Sequential Siamese VE Tables

Main site: http://jbperf.com/index.html
nedlohd
Posts: 14
Joined: Sat Mar 24, 2018 6:54 pm

Sequential Siamese VE Tables

Post by nedlohd » Sat Mar 24, 2018 9:07 pm

Hi Jean
I have already posted this on the MSExtra forum but thought I might post it here as well in case you saw it first as I'm sure you can give me a definitive answer.

I read that when running the MS2 Sequential Siamese Code it is possible to change fueling individually for the Outer and Inner Cylinders.
I do understand that Injection Timing Table 1 applies to the Outer Cylinders and Injection Timing Table 2 applies to the Inner Cylinders.
However, as I have been unable to find much detailed info on setting up Full Sequential Siamese in Tunerstudio, my assumptions are as follows.

Assumption 1: When running Dual Tables, VE Table 1 applies to the Outer Cylinders and VE Table 2 applies to Inner Cylinders?
Assumption 2: When running Dual Tables, AFR Table 1 applies to the Outer Cylinders and AFR Table 2 applies to the Inner Cylinders?
Assumption 3: VE trim Table 1 applies to the Outer Cylinders and VE Trim Table 2 applies to Inner Cylinders?
Assumption 4: VE Trim Tables 1 and 2 are used to tune for the Outer/Inner Cylinder imbalance? ie Set VE Trim Table 1 to 100% across all cells then tune VE Trim Table 2 to achieve balanced fueling?

Can anyone please confirm or otherwise correct my assumptions?
Any help would be greatly appreciated.
Regards
Dave Holden
Adelaide, Australia.

1992 Mini Cooper. Blow through supercharger setup. MS2 V3.57. Wasted spark.

jbelanger
Posts: 3860
Joined: Sat Oct 03, 2009 12:24 pm
Location: Quebec, Canada
Contact:

Re: Sequential Siamese VE Tables

Post by jbelanger » Sun Mar 25, 2018 12:27 pm

Assumptions 1,2,3 are correct. As for 4, it is partly correct but it is a bit more complicated than that. To balance the outer/inner cylinders, you want to first make sure you have the correct injection timing which is more critical for the outer cylinders. Then you can balance through either the VE table if you're using dual tables or through the VE trim tables. There is no need to use both dual tables and trim tables because you then have redundancy.

And yes, when using trim tables you would initially set 100% across the whole table and then adjust table 2. But you can also adjust table 1 if for some reason that makes more sense with your setup or what you encounter. And since you're limited to +-12.8% of trim, you might need to use both tables if the imbalance is higher than that (lower one trim table and increase the other).

Jean
Image

Rod S
Posts: 310
Joined: Mon Sep 06, 2010 4:03 am
Location: Rural Suffolk, England, UK

Re: Sequential Siamese VE Tables

Post by Rod S » Mon Mar 26, 2018 5:49 am

I haven't seen your post on the MS-E forum - I do check occasionally for anything related to the siamese code as I use and it's always useful to see what others are up to - but just to add to what Jean has said, of the people I know who are running the siamese code, I'm not aware of any who are using the trim table option.

That is not to say there is no-one, just none among the people I know, even though it is a valid option.
I just use "dual" tables, one VE table for the outers and one for the inners.

As Jean says though, it is absolutely essential to get the injection timing spot on first before using any of the tables that vary the amount of fuel, if it's not going through the expected/intended inlet valve in the first place, no messing about with individual VE tables or trim tables will fix it.

There is a third alternate which I have seen used successfully - but I don't agree with it for a variety of reasons - which is to have just the usual single VE table and deliberately set INJ timing 2 (the inner cylinder) slightly late so some of that pulse is dragged (or pushed) into the outer cylinder during the valve overlap period.

One final thing re. limited documentation, if you weren't aware, a lot of the development of this part of the code was done in response from requests to Jean on the turbominis forum (turbominis dot co dot uk slash forums slash index dot php) so a lot of the info/knowledge/experience is scattered about on there.

nedlohd
Posts: 14
Joined: Sat Mar 24, 2018 6:54 pm

Re: Sequential Siamese VE Tables

Post by nedlohd » Mon Mar 26, 2018 8:25 am

Many thanks to both of you Jean and Rod.
I have spent hours trolling about on the MSExtra forum and found conflicting information.
For example there is one post which states that "inj timing 1 happens first on the inner cylinder 2 (or 3), not cylinder 1 (or 4) followed in quick succession by inj timing 2 on the outer cylinder (1 or 4). So it appears reversed to normal where anything timing 1 would be cylinder 1. It isn't the way you would expect." http://www.msextra.com/forums/viewtopic ... 01&t=65298
This is at odds with the answer you gave me Jean. Bit confusing for noobs like me so it's good to hear it all straight from the horses mouth.

