sim's Volvo 245 comments

All home-built FreeEMS implementations without a forum of their own, usually TA-based.
User avatar
Fred
Moderator
Posts: 15431
Joined: Tue Jan 15, 2008 2:31 pm
Location: Home sweet home!
Contact:

Re: sim's Volvo 245 comments

Post by Fred »

Sim, I've been thinking about and working on decoders today and I've decided that I want you to stay on the 24/1 code until you are injected. Reason being that it's too much of a pain in the arse for you to prevent fuel flowing until then, and testing other decoders and changes really requires that. I know that you don't care about that right now, I just wanted to make a note of it.

In contradiction to that, if you swing past the junk yards, see if you can get yourself a 4/1 disk for your CAS at some point :-) I don't want you using it yet, but it would be nice to get some testing on that.

Cheers!
DIYEFI.org - where Open Source means Open Source, and Free means Freedom
FreeEMS.org - the open source engine management system
FreeEMS dev diary and its comments thread and my turbo truck!
n00bs, do NOT PM or email tech questions! Use the forum!
The ever growing list of FreeEMS success stories!
User avatar
sim
LQFP112 - Up with the play
Posts: 112
Joined: Thu Jun 02, 2011 8:17 pm

Re: sim's Volvo 245 comments

Post by sim »

Fred wrote:Sim, I've been thinking about and working on
decoders today and I've decided that I want you to stay on the
24/1 code until you are injected. Reason being that it's too much
of a pain in the arse for you to prevent fuel flowing until then,
and testing other decoders and changes really requires that. I
know that you don't care about that right now, I just wanted to
make a note of it.
It is a pain to shut off the fuel with a carb set-up. Why can't
cars have petcocks like bikes do? I was wishing for injectors I
could unplug while getting my timing dialed in yesterday.

Injection is phase two, and will be coming soon. I have
everything I need for the car, except a fuel pump.

I need to build a new board for the TA to run the injectors and
the rest of the sensors I will need for fuel. I guess I should
see if I can get a couple of those connectors that DeuceEFI
found.

I may pull the head and swap it also, the carb head has coolant
ports for the manifold. The intake manifold is actually heated.
It helps atomize the fuel, but it's kind of like a reverse
intercooler. I have a couple of B21FT heads that don't have the
coolant ports and do have sodium filled exhaust valves. It's a
good chance to install a MLS head gasket too.

I guess it will likely be a few months before fuel happens. I
hope to have it sorted for the new year though, there is likely a
several thousand kilometer roadtrip on the schedule for early
next year.
Fred wrote: In contradiction to that, if you swing past the
junk yards, see if you can get yourself a 4/1 disk for your CAS
at some point :-) I don't want you using it yet, but it would be
nice to get some testing on that.

Cheers!
I will grab one if I get the chance. I'm keen to get a "FreeEMS"
disk also, though I suspect the only way to make that happen is
to get a run made up. The 24and1 will be fine for the time being.
<@TekniQue> but in the end, it's code that makes a computer useful
User avatar
Fred
Moderator
Posts: 15431
Joined: Tue Jan 15, 2008 2:31 pm
Location: Home sweet home!
Contact:

Re: sim's Volvo 245 comments

Post by Fred »

sim wrote:Injection is phase two, and will be coming soon. I have
everything I need for the car, except a fuel pump.
Cool! :-)

Beware, there are more features missing from injection than from ignition :-)
I need to build a new board for the TA to run the injectors and
the rest of the sensors I will need for fuel. I guess I should
see if I can get a couple of those connectors that DeuceEFI
found.
Good idea, to sandwich it 3 layer style, right? :-) VERY DIYEFI!!! :-)
I will grab one if I get the chance.
Thank you! :-)
I'm keen to get a "FreeEMS" disk also, though I suspect the only way to make that happen is to get a run made up.
I don't know about there, but at home, if you supply the design file, it's pretty inexpensive to have done. Perhaps you could approach Josh and see if his supplier can do a one-off prototype for a reasonable price? BTW, I can't guarantee that that disk design will work with the DSM CAS. The DSM CAS is not as precise as the Nissan CAS in terms of edge angle alignment. It may be necessary to make the outer windows smaller and the solids between them larger as the effect is to widen the windows. It could screw up the counts too, as the ID windows would be wider also. I could do some calculations and beg Jammi to do us a guessed Mitsu one to try, though. It's hard to say if it will work well or not. With a Nissan unit, it would work as is perfectly fine.

