[Duk]

Next question is would it be possible to predict - up to a point - how to compensate for relatively small modifications such as the additional .5 compression on a modified chip? Meaning stock cams, etc.? Or is the only way to really know is with the car on a chassis dyno?

Most/all factory ECU tuning is pretty conservative. As long as you are using an appropriate octane fuel, 0.5 extra compression wouldn't require any ignition map changes. There would be more requirements for ignition map changes if you played with the settings of the variable valve timing.

 

[festy]

Speaking of VVT (and if you'll forgive my ignorance), how does it work?
From the few threads I've read on this forum this morning I get the impression it's load based, and either on or off - i.e. it doesn't add varying degrees of advance, it's all or nothing.
Is that the general consensus?

The reason I ask is that I've found what looks to me like the VVT actuator map - exept the "turn on" is staggered across a few engine load cells.
This could be to smooth the progression (i.e. add some hysteresis) - or it could be to gradually bring the cam advance in, rather than a savage hit?

Then again, I could be wrong as to what the map is for, but it's so smooth up the rev progression that it really does look like an ancilliary control map.

(vertical axis is formatted here as VVT actuator %)

 

[Duk]

Jim has already said that it is On or Off rather than having any sort of progressive characteristics. I think that is accurate. If you look at continuously variable cam timing engines (just look at a BA onwards Falcon 6 for us Aussies) you will see sensors to detect camshaft position and the camshafts themselves will have trigger edges on them in the same way a crank angle sensor would have. This would give the ECU closed loop control (ie feed back) to make sure the camshafts are where they are suppose to be at any given time.
The solenoid valve(s) would also be either open or closed to control the oil flow in a switched, single step system and there would be pulse width modulation of the solenoid valve in continuous VVT systems.

You have brought up an interesting point about the VVT being load switched. With an air flow meter equipped engine management system, you can't get big AFM signals (signals that come from large air flow rates) without highish engine speeds. It's easy to get high load signals from a MAP sensor or TPS based load sensing, but with AFM load sensing, it can't be done so easily. So triggering the VVT based only on the AFM's signal could be the way the Motronic 4.1 did it.
Alternatively, there is also an RPM setting inside the code and some sort of data based comparator is used.

BTW, awesome work so far digging through this stuff and sharing it with the (Alfa) world

 

[festy]

Maybe the table value is added to another parameter to determine whether to activate vvt?
So the mid-range values might turn it on depending on another condition, and the 100% cells garantee it's on in those areas?

 

[Duk]

Maybe the table value is added to another parameter to determine whether to activate vvt?
So the mid-range values might turn it on depending on another condition, and the 100% cells garantee it's on in those areas?

It sounds like you are getting to the point where you need to have every thing set up on a bench. With some means to simulate engine speed (either some sort of frequency generator that replicates the behavior of the crank angle sensors output and signal shape) or a 60-2 tone wheel attached to a variable speed drive and Bosch sensor.
Then it's an emulator in the original ECU, replace the analogue inputs (AFM, inlet air temp. and coolant temp.) with easy to adjust potentiometers. Replace any PWM or switched outputs with current limited LEDs. Then you can put a duty cycle equipped multimeter on the outputs and see what effect changing that part of the code has while simulating 'X revs' and 'Y load'.
This would help you to make absolutely sure that you are looking at things like the injector map(s), and the (probably need an oscilloscope for this one) ignition map(s).

Another thing which will be in the code is the ECU's calibration table for the analogue inputs. Nissan had dedicated calibration codes for their AFM tables and they were different for different AFM's.
The trouble is, if you start adjusting it, it would effect the out behavior of the injectors in the same way a lot of interceptor tuning devices change the main load sensor's signal.
So it may be easy to misinterpret the AFM calibration table as the main injector map. You modify this map and the injector duty cycle changes (much easier to see that than see innition timing changes when stuff is simulated on a bench).
The big problem, if that happens, is that while you may be able to achieve the desired air fuel ratios when tuning, the ignition timing would be going all over the place as a result. That is, the ignition map would be accessing area's that the tuner isn't intending it to access. So while you think you are pulling fuel out to correct an overly rich mixture at 1 rev/load point, you would also be causing the ECU to access a lower load ignition timing point which is typically more advanced. That extra ignition advance MAY cause engine knock or push the engine past its best ignition angle and loose a bit of torque as a result. The tuner may interpret that as a fuel deficiency.
"We leaned it out from a (typically) rich condition and it lost torque. So the engine needs to run rich." (based on readings from an AFR meter and the dyno).
But because they (the tuner) are effecting 2 things at once (injector open times and ignition timing) they could, and probably are, barking up the wrong tree about the requirements of the engine.

