153 |
|
| Author: | Rafa_bmx [ Sat Jun 13, 2020 11:51 am ] |
| Post subject: | Re: o5e firmware |
And is that only ported for their very expensive ecu's or is it possible to use it with other MCU's/hardware |
|
| Author: | mk e [ Sat Jun 13, 2020 12:19 pm ] |
| Post subject: | Re: o5e firmware |
To me, it was worth every penny, no compiling, see/log every internal calculation for bug hunting, just nice. This it the AEM infinity series ecu....they license from enginelab and I suspect it's that arrangement that means only the highest end, infinity 7, is available directly from enginelab. Its basically a motec M150 at a fraction the price. Way back we had talked about something a bit similar for o5e....but never acted on it. Everyrhing a library function you call. We tried to keep the code a user would mess with simple and that would have been a next step...but we never got there. There was another option, an adaptronic 1280 but I thing they dropped it. It used an interface like labview, but didn't quite have everything I needed so ibreplaced my M150 with enginelab once I fully realized the $4k I paid for the motec was just a deposit and I'd be out at least 10k, probably 15 before it did everything I wanted....engine lab does it all for 2k so a bargin to me. Anyway, all my latest control logic is written in enginelab, i wrote a user manual to explain everything and you're welcome to use anything that helps. |
|
| Author: | Rafa_bmx [ Sat Jun 13, 2020 4:12 pm ] |
| Post subject: | Re: o5e firmware |
It looks like a good solution altough im looking for somehting i can use on other. MCU's with our own hardwa4e, if you know how to port all of that to c++ for a decente MCU and you want to earn a couple thousands get in touch |
|
| Author: | mk e [ Sat Jun 13, 2020 6:29 pm ] |
| Post subject: | Re: o5e firmware |
I know exactly 0 C++, I know enough C to be dangerous but not enough to be very useful. We abandoned o5e because we realized there were just too many limitations with the eTPU code and no cheap way to change it. I'm convinced that the right way to do a low cost ECU is a an ARM+FPGA but I'm not a programmer so I can't do it. For a while it looked like the rusEFI guys might as they realized their code didn't have the accuracy they thought it had....but they made some improvements and it looks like have decided its good enough and honestly, for most people it probably is, but for sure its not OEM accurate so not something I'd personally consider knowing 2k buys me something that is OEM accurate and fully configurable....but I understand wanting a good yet low cost option, which is ARM+FPGA |
|
| Author: | apalrd [ Sat Jun 13, 2020 6:42 pm ] |
| Post subject: | Re: o5e firmware |
I think if you can get the eTPU development done, it becomes a building block that you don't need to modify as much. NXP has some better eTPU functions for engine control than they did a few years ago, but it's still not great. I agree that the eTPU is a pain to develop for. The good compiler costs $2800. If you jump up to ~$5k you get a simulator too which makes development possible. Having done both eTPU and FPGA work, I'd much rather work with the eTPU than an FPGA, but some of that is how awful Xilinx Vivado tools are. You can now get a dev board for a 5777C for ~$250, and that's a dual core PowerPC e200 (@260Mhz) and three eTPUs. Not that you need anywhere near that to get an ECU with good accuracy and easy development. I think the biggest hurdle to a good open source ECU is a good open source graphical environment for end users to develop in, or people willing to build that. It's arguably more work than the embedded code. |
|
| Page 2 of 6 | All times are UTC - 5 hours [ DST ] |
| Powered by phpBB® Forum Software © phpBB Group https://www.phpbb.com/ |
|