Jump to content
jetape

ECU afr not matching stand alone gaugue

Recommended Posts

This is making me a little nervous, my dynotune nitrous digital gauge is way off from what I see in the ecu.  By way off I mean if the ECU shows 13.5 the gauge will show mid 14s.

I can't figure out what is wrong, I think I have everything configured correctly.  I'm using two outputs from my LC2 controller, one goes to the gauge and one goes to the ECU.

LC2 output #1 goes to the gauge, #2 goes to the ECU.  I found gauge documentation at http://www.dynotunenitrous.com/store/INSTRUCTIONS/GAUGE AIR FUEL RATIO WIDEBAND.pdf which says 1V = 10 AFR and 2V = 20 AFR

The controller is wired directly to the battery since it requires a lot of current for the heater.  Gauge is also powered from the battery.

What am I missing here?1

 

 

lambda-calc.PNG

lc2-1.PNG

lc2-2.PNG

Share this post


Link to post
Share on other sites

I'd trust the ECU reading more than the gauge, a 1V span for your AFR is quite small and will be massively impacted by any ground offsets.  When you connect to the LC2 with the laptop, how does the reading in the innovate software compare to what you're seeing in PClink?

Share this post


Link to post
Share on other sites
3 minutes ago, JMP said:

I'd trust the ECU reading more than the gauge, a 1V span for your AFR is quite small and will be massively impacted by any ground offsets.  When you connect to the LC2 with the laptop, how does the reading in the innovate software compare to what you're seeing in PClink?

Yeah I definitely trust the ECU more but I got a little paranoid that I may have screwed up some config in the ECU.

I wasn't aware Innovative software had the option to see real time data, I just looked at the programmer tool.

Share this post


Link to post
Share on other sites

So I did some testing, I programmed LC2 to output 2.5V and the ECU saw 2.53 with the engine off and 2.34 with the engine running. Did very minor calc tweak to account for the 0.03V difference.  Is it normal to have that kind of a difference between running and not running?

I used an empty, but recently used, gas can and I could trigger different AFR values depending how far I put the sensor in the can :)  With some tweaking on the gauge LC2 output I got it to track the ECU, it seemed to have been off by 0.4-0.5 AFR point.

I also checked with LogWorks and it's matching the ECU.

I'm just not sure what to make of the voltage difference the ECU sees when running vs off.  Isn't that throwing off the AFR?

Share this post


Link to post
Share on other sites

The Ground offset is always a problem with the innovate wide bands.  Usually you will get the best results if the innovate ground wire is connected direct to the engine block or cyl head,

Share this post


Link to post
Share on other sites

It's grounded to the battery, I have a kind of a Y splitter at the batter ground thats 3-4" long and from that a few ground quick connects.  Would running it to the block be better?

Share this post


Link to post
Share on other sites

The ecu would usually be grounded to the block so if you ground the innovate to the block also you should have the best chance.  Having said that, many people have had issues with getting consistent readings from the innovate. 

Share this post


Link to post
Share on other sites

Oh man, so I ran a 16gauge ground wire, less than a meter, from LC2 to the same spot on the block where ECU has it's ground.  It made no difference, with motor off the source 2.50 ends up being 2.54 at ECU, at idle1800 RPM it's 2.34v.

There was literally not even a 0.01 volt difference with the engine running using this new ground spot.

I did a quick throttle blip and it doesn't seem to change with RPM.   Maybe I'll be just stuck with making a calc correction for this unless there's something else I can try.

 

Thanks

Share this post


Link to post
Share on other sites

I checked signal voltage using a multimeter directly at the LC2 where factory wiring terminates.  I grounded the multimeter at the motor, with LC2 grounded at the motor voltage was 2.57x and with it grounded at the battery it was 2.58.  There was only ~0.01 V difference between engine running and not running.

All this was again with 2.50 V coming from LC2.

 

Could there be some interference on the signal between the LC2 and the ECU?  My signal wire is 18g about 1.5m in length, it's in a split loom with a ground and +12 but those aren't in use.

 

 

Edit: Just realized I could have checked signal just before it goes into the ECU, tomorrow.

Share this post


Link to post
Share on other sites

Onboard lambda is the correct way...no fucking about with CAN or Innovate....the latter being the most troublesome.

 

The only way to do Innovate...is to manually test and calibrate over a sensible voltage range.

 

Flatline the LC2's output at 0.5v, 1.0v, 1.5v etc etc and test what you see on the ecu for each voltage. Then DIY your calibration vs AFR to suit.

 

And do it over as small an AFR range, for as wide a voltage range as you can....ie use the full 0-5v, and limit the AFR range to what you need, and is also still functional.

That way small voltage errors or offsets will have less impact on the reading

Share this post


Link to post
Share on other sites

So I tested the voltage as close to the ECU as I can.  My vipec has short, maybe 6", expansion port cables where checked the voltage at the same time as the ECU with a multimeter.

Both LC2 and multimeter grounded to the engine.

With engine off there was ~0.01 - 0.02 difference, 5.57 - 5.54

With engine running the voltage at connector barely changed, maybe 0.02V so it was showing about 5.56-8.  The ECU on the other hand reports the usual 2.34V

 

 

 

20180815_183301.jpg

Share this post


Link to post
Share on other sites

I also did a short log whats interesting is it doesn't look to be tied to the engine actually running.  The log shows that it took the voltage almost 4 seconds after the engine stopped to recover.

I think I'm down to trying to work around this.  Set LC2 to wide voltage range like 0.10 - 4.90 and AFR 11 - 16 to minimize error and then adjust for the what it seems to be 0.16V offset.

Anyone have a offset formula calculation I can apply to my numbers?

 

 

groundoffset.PNG

Share this post


Link to post
Share on other sites

From that log, it looks like you might have some of your wiring or a relay/fuse a bit too small for the amount of current being drawn, which is causing some voltage drop. What does the wiring look like on the positive side of the LC1 and the ECU? Everyone focuses on grounding but voltage drops on the positive side can cause similar offsets if the sensor does not have any internal voltage regulator. At a guess you have some device fed from the same 12v fuse/relay as the ECU or LC1 drawing more current than the wiring can handle, and 3-4 seconds after the engine stop, this device turns off and so stops overloading the wiring.

You can measure voltage drop across a wire by setting your multimeter to volts, and sticking one end of the meter at each of the wire while the circuit is live - eg one terminal on the battery negative, one into the back of the ground pin of the LC1. Then try between the engine block and the LC1 ground pin. If you see any meaningful voltage difference, you have some resistance on that path. If you see bascially zero, there is no issue with the level of current passing through (hence why you have to do this running as you're basically testing if the wire is big enough for the current its flowing). Then test the same thing on the ECU grounds > engine block, then LC1 positive to alternator, then LC1 positive to battery. 

How did you stop the engine without having the ECU power off to generate that log? This might be relevant to why it took a few seconds for the voltage to return to normal. Also, can you try the same checks on both sides of the ECU as well? It's less likely but possible the LC1 wiring is up to scratch, but the ECU wiring is not and this is causing the relative voltages to report wrong. If the voltage reported by the ECU is always the same 2.34V, try pulling out the ANVolt pin and measuring the end of the wire where it should go into the ECU - do you see 2.34V or 2.57v?

Share this post


Link to post
Share on other sites

The vipec ECU is basically a plug in, it plugs directly to the stock jetski harness just like the stock ECU.  No wiring is touched.  I can't get directly to the vipec pins, this ecu is totally waterproof, it's sealed and the inside is actually filled with something so no water will ever touch the electronics.  The expansion ports are the only accessible connections.

LC1 is powered via relay at the battery, the only other thing on that relay is a tiny digital gauge.  I believe I have a 5amp fues there, all my connections are solid, I use waterproof connectors and solder all the pins to wire.

 

OK as I was writing this I had another look at the log to see if I can find anything to match the drops, and I did.  It matches exactly the fuel pump ON/OFF status. Now I wonder if the stock setup even has a fuel pump relay, could be what is causing the current draw at the ECU.

Ignore the large drop to 1.3v at the beginning, that's the sensor heating up.

groundoffset2.PNG

Share this post


Link to post
Share on other sites

You could try running some rather large wires from battery to the fuel pump relay, then from the relay to the pump, and from the pump to ground. I'm not sure about jetskis but its relatively common in cars for the factory fuel pump wiring to be a bit limiting and people regularly report a volt or 2 lower output across the fuel pump than they expect. 

Is it possible the ECU and fuel pump share a common power or ground wire somewhere?

Can you post the log file, or a screenshot showing battery voltage during the same time window?

Share this post


Link to post
Share on other sites

Here is the log.

There is no relay to trigger the stock fuel pump.  It's triggered by the ECU via ground output.

Pretty sure I need to add a relay so that the ECU triggers the relay which in turn would pass ground to the fuel pump instead of passing ground to the pump directly causing the voltage drop at the ecu.

 

groundoffset.llg

fuelpump.png

Share this post


Link to post
Share on other sites

Success! It was the darn fuel pump, I put a relay between the ECU and the pump and no more voltage drop.  I guess the stock ECU was design to handle the current draw without a relay.

Ground offset now looks a lot more acceptable.

 

groundoffset3.PNG

Share this post


Link to post
Share on other sites
40 minutes ago, jetape said:

Success! It was the darn fuel pump, I put a relay between the ECU and the pump and no more voltage drop.  I guess the stock ECU was design to handle the current draw without a relay.

 

1

The ECU is designed to drive the fuel pump directly.  It has large power mosfets in it to drive the fuel pump, they are rated for 120A continuos, I think you would melt the pins off the connector before the mosfets gave trouble. I suspect more likely the cause is one of the main ECU grounds is marginal due to corrosion or something so with the load of the fuel pump going through them it causes the large volt drop. 

Removing this load by fitting an external relay with a good ground is an acceptable fix provided the assumed bad ECU ground doesnt deteriorate further in the future.  These ECU's also have built in ignitors so all of the ignition current must pass through the ECU grounds also.

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

×