Raising Boost ...

[Dandu 0]

Hi mates,

 

does anyone feel familiar with the following maps (Motronic MP3.2 belonging to a XU10J2TE, 2L Turbo):

 

I suppose, they are both important for setting up the wastegate duty cycle. But I can't figure out the need to set up 2 wastegate duty cycle maps. One for detonation? The other for no detonations coming up? One for premium unleaded fuel (95)? The other for super unleaded fuel (98)? My assumptions are probably mistaken. Does anyone have a better idea?

 

Cheers Dandu

Edited June 2, 2011 by Dandu

[rallysteve 22]

Not hugely familiar with the management used in these engines and whether they use the electronic speed sensor, but I know saabs and some other turbo fwd cars back off the boost at low road speed (i.e in first gear) to limit wheelspin when pulling away.

[Dandu 0]

Hey rallysteve,

 

I appreciate your comment, but I'm pretty sure that my ECU does not evaluate the VSS signal. The VSS is unplugged. Thus, I don't suppose any influence of the VSS on the maps.

 

My assumptions for the need of 2 maps: knocking vs. no knocking, cold start vs. warm operating mode ...

 

Cheers Dandu

[wicked 103 3 Cars]

What makes you think that there are duty cycle maps?

I think the ECU has pressure targets and a PID kind of controller to get to that pressure.

 

As far as I know, the ECU does evaluate the VSS for the mixture control (richer at low speed and low load, to keep the engine cooler).

[Dandu 0]

Hey wicked,

 

much thanks for your comment. There have been quite a lot of discussions (here) about the necessity of a VSS. Some say it's important, others do not. I agree with the later. I don't pick up / evaluate the vehicle speed - that means, there is no VSS connected to my ECU.

 

As far as I know, turbo boost control is little bit tricky. Of course it's not that difficult to understand the basic principles. But if you want to influence boost control (raise boost) you have to crawl deeper into the burrow. Thus, my idea is the following: There is a map indicating target boost (in dependence of engine load and rpm). Secondly there is a map indicating the wastegate duty cycle (in dependence of engine load and rpm). The wastegate duty cycle map determines (for the given engine state) the amount of time the boost solenoid is supplied with current (thus releasing boost into the inlet manifold). At the same time the map determines (the remaining time of the cycle) how long the boost solenoid is without current, therefore releasing boost to the wastegate. Usually the boost solenoid is bleeding a lot of air into the manifold as soon as you push the throttle. The maps I refered to give a proper hint. The more the engine is getting speed, the smaller the values for the duty cycle (therefore forcing the solenoid to open longer in direction to the wastegaste, therefore limiting boost more and more). You could stop boost control at this moment. Following wastegate duty cycle, that's it. But actually the ECU wants to be sure the desired boost value is actually reached. That's the reason the target boost map comes into play. The ECU permanently compares actual boost with targeted boost. In case the values differ the wastegate duty cycle is adapted. As far as I know there is a map indicating the adjustments depending on the boost error (+/-).

 

In our case there are 2 maps for wastegate duty cycle and 2 for boost target, one WGDC map per one boost target map. And I'd like to know the reason, why the ECU needs 2 pair of maps? I suppose, they stay in relation with different engine states. Being cold vs. being warm, having detonations vs. no detonations etc. ... but it's only an assumption. Perhaps the boost control of the Motronic MP3.2 is a bit more complicated than depicted above. Perhaps there is an initial wastegate duty cycle map and a maximum wastegate duty cycle map (indicating the max values the duty cycles are allowed to reach after the adaptations / adjustments are made). Then I'm still in need to know the reason, why two rather similar target boost maps exist.

 

Bye Dandu

Edited June 6, 2011 by Dandu

[wicked 103 3 Cars]

Sounds like you did already a lot of research on the MP3.2 ecu... I don't have detailed knowledge of the implementation of these ecu's but only a fair understanding of how the stuff should work from control theory. I never heard of duty cycle maps in ecu's but if you say so, I'll take that for granted.

 

A possible theory, not underpinned with factual evidence :

The boost control is a closed loop control, mainly driven by the boost error. To speed up the control over changing rpm and load (resulting in different dynamics in the control loop), you can add feedforward control by presetting a duty cycle based on rpm and load. That's duty cycle map 1. Given that the dynamics of the boost control change over engine state (can generate boost faster at high rpm), I think the second map is the correction factor for the duty cycle that is multiplied by the boost error and added to duty cycle of map 1.

 

 

The fact that there are 2 boost target maps, does not imply that the duty cycle maps are 1 to 1 related. Could be a wrong assumption?

I don't know the reason of having two boost target maps. Do they differ a lot?

Edited June 6, 2011 by wicked

[wicked 103 3 Cars]

But in the end; why do you want to know about the duty cycle maps? To raise boost, you should modify the boost target, not the duty cycle maps....

[Dandu 0]

Hey wicked,

 

I'm glad that you spend so much time in helping me.

 

I agree with you. Basically I'm in need to alter the target boost maps, that could be the end of the story. But I consider it helpful if I can adjust the process of raising the boost, too. I mean, a wastegate duty cycle map determines the direct physical influence on the wastegate. It's the prime element for setting up the amount of air going to the wastegate. If WGDC and target boost are properly tuned, there is no need for taking any kind of boost error into account. I mean, changing the WDGC map hopefully reduces time the system needs to adjust the cycle, therefore I expect a better turbo response.

 

Your theory of boost control is seductive. In short it means: M1 = initial WGDC, M2 = WGDC correction. But there a some contradicting points. Both maps actually have the same x- and y-axis. I mean, the code of the Motronic (in front of every map) is telling me that the basic input (x,y) for the map is exactly the same (regarding units and values). If M2 would be a map for WGDC correction, I would have expected different input parameters - something like x = boost error, y = rpm for instance (z = value for correction). Although I've to admit, that any dependence on rpm seems to be inappropriate. Either you leave it out (2D map) or you take load into account as well (4D map?). I prefer the first idea. Additionally I expect a WGDC correction map to deliver negative values, too (in case too much boost has been recognised).

 

The boost target maps do not differ a lot. They are nearly (60%) the same. The values for low load and high rpm are different, the values of the second map are slightly higher. The maps can be found here:

 

 

What about the theory, that there is really one WDGC map for one target boost map? What could make the difference? The transmisson system (auto, manual)?

 

Thus, we've got 3 theories for the maps given at the beginning of the thread.

1. M1 = initial WGDC, M2 = max WGDC

2. M1 = initial WGDC, M2 = WGDC correction

3. M1 = WGDC 1, M2 = WGDC 2 (refering to different engine states, transmission systems etc.).

 

Cheers Dandu

Edited June 8, 2011 by Dandu

[wicked 103 3 Cars]

The 2 WGDC maps from your first post look quite different from each other an that's what makes me think that they are not just copies for different condition.

 

Acc. to my posed theory, the WGDC is calculated like this:

 

WGDC = M1(rpm, load) + M2(rpm, load)* (Boost_target(rpm,load)-actual_boost).

 

The boost error will change sign on over/underboost and adapt the WGDC accordingly.

The reason to implement it like this could be that the engine state (rpm,load) can change faster than the boost error.

 

 

Note that you always need to modify your boost target to raise boost, because the close loop system will in the end always regulate towards that value.

Altering the duty cycle maps is a bit risky if you are not 100% sure of what you're doing. Altering gains in a closed loop system can cause overshoots and instablility of the control loop.

 

 

Btw, 2 boost target maps don't need to imply that there are 2 different engine states that select both a dedicated map.

It can also be the case that the first map is the map is for instance at low IAT and the lower map for high IAT.

For the tempertures in between the values can be linear interpolated to the right value for that temperature.

 

FOR all above: this is based on reasoning, not facts or implementation knowledge.

 

Seems that you're quite far in disassembling the rom images....

I would try to alter the boost target first. Implicit the boost error increases, so the duty cycle adaption will be more agressive as well.

[Dandu 0]

Hey wicked,

 

I like your theory more and more. Much thanks again for your invaluable contribution. Unfortunately it's difficult to prove the theorys (the others as well) correctness.

 

2 different target boost maps for 2 different IATs? Could be possible. But isn't it more likely that the IAT (in front of the turbine wheel) would / should influence the target boost more accurately? Something like: final target boost (rpm, load) = initial target boost (rpm, load) + boostadjustment (iat_actual-iat_norm) [2D map]. Thus, different states in IAT result in more (than 2) different target boost adjustments.

 

The same prevails for atmospheric pressure. Something like the following could be valuable: final target boost (rpm, load) = initial target boost (rpm, load) + boostadjustment (ap_norm-ap_actual) [2D map]. Thus, resulting in a formula like: final target boost (rpm, load) = initial target boost (rpm, load) + boostadjustment (iat_actual-iat_norm) + boostadjustment (ap_norm-ap_actual) ...

 

I still prefer the idea, that both target boost maps belong to 2 different transmission systems. Unfortunately I can't figure out, why one of both might take advantage of higher target boosts in low load / high rpm conditions. It has to stay in relation with gear change. What does automatic systems do differently in detail? The question seems a bit stupid ...

 

My next step will be to alter the target boost maps only. I'll will report, how smooth / fast etc. the boost increases.

 

Cheers Dandu

Edited June 9, 2011 by Dandu

[wicked 103 3 Cars]

The usage of IAT was an example; it can be any parameter.

Some Audi's and Volvo's for instance increase the boost as function of the outside pressure, to avoid power loss in the mountains.

(This ECU does not have a sensor for that) It was to indicate that you can use 2 maps at the same time by interpolating, since you were quite biased on 2 distinctive scenario's.

 

But the boost maps can still be for 2 distincive scenarios. The ecu has 2 pins that indicate destinctive modes: Inhibit pin for automatic boxes and an airco switch.

 

If you are capable of modding the maps; change 1 map and play with these kind of inputs.

[Dandu 0]

Very interesting. But unfortunately I don't know, what an "inhibit pin for automated boxes" means. Can you describe it a little more in detail?

 

In the meantime I adjusted both target boost maps only (especially the values for high load) and made a testdrive. My changes resulted in a higher driveable boost around: 13 psi. The boost is built up "normallly": 4th gear, 2.5k-3k rpm at the beginning of acceleration ... full throttle ... it takes the system approx. 2-3 secs to reach a boost of 13 psi. While continuing acceleration the boost gauge depicts oscillations (11psi to 13psi). At 5000 rpm the engine responds rather harshly, it seems to stall. Boost stays constant, lambda values go up. It seems like the fuel cut sets in. At 5000 rpm? I suppose, the system reacts to the heap of boost errors. Is there a chance? The stalling was reproducible. Bytheway I've to admit there wasn't a lot of fuel in the tank. But it was definitely enough to ship me home (10 kilometres) without any problems.

 

In the next step I tend to alter the first WGDC map only, I'll post my results.

 

Bye Dandu

Edited June 12, 2011 by Dandu

[wicked 103 3 Cars]

Probably you run into some physical limitation:

 

- map sensor goes up to 14psi; maybe you hit the max an fuel cut is activated

- injectors go to max duty cycle

 

You can try to replace the map sensor with 3 bar version, since you seem handy with electronincs...

[wicked 103 3 Cars]

On second thought; try to increase both boost maps (with same scaling factor) and check if the behavior is similar...

[Dandu 0]

I know, how it feels, when the system tries to surpass the boost limit (1 bar, 14.5psi). The engine diagnostic light immediately comes into play, the ECU goes into limp mode ... the boost goes down (not that harshly). In my case the engine diagnostic light did not even flicker and the driving experience was rather different. It felt like revving the engine too much (resulting in fuel cut setting in). Thus, it's probably not reasonable to change the MAP sensor. Although I've already done that once in the past (without wanting to show off ... it's not that difficult).

 

I'm pretty sure, that any kind of malfunction or misuse of engine components would have been indicated by the engine diagnostic light. I know, there is another term for "engine diagnostic light", but I can't remember.

 

Can you help me with the understanding of "inhibit pin for automated boxes"? Why do you suggest to increase both boost maps with the same scaling factor (althought my alterations have been quite similar yet)?

 

Bye Dandu

Edited June 12, 2011 by Dandu

[wicked 103 3 Cars]

Altering 1 or both maps is to find out if it using a single map or both maps with interpolation....

Kind of "find the difference and color the picture"-game, like you had on primary school....

 

The INH pin is only used on automatic boxes. On manuals it's left open.

I don't know the exact use of it, but i expect it to be related with selecting the gears (putting it in 'D' or direct gear selection) or kind of kick-down functionality, but I don't if PSA implemented that... The high boost at low load is bit strange and cannot find a good reason for it...

[Dandu 0]

What? Interpolation does not stay in relation with "Interpol"? I considered it to be dangerous and secret. ... Both target boost maps are nearly the same when it comes to full load. There is actually nothing that can be interpolated.

 

Much thanks for your persistent help! Cheers Dandu

Edited June 12, 2011 by Dandu

[wicked 103 3 Cars]

Ok, I didn't realize that they were that close to each other...

 

But you raised 1 map or both?

Std it's at 0.6bar or so.. So if you modify 1 map to 0.9 bar, it will run on

a) 0.6 bar

0.75bar or so...

c) 0.9 bar

d) oscillate between 0.6 bar and 0.9 bar

 

It's way of finding out which map is used when...

 

Replacing the MAP indeed is a 5 minute job, but to get a nice running engine is second thing and you don't get it if you're not able to mod the maps....

In the past I've start digging into ECU but never spent enough time to get deep enough to mod it properly.

[Dandu 0]

I raised both target boost maps. Following the assumption that they differ only because of the applied transmission systems. You are on the right path with your strategy (in getting to know the reason these 2 target boost maps exist). But at first I'd like to neglect their difference. I'd rather concentrate on the WGDC maps now. The next software version (modded WGDC map 1) is ready for use. I suppose, I can comment on it tomorrow.

 

In think replacing the MAP sensor in order to reach boost targets, higher than 14.5 psi, is something for the future ... if any. I consider it to be rather difficult to adjust the fuel injection timing. It will be easy to find the maps, it will be easy to adjust them ... but I suppose, you have to do it right on spot. There will be a lot of different versions / variations wanting to be tested. I'm pretty sure my ECU will not survive it. It would be great if you could tune the Motronic MP3.2 live or kind of semi-live (having a laptop on the passenger seat, being connected to a special socket on the ecu, showing the maps ... every alteration is immediately transfered to the ECU). Do you know of such equipment?

 

Although it takes dozens of hours (in thinking, finding answers, modding, driving, evaluating) I'm willing to go deeper into the burrow. And you are welcome to join me. I'll post my results soon.

 

Cheers Dandu

Edited June 14, 2011 by Dandu

[Dandu 0]

There was not a lot of time to check the new software version today. But it was enough to notice a normal acceleration in 3rd, 4th ... every gear. The fuel cut is not present anymore. Thus, it seems like the first WGDC map determines the initial wastegate duty cycle - a kind of feedforward control.

 

I'll check the software more elaborately soon. In the meantime I'll celebrate the success. Much thanks for everyones help! Cheers Dandu

[wicked 103 3 Cars]

Nice!!

 

There are emulators available to replace your eeprom:

Emulators

 

Didn't open the MP3.2 yet, so I donno if they used compatible eeproms, but I wouldn't be suprised if they did.

[wicked 103 3 Cars]

This is the newest Ostrich, the 2.0 unit. It is extremely small and compact, durable, versatile, and powerful. Full 921.6k speed, from 4k (2732A) up to 512k (4mbit 29F040) and everything in between (27C128, 27C256, 27C512). Also works in conjunction with the FordEmu interface.

 

The xxx219 ecu uses the 27C256, the later version 27C512.

Which one do you use?

[Michael 0]

Hi Dandu,

 

here you can also find an Eprom Emulator which would be able to simulate 27C256 Eproms.

 

Best Regards

 

Michael

[Dandu 0]

@Michael: Much thanks for your link. The "WICE-M4 USB" is surely capable of emulating a 27C256 chip (that's the one I use). But ...

 

@wicked: I consider your emulators a little bit better. It's seems to me that the "Ostrich 2.0" has got some special advantages: 1. lower price, 2. small connector (it could be possible to connect the ostrich to the ECU without being in need to leave out the cover), 3. tracing capabilities (in conjunction with TunerProRT and so on). But ... the ostrich 2.0 is hard to get in Germany.

 

Today I got the chance to check out my software version a little more in detail. It was fascinating to watch the cars getting smaller in the mirror while pushing the throttle. In 5th gear I never had any problems with the boost pressure. But sometimes (it seemed to come from nowhere) the 3rd and the 4th gear gave me trouble. Sometimes the boost pressure did not exceed 8 psi. Strange. Probably the boost error (boost control) mechanism is still in need to be tuned. But how? It appears to me that I have to check some more new software versions, ideally with the help of an eprom emulator.

 

Cheers Dandu

Edited June 17, 2011 by Dandu

[Dandu 0]

The Ostrich 2.0 has been ordered. I'll report about its applicableness. Cheers Dandu

Edited June 21, 2011 by Dandu