Honda J Series 6 Cylinder

For discussing and developing different RPM/Position decoders using our superior modular architecture! One thread per pattern, please.
Post Reply
Peter
LQFP144 - On Top Of The Game
Posts: 268
Joined: Tue Dec 27, 2011 5:37 am

Honda J Series 6 Cylinder

Post by Peter »

I have a J-series Acura/Honda engine that needs a better engine management system.

Here's a picture of the cam gear?/sprocket/wheel/thingy, and sensor plate. The two shiny caps on the left plate are the sensors. The four darker areas that are pressed in on the lower spokes are the sensor's "triggers?".
Image

The firing order is 1-4-2-5-3-6.

Code: Select all

if(LeftSensor => ThresholdValue and RightSensor => ThresholdValue)
  either 1 or 5 is on TDC so test it against the previous value if there is one or wait for the next high reading to center on 4 or 6
else if(LeftSensor => ThresholdValue)
  if(previousCylinder == 1)
    CurrentTDCcylinder = 4
  else if(previousCylinder == 4)
    CurrentTDCcylinder = 2
else \\No need for an if?
  if(previousCylinder == 5)
    CurrentTDCcylinder = 3  
  else if(previousCylinder == 3)
    CurrentTDCcylinder = 6
My attempt at hopefully correct logic in pseudo code.

I don't know if they're "Hall" sensors, but they're magnetic so they're some derivation of a Hall sensor.
This is one of the cam sensor's output in 5V and 20mSec divisions. The output on the cam sensors, and the crank sensor was about 13.2V peak to peak. The battery voltage I'm guessing.
Image

Over 5mSec divisions.
Image

2V and 2mS
Image

5V and 0.1mS
Image

The crank trigger wheel has 12 high areas on it, and the low areas are about twice as long as the high areas. I don't have one on hand, but I could get one fairly easily if I need to get some measurements off of it.
Crank Sensor at 5V and 20mSec divisions
Image

5V and 2mSec
Image

If you want, you have my permission to turn those pictures into urls if you think they're taking up too much space. They are kind of redundant, but I like pictures. :)
I have a TA card that I could wire up to test with.
:-p
User avatar
Fred
Moderator
Posts: 15431
Joined: Tue Jan 15, 2008 2:31 pm
Location: Home sweet home!
Contact:

Re: Crank+Cam based input pattern support requests thread

Post by Fred »

Awesome post! :-)

Firstly, those are almost 101% certainly VR sensors. The magnet is there to create flux in the coil, and when the steel passes the coil and magnet it changes the flux generating a voltage. As you can see there is a very exact zero crossing as you go past the teeth and then a gradual transition with a very imprecise zero crossing in between two teeth. This is why it's crucial to wire them correctly even if there are no missing teeth.

Secondly, that's a 6-2 pattern! The fancy new missing tooth code can probably run your engine from only one of the cam sensor inputs alone ignoring the crank and other cam signal.

EDIT: I totally missed the text describing the crank sensor, it's VR also.

To run using 2 or 3 of these sensors will involve shuffling existing code around and adding some new code or writing fresh code (but I love reuse...). If you want to get some relative edge signals, you'll have to wait until I implement the fancy new input time stamp logging functionality. You'll also need a VR interface chip such as LM1815 or MAX99xx to read the cam signals. Jean's one is expensive and the other chips aren't that great, but are good enough (engine 9 runs one). For a few reasons, I think this will likely end up only using one of the cam sensors and the crank sensor in the end. It should run OK without the crank sensor, but accuracy will be pretty average/poor during cranking and dynamic RPM change situations.

Re the pictures, I like them too, and the 7kb/s droid connection didn't choke, so it's fine. They're suitably low res and in suitably low quantity, good work.

If you don't mind, I'll split this into its own thread with an appropriate title chosen by you after I do that by setting the subject in the first post.

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!
Peter
LQFP144 - On Top Of The Game
Posts: 268
Joined: Tue Dec 27, 2011 5:37 am

Re: Honda J Series 6 Cylinder

Post by Peter »

I would prefer to have all 3 sensors hooked up so that one or two of the sensors or the circuits behind them can die, and the engine will still run.

What should I try and get in the mail today?
3 - LM1815(I assume), or MAX992x, or NCV1124
Any specialized components those chips need to run
6 - LM1949 and the stuff for that

I've already put a 5V 1A source on the TA card. So I need one resistor and two capacitors
Fred wrote:3.3k, 6.8nF and 680pF (critically damped) Fred's arbitrary recommendation
for the PLL which I might already have depending on what tolerance the resistor needs to be.

On the second page http://support.technologicalarts.ca/doc ... 512dat.pdf it says "Optional precision Voltage reference for Analog-to-digital converters. I assume that would be a 4.096V source like a LM1541. Which I already have around for an MCP3208. So I was wondering if that would be a good use of my time since one of the ADC's resistors already needs to be bridged.

Anything else you can think of?
:-p
User avatar
Fred
Moderator
Posts: 15431
Joined: Tue Jan 15, 2008 2:31 pm
Location: Home sweet home!
Contact:

Re: Honda J Series 6 Cylinder

Post by Fred »

You might prefer it, but it's not possible for the time being, and may not be officially possible ever, hard to say at this stage. You will always be able to hack your own support in, though, no problem there, if you're up to the task. The only real tangible benefit that the extra sensor gives you, which is why Honda threw it on there, is faster starting. The reliability thing is a non issue, VR sensors have no moving parts. The only thing that can go wrong is the wiring, same goes for your injectors but I doubt you'll suggest wiring 12 injectors up, just in case 1 stops :-)

Hmmm, thinking again, you might be able to do three sensors on the current setup, but it might involve some magic code to make it work. Copy the RavAGE RPM input schematic to try that approach.Will it get you run on reliability if you cut one wire? Maybe. Think about an intermittent connection, though. You're going to be receiving a dodgy signal that tells you things that you don't understand, your only choice is to run on and risk engine damage, or give up and throw a sync error and be safe. I don't support having optimistic sync code in the firmware for obvious reasons.

Something might be doable, but it would be overly complex and have little benefit. For a start roll with 2 of them, or even just one (available now) and improve about it later if you still want that.

The max99xx chips are the best of the bunch, but very fine pitched, unless you grab a prebuilt board.

You definitely do not want or need the precision v ref, instead you want to perform the mod as described and reference your sensors against their supply. MAP is ratiometric in most cases, for example, as are all the temp sensors, tps, etc. Referencing from a different source = errors.

By the sounds of it you're planning fuel-only control? If you weren't planning that, see this thread: viewtopic.php?f=54&t=1084

You'll need components for conditioning temperature sensor inputs and similar.

I'm not sure what else, however you should probably start a new thread in DIY hw if you want to plan a build and leave this for decoder discussion :-)

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: Honda J Series 6 Cylinder

Post by Fred »

Thanks for pointing out that window in the wheel. Yes, it likely does change the exact sensor centre timings. I'd still like to see if the missing tooth code has anything to do with it, though. Wait, found a definite issue with the header generation that happens on the 41st event angle, which matches the logs, so you're probably fine code wise. Hard code those signals and claim 10 :-)
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: Honda J Series 6 Cylinder

Post by Fred »

Yep, 100% confirmed typo in header. Pushing now.
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!
Peter
LQFP144 - On Top Of The Game
Posts: 268
Joined: Tue Dec 27, 2011 5:37 am

Re: Honda J Series 6 Cylinder

Post by Peter »

I've been trying to get my car back together today, and I noticed something kind of interesting.

This picture is at TDC on cylinder#1 the cam tooth is at the zero crossing point, but the crank tooth isn't at a zero crossing point. I'll can try and figure out how much offset it has if it's important for a crank with cam sync decoder.
Image
:-p
User avatar
Fred
Moderator
Posts: 15431
Joined: Tue Jan 15, 2008 2:31 pm
Location: Home sweet home!
Contact:

Re: Honda J Series 6 Cylinder

Post by Fred »

It's probably not super important, but it would be interesting. Only if it's not a major hassle :-) Thanks for sharing!
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: Honda J Series 6 Cylinder

Post by Fred »

This is now official! Thanks Peter for both the hacking and testing :-)

Available from master as of now.

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: Honda J Series 6 Cylinder

Post by Fred »

Quick hack in the BenchTest decoder provides this pattern now, see this post: viewtopic.php?f=56&t=1180&p=33804#p33804
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