I do however have one other question before I stop bothering you.
In the above mentioned post it also states "the injection timing figures are relative to TDC of the actual cylinder when the inlet valve is beginning to open. Note, it's relative to the individual cylinders, not just TDC on No 1 like the normal code for an 8 port engine."

Jean, the document on your website states "For sequential siamese mode, the timing is with respect to TDC on the intake stroke which is the start of the injection window for the outer cylinders."

My question: Is Injector Timing 2 relative to TDC Intake on the Outer Cylinders? Or is Injector Timing 2 relative to TDC Intake on the Inner Cylinders?

Thanks so much again for taking time to respond.
Dave

Rod S
Posts: 310
Joined: Mon Sep 06, 2010 4:03 am
Location: Rural Suffolk, England, UK

Re: Sequential Siamese VE Tables

Post by Rod S » Mon Mar 26, 2018 8:58 am

I think I probably confused things a bit in the answer I gave initially on the MS-E form (I'm "Retired" on MS-E) and trying to explain it better later didn't work (although he got there in the end...).

What I was trying to say is the first injection event, on a single channel of the MS2, occurs on the inner cylinder. So the two you see in quick sequence is 2 then 1 (or 3 then 4). But the tables are still 1 (outer) and 2 (inner), they are just in the reverse order to what actually happens.

The answer is in the scope screenshot I put in that thread.

The timing values which you would normally start with as something like the 45 and 90 BTDC in the scope screenshot (depending how close your injectors are to the valves), are relative to the inner and outer cylinders independently (which is different to "normal" things in MS2 where they are all relative to No.1).

It is doubly confusing because when you first select the siamese sequential mode in TS with a new firmware install or new project, the default values for fixed timings come up as 90 / 270 - but this is just a hangover from the original code as Jean wrote it for an attempt at making it work on an A series in semi-sequential, no cam sensor, and everything relative to No1 TDC.

Once you select siamese sequential (proper sequential) the injection timing figures, whether you choose fixed or tables (I recommend tables) become relative to their individual cylinders, table 2 being inner but happens first. I even have my three AFR gauges, logs, everything in the order 2, 1, common to avoid confusing myself....

EDIT - Note, that statement applies to the "normal" siamese sequential mode, if you dare to select the "hybrid" mode (which I currently use), you get a third injection timing table, a third VE table, and a new set of rules.

With Jean's agreement, I did start a write-up of all of this, because it has never been fully documented, but, as usual, life got in the way....

nedlohd
Posts: 14
Joined: Sat Mar 24, 2018 6:54 pm

Re: Sequential Siamese VE Tables

Post by nedlohd » Mon Mar 26, 2018 7:03 pm

Sorry Rod.
I'm still confused as to whether Injector Timing 2 is relative to TDC Intake on the Outer Cylinders or the Inner Cylinders?
Jean's writeup seems to point to all Cylinder timing being relative to TDC Intake on Outer Cylinders only. Is my understanding of Jean's statement correct or not?

I am going down the end of pulse path.
Have worked out the timing with transient time included as a start point. (-175deg End of Pulse, relative to individual cylinder)

Dave

jbelanger
Posts: 3860
Joined: Sat Oct 03, 2009 12:24 pm
Location: Quebec, Canada
Contact:

Re: Sequential Siamese VE Tables

Post by jbelanger » Mon Mar 26, 2018 9:21 pm

Timing is with respect to each cylinder TDC, i.e., the pulse for cylinder 1 is timed with respect to TDC cylinder 1 on the intake stroke, the pulse for cylinder 2 is timed with respect to TDC cylinder 2, and so on. So timing 1 is for outer cylinders and each pulse is timed with respect to its own TDC and timing 2 is for inner cylinders and each pulse is time with respect to its own cylinder. If you choose to use a single timing value, this still applies but you have a single timing number shared by all cylinders (still with respect to their own TDC).

