For anything you'd like to see added to Speeduino
By Trevor Getty
#51428
LOL - I will return your good news - with some more....
Its ALIVE I tell ya - ALIVE... :o

Thanks very much, we have just got the worlds first Rover T16 Speeduino powered car running, brilliant work - decoder works a treat.

I chased down some bad earths to the starter motor, and got it turning over quicker, and set my running offset to 10 degrees for now, to try to get it running a bit better ( want fixed angle to check timing when running ).

But she kicked into life - now it does sound rough or RUFF, whatever you prefer lol - but it is alive.

The decoder is very, very close to complete now I recon, the cam sync seems to be very reliable now.
NB - I will check at higher rpm later, as It would not be safe to do this yet, is there still a <6k issue in the code or is that gone?.

Next:
I will try to dial in the ignition advance properly, then Mike I will pick up your new code update, and try again, to dial in ignition trigger. ( we recon is should be around 10 degrees now ).

Once this is dialed in, I will try to get it idling - as currently I have just jamed the throttle body open a little but if I can get it to idle correctly at say 5 percent throttle using ignition timing and get my fueling ok, then I will try to add the IACV once I know the rest is dialed in correctly.

I have attached the running log file, running at the 360 trigger, and (5 cranking , 10 fixed running ). It does run at both 5 and 10 degrees running, but sounded better at 10 ). Note the MLG file is again renamed with *.txt on the end to allow it to upload.

The logs contains some different load points of the car, and it responds ok, so far ( early days ).

I defo have not lifted RPM above say 2k at moment, as it sounds like we are massively spooling the turbo up ( like launch control ) lol, as its running at 6 - 10psi even at around 1-2k rpm speeds ). I hadn't set the duty cycle on my external boost controller, so I may disconnect that for the moment, and let the wastegate just keep it to the factory 9psi max as I dont want any head popping accidents just yet :p

It appears to be running lean, at the moment, but then again I have no AFR closed loop enabled below 1200 rpm where it spends most of its time at the moment.
Once I get the timing dialed in at idle, and we are sure its correct then I will proceed up the table lol.

I have narrow band O2 that I will try to add to the log file output also - just to make sure the calibration is correct on the WB controller.

Regarding the VE table, I see that it has a scale up to 100kpa ( is this not atmospheric pressure normally or close to it? )
My question is that you can see I am around 90kpa unless I press accelerator and it drops quickly and picks back up.
Is my scale correct for a boosted car)0-100kpa) which will be running up to 15psi boost?

Its just if it thinks that 90kpa is under heavy load at the moment you can see from the TPS values it was really at 0% - 5%.
Might it be better to base the VE table off the TPS initially, I am not sure the best way to setup a brand new tune, if there is any good reading I should do, or videos to watch on best way to start a tune from scratch - it would be handy.....

If you have any ideas of what else looks wrong in the logs, and any other areas I need to fix or look at before progressing let me know.
Attachments
(167.47 KiB) Downloaded 284 times
User avatar
By PSIG
#51434
Congratulations! Image

Of course, it looks horrible due to being a first-start. It will get better when you can fully warm it and begin making large adjustments. Log everything, and sit with the logs to "see" what the engine responses are trying to tell you.

The first concerns I had that need attention to allow smooth transition to tuning were 127 sync losses and poor voltage regulation. Of course, your battery is really low and needs charging. But, most of the sync issues were not while running or cranking, and randomly saw several in a second. This indicates an electrical or connection issue.

Also electrical-related is your running voltage regulation, bouncing from 14.2 to 12.5V and back, several times per second. This rapid fluctuation will throw-off your fueling as it calculates voltage corrections, but then the voltage is different by the time it actually injects. I'd look into why your charging voltage is not stable.

Finally, your MAP may be high simply because it's running so poorly at this point. However, if you get it warmed and do some idle tuning but find the MAP is still high (guessing on that engine type anything over about 30kPa on a tuned warm idle), then look for issues (calibrations, MAP line leaks, engine issues, etc). Your logs will help you in what direction to look (diagnostics) at that point. Have fun! 8-)
By Trevor Getty
#51457
Thanks very much for the hints and tips.
the voltage did need a look - it was on a charger at the time which was causing some external fluctuations and noise in the lines.
I got a bigger battery now to help with this.

Next, I will try to home in on the base ignition timing, and once stable I can put the IACV onto the car and get a good stable idle which will help me to warm it up more easily ( :p ).
User avatar
By PSIG
#51482
The decoder is working, so it's time to end this thread for things not-decoder. I'll quote your last post to your Rover Speeduino setup thread from here. ;)
By Trevor Getty
#51486
I understand taking issues into the other thread to rover setup. That makes sense

Although this thread isnt finished its covering the rover decoder and we have only started testing on one of the cam crank patterns out of 4 different combos.

We have flipped the crank rotation bit I will report back once this is tested and we can tick one car off the list.

2 gap crank covered. T series early. (Tested)
4 gap crank - t series late
K series - vvc
K series - non vvc.

Mike what would you prefer to do. Close this down now and raise new threads for each of the other patterns to be tested?
By Trevor Getty
#51487
I would like to just finish and test the new code fix u got up on github.
We could then close off this thread and get the current rover option you have into the main code base that would be great.

We could open new threads for the other testing patterns.
It at least closes this thread with a nice conclusion and stops it being open forever.

Ill try to test this tonight and get back ASAP.
By miker
#51488
Due to trying to be clever I've created a problem. As the code covers 4 different variations of the rover mems fly wheel pattern we could be here for months.

I like the idea of opening a new thread for each new variation. Im thinking this will then appear in a search so if anyone in future is looking for information it will be easier to find

Does mean Josh has a decision on when to include the pattern in the full build. As the base code is generic and the difference is the pattern matching hopefully he'll be ok with including it earlier.
User avatar
By PSIG
#51489
I certainly didn't mean to imply closing the thread, but trying to keep conversation about the decoder, and non-decoder conversation elsewhere. I probably should have said your version of the code works… :lol:

Your points are valid, and there is a line that needs to be defined for Speeduino. There are so many different patterns out there, but fortunately they are all digital. At some point, Speeduino will have to (or should) transition to a hash code or array code system. This would involve a TunerStudio or standalone (executable? web page?) calculator where you enter the specifics of your wheels so it can generate an array code. Something Speeduino could use to know exactly what the wheel—any wheel—looks like. That code would be entered into TS for Speedy to read.

It's not that dissimilar to other systems of entering the total and teeth that are missing, e.g., 36 teeth, 5 missing, located from sensor TDC as teeth 5, 20, 21, 31, and 32. Conversely, every (rising or falling) tooth edge by degrees, from zero. Same for the cam wheels. Generate the code with a calculator and load 'er up. Involved? Probably. Capable of handling almost any trigger combo on Earth? Hopefully.
By Trevor Getty
#51492
Thanks Mike. I will get back to you with confirmation of this last code change and if working hopefully we can request it be merged. I know this does lay with josh.

As you have done the hard work to create a generic wheel decoder. If this was in the mainstream branch, we could start thinking about a decoder pattern generator like that mentioned by PSIG as it could produce the bitwise array used in the new decoder and ultimately most existing decoders could be replaced by array combinations. While also allowing new unknown types. I understand its all building blocks to something like this.

I will raise another feature request for a k series variant next and cover the already known information for that type.

Many thanks guys.
By miker
#51495
Final comment - I've written a version of the completely generic decoder covering crank and cam. Unfortunately covering 90% of the possible patterns involves a lot of code and memory. I couldn't get it to run reliably in a mega (performance with so many things to process) and it needed more free memory than we have so stopped.

Grab me on slack if your want to discuss it further
  • 1
  • 7
  • 8
  • 9
  • 10
  • 11
  • 12

It looks like you have a fuel supply issue. readin[…]

Will this have an updated version about this featu[…]

Vr Conditioner Noise when cranking

I did the fix, but it's not work. I put my setti[…]

Perhaps some different points and pictures. Instr[…]

Still can't find what you're looking for?