RPM lags MAP by 0.4 seconds

[DHH]

The board itself is grounded to the battery via a dedicated line for sensor grounds as well as power grounds.

That's not how the sensor ground works. You do not want it to be connected to an external ground. You only connect the sensors to the sensor ground and that connects them to the board reference ground. An external connection to a ground defeats the purpose of using the board reference ground and can easily introduce noise.

Jean

But they are all tied together internally, so when you ground just one of the wires, you are grounding them all (sensor grd, spare grd, power grd) to the same location. Don't fully understand why Jim's setup would make a difference?

[turbo conversion]

TPS wiring close to injector wires could be an issue.

Also is the sensor grounding correct, to the MS sensor ground not block or head or firewall etc?

David

Dude, you have got to start reading the manual and listening to good advice and stop assuming.

If not everyone will stop trying to help you and that is not what this forum is about.



David

[racingmini_mtl]

The board itself is grounded to the battery via a dedicated line for sensor grounds as well as power grounds.

That's not how the sensor ground works. You do not want it to be connected to an external ground. You only connect the sensors to the sensor ground and that connects them to the board reference ground. An external connection to a ground defeats the purpose of using the board reference ground and can easily introduce noise.

Jean

But they are all tied together internally, so when you ground just one of the wires, you are grounding them all (sensor grd, spare grd, power grd) to the same location. Don't fully understand why Jim's setup would make a difference?

This is not a static thing. You don't want the portion of the board reserved for sensor ground affected by noisy high current which is what you will get if you connect a sensor ground pin to the battery.

Think of the whole thing as hoses carrying water with the board being a small container with many hoses on opposite sides: one side being the high current ground draining into a large pool (battery) and the other side having small hoses draining into the container (sensor return ground). If you add a big drain into the pool just besides the small hoses, that will affect the water flow from the small hoses into the container. That's not a perfect analogy but hopefully that allows you and other to visualize this better.

Jean

[Jim_Blackwood]

V3.0 board so IIRC all of the pin-outs for ground do connect to the internal ground plane of the board. Is this incorrect? Because I've researched this and never found any indication anywhere that states definitively that the sensors are on a different ground plane. If that's the case we're sort of limited just because they are tied together.

Jean is perfectly correct in his explanation, and this is why I do have an external sensor ground terminal. I have most of the ground leads tied together and feeding to my main ground for power. Then I have one lead that goes to the sensor ground post. That post also has a 12ga wire directly to the battery. So power in general, loads are carried on the main ground cable but sensors are connected directly to the battery (sink) by their own isolated lead. I do not know if this is enough to damp out the tie-in at the board's grounf plane but it was the best I could come up with.

In the case of the TPS, it had it's own ground lead coming off the main connector and that's the one I'm using. It is working. Connecting it to the external sensor ground lug causes noise spikes, maybe due to some load sharing between the two grounds.

I have not identified the source of the noise. It appears to be random.

Jim

David, I'm sorry you feel offended at my approach. Logic dictates that I analyze what is going on and use the most reasonable approach. Pulling the injectors out, removing the rails, and cobbling up some sort of pseudo in-car injector bench using the MS controller and the car fuel system to flow test the injectors in situ would be adding a tremendous workload to do something that can be more easily done in an alternative manner. I get that you are offended that I have not analyzed the DT period, but in relation to the idle PW the potential error is small and more importantly, DOES NOT CHANGE. It can therefore be tuned out, and will not induce an oscillation.

It is NOT a requirement of an EFI system to know exactly what the flow characteristics of an injector are under a specific set of conditions, to get the engine to run acceptably. It's an engine. All it wants is reasonably accurate advance and a fairly close balance of air and fuel. Carbs did that fine for decades and still do. You can do it with a squirt bottle standing and spraying fuel into the intake. All the flow numbers give you is a ratio between what the engine gets and what you see on a map. Except for DT it is all multiplication, so 1 = unity and does not change the result. DT is an adder so unity + DT. It's just that simple. If you make one multiplier larger another has to be smaller to get the same product. That has NO EFFECT on the amount of fuel delivered. So as VE and req fuel are both multipliers, for the same fuel delivery if VE goes up req fuel MUST go down.

How does that make me some kind of outcast?

Jim
(Caps added for emphasis, not intended as yelling.)

[turbo conversion]

Your grounding is not correct (among other things) but you continue to say I don't see the issue.

As I said you need to spend some time reading the manual, it will answer all your questions and doubts.

I am not offended just trying to help.

Good Luck, David

[billr]

Please, let's all take a deep breath and calm down; no need to make this experience more aggravating for the OP (or anybody else) than necessary. Now on to a couple of specific points:

"Req fuel is a multiplier. Pretty straightforward too it seems. Therefore the value just moves the end result up or down. If it is set wrong the VE map scale will be off. But I have a question: If you already know the acceptable range of VE values then can't you use the installed MS-EFI system as an injector flow bench? Set your map and work backwards from there? Adjust the req fuel until the engine runs well with those values, copy them down, and then work the injector flow number to give that req fuel? At the end you should have injector flow rates that match what you bought." The problem with that theory is that "the acceptable range of VE values" is usually unknown for any specific engine. We just know that your VE table is way off for most any engine, not what it should be. Backing into the req_fuel by setting VE might well lead you to needing a different req_fuel for each cell in the VE table. In contrast, the req_fuel is very definable. Engine displacement is fairly easy to measure precisely, and counting the number of cylinders/injectors is "child's play". Determining the stoich value for your fuel is also fairly easy, although gasoline is a blend of "different stoichs", so we have to use 14.7 as an approximation. Likewise, common E10 is even more variable, as the ethanol % isn't precisely fixed. Even so, stoich tends to be a "single value" that can be compensated for, not bouncing around like VE with MAP and rpm. Lastly, of course, it the injector flow rate. That is an effort to determine, but can be defined fairly accurately. The biggest compromise there is that no two injectors flow exactly the same. That is why injector "trims" can be set that effectively change req_fuel slightly to match each injector. I'm rambling, but point is that setting VE and horsing with req_fuel means you will end up wanting a req_fuel "table", which isn't available and would add the complexity of working one table against the other. Setting req_fuel simplifies that, requiring only one table and, maybe, some simple injector trims.

I think you are not appreciating the importance of DT. Problems with that most often appear when trying to get a stable idle... sound familiar? Your idle PW seems to be down about 2 msec. Just a couple of tenths of a msec amounts to 10% variation in the fueling, and that when compression pressures are fairly low and combustion characteristics can be poorer. I don't recall if we have talked much about your injectors (I'll go back and peek through the thread). I may want you to get specific about what they are and how they are wired so I can guess if your DT may be far off.

Finally, feel free to question what I say, as long as we keep it "friendly". The fact that I have a lot of posts does not mean I am infallible, just that I have time available to try to help...

[racingmini_mtl]

Jean is perfectly correct in his explanation, and this is why I do have an external sensor ground terminal. I have most of the ground leads tied together and feeding to my main ground for power. Then I have one lead that goes to the sensor ground post. That post also has a 12ga wire directly to the battery. So power in general, loads are carried on the main ground cable but sensors are connected directly to the battery (sink) by their own isolated lead. I do not know if this is enough to damp out the tie-in at the board's grounf plane but it was the best I could come up with.

I'm sorry but my explanation is for showing why you do not want to have a sensor ground connected to the battery. You never connect the sensors to the battery because the CPU is not at the battery. Ideally, you would have a sensor directly connected to the CPU at the 5V and ground references. That's obviously not practical so you have wires to extend those and you connect the sensor to Vref and the sensor ground. The sensor ground is not connected to anything else (battery or other grounds) because that's not the voltage reference that the CPU sees.

Even with a single ground plane on the main board what I mention above still stands because the different circuits will use different parts of the same ground plane. If you connect a battery ground (or any other ground) to the sensor ground, that becomes a path for the noisy high current components and you compromise all the sensor path.

So remove your battery ground from the sensor ground post. That is a fundamental mistake.

Jean

[EspeNS]

I guess nobody have mentioned it before, but ought to figure out the deadtime for your injectors.
You are wrong assuming deadtime have no effect on idle.
Wrong deadtime is most noticeable on idle, and low load.
Almost no injectors have exact 1ms deadtime, and can in fact differ from one injector to another in a "matched" set.
I have measured differen high and low impedance injectors, all from about 0.4ms to about 1.2 ms deadtime. Including a "matched" set of Siemens injectors with deatimes from 0.55 on the lowest to 0.7ms on the highest. Correcsions of the deadtimes made a big differece on idle. Most injectors are fortunately not that mismatched.
But the higher cc injector to cc cylinder the more accurate deadtime needs to be.
I never did get good consistent idle on engines before I startet using the deadtime constant for what it's ment for with bigger injectors.
When measuring deadtime it's important to use injector drivers similar or the same as the ecu uses, because different drivers give different deadtimes.
Example, pcb. V3.57 mainboard drivers and MS3X drivers don't have the same deadtime on the same injectors.

Wish you good luck figuring out deadtime and grounding issues
Best regards Espen

[billr]

I gotta ask... have you read through this thread? DT has been mentioned/discussed several times!

[panel]

Are these the High Current Ground: pins 15-19 on a V3.0 board that shouldn't go to the battery or ?

[racingmini_mtl]

Are these the High Current Ground: pins 15-19 on a V3.0 board that shouldn't go to the battery or ?

No. The sensor return grounds (pin 1,2,7 as per the wiring schematics in the manuals) are the ones I mean.

Jean

[panel]

No. The sensor return grounds (pin 1,2,7 as per the wiring schematics in the manuals) are the ones I mean.

Jean

Ah , thought so.....just triple checking.

[DHH]

Jean, I like your analogy. The thing I am having a problem wrapping my brain around is that it doesn't matter where the large drain is put in the container, it will still affect the small hoses the same, because they are all connected to the same container. The way I see it, is you don't want to run the sensor grounds directly (to lets say the) block and ground the ECU someplace else. That can cause ground loops. But if all the grounds are connected internally to the same bus, then it doesn't matter if all the grounds are to the same location. There will be no ground loops, because there is no "loop". The manual states "pins 1,2,7-19 are all interchangeable grounds". In my view....and keep in mind, I'm only explaining how I see things, not saying this is the right way....the important thing is not to ground the sensors and ECU at different locations.
In my setup....which is not yet complete, I am using the "sensor" ground and the "spare" grounds for the sensor grounds. The "power" grounds are all going directly to the battery. But I also don't see a problem with tying all those grounds together (which they already are, internally).

[racingmini_mtl]

Jean, I like your analogy. The thing I am having a problem wrapping my brain around is that it doesn't matter where the large drain is put in the container, it will still affect the small hoses the same, because they are all connected to the same container. The way I see it, is you don't want to run the sensor grounds directly (to lets say the) block and ground the ECU someplace else. That can cause ground loops. But if all the grounds are connected internally to the same bus, then it doesn't matter if all the grounds are to the same location. There will be no ground loops, because there is no "loop". The manual states "pins 1,2,7-19 are all interchangeable grounds". In my view....and keep in mind, I'm only explaining how I see things, not saying this is the right way....the important thing is not to ground the sensors and ECU at different locations.
In my setup....which is not yet complete, I am using the "sensor" ground and the "spare" grounds for the sensor grounds. The "power" grounds are all going directly to the battery. But I also don't see a problem with tying all those grounds together (which they already are, internally).

You're forgetting the part where I said things are no static. You have the CPU which has the ground plane as it's reference for the analog inputs (sensors) and you can have a rather direct path on the ground plane from this to the sensor return pins. Then you have the high current circuits on the other side of the board and ground plane which have a direct path to the power ground pins (pins 15-19) which does not cross the other path. And this high current path is noisy because it is high current and on and off (injection pulses, ignition pulses, PWM, ...). Also the paths have different directions.

So if you don't have wires directly from the sensors to the sensor return pins, you have the potential of having the sensor signal compromised by other current flows. And having a common ground external to the ECU means you've just compromised you sensor signals because you've eliminated the segregated paths.

Having said that, you can share the sensor return wires between several sensors because the current is very low and the signals usually change slowly.

A good example of the issue is when people connect the WBO2 ground to the sensor return with a controller that has only one ground. Since the ground also serves as ground for the sensor heater and this is a high current (1-3A) noisy (high frequency PWM) signal, that results in a completely compromised set of sensor data. And when you ground this to the battery, you often get a ground offset on the O2 signal.

Jean

[Jim_Blackwood]

Might be a good time to mellow out a bit at that.
How about If I just test DT settings from 0.4 to 2.0 and see what that does. That'd be quick and easy. (yes I realize it won't make you happy)
I don't think it'll cure the oscillations though, or even make much if any of a dent in it. But if it does I'll be the first to say I was wrong. I don't think you guys truly appreciate just how persistent those oscillations are. I can get the rest of the tune, no problem. And I think my VE values are now in the range where you guys would be OK with them. (Tell me what you are expecting to see maybe?)

I'm about done with the grounding discussion, since it now works and TPS is super clean. Hard to argue against success. But about the battery and cables. First, the battery is in the trunk, and a battery is an absolutely HUGE sink for noise. So the noise pulses come in on a long heavy cable and the signal ground comes off on another long cable, at one time it was even connected to a different post. So the cables double the signal path for the noise and put the filter (battery) half way down the line. Standard filter theory at work. But, I'm not convinced it actually works, so I'm not fighting that battle.

As for which leads to use for power grounds and which to use for signal grounds, maybe if you had the mask for the ground plane you could guess which would be better for what but I doubt that much thought went into it at the time. I remember those days. I've looked at the mask patterns. And I think they are just all tied into the same area, I know they were on the v2.2 board.

Really, the strength of these oscillations are extremely strong. It's not just a little wobble here. The engine absolutely will not run below 1200rpm. It dies because the downward excursion goes below the cranking speed of 200. I'm telling you, it is a STRONG oscillation and it comes in like a ton of bricks. No maybes, no waver, just boom, it's there. And if the engine manages not to simply die on the first couple of cycles then it very aggressively sweeps a 1000rpm+ range. It's going to take something significant to affect that.

And to me the obvious thing out of place is the location of the MAP signal in relation to those oscillations. We KNOW the MAP is going to affect the PW. We know the PW slightly leads RPM. And we know that in THIS system MAP leads PW by a significant margin. Doesn't logic lead you to think that too much delay from MAP to PW would lead to oscillation? It does me. I think it's glaringly obvious. I know you all want to exhaust every remedy that has been proven effective in the past and that's admirable. But you really don't understand what I'm dealing with here. You can see it on the datalog sure, but that doesn't really convey the absolute Robustness of the oscillations.

So I'm going to try Alpha-N. I have very little to go on in using it, but I do remember some guys were messing with it when it first came out on MS and they seemed to be having some success. That's just about all I know. But if I get it wrong it'll just run bad and it already does that so why not? On the other side of the coin, it could work. And then I just have to sort out how to tune an Alpha-N system.

Jim

[racingmini_mtl]

This is going to be my last post the the ground subject. Jim, look at the board layout and how all the high current components are located on one side of the board (the heat sink side) and where the CPU is located. So unless you provide a ground on the opposite side of the board, the high current will flow out on the DB37 on the same side and the sensor return will have a clean path to the CPU. Saying that it will go to the battery and be filtered there is nonsense because they won't be segregated since you've provided 2 paths for the high current signals that will get mixed with the sensor return at the board itself.

And can you get a clean signal with your setup? Obviously yes because that's what you have. Will that remain so for long? I don't know but if you get corrosion or a loose connection or any other normal wear and tear, you may get the noise back. Starting from the correctly routed wiring, you'll likely have a more reliable setup. And do you still need shielding with a corrected wired setup? Possibly because the car environment is harsh and there's a lot of potential interference sources.

So keep your ground wire to the sensor return if you want and continue wondering if the board design works as I say, I don't care (also remember that the v3.0 board is a 4-layer board with a 'real' ground plane while the v2.2 is a 2-layer board with all the compromises it entails). I also don't care if the board layout was intentional or not but the physics says that there is configuration that is less prone to noise and that is the one that is shown in the wiring diagrams in the manuals (the ones here; I'm not sure the Megamanual (pdf file and web site) is correct and it should be ignored anyway). And boy was that a waste of time and effort trying to educate a couple of people; I understand being skeptical but that seems over the top.

Jean

[jsmcortina]

Jim,
Unless you listen to and follow the advice given, it almost seems like you are trolling.

You seems to be desperate to focus on irrelevant things. Fix your sensor inputs and adjust your fuel and spark tables.

James

[EspeNS]

I gotta ask... have you read through this thread? DT has been mentioned/discussed several times!

Yes indeed
My first sentence was irony. Because deadtime have been ignored by the topic starter, among other basic things, the whole thread. I know, irony is hard to convey written, but I think it got noticed.
Have faught hunting idle, ground issues and noisy signals myself. Ignoring the basics in the manual don't help, and generaly ads to the problem.

[billr]

The last MSQ I see is dated 9/8, is that still current? If not, post the current MSQ.

What is the resistance of your injectors?

You have tried "fixed" spark timing, but have you verified timing really is fixed, using a strobe light? I'm wondering if EDIS is keeping timing "as commanded".

As I recall, there is a picture showing the blower on top of the intercooler "box". Are there discrete runners in that box to each cylinder port, or is it one plenum common to all ports on both heads?

The VE table in that last MSQ is still very strange, and my understanding is it was obtained using the VE "generator". After playing with that generator a bit I realize there are several "estimates" you have to enter; you can make it produce just about anything by moving those data inputs around a bit. How about posting a screen-shot of what you used for the current MSQ?

PS: Post a video of it doing the idle oscillation.

PS2: Somebody suggested disconnecting the vac to MAP sensor, making MAP "fixed" too. Has that been tried?

[Jim_Blackwood]

First thing, I agree with Jean. I do. I really do. Running your noisy current through your signal handling area is never a good idea. Almost like running your sewage through your drinking water. It would have been great if the MS wiring diagrams had included a comment about how they dealt with that. Primo wire and gas sealed connectors also are a wonderful development. And a battery is just like a huge filter capacitor in terms of noise removal, so why not use it? A couple of chokes and you'd have a Pi filter, but just a single cap is often highly effective. In terms of filtering there is no difference between a battery and a cap. Construction is almost identical. Yes it has an electrolyte instead of an insulator, but think about it. The insulator is there so you don't lose your signal to ground. And an isolated lead that carries no or minimal current is not going to bring in much if any noise. A choke on that line would get rid of anything that got by the cap (battery). That's the theory anyway, and that's why it should work. Does it? IDK, I haven't proven it. But the theory says it does. It's all mucked up by the ground loop inside the 4 layer board so who knows? In theory that's not all that much different from saying that running all currents through the same bus but only putting the noisy ones on one side and the quiet ones on the other, separated by a couple inches of conductive material is going to get you signal separation. It sort of defies logic, but at very high frequencies it would clearly work. If you really wanted to isolate the signal ground you'd have separate ground planes and use a Pi filter to connect them.

Really, I'm not intentionally trying to be difficult. I just never was very good with blinders. You guys have worked very hard to develop an approach and a system that works well and would like everyone to go down that path. I get it. It adds consistency and it makes everything easier. Right up until it doesn't, and all I'm saying is that this looks very much like one of those zingers where it doesn't. And guys, just so you know, that's where I live and work. In the realm of Outliers. Have all my life, and as such I do have a finely tuned ability to recognize the zingers when they go flying past. And I'm telling you, this has all the characteristics of a very classic and robust outlier. But we'll satisfy convention first, just to be sure. At this point there is very little left to eliminate.

I'm not ignoring anyone. But you have to choose your battles. Do you realize the list of things you have all asked me to address? It is rather extensive, and correct me if I'm wrong, except for what we're discussing now I'm pretty sure I've addressed every single one of them. As I will the DT, as I will the noise on the CLT line, and as ultimately in one way or another I will conclude the injector flow rate question. Maybe not to your complete satisfaction, that is not the goal. The goal is to eliminate variables. Please try not to get too frustrated when my approach deviates from the conventional one. I assure you it will be supported by perfectly good theory when I do otherwise I wouldn't bother. I've been doing this kind of thing for around 50 years now and I'm no novice. I have perfectly good reasons for the order in which I am dealing with these issues, but you'll note, I am not inflexible. I think I've demonstrated that.

But again, I do greatly appreciate all your help. You have all been outstanding. I like you guys. Wish we could meet in person sometime.

Jim

BTW, birthday today/tomorrow so likely an interruption in testing, but not for long.