In the event that the 144/72 style wheel wouldn't work with the DSM CAS, we cold design a more coarse one with similar benefits for that CAS.

Fred.
DIYEFI.org - where Open Source means Open Source, and Free means Freedom
FreeEMS.org - the open source engine management system
FreeEMS dev diary and its comments thread and my turbo truck!
n00bs, do NOT PM or email tech questions! Use the forum!
The ever growing list of FreeEMS success stories!
User avatar
Fred
Moderator
Posts: 15431
Joined: Tue Jan 15, 2008 2:31 pm
Location: Home sweet home!
Contact:

Re: sim's Volvo 245 comments

Post by Fred »

sim wrote:It works!
HOORAY :-) Congratulations!
rewiring the CAS using two lengths of coaxial antenna wire
LOL, fantastic! :-)
With a flat ten degree advance map, the engine started right up
and ran. Revving the engine causes the timing to flutter
slightly, no more than 1 degree back and forth.
But it's rock solid stable while at any given fixed RPM, right? It should be. Any jitter there is CAS drive slop, most likely. The error while revving suddenly is due to a difference between reality and last calculation, but for you, that difference should always be minimal. One day we'll add an option to predict, but I suspect we're better off without it anyway.
I created an initial ignition map
I still think that is too aggressive in the medium pressure ranges. If I were you I'd back timing out of it there and bring it back in as you tune. You're safe at about 32* from 3k up at any pressure level, and you can add more at lower pressures, but I think it's too steep from 100kPa to 60kPa. I could be wrong though, just urring on the side of the conservative! :-)
With the initial ignition map, the engine runs very smoothly.
There is a noticeable difference from the old points set-up.
Is the noticeable difference that it's more smooth or what? Just clarifying. It's good to hear good things about the system, that's all :-)
The car ran very well, except for occasionally losing sync.
Great to hear! :-)
The sync loss happened fairly randomly, but seemed to occur more often at idle, and when well loaded.
Most likely that was just a perception thing. It's *much* more obvious when under load or at idle.
By the top of Burnaby Mountain (about 300 metres of climbing) it had stumbled enough times to generate a considerable backfire. Somehow, this caused the #1 plug wire to come loose.
You've got a datalog, though, right? If you're going to do significant testing, let me know and I'll send you a good custom datalog generator file to use so they are maximally beneficial.
The next step is to get set up for datalogging and diagnose/fix the sync loss.
OK, that answers my last question! Yep, the logs record where sync was lost and why, so we'll get some good info out of that. Most likely your wiring could still do with some improvement. You could add a very small cap on each of the inputs from the CAS too, but when I say very small, I mean it. The phase shifts from larger caps are not acceptable at higher RPM. We can discuss this when you're next available. Another thing you could do is stiffen the pull up resistors that the CAS pulls down. That will allow a larger cap to be used and in and of itself increase noise immunity. Another thing you could do is add a Schmitt trigger on each input. If the noise is bad enough, though, this won't help. http://en.wikipedia.org/wiki/Schmitt_trigger

So then:
  1. Wiring first
  2. Tiny caps second
  3. Pull ups third
  4. Schmitt trigger fourth
We really need you to have a clean signal so we can develop the scheduler and decoder further. Once it's in good shape it will make sense to revert and try some software fixes for noise, but not until after the foundations are really solid.

At the end of the day, though, the software can only do so much, shit in, shit out, or in our case, shit in, nothing out, briefly. Shit out is bad for rods and pistons and headgaskets :-)
Video is coming soon.
Hooray! I hope your soon is sooner than my soon... I have many FreeEMS videos to upload... one of these days soon... :-)

Fred.
DIYEFI.org - where Open Source means Open Source, and Free means Freedom
FreeEMS.org - the open source engine management system
FreeEMS dev diary and its comments thread and my turbo truck!
n00bs, do NOT PM or email tech questions! Use the forum!
The ever growing list of FreeEMS success stories!
User avatar
jharvey
1N4001 - Signed up
Posts: 1607
Joined: Tue Jun 10, 2008 5:17 pm

Re: sim's Volvo 245

Post by jharvey »

Congrats congrats. I'll eagerly await the video.

Hey, why can I post here? Fred can you move to the comments thread please.
Last edited by Fred on Thu Sep 08, 2011 9:05 pm, edited 1 time in total.
Reason: Done
User avatar
Fred
Moderator
Posts: 15431
Joined: Tue Jan 15, 2008 2:31 pm
Location: Home sweet home!
Contact:

Re: sim's Volvo 245 comments

Post by Fred »

sim wrote:On a hunch, I disconnected the power ground for the coils (not the logic ground, it's a separate pin on LS1 coils) and rounded them instead to the cylinder head.

The engine ran much better, no longer losing sync.
Good to hear. Make sure that you can put it back to bad so we can verify that the sync loss stuff works properly. Don't forget, you're testing brand new code and it's going to need work, and testing, and so on :-)
My theory is that the high voltage from the coil had to go
through the coil ground in the loom and then back through the
block ground to the plugs. This induced a current in the power
line to the ECU which was routed alongside the coil ground. The
spikes overloaded the voltage regulators, which cut out and reset
the ECU.
Much more likely is that it did the same thing to the CAS signals instead :-)
It looks like this problem can be chalked up to my inexperience
in wiring a vehicle ECU rather than any flaw in the code or the
design of the ECU board.
You're not the first to make wiring layout mistakes and you won't be the last :-)
I'm now confident enough with the set-up to drive the car into town for work.
Great! :-) Two FreeEMS daily drivers in the world! Take cover, quick lol.
The next step is making a set of det cans to listen for knock and then playing with MegaTunix and the spark table.
May I respectfully disagree and suggest some alternatives below.
Big thanks to everyone on the IRC channel for the support and advice.
No worries! :-)
I really feel like I'm standing on the shoulders of giants.
Be careful not to fall off! :-)

OK, I just took a look at your log file, and there are few concerns in there. Some my problem, but for you to test when I get around to that work, others your problem, and for you to fix before the ignition fine tuning begins.
  • Noisy MAP - my fault, samples on every event, and you can actually see the per cylinder pressure waves in the log. I will fix this in the future, but it will affect the tune.
  • MAP values wrong (110kPa) - our fault, I need your help to verify what the cause is, and get it solved, either with settings or with circuit stuff, or with code change. Work with me on this as soon as you can. This obviously affects the tune.
  • Dead readings during cranking - my fault, I need to understand what happened and fix that. I'll let you know if I need some help testing that. Caused by a single sync loss during cranking. Flash back! Same as Josh's volvo with the CAS, It needed a larger tolerance than other engines for RPM variation during cranking. You can hack that in to the decoder yourself, if you want. Lines 108 and 111, change 1500 to 2000 and 667 to 500 and see if that works better. I've got a feature upgrade coming for that check, so it'll be temporary. Keep them as low and high respectively (closer to the old values) as you can. Or tweak them and forget about it and wait for a real fix.
  • Wrong and noisy BRV readings - your fault, or your car is not healthy, or both. Readings fluctuate from 12 to 11 Volts at high speed, dwell follows, though you'd never know because your coils are overkill. However if the Voltage isn't as low as the ECU thinks it is, then the coils will be getting warm. Check that. Firstly, though, check the actual battery voltage while running, after a drive, idling. See if it appears to fluctuate, and what level it is at. It seems likely that your wiring is too light and voltage is being induced by something. A healthy alternator should push out 13.8 - 14.4 when spinning at 2k or so. Your voltage seems to drop when you stand on the brake pedal too (lift off shown by low map) which makes me wonder if you're really getting a clean feed of batt V from the battery directly, or if you're cheating and taking it from some dodgy part of the loom.
  • Noisy RPM (pretty minor) - fault unknown, need to look at code to place blame. Will let you know if I need help.
Note, fixing the MAP sampling and configuring where to take samples WILL affect the tune a bit. That fix is a little bit of time away. I'd suggest that you keep your tuning to starting, idling, trailing throttle higher RPM tip in, and other drivability concerns for the time being. Use the det cans to make sure you're not knocking, don't use them to push up against the knock threshold. That can be counter productive on a low stress engine such as yours anyway.

These pics contain:
  • Noisy battery (lower green) and exactly as noisy dwell (lower red)
  • Cylinder pulsy MAP (upper orange)
  • Flat line just before starting (all colours)
Image

Image

OK, good progress there! :-)

Fred.
DIYEFI.org - where Open Source means Open Source, and Free means Freedom
FreeEMS.org - the open source engine management system
FreeEMS dev diary and its comments thread and my turbo truck!
n00bs, do NOT PM or email tech questions! Use the forum!
The ever growing list of FreeEMS success stories!
User avatar
Fred
Moderator
Posts: 15431
Joined: Tue Jan 15, 2008 2:31 pm
Location: Home sweet home!
Contact:

Re: sim's Volvo 245 comments

Post by Fred »

Sim, the sync loss during crank IS exactly what I said it was. With the fixed up OLV it's obvious :-)

Image

You get a "sync lost how" code for each loss, and this one was 2, and 2 is the time checking code.

IE, the numbers change will solve it, you can relax on the logging endeavour that I asked for in IRC. I'm 100% confident that it's this.

Fred.
DIYEFI.org - where Open Source means Open Source, and Free means Freedom
FreeEMS.org - the open source engine management system
FreeEMS dev diary and its comments thread and my turbo truck!
n00bs, do NOT PM or email tech questions! Use the forum!
The ever growing list of FreeEMS success stories!
User avatar
sim
LQFP112 - Up with the play
Posts: 112
Joined: Thu Jun 02, 2011 8:17 pm

Re: sim's Volvo 245 comments

Post by sim »

Fred wrote: [*]Wrong and noisy BRV readings - your fault, or your car is not healthy, or both. Readings fluctuate from 12 to 11 Volts at high speed, dwell follows, though you'd never know because your coils are overkill. However if the Voltage isn't as low as the ECU thinks it is, then the coils will be getting warm. Check that. Firstly, though, check the actual battery voltage while running, after a drive, idling. See if it appears to fluctuate, and what level it is at. It seems likely that your wiring is too light and voltage is being induced by something. A healthy alternator should push out 13.8 - 14.4 when spinning at 2k or so. Your voltage seems to drop when you stand on the brake pedal too (lift off shown by low map) which makes me wonder if you're really getting a clean feed of batt V from the battery directly, or if you're cheating and taking it from some dodgy part of the loom.
This is certainly my fault as I haven't built a circuit
to measure the battery reference voltage. You are seeing the
input floating.

I feel a little sheepish about this, I had thought about it at
one point, and then completely forgot.

Sorry for the confusion.

I'm wiring this up today.
<@TekniQue> but in the end, it's code that makes a computer useful
User avatar
Fred
Moderator
Posts: 15431
Joined: Tue Jan 15, 2008 2:31 pm
Location: Home sweet home!
Contact:

Re: sim's Volvo 245 comments

Post by Fred »

LOL, it's all good. I was actually surprised that it was floating, as it looked pretty stable (within 1v from 24v range), too stable to be floating, I thought, but too rough to be correct either. There must be some stray capacitance and/or resistance there. A bit of flux residue maybe? Did you wash your boards with fluxclean? It's conductive, esp with humidity.
DIYEFI.org - where Open Source means Open Source, and Free means Freedom
FreeEMS.org - the open source engine management system
FreeEMS dev diary and its comments thread and my turbo truck!
n00bs, do NOT PM or email tech questions! Use the forum!
The ever growing list of FreeEMS success stories!
User avatar
Fred
Moderator
Posts: 15431
Joined: Tue Jan 15, 2008 2:31 pm
Location: Home sweet home!
Contact:

Re: sim's Volvo 245 comments

Post by Fred »

BTW, quick question, did you clean up that 24/1 disk with some sand paper before installing it? The ones I saw were slightly rough around the edges which could possibly explain the slight wobble in the RPM trace. Just a thought. It could be other things too.
DIYEFI.org - where Open Source means Open Source, and Free means Freedom
FreeEMS.org - the open source engine management system
FreeEMS dev diary and its comments thread and my turbo truck!
n00bs, do NOT PM or email tech questions! Use the forum!
The ever growing list of FreeEMS success stories!
Post Reply