 

[Jim K.]

Consider the fact that during the two VVT positions, you are certainly having two entirely different ignition maps, one for very low compression ratio with an average cam LC=128* (!!) and one for higher CR with an average LC=97*. In my mind, switching between the two maps will have to follow the VVT solenoid operation; hmmm, I suppose this will complicate matters when attempting to map an aftermarket ecu, won't it?
Jim K.

 

[festy]

Why would you need two ignition maps?
If both the spark and VVT tables are rpm vs load, you'd only need one map wouldn't you, as the load points interpolate?

 

[Jim K.]

Yes, but the base map depends on some basic engine constants like chamber shape and CR and this, is separated from any load signal like afm/map. Unless you have an active system always searching for the max advance value (brink of detonation) you must make allowance for CR in your base map. Visualize this: two identical load conditions, one with 8:1CR and one with 10:1CR. Wouldn't they warrant two different maps (remember, no knock sensor magic)?
Jim K.

 

[Duk]

Yes, but the base map depends on some basic engine constants like chamber shape and CR and this, is separated from any load signal like afm/map. Unless you have an active system always searching for the max advance value (brink of detonation) you must make allowance for CR in your base map. Visualize this: two identical load conditions, one with 8:1CR and one with 10:1CR. Wouldn't they warrant two different maps (remember, no knock sensor magic)?
Jim K.

Peak torque isn't always developed just prior to detonation. At part throttle, where the effective/dynamic compression ratio is much less than the static compression ratio (regardless of cam timing), the engine will require more ignition advance to achieve best performance (Jim, I know that you know that). But it is quite possible to advance the ignition timing even further and have the engine start to loose torque/power without encountering detonation. So this whole theory that an engine management system is constantly trying to advance the ignition timing to the point of knock and then back it off is both misleading to the average Joe and is potentially giving the ECU manufacturer's a bad reputation and technically to much credit.
Maybe/MAYBE(!!) if we used an engine with high static compression ratio and excellent breathing characteristics (and so an high dynamic compression ratio), run on pissy low octane fuel, but really, if you own a genuine achiever in the torque per litre ratio, what are you doing running on low octane fuel?
IE: A seriously octane limited engine, where in any given rev and load point, the engine could achieve better performance if the engine had a higher octane fuel and ignition mapping to get the most from it.
With a normally aspirated road car engine running on 95-98RON fuel, that isn't going to happen at less than full throttle.

I think this idea (ECU advances until knock then backs off) is a lot like the supposed principle of the bad shift characteristics in an Alfa trans axle car. The shift sux because the 'gearbox is a long way from the gear lever'. Written in a magazine once, by someone who was trying to sound like they knew what they were talking about and repeated ever since.

 

[festy]

The base map is a 3 dimensional table, it is tied to engine load - this I'm 99.9% sure of.
But the numbers are just starting points for the ECU's calculations, they're not the final commanded timing. There's many other environmental conditions that are applied to the calculation to get the finished product - but it has to have a number to start with.

If your CR changed by a known amount based on RPM/load, your ignition map would cater for that by having suitably scaled values in those load ranges. A 2d RPM-only map couldn't do this.
Your example doesn't fit the VVT scenario though - you can't have the same RPM and load point with VVT on and off - it's one or the other.

When the ignition table was populated by the designers, they already knew which cells would and wouldn't have VVT active.

Kind of like where the ignition map I posted last night suddenly drops the advance at load point 50 throughout the rev range - that to me looks like it expects the engine characteristics to change at that point and/or another parameter to start having a significant impact on the final calculated advance.

 

[Jim K.]

I didn't say you can have both the same load and rpm with VVT on/off, that is impossible.
What I mean by 'brink of detonation' -unfortunate phrase on my part- is achieving max cylinder pressure after firing spark (which occurs somewhere ~12* ATDC) and the way this can be researched with specialized plugs like the Kistler plug/sensor. A proper ignition advance map can then be generated for various engine operating conditions, even by individuals attempting to map their engines on the brake.
Yes, all ecu ignition maps are 3D!
Jim K.

 

[Biba69]

Hey festy, are our chips about ready? I'm sure this must be an extremely brain intensive project but it sure would be nice if there was an open source program for the 75 TS ECU.

To verify, is there only one ECU model for the 75 TS? My Bosch number is 0 261 200 108 Sound about right?

 

[festy]

The '108 is the ECU that my ROM image came from.
I'm still working through all the code, I now know far more about the TS VVT and spark calculations than I ever really wanted to - but I'm still a while off having a DIY-tune platform ready.

 

[Biba69]

festy, it is great to hear that you are still at it. I was afraid that since all has been quiet on this thread you might have lost interest.

 

[InstiTS]

wow, what a fantastic discussion!

I haven't as of yet, but I do plan on reading though all the linked resources, as this is an area I have always wanted to dive into. I am nearing the end of my 'Computer Science' degree (it is a VERY broad field these days) and have found this thread very inspiring.

I currently have two TS 75s at my disposal (the term loosely used!). One ECU is already chipped with the Squadra chip + fuel quality plug changed to 97RON, the other is stock. For what it is worth I have swapped both ECUs around a few times to get a feel for the different mapping, there is most definitely a notable difference! (assuming many people don't get this opportunity to switch between, as it is a soldered EPROM)

May I ask how you both (Duk, festy) became involved/interested in such tuning in Australia?

Thanks for the awesome insight.

 

[Biba69]

InstiTS, were there any modifications to either TS, other than the Squadra chip? And did you lose anything, like low rpm torque, poor idle, or poor gas/petrol/benzina mileage?

I gather you feel the chip is worth getting?

While I would much prefer not to have to, I would be wiling to add an additive, providing I can find one that really works. Many years ago when I lived in Pasadena (about 12 miles away) a station sold 98 octane whose cost was about double the rate of the next highest octane (92). Assuming they still sell it, I don't believe I'd be willing to drive that far to fill up.

 

[festy]

97RON is pretty close to what you guys call 92 octane, you shouldn't need an octane booster to use the 97RON setting.
Your 98 would be around 102RON over here - I'm not surprised it cost twice as much!

Does the 97RON plug just add ~3 degrees of advance to the ignition?

@InstiTS - My interest comes from having recently converted an alfetta to EFI using mostly GM (holden) parts (see this thread), and having a bit of experience in embedded systems programming/design helps.
When the Motronic discussion started, I was surprised at how little knowledge of the inner workings there was in the community, so thought I'd poke around a bit.
I've been doing some 8051 dissasembly recently, so was familiar enough with the processor's instruction set to make some sense of it all.
I've never seen a TS motor or ECU, so I'm really flying blind here
If you weren't on the other side of the country, I'd be grabbing my test gear and paying you a visit!

 

[InstiTS]

InstiTS, were there any modifications to either TS, other than the Squadra chip? And did you lose anything, like low rpm torque, poor idle, or poor gas/petrol/benzina mileage?

I gather you feel the chip is worth getting?

While I would much prefer not to have to, I would be wiling to add an additive, providing I can find one that really works. Many years ago when I lived in Pasadena (about 12 miles away) a station sold 98 octane whose cost was about double the rate of the next highest octane (92). Assuming they still sell it, I don't believe I'd be willing to drive that far to fill up.

Both cars are running on 98RON octane (Australia).

Red TS has full CSC mild steel exhaust with oxygen sensor, cat delete, k&n pod (bad choice outside of 3am drives or Winter!). This is the car with the Squadra chipped ECