Donations towards server fund so far this month.

 
£0.00 / £100.00 per month
Page:
Home > MS Trials & Testing > MS3X MPi Trials

dan187

User Avatar

747 Posts
Member #: 6724
Post Whore

Wootton Bassett

I've rebuilt an MPi engine for use with an MS3X. The intention is to get it running and tune it N/A before later adding bespoke manifolds and a GT15 turbo.

I've managed to get the basic setup to give me a running engine. Needs work to balance the AFRs but the image (or linked datalog) shows my state of play after this evening getting it going. It's all very exciting.



https://dl.dropboxusercontent.com/u/1566128...25_23.55.26.msl

1275 N/A Sprite, 998 T2 Turbo Mayfair
1275 EFi Turbo


Paul S

User Avatar

8544 Posts
Member #: 573
Formerly Axel

Podland

You might want to make the two AFR scale ranges the same - it will make the AFR distribution look a lot better.

Also you may want to use a spreadsheet to analyse the data:

http://www.turbominis.co.uk/forums/index.php?p=vt&tid=319675

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."


Rod S

User Avatar

5814 Posts
Member #: 2024
Formally Retired

Rural Suffolk

As Paul says, both matching the scales in the "pictures" helps and working with a spreadsheet is much easier (as per our emails of a couple of days ago).

The dropbox link has corrupted the raw data, well at least the table headers, they are over the wrong columns for some reason, which makes it a bit of a guess reading it.

But most of the AFR matches are very good but at certain RPMs, bad.

Without being 100% certain what the data columns are it's difficult to be sure but (assuming this isn't on the road yet) I would guess at AE.

Also, what widebands are you using ??

Looking at the raw data, with the engine not running, they both appear to read the same offscale rich (7.5) but read different for off scale lean (17.2 and 22.3) but there may be a good reason for that, the data block isn't long enough to see what was happening before the off scale lean readings.

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


dan187

User Avatar

747 Posts
Member #: 6724
Post Whore

Wootton Bassett

ah thanks yeah excel would be much better. sorry about the scales, it was late and i just wanted to get a quick bit of info up.

So I want to confirm how you'd use the cylinder trim to compensate this setup (if you can)... I'm running 4 cylinder sequential on the two standard MPi injectors.
I've connected up A & C to the primaries. With required fuel based on twice the engine capacity.

I think i'm correct in saying that this injects the fuel so that the start of the pulse goes to the inner cylinder and rest goes to the outer cylinder once the inner valve has shut?

Presumably then I can't use the cylinder trim functions in this setup. Is it best then to run each injector off of two ECU outputs so that A,B,C,D are all connected and then I cant trim A independent of B (for example).

1275 N/A Sprite, 998 T2 Turbo Mayfair
1275 EFi Turbo


dan187

User Avatar

747 Posts
Member #: 6724
Post Whore

Wootton Bassett




On 26th Jun, 2013 Rod S said:


The dropbox link has corrupted the raw data, well at least the table headers, they are over the wrong columns for some reason, which makes it a bit of a guess reading it.




if you right click the link and "save target as" you can download the msl file

On 26th Jun, 2013 Rod S said:


Also, what widebands are you using ??

Looking at the raw data, with the engine not running, they both appear to read the same offscale rich (7.5) but read different for off scale lean (17.2 and 22.3) but there may be a good reason for that, the data block isn't long enough to see what was happening before the off scale lean readings.


Both AFR sensors are Innovate MTX-L.

The off scale rich i think is before they are up to temperature (so the gauge otputs a rich reading) and I have assumed the lean is the different conditions in the exhaust once the engine stopped. (i'll keep an eye on it and confirm)

1275 N/A Sprite, 998 T2 Turbo Mayfair
1275 EFi Turbo


Paul S

User Avatar

8544 Posts
Member #: 573
Formerly Axel

Podland



On 26th Jun, 2013 dan187 said:
ah thanks yeah excel would be much better. sorry about the scales, it was late and i just wanted to get a quick bit of info up.

So I want to confirm how you'd use the cylinder trim to compensate this setup (if you can)... I'm running 4 cylinder sequential on the two standard MPi injectors.
I've connected up A & C to the primaries. With required fuel based on twice the engine capacity.

I think i'm correct in saying that this injects the fuel so that the start of the pulse goes to the inner cylinder and rest goes to the outer cylinder once the inner valve has shut?

Presumably then I can't use the cylinder trim functions in this setup. Is it best then to run each injector off of two ECU outputs so that A,B,C,D are all connected and then I cant trim A independent of B (for example).