EDIT: ignore this paragraph; due to a brain fart I inverted the sign of the timing value. And using -175deg seems like an extreme number. For this to be even close to correct you would have to have a fuel travel time equivalent to about a full revolution since the intake stroke is from 0 to 180 degrees (ignoring the actual cam timing and simply using the piston travel) and using end of pulse means the pulse ends at your chosen timing value.

Jean
Image

nedlohd
Posts: 14
Joined: Sat Mar 24, 2018 6:54 pm

Re: Sequential Siamese VE Tables

Post by nedlohd » Mon Mar 26, 2018 11:38 pm

Thanks Jean.
Timing concept perfectly understood now. :idea:

With regard to my timing, I came at that figure (175deg ATDC) as follows.
Intake valve closing angle = 240deg ATDC Intake stroke
Transient time = 64.2deg
End of Pulse angle = -240 + 64.2 = 175.8deg ATDC
I was going to try the Inner Cylinders at 90deg ATDC

Taking another look at it after what you've said though, I can see why 175deg is a little extreme.
I might try advancing it to 145deg ATDC and leave the Inners at 90deg ATDC for starters.

Cam is MD276. Intake max lift at 106deg. Duration 270deg. (29deg BTDC to 241deg ATDC).
Outer Cylinder End of Pulse timing of 145deg ATDC is aimed at roughly 0.05" from Outer Valve closing.
Inner Cylinder End of Pulse timing of 90deg ATDC is aimed at point where Outer Cylinder Valve is leaving it's seat.

Am I barking up the right tree Jean?
As you have no doubt gathered I am flying by the seat of my pants trying to gather enough information to make this work.
Still, I like a challenge and are doing it more for personal interest than anything.

Thanks again
Dave

jbelanger
Posts: 3860
Joined: Sat Oct 03, 2009 12:24 pm
Location: Quebec, Canada
Contact:

Re: Sequential Siamese VE Tables

Post by jbelanger » Tue Mar 27, 2018 12:20 am

I'm sorry my previous message was wrong because I had a brain fart. I inverted the sign on the timing values so the intake stroke should be 0 to -180 degrees so my comment about -175 was wrong.

Having said that, my comment about -175 being extreme is still valid. Your conclusion would make sense if your transient time was valid and you could use the intake event for the full advertised duration. I don't think that it is a safe assumption and you will want to keep a larger margin. Also I don't know how you came up with your transient time value but from what others have found, it seems to be much more than that. Moreover, it will vary with load and RPM so you should plan from the start to use the timing table. As for what the actual value is, you will have to get your answer from actual users because while I know all the theory I haven't actually tuned a 5-port engine and I don't remember from the top of my head what was reported nor do I have data available (at least not without searching forums). You may want to have a look at the turbominis forum.

Even with data from others, you will still need to spend some time tuning the injection timing. Unless you have the exact same physical setup as someone else, you will need to find out what the actual transit time is for your setup and that will depend on the injector location/orientation, the injector type/spray pattern, the geometry of your intake manifold and a bunch of other factors. So any timing value you start with will only serve to get the engine started (either through some complex computation, taking someone else data or using some WAG data) and then you'll have to tune from there using the actual engine data (mostly AFR from the inner and outer cylinders).

Jean
Image

Rod S
Posts: 310
Joined: Mon Sep 06, 2010 4:03 am
Location: Rural Suffolk, England, UK

Re: Sequential Siamese VE Tables

Post by Rod S » Tue Mar 27, 2018 1:33 am

Jean's description of timing probably explains it better than I have done, if I ever finish my write-up I shall copy those words exactly.

When you refer to Jean's write-up, assuming you mean this one http://jbperf.com/sequential/index.html it covers all the options so you need to pick out the specific bits for siamese sequential (and ignore any reference to siamese semi-sequential).

I also agree a timing figure of -175 seems way out.

The figures you enter, whether it be into a single value field or any of the cells in a table are all before TDC on the intake stroke. Even if you use an end of pulse setup (I don't) a negative number is after the valve has already opened.

Even with the timing figures for your cam, what you have to bear in mind is your injector(s) must be sized large enough that the individual pulses are less than 25% duty cycle or, put another way, equate to less than 180 degrees of engine rotation, or you have to separate them by more than 180 degrees (put larger numbers in timing 2 fixed/table) because you must not let them get too close to each other unless you are going to specifically invoke the "hybrid" (merged pulse) part of the code. If they get too close together, (a) they don't merge nicely, (b) you lose control of the dead time, and (c) you don't really know which valve is going to be favoured in the overlap period because, as you will find out once you are running this, wall wetting transients play a large part in getting these setups right.

Finally I concur totally with Jean, every engine is different for all the reasons he has listed and more. The figures I started with are as per the scope shot I posted on the MS-E forum, they are positive numbers (but I use mid-pulse which Jean included in the code for ease of use with siamese, as per his write-up) and are much more positive in my tables as load/speed rises. And they are completely different to say Graham T's (user name on here and the turbominis forum) whose engine is the one closest to mine that I have been involved with.

nedlohd
Posts: 14
Joined: Sat Mar 24, 2018 6:54 pm

Re: Sequential Siamese VE Tables

Post by nedlohd » Tue Mar 27, 2018 1:57 am

Haha! No problem Jean.
I calculated my transient time using Marcel Chichak's method. http://www.starchak.ca/efi/speed.htm
While the transient time varies about 8.9ms depending on revs, the corresponding angle only varies by about 0.3deg due to the increase of air velocity with increasing revs.

At this stage I am putting together as much as I can to achieve a start and idle.
Then, hopefully, I can start tuning.
The plan is to tune using two WB sensors (Outer and Inner exhaust tubes), then run on one sensor behind the collector.

Dave

nedlohd
Posts: 14
Joined: Sat Mar 24, 2018 6:54 pm

Re: Sequential Siamese VE Tables

Post by nedlohd » Tue Mar 27, 2018 2:08 am

Thanks Rod.
One thing though. You say that the timing values must be BTDC even if End of Pulse is selected because ATDC the valve has already opened.
I understand that, however shouldn't the actual End of Pulse line up more with when the valve is closing? Or do we need to get the charge in before the valve reaches max opening?
I don't mind trying it but I'm one who needs to understand why.
Sorry about that.
Dave

jbelanger
Posts: 3860
Joined: Sat Oct 03, 2009 12:24 pm
Location: Quebec, Canada
Contact:

Re: Sequential Siamese VE Tables

Post by jbelanger » Tue Mar 27, 2018 2:21 am

Unfortunately Marcel's method is flawed and has not been proven by actual measurement. All the actual timing values from real working setups are much larger than that. I don't know what exactly is not correct but there are more factors that come into play than the simple model he uses.

That is why Rod mentions BTDC values. You want the fuel to enter the open valve but for that to happen you need to start the injection pulse much earlier that when the valve starts to open. It is possible that with the end-of-pulse timing you chose (as opposed to the middle-of-pulse most setups use), you may need to use ATDC values at very low load RPM but you definitely will end up with mostly BTDC values.

Jean
Image

Rod S
Posts: 310
Joined: Mon Sep 06, 2010 4:03 am
Location: Rural Suffolk, England, UK

Re: Sequential Siamese VE Tables

Post by Rod S » Tue Mar 27, 2018 2:29 am

My bad wording again.

If you are going to use end of pulse then the figures can be negative (ie, ATDC) but -175 just seems too extreme.

If you look back at the plot I posted on MS-E (I'll have to get it re-hosted to post it directly here) http://www.msextra.com/forums/download/ ... 604532d710

Although the timing figures are positive, mid pulse, the end of the second pulse is marginally after TDC on the outer cylinder. But no-where near 175 degrees.

That plot was from my actual starting figures which are still pretty close to what I use now at low load/RPM but a simulation on the Jimstim pushing the pulsewidths to there maximum before they clashed (which you do not want to happen). And if you saw my actual current tables and converted them to end of pulse the high load cells would definitely not be ATDC

And as Jean says re. transit time, whatever Marcel's logic was, it isn't what happens in practice.

nedlohd
Posts: 14
Joined: Sat Mar 24, 2018 6:54 pm

Re: Sequential Siamese VE Tables

Post by nedlohd » Tue Mar 27, 2018 4:14 am

Ok Gentlemen.
It is making sense. The fuel has to be there when the valve opens more or less.

