References values MEMS3

Hello!

Do you know were à can find the references values for the mems3 and the range values please?

Thanks

Ludo

Having reference ranges would be hugely helpful - but I don' t think that this information is widely available on-line? Perhaps it would be good to set up a data repository?

I agree. Reference ranges for all MEMS versions would be very useful.

I did start compiling a spreadsheet for my MEMS2J car some time ago but that is not to say that my values could be used as reference values of course. I now have a MEMS1.9 car so could record some values for that if it would help.

I guess if we had values from say 3 cars of each MEMS version we would have a 'reasonable' view of expected values. I've got a spreadsheet set up if anyone would like to send me some screenshots - or post them here, making sure you note which MEMS version you have. I'll populate the spreadsheet and post it back here if it would be helpful.

Presumably, we would want to capture the values with the car up to temperature and at idle, at least in the first instance?

Hi Robin,

Sounds like a plan. You'll need different engine states to be recorded: off, cold idle, warm idle would be minimum dataset I would think.

What does your spreadsheet look like?

Note that sensors will generate a range of values, not just one. The way I am thinking to handle this is to use the live data logging facility to capture 30-60 seconds of data in the three states mentioned. Each data point could then be reduced to a mean value and a standard deviation

Hi Rob

Live data logging wasn't available when I started playing with pscan - still isn't for Linux I don't think but is imminent (Edit: available for 64bit Linus - 32bit imminent).

Don't think Excel spreadsheets are allowed uploads but here's a .pdf and .docx version to show what I was doing. This was 4 years ago so pscan has probably changed quite a bit since then. When I get a chance I'll plug mine in and do a couple of screenshots to check that my headings are still the same.

Hi Robin .
This sounds really good and your spreadsheet is great . My TF is sorn'd till spring and not convenient for me to start playing with Pscan but our Classic Mini is in the garage so I can have a play with it this weekend. It won't be of much use to folks on here being the A series engine but it is MEMS2 and may be of interest to MPI Mini owners .
John

As I suspected, pscan has changed in the four years since I built this spreadsheet. The revised spreadsheet is as follows.

As before, I can't post this as an Excel sheet but .ods seems to be allowed so I'll try that. It should open in Excel or LibreOffice anyway. All feedback - negative or positive - would be appreciated. Is it worth pursuing this or is it just of minor interest?




The only value that I could see that changed significantly over time from the log capture was the Oxygen Sensor Voltage and this varied between 0.02 and 0.84 on my car.



Live logging is interesting - I may have a go at it while driving the car around to see what else it might show.

Thanks once again to Philip for a superb tool for our pre OBD2 cars and for continuing its development.

cairnsys wrote: The only value that I could see that changed significantly over time from the log capture was the Oxygen Sensor Voltage and this varied between 0.02 and 0.84 on my car.

Working as it should, wish mine was.

My PC appears to be struggling with the .ods file for some reason, but the PDF looks good.

For emissions-related functions (Lambda etc) - then it would be good to correlate these figures with the last MoT emissions data. My guess is that one of the commonest reasons why folks will be using the MEMS data download will be to work out why their car has failed its MoT!

Therefore, it would be good to collate both known "good" data and known "bad" data.

The former is a big enough project, so let's start with that, concentrating on the "good" data

Are you using Microsoft Office software to open the .ODS file?
You're better of using Open Office for this, you can download it for free.
IMHO it shits all over the Microsatan rubbish.

Here are the Lambda logged data from a MEMS3 equipped Freelander 1.8 K-series. The engine was pre-warmed, but switched off for about 10 minutes to allow the exhaust temperature to cool.

This appears to be completely normal operation. There are two lines of data - the pre-catalyst sensor and the post catalyst sensor.

Both traces start at the same point when the ignition is switched on, and then diverge as the exhaust warms up. The Lambda sensor has an operating temperature of 300C - which explains the flat trace before the signal starts to oscillate (blue line). The post oxygen Lambda tends not to oscillate - which is normal. If it were to oscillate in rhythm with the pre-cat sensor, then the catalyst has failed...

I recorded these traces with just two channels - for both the Lambda sensor. The re-fresh rate is then somewhere between 6-8 times per second. If you have more than this open, the re-fresh rate plummets - add engine speed or oil temperature for example, and the re-fresh rate drops to 3-4 times per second. Open all channels for recording, and the refresh rate appears to be about 3.5 seconds! Less is more if you want some fidelity of the Lambda trace (and arguably, you would )

Note that I did the recording both at natural idle and at a fast idle of between 2,500-3,500 rpm. This replicates the MoT conditions. The cold idle will have a lower exhaust temperature, the fast idle a much higher exhaust temperature - and it is under the latter condition that the Lambda will likely give the most reliable data.

What struck me is that the switching frequency varied so much between natural and fast idles. The ECU feedback cycle is slow when the idle speed is low - just once evry 3 seconds or so (approx. 0.3Hz), but increases significantly with a fast idle to once a second (approx. 1.0Hz). I don't think I've ever seen this on a trace published online - so was quite fascinating to me.



This data telemetry would be very helpful in diagnosing MoT emissions fails - if you repeat the above with an oil temperature of at least 70C, natural and fast idle (2,500-3,000 rpm), you will see whether the Lambda is behaving itself. An old Lambda will have a lower mean signal (lower peak, lower minimum) and may also oscillate at a lower frequency. Old Lambdas work better at warmer temperatures - so at idle there may be a poor signal that is recovered by fast idle (hotter exhaust gas temperature, and therefore a hotter Lambda sensor). A failed catalyst will be picked up on MEMS 3 through the post cat signal that will oscillate with the pre-cat sensor.

Hope this is of interest!

I have also captured about a minutes worth of warm idle data from my 2001 Freelander (1.8S) with MEMS3. This data I have averaged, and calculated the standard deviation. I have done this for both ISO and BMW proprietary modes (they're slightly different).

ISO MEMS3:

Variable	mean	SD
Ignition	on
Control unit status	enabled
Coolant temperature (°C)	85.26	0.55
Oil temperature (°C)	79.22	0.18
Intake air temperature (°C)	40.40	0.13
Ambient air temperature (°C)	-40.10	0.00
Intake manifold pressure (kPa)	33.06	0.23
Torque conv clutch pos sens voltage	0.59	0.00
Engine speed (rpm)	773.81	12.34
Battery voltage (V)	13.97	0.12
Vehicle speed (kph)	0.00	0.00
Vehicle motion	not moving
Adaption of throttle valve act motor	29.00	0.00
Idle stabilisation	completed
Idle control	active
Idle conditions	ok
Desired idle speed (rpm)	775.00	0.00
Throttlealve position	closed
Throttle valve servo motor (steps)	37.35	0.61
Idle speed control	-3.58	17.16
Injection time cylinder 1 (msec)	2.69	0.05
Injection time cylinder 2 (msec)	2.69	0.05
Injection time cylinder 3 (msec)	2.69	0.05
Injection time cylinder 4 (msec)	2.69	0.05
Injection time correction (msec)	0.50	0.00
Additive mixture correction IV error	101.89%	2.40%
o2 sensor voltage bank 1 sensor 1 (V)	0.43	0.23
o2 sensor voltage bank 1 sensor 2 (V)	0.72	0.01
Lambda control	active
Mixture	lean/rich
o2 sensor ready	ready
Lambda probe function	normal
o2 sensor heater (bank 1 sensor 1)	100.00%	0.00%
Crankshaft signal	detected
Ignition angle (degrees)	13.49	0.86
Ignition angle correction (degrees)	0.00	0.00
Fuel evap controlalve duty cycle	7.00%	0.00%
Fuel pump relay	on
Exhaust gas relevant misfire total	0.00%	0.00%
Misfiring Cylinder 1	0.00%	0.00%
Misfiring Cylinder 2	0.00%	0.00%
Misfiring Cylinder 3	0.00%	0.00%
Misfiring Cylinder 4	0.00%	0.00%
Power steering pressure sensor	14776.70	2.40
Generator load	0.59	0.01
Aircon request	off
Airconap temp (°C)	-273.20
Aircon compressor release	inactive
Radiator fan relay 1	inactive
Radiator fan relay 2	inactive