The V8 Owners Forum -

So a couple of months ago I was merrily driving along the a3 when the Land-rover started to surge. This was running on singlepoint LPG, with the closed loop system working via a simulated narrowband output from my LC1 driving a stepper motor controlled power/economy valve.

Datalogs of the problem showed that the AFR was like a sine wave with the peak at 22:1! I suspected that this was either down to:

-Dodgy readings from the wideband itself, causing the closed loop system to go bananas.
-Problems with the black box that controls the closed loop.
-Problem with the stepper motor not doing what it was supposed to.

I disabled the closed loop, and went back to no automatic mixture control, life was good, but expensive.

I took some more datalogs of the problem through my Megasquirt (controlling ignition only at the mo) with the car running open loop but the wideband in and sniffing exhaust. These still showed some degree of bouncing AFR's with the readings moving up and down by a whole integer value or more.

I would have suspected noise to be the culprit, but the MS and LC1 are powered and earthed at the same locations (bus bars in the battery box) and the MS is perfectly happy- no resets etc. Also the period of the sine wave seemed to be around 1 second, too slow for noise surely?

Screenshots here:

http://forums.lr4x4.com/index.php?showtopic=70067

Next up I decided to lashup the LC1 so that it was powered from a couple of 3S (11.1V) Lipo batteries that I have for my various radio controlled toys. These were wired in parallel, each rated at 2200mah with burst discharge rates of up to 50C. The LC1 was therefore fully independent of the 90's electrical system.

The downside to this was that I could not log from squirt as the potential differences involved between LC1 and MS would give me problems. I am therefore stuck with working only with the logworks files, which give no indication of load, rpm etc.

I took these logs this afternoon- I still think that they are too spiky. And at one point the gauge went berserk, again bouncing from rich to massively lean with the car showing no undue effects- i put this down to a problem with the readings, and not an actual problem with the mixtures.

So- can someone take a look at these and tell me if they think they look bad? Apologies for the rubbish file host- I hope they download ok.

You will need logworks to view them.

Engine running (I stopped for LPG at one point with the LC1 still powered, engine off):

http://www.filedropper.com/lipos

Engine switched off (looks good at the start but gets progressively worse):

http://www.filedropper.com/engineoff

I even tried covering the tip of the exhaust on this one to try an negate any impact of the wind blowing.

I have also tried multiple free air calibrations, and heater calibrations, all with the same results.

So the big question is this- do these look normal to you? (ignoring the actual AFR's and focusing on the consistency of readings for now)

Thanks in advance


Jamie

Can you get hold of a battery powered oscilloscope? If so, you could monitor AFR, the crank sensor, etc etc, to see if it is intermittent.

Chris.

As a rule of thumb, each and every sensor ground should be run back to the DB37 plug of the MS individually - not to where the MS is grounded to the car electrics. Basically, to prevent the high currents which flow in the MS ground interfering with the tiny sensor currents.

DaveEFI wrote:As a rule of thumb, each and every sensor ground should be run back to the DB37 plug of the MS individually - not to where the MS is grounded to the car electrics. Basically, to prevent the high currents which flow in the MS ground interfering with the tiny sensor currents.

Is that correct Dave? I was under the impression that all the earths should go to a common ground on the engine. Of course, I've been wrong before

The common ground point should be the MS itself. If you look at the MS external wiring diagram, it's pretty clear about this.

http://www.megamanual.com/ms2/V3assemble.htm#ew

If you're using an MS relay board, follow the instructions for that.

The important thing is to avoid high currents flowing in the sensor grounds. If you ground them elsewhere this may prove impossible.

Of course it's all a bit of a black art - you might well get away with another arrangement. However, factory EFI - even the flapper - brings back each and every sensor ground individually to the ECU. Which is then grounded to the engine block via separate wires. Exactly as the MS schematic shows.

Seems there may be some confusion about MS grounds!

The info i've used:
(http://www.extraefi.co.uk/wiring_basics.html)
says to wire the sensor grounds back to the engine to the same point where MS earths using individual wires. I have done this and works fine.

Regarding the erratic O2 issue, if it was running fine then suddenly started going erratic I would first check your wiring and connectors. Double check the grounds for your O2 sensor, should be the heating ground (thick cable) sensor ground and maybe a differential ground all going back to where MS grounds. Or going back to the MS the back to where it grounds (engine) via indiviaual wires.

How old is the sensor? They do have a 'working life' which can be limited if their running conditions aren't correct (too rich, too hot etc)

Tom.

I've seen this before, and while it may well work is not good practice IMHO. No car maker I've ever seen does it that way. They all bring back the individual sensor grounds to the ECU, which is then in turn grounded to the block. So the 'star' of the grounds is the ECU itself - not elsewhere.

Again IMHO, it's an easier way of doing things, and I've built several looms. But don't use the MS relay block.

I'm not using the extra code, but B&G, so followed their wiring. Which was exactly what I'd expect from my experience as an audio engineer.

If you want to control a narrowband setup with a sensor....use a narrowband sensor.

I wouldnt trust a simulated output, and the ones that do simulate, often dont do it as good as a NB sensor would.

And whilst I cant see the readings. Sensor life with the LC-1 isnt measured in years most of the time. They can be very temperamental and do need replaced.

Thanks for all the replies guys, was working fine over last few months then just went nuts! I had been meaning to take it out as all twiddling with fuelling was finished a while ago. I will try another sensor and go from there. In the meantime I've put the narrowband back in and recomissioned the closed loop control on Lpg.

Followup to this post - I recently bought a new sensor and this threw an error code of 8 as well!

Got on the interweb and did some trawling of the innovate forums, seems that they can generate this error for any number of reasons...

Anyway I flashed to the version 1.2 firmware (the beta version) and all is now working.

At least I have a spare sensor now for the day it really does die

Quagmire wrote:Followup to this post - I recently bought a new sensor and this threw an error code of 8 as well!

Got on the interweb and did some trawling of the innovate forums, seems that they can generate this error for any number of reasons...

Anyway I flashed to the version 1.2 firmware (the beta version) and all is now working.

At least I have a spare sensor now for the day it really does die

Make sure that you keep your LC1 away from anything that creates HT voltages, they really don't like that! (I've totally blown up one unit because of this, there is a lot of info on the web on how badly they are shielded from HT)