EFI Testing - Dyno Day 6: 1.5:1 Ratio rockers
Turbo Mini Forums

the map window thing i was thinking may be affecting the signal ..

Attachments:

275.71kb

Medusa + injection = too much torque for the dyno ..https://youtu.be/qg5o0_tJxYM

’77 Clubman build thread
http://www.turbominis.co.uk/forums/index.php?p=vt&tid=618189

Siamese 5 port EFI testing
http://www.turbominis.co.uk/forums/index.php?p=vt&tid=611675

Low pre-load on the actuator could be the problem.

If your boost controller can work below 5psi boost, then I think that you must have a very low pre-load setting. Assuming that the boost controller controls a solenoid valve that bleeds off pressure from the signal line to the actuator.

I've got boost control on my car but I will not use it until I've got everthing sorted at a nominal 0.8barg boost on the actuator alone.

Saul Bellow - "A great deal of intelligence can be invested in ignorance when the need for illusion is deep."
Stephen Hawking - "The greatest enemy of knowledge is not ignorance, it is the illusion of knowledge."

’77 Clubman build thread
http://www.turbominis.co.uk/forums/index.php?p=vt&tid=618189

Siamese 5 port EFI testing
http://www.turbominis.co.uk/forums/index.php?p=vt&tid=611675

And those "AFR offsets" were also fairly much comparable with The EFI dyno Day back in August 2017

’77 Clubman build thread
http://www.turbominis.co.uk/forums/index.php?p=vt&tid=618189

Siamese 5 port EFI testing
http://www.turbominis.co.uk/forums/index.php?p=vt&tid=611675

Interesting.

So how does the MS2 receive the data, digital or 0-5v?

Saul Bellow - "A great deal of intelligence can be invested in ignorance when the need for illusion is deep."
Stephen Hawking - "The greatest enemy of knowledge is not ignorance, it is the illusion of knowledge."

It's digital Paul

Edited by Graham T on 5th Jan, 2018.

’77 Clubman build thread
http://www.turbominis.co.uk/forums/index.php?p=vt&tid=618189

Siamese 5 port EFI testing
http://www.turbominis.co.uk/forums/index.php?p=vt&tid=611675

I assume that the Innovate is also digital?

Saul Bellow - "A great deal of intelligence can be invested in ignorance when the need for illusion is deep."
Stephen Hawking - "The greatest enemy of knowledge is not ignorance, it is the illusion of knowledge."

Honestly, I can't answer that.
I've not got involved in the innovate (LM1 ?) and how it works.

’77 Clubman build thread
http://www.turbominis.co.uk/forums/index.php?p=vt&tid=618189

Siamese 5 port EFI testing
http://www.turbominis.co.uk/forums/index.php?p=vt&tid=611675

A few tests over the weekend…


Firstly to establish any offsets between sensors

Iox.Air filter Pressure sensor, which is on the Input/Output Extender (Iox) board, Not the MS2) connected to manifold.
Iox.Air Pressure Post intercooler, again on the Iox board. This trace is just to show a clean sensor signal on the Iox.
MAP, this is the MS2 MAP sensor.
Logging was started before starting the engine, giving readings at Atmosphere.
After start up, MAP and Iox.Air filter Pressure sensor pretty much matching up, obviously taking into account the 0.54Kpa offset at atmosphere.








So then disconnecting the MS2 MAP Signal pipe from the inlet Manifold.
The first part of the log shows the MS2 MAP sensor to atmosphere.
MS2 MAP then dips to 80KPA with the syringe attached and a little suction added.
Finally, positive pressure put on the MS2 MAP sensor.






So it appears there is no electrical noise on either the MS2 MAP sensor or the Iox Pressure sensors.




All that established, a few logs from a drive or 2 down the road.







I was not on 100% Throttle until 3100RPM in the above, hence the initial boost is slightly low and possibly slow coming in.


Now a 4th Gear pull from the same drive out, 100% throttle from 1700RPM.




Muuuch better. Boost coming nice and early…

Max boost recorded by the MS2 MAP sensor, for 4mm Wastegate actuator pre load was 138Kpa (MS2MAP) or 5.2PSI


The next drive out was with 7mm Preload on the Wastegate actuator.
I only managed to get a reasonable 4th Gear pull logged, hence only low revs.






Max boost was at 144Kpa, or 6.2PSI in other parts of the log.



The next test was with a different MS2, just to see how the MAP compared with my MS2.
So again with 7mm Preload on the Wastegate actuator.
I managed to get a good 3rd gear pull this time…






So that MS2 is pretty much showing the same differences in MAP readings as my MS2 does, compared to the Iox pressure readings taken from the inlet manifold.
Also, it looks like the Iox pressure readings from the inlet manifold closely resemble those of the Innovate pressure readings from the Dyno day.

Next up, the MS2 settings for MAP sampling…

’77 Clubman build thread
http://www.turbominis.co.uk/forums/index.php?p=vt&tid=618189

Siamese 5 port EFI testing
http://www.turbominis.co.uk/forums/index.php?p=vt&tid=611675

Changing the figures for the MS2 MAP Sampling settings, as suggested by Robert and Rod, resulted in some interesting, and I must admit for my part, unexpected results.




I could not set the MAP Sampling window at the 360 Degrees that Robert suggested, as there is a maximum of 50Deg, so I set it at that.





Setting the MAP Sampling window back to 10 Deg and then changing the MAP Sampling Angle to a random 120Deg:






Not only can you see the result of the MAP reading much lower, but it was evident in the car also, I started getting miss firing, which I assume would be detonation. Presumably brought on because the MS2 was seeing 136KPa when the engine was actually getting 144KPA, so it was getting somewhere around 2.5Deg more ignition advance for the same RPM at the same Boost as it was in the previous run.


So, I now need to read the whole MS2 MAP sampling thing yet again to understand the MAP Sampling angle, which I believe is supposed to be set “just before the Valve closes”…

’77 Clubman build thread
http://www.turbominis.co.uk/forums/index.php?p=vt&tid=618189

Siamese 5 port EFI testing
http://www.turbominis.co.uk/forums/index.php?p=vt&tid=611675

just had a thought,

how are the sensors calibrated? if its a 3 point callibration then if the sensors not very inear it might be out at varios points,

actually Ive just checked in TS and its a 2 point cal (0 & 5v) so assumes a linear sensor, which accourding to the datasheet it is, but it looks like sensors could vary up to 15kpa from each other from the datasheet graph, also asuming the mpx4250, there are 2 versions, an A and a D, the A cuts off at 20kpa which would skew the low end readings,


https://www.nxp.com/docs/en/data-sheet/MPX4250D.pdf

https://www.nxp.com/docs/en/data-sheet/MPX4250A.pdf

wow that made a bit of a difference .

Medusa + injection = too much torque for the dyno ..https://youtu.be/qg5o0_tJxYM

Interesting observation Joe.
The version specified for the MS units (standard fitting) is the "A" ie, the absolute pressure version and is described on the datasheet as being for automotive MAP readings and I would be reasonably confident the standard calibration loaded in MS (unless its accidentally been changed to a custom calibration for a higher pressure sensor) will match the "A" graph.

The ones I've been purchasing are definitely the "A" version and I'm 99.9% certain Graham has been using the same Farnell order code.

To my mind, the comparison between the ones in two different MS2 units vs one in an IOx box - which doesn't use a timed window sample regime because, like Robert's setup, it doesn't know the crank position - means the sample window is the more likely explanation.

So I think it becomes a case of understanding the reasoning behind the sample regime and optimising it.

But, in an EFI setup, providing the chosen regime is giving consistent and repeatable readings that follow a linear scale, you will just end up with fueling and timing tables that suit whatever that regime is reading for a given manifold pressure even if it isn't quite the same as the actual average pressure (if that makes sense). To my mind, it's only when comparing to externally sourced readings, or trying to do any efficiency calculations (using the other pressure readings around the cycle) that the real value becomes an issue. And if a "real" reading is needed for other calculations/comparisons, use another independent one wired to one of the remaining ADC inputs on the IOx where the whole group will be calibrated the same and all independent of the engines rotational position.

Schrödinger's cat - so which one am I ???

Ah,

I didnt see ypur last post Graham, I was probably looking at the datasheets as you posted,

yep that MAP angle is really making a difference!

Hmmm thats very interesting Rod, I hadnt considered that the Iox would'nt know crank position, I was planing to use one to measure stuff like TIP, but might have to rethink that.

Yes, I'm using the A version MPX4250.

The MAP calibration is set as default for the MPX4250 (from the pull down list):
Value at 0.0Volts(%) = 10
Value at 5.0 Volts(%) = 260

Which seems not correct according to the data sheet?

’77 Clubman build thread
http://www.turbominis.co.uk/forums/index.php?p=vt&tid=618189

Siamese 5 port EFI testing
http://www.turbominis.co.uk/forums/index.php?p=vt&tid=611675

More interesting data, thanks for posting. My analysis:

The IoX data is showing cycle average values upstream of the carb. The Innovate data is showing cycle average values from the manifold runner. The MS is sensing the manifold runner pressure during a specific cycle angle window.

The variation between the Innovate and MS data is to be expected. At boost of say 1.0 Barg, the pressure in the runner will vary from 0.5 to 1.5 Barg. You are seeing the pulse movement by changing the map sampling window settings. That's why you can see higher MS values at low revs and then they drop as the revs climbs and the high point of the pulse moves away.

There's nothing wrong here.

Edited by Paul S on 8th Jan, 2018.

Saul Bellow - "A great deal of intelligence can be invested in ignorance when the need for illusion is deep."
Stephen Hawking - "The greatest enemy of knowledge is not ignorance, it is the illusion of knowledge."

Schrödinger's cat - so which one am I ???

it looks like what they have done is extrapolated the average line out to where it would cross the 0 and 5v ponts, on the assumtion that the engiine is never going to hit these points.

Schrödinger's cat - so which one am I ???

Actually, Ignore me.


The Sensor sensitivity is 0.02V/KPA
(Sensor output voltage / sensitivity) +10 KPa seems to come in line with the graph on page 5 of the MPX4250A datasheet.
Which is good, because that’s what I have been using as the folmula for the Iox Pressure channels.

’77 Clubman build thread
http://www.turbominis.co.uk/forums/index.php?p=vt&tid=618189

Siamese 5 port EFI testing
http://www.turbominis.co.uk/forums/index.php?p=vt&tid=611675

Thanks everyone for the feedback.

I’ve just done a little more reading on the MS2 forums and the MS3 manual with regard to MAP Sampling.

As per Rod’s comments above, it seems repeatability is required over accuracy for MAP sampling, which makes sense.

The general advices appears to be to make the Sampling angle “just before the valve closes”, so for me that would be somewhere around 124 Deg BTDC.
That is a bit simplist though and is more in depth advice on setting MAP sampling in the MS3 setup Manual.


The version of code I am using is limited to only using the “Timed min” MAP sampling strategy, whereas later MS2 and MS3 codes have an “event average” MAP sampling strategy, which appears to be the Megasquirt preferred sampling strategy.
The version of code I run also means I do not have the MAP logger available to do more in depth analysis and setup.


The “event Average” Sampling strategy might offer a more accurate MAP reading, but that would mean I would lose Traction Control as I currently have it implemented, and possibly gain me little to nothing with regard to tuning and everyday driving.
Ultimately I need to look at how else I can implement Traction control, and probably do a bit of testing with a newer MS2 code release.


However, I think maybe for now I will continue to use the Air filter pressure sensor for a “true” reading of MAP as an external comparison, or install a 4th MAP sensor for Iox pressure measurements.

What I will also do as a test is to add a fuel filter inline from the Inlet manifold to MAP sensors to see if that helps smooth the MAP out.

Edited by Graham T on 8th Jan, 2018.

’77 Clubman build thread
http://www.turbominis.co.uk/forums/index.php?p=vt&tid=618189

Siamese 5 port EFI testing
http://www.turbominis.co.uk/forums/index.php?p=vt&tid=611675

it does look like with a bit of jiggery pokery it will match the innovate most of the time .

Medusa + injection = too much torque for the dyno ..https://youtu.be/qg5o0_tJxYM

A decent buffer on the signal line should sort it out.

Here's on old plot from a simulation showing the variation of pressure in the inlet runner:



The peaks of the pulses ocurr at different times in the cycle depending on the engine speed. The pulses move at the speed of sound and are not synched to engine speed. So with your map sampling you will be picking up the MAP during the intake event, but it will not be the same as the cycle average.

Saul Bellow - "A great deal of intelligence can be invested in ignorance when the need for illusion is deep."
Stephen Hawking - "The greatest enemy of knowledge is not ignorance, it is the illusion of knowledge."

http://www.jbperf.com/