Rod, on your plot the red zig zags represent the cylinders going between TDC and BDC, correct?
I can see now how the injection timing offset.

Keep up the good work.
Cheers
Dave

Rod S
Posts: 310
Joined: Mon Sep 06, 2010 4:03 am
Location: Rural Suffolk, England, UK

Re: Sequential Siamese VE Tables

Post by Rod S » Tue Mar 27, 2018 5:51 am

Dave, yes the zigzags follow the cycles of No1 and No2 cylinders.
Ignore the numbers 3 & 4, with hindsight (as usual) it just confuses things.
If I ever get to do the write-up I will make it much clearer.
So ignoring the numbers 3 & 4, the phase marker on my JimStim is at 120 BTDC on No1 at the commencement of its firing stroke. That is where it is recommended you physically set it on an actual engine so is also where my real cam signal is. Then just follow the zigzags from the circled (1) or (2) for Firing, Exhaust, Intake, Compression as per the letters.
The sloping lines (left) represent the pulse width being slightly smaller than the 180 degree intake window (obvious real valve opening is more than 180 but it represents what we are trying to achieve with less than 25% duty to keep the pulses from colliding) and the vertical lines plus red bars (right) represent the mid pulse timing as degrees BTDC on the induction strokes (which is where they actually are on my setup at idle).
Because its from a real msq you see the second pulse (outers) is slightly longer because the numbers in VE table 1 are slightly larger (I use 2 VE tables rather than trim tables).
Well, actually I use 3 VE tables (and 3 timing tables) because I'm now running the hydrid (merged pulse) mode above 2K RPM but let's not go there just yet.....

nedlohd
Posts: 14
Joined: Sat Mar 24, 2018 6:54 pm

Re: Sequential Siamese VE Tables

Post by nedlohd » Tue Mar 27, 2018 8:11 am

Gee thanks Rod. That is really helpful.
I have been working with a circular chart but that is way easier for me to visualise what is going on.
I don't suppose you would consider sharing your .msq or at least tables with me to use as a baseline.
End of Pulse Timing was my preferred method mainly because, in my mind at least, it seemed the easiest to tune. (End of pulse is set then any tuning only increases or decreases the PW at one end).
However, after all you've said in your previous replies, I am now leaning towards Mid Pulse Timing.

Hybrid Mode is something I will definitely approach once I have the basics up and running.

Thanks again for being so generous with your time.
Dave

nedlohd
Posts: 14
Joined: Sat Mar 24, 2018 6:54 pm

Re: Sequential Siamese VE Tables

Post by nedlohd » Tue Mar 27, 2018 8:13 am

Also, my cam signal is currently set at 80deg BTDC.
Can be adjusted if need be.
Dave

Rod S
Posts: 310
Joined: Mon Sep 06, 2010 4:03 am
Location: Rural Suffolk, England, UK

Re: Sequential Siamese VE Tables

Post by Rod S » Tue Mar 27, 2018 9:21 am

First, you definitely need your cam signal earlier than 80 degrees if you are using a traditional 36:1 or 60:1 (or similar) primary wheel with the missing tooth set at the traditional 90 BTDC - that is because the MS code (I think all versions) needs to see the second trigger just before the primary trigger reference. If you have it later than the primary trigger, everything will be 1/2 cycle out to what we would be expecting to see (and that will give you real problems as everything will end up in the inner cylinders).

Re. tables etc., give me a while, we are experiencing lots of random power cuts in my part of the UK at the moment so I'm loath to start opening too many files on my PC in case anything gets corrupted (I really need to replace the battery in my UPS, the warning light has been flashing at me for over 6 months....).

More later.

nedlohd
Posts: 14
Joined: Sat Mar 24, 2018 6:54 pm

Re: Sequential Siamese VE Tables

Post by nedlohd » Tue Mar 27, 2018 7:36 pm

Sorry Rod.
My mistake. Missing tooth is at 90deg BTDC No 1 cyl.
Cam trigger is at 120deg BTDC as per James Murray's MS3X/V3.57 Hardware Guide page 143.
I know all about blackouts. I arrived from the UK during a statewide power failure in Sep 2016.
It was a chance to see humans at their worst behavior. Scary how much we rely on it.
Dave

Post Reply