I strongly suggest, in the first instance, you use the injection timing to equal the AFR distribution as much as possible. This will be the simplest way to get your setup running well. You can also get the hang of WUE, ASE, AE and all the other secondary fueling settings that make for good driveability.

Once you have a "good" VE table and a handle on the injection timing, then you may wish to try using all four primary channels, cylinder trimming and O2 feedback loops on the inner and outer cylinders. BUT it's altogether unproven theory at this stage.

I might be trying this myself in the not-to-distant future.

Edited by Paul S on 26th Jun, 2013.

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."


Rod S

User Avatar

5814 Posts
Member #: 2024
Formally Retired

Rural Suffolk

On 26th Jun, 2013 dan187 said:
if you right click the link and "save target as" you can download the msl file


Schoolboy eror - I should have downloaded it rather than assume Firefox was displaying it correctly.
Formatting is fine once downloaded and into excel.


On 26th Jun, 2013 dan187 said:
Both AFR sensors are Innovate MTX-L.

The off scale rich i think is before they are up to temperature (so the gauge otputs a rich reading) and I have assumed the lean is the different conditions in the exhaust once the engine stopped. (i'll keep an eye on it and confirm)

The 7ish is fine, so is the 22ish according to their datasheet.

It takes a while for both exhausts to go fully lean on shutdown but, on mine, the two Techedges do keep reasonably in step getting there.

So, yes, just keep an eye on it to see they both reach 22ish once the exhaust is full of air.

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


dan187

User Avatar

747 Posts
Member #: 6724
Post Whore

Wootton Bassett


On 26th Jun, 2013 Paul S said:


I strongly suggest, in the first instance, you use the injection timing to equal the AFR distribution as much as possible. This will be the simplest way to get your setup running well. You can also get the hang of WUE, ASE, AE and all the other secondary fueling settings that make for good driveability.

Once you have a "good" VE table and a handle on the injection timing, then you may wish to try using all four primary channels, cylinder trimming and O2 feedback loops on the inner and outer cylinders. BUT it's altogether unproven theory at this stage.

I might be trying this myself in the not-to-distant future.



Yeah sound sensible to me, will do. Also was i correct in my description of how my current setup (i.e. using injection timing) works... saying that it injects the fuel so that the start of the pulse goes to the inner cylinder and rest goes to the outer cylinder once the inner valve has shut? (and presumably that's why it is easier to use mid-pulse timing?)

I look forward to your results, hopefully I will be able to contribute once I get this running well.

1275 N/A Sprite, 998 T2 Turbo Mayfair
1275 EFi Turbo


jbelanger

1229 Posts
Member #: 831
Post Whore

Montreal, Canada

Yes your understanding is correct. And the mid-pulse timing is not only easier but necessary. That's the only way to make sure the fuel distribution stays the same (or close to it) when conditions change for the same load and RPM resulting in a different pulse width.

http://www.jbperf.com/


Paul S

User Avatar

8544 Posts
Member #: 573
Formerly Axel

Podland

Simply advance the injection timing to richen the inner cylinders and vice-versa.

You could first try to get rid of the divergence exhibited in that plot above by reducing the injection timing at the low map sites.

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."


minivan63

113 Posts
Member #: 9645
Advanced Member

South Wales

Well done for getting it running particularly with the Mpi crank/cam position sensors!

I'm still some way behind!

Are you using the Mpi inlet manifold and injector set up or have you fabricated something? If you are using the Mpi inlet do you plan on using this for the turbo as well?

I guess you might have to do quite a bit of tuning of the injection timing map due to differences in the camshaft and injection set up that Paul used to develop the initial map.


dan187

User Avatar

747 Posts
Member #: 6724
Post Whore

Wootton Bassett


On 27th Jun, 2013 minivan63 said:
Well done for getting it running particularly with the Mpi crank/cam position sensors!

I'm still some way behind!

Are you using the Mpi inlet manifold and injector set up or have you fabricated something? If you are using the Mpi inlet do you plan on using this for the turbo as well?

I guess you might have to do quite a bit of tuning of the injection timing map due to differences in the camshaft and injection set up that Paul used to develop the initial map.


Hello,

I've been meaning to pm you to see how you were getting on. Where have you got to in the build?

Yeah, it is the basic MPi setup so original manifold and injectors. For the turbo setup I am fabricating my own inlet (and exhaust) manifold.

The quick bit of investigation i've done certainly shows my AFR balance isn't great. So yeah I think like you say, this is due to the build differences.

1275 N/A Sprite, 998 T2 Turbo Mayfair
1275 EFi Turbo


Paul S

User Avatar

8544 Posts
Member #: 573
Formerly Axel

Podland

Please bear in mind that the injection timing map was developed in about one hour of driving on the road and was just done for safe WOT on boost.

The part throttle areas still needed some work.

Having said that, I don't think that it will be too far off. I was running the same injectors and a road cam. It's more about air and fuel speed in the manifold and port, which won't be much different.

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."


dan187

User Avatar

747 Posts
Member #: 6724
Post Whore

Wootton Bassett

Yeah no worries, sorry Paul wasn't meant to sound negative about your efforts. They are much appreciated.

So Had another 30 minute play this evening. Originally I'd taken 360 degrees out of Paul's injection timing map (to bring the figures circa 140 degrees), I guess due to cam timing. Then to get the cylinders even i've put -20 across the map, which seems to have balanced them.

Accel Enrichment appears to be a little bit off, the outers lean off so I'll have a look at that over the weekend. Otherwise it sounds lovely. I'll get busy over the weekend and book an MOT for next week (I have the week off too so should hopefully gets lots done).

1275 N/A Sprite, 998 T2 Turbo Mayfair
1275 EFi Turbo


minivan63

113 Posts
Member #: 9645
Advanced Member

South Wales

Not really made much progress, been tied up with work over the last 6 months. I have all the bits ready for a NA setup like you have got running though.

Problem is that I have to do quite a bit of welding on my shell first before I can get onto the exciting bit :( I hope to have this done by the end of the year as have a bit more time now.

I'm pleased you got it running from Paul's map with little effort (thanks again Paul for your work!) there is some hope for me yet!


Paul S

User Avatar

8544 Posts
Member #: 573
Formerly Axel

Podland


On 27th Jun, 2013 dan187 said:

So Had another 30 minute play this evening. Originally I'd taken 360 degrees out of Paul's injection timing map (to bring the figures circa 140 degrees), I guess due to cam timing. Then to get the cylinders even i've put -20 across the map, which seems to have balanced them.



I'm not sure which .msq you started with, but timing figures of around 140 degrees is what I was using with the MS2. The MS3 values should be 360 degrees higher as it uses a different TDC datum.

On our 1293, I put the cam sensor trigger in the wrong place and had to revert to the MS2 values to compensate.

The fact that your engine idles fine at circa 140 degrees could be due to one of two reasons. Firstly you could have reversed the connections at the injectors or secondly (and most probable) the timing of the MPi cam sensor is 180 degrees out from the preferred MS3 setting.

Obviously, the wasted spark coil doesn't care that its signals are 360 degrees out.

Whatever has happened is of no consequence though as it is all working fine. However anyone reading this and using the sensor setup in the MS3 manual should expect to have to use injection timing circa 500 degrees.

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."


dan187

User Avatar

747 Posts
Member #: 6724
Post Whore

Wootton Bassett

MOT passed today with flying colours. I've been doing a little tuning on the way to and from test station (still need car tax, and the tracking needs sorting before i do some serious miles).

I purchased the Megalog Viewer license to gain access to the 3D scatter plots (saves me writing something myself). Very useful plots and is helping greatly with working out what needs adjusting.

Below image shows: on the left the outer cyl AFR and the right shows the difference between inner and outer AFR (positive means outers are leaner).

As you can see some larger differences at the top end and a bit near cruise.

edit: the image is detailed so click to load it in a new tab and see the full detail

Edited by dan187 on 1st Jul, 2013.

1275 N/A Sprite, 998 T2 Turbo Mayfair
1275 EFi Turbo


Paul S

User Avatar

8544 Posts
Member #: 573
Formerly Axel

Podland

That's a good way of looking at the results.

A bit more fuel and less injection timing at the top end needed before you push it too hard.

EDIT: Scratch that. Thought I was looking at inner AFR. Fueling looks close, just need to get more into the outer cylinders at the top end by delaying the injection pulse.

Edited by Paul S on 3rd Jul, 2013.

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."


dan187

User Avatar

747 Posts
Member #: 6724
Post Whore

Wootton Bassett

Yeah, tah. I have been working on this as well as fixing some general bugs. Latest plot below...

Again, on the left the outer cyl AFR and the right shows the difference between inner and outer AFR (positive means outers are leaner i.e. take this value away from the left plot to get inner AFR).

Getting closer, it's difficult to get the top end correct. Plus I don't want to put the engine up there for long as it's freshly built. Drives very nicely now. The gear and final drive choice is good for round town. Very pleased.

Edited by dan187 on 4th Jul, 2013.

1275 N/A Sprite, 998 T2 Turbo Mayfair
1275 EFi Turbo


Paul S

User Avatar

8544 Posts
Member #: 573
Formerly Axel

Podland

How are your pulse widths looking?

The holy grail of 5-port injection is to get enough fuel into the outer cylinders, but you need a lot of injector capacity and small pulse widths.

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."


dan187

User Avatar

747 Posts
Member #: 6724
Post Whore

Wootton Bassett

Sorry not had much chance to look at this over the last week. Plus was chasing my tail a little... the timing and ve kept needing adjustment. Turned out my vac pipe into the MS was slowly splitting and caused me to constantly adjust.

Fixed that and I've got it running fairly well except for a few issues:

as mentioned above my timing figures are in the region of 100 degrees. at full throttle and over 3500 rpm the inners go very lean. I have tried removing more timing but it seems to make no difference. in this region the timing needs to be over 100 degrees lower than idle. hence i end up with a negative figure. Is that an unexpected value for the MS3? should i enter, for example, 710 degrees rather than -10? (i just removed timing using the take away function in the table and tunerstudio made it negative, rather than rolling round to 720).
I'll check my injector wiring to make sure i haven't put them backwards. May be worth swapping them anyway though and get values near 500?

slightly less critical items:
when the cooling fans kick in the AFRs go out of balance by around 0.9 (noticeable at idle) not a huge problem but does make it struggle a little.

During warm up the timing needs to be different. is there anything to allow for this?

Paul, i'll post up a scatter plot of pulse width when i'm home.

1275 N/A Sprite, 998 T2 Turbo Mayfair
1275 EFi Turbo


Paul S

User Avatar

8544 Posts
Member #: 573
Formerly Axel

Podland

Negative numbers are fine. I think that it will accept from -360 to +720 degrees.

Are you sure that it is the inners going lean at higher revs. I would expect it to be the outers as shown in the plot above.

I suspect that you do not have enough injector capacity to get richer than stoich at the top end. I was using four MPi injectors and was topping out at around 120hp. That's why I was asking about injector pulse widths. Any higher than 10ms and you will struggle.

As for the problem with the cooling fan kicking in, then assuming that it's not a instrument problem due to grounding differences, you probably need to increase the voltage compensation.

As it stands, you will find it impossible to get even AFRs on a cold engine if the setup is optimised for a warm engine. Wall wetting is far higher on a cold port. Unless someone can write the code for warm-up injection timing correction, then we have to live with that.

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."


dan187

User Avatar

747 Posts
Member #: 6724
Post Whore

Wootton Bassett

Sorry, yes you are correct, i've written the wrong thing... inners are richer.

PW below. at 4000 rpm it tries to use more than 10 ms of pw. So perhaps you are correct. I'm expecting the engine to be 60 - 70 bhp at the moment (N/A) so this could be on the limit

Fixing the broken vac pipe has dramatically changed the results (see below). This time i've just plotted AFR1 (outers) and AFR2 (inners).

A fair amount of retuning is required




1275 N/A Sprite, 998 T2 Turbo Mayfair
1275 EFi Turbo


jbelanger

1229 Posts
Member #: 831
Post Whore

Montreal, Canada

Paul is right about the range of -360 to 720 degrees. And never go from 0 to 720 in your table; even though they represent the exact same engine position, the table interpolation will completely mess up your timing between these two.

For example if you have a cell with 10 degrees and a following one with 710 degrees, you would expect that the advance in the middle of those cells to be 0 or 720 degrees but the math will give 360 degrees which is completely wrong.

By the way, the same would apply to -360 and +360 degrees. That's the reason the whole possible range is more than 360 degrees: you never want to have a discontinuity in your advance values.

Jean

http://www.jbperf.com/


dan187

User Avatar

747 Posts
Member #: 6724
Post Whore

Wootton Bassett

The tune is getting there. I'll post up a pic later but it's quite variable with temperature. Especially in this heat start of day vs end gives a change to afr1 vs afr2.

I think I'll try changing it to fire the same injector twice once I get chance in the hope to get it much more robust. I'd love to know how a well tuned carb would compare. Did you post that info somewhere Paul?

If I fired the same injector from two outputs do I need to put in some diodes after the FET?

1275 N/A Sprite, 998 T2 Turbo Mayfair
1275 EFi Turbo

Home > MS Trials & Testing > MS3X MPi Trials
Users viewing this thread: none. (+ 1 Guests)   Next ->
To post messages you must be logged in!
Username: Password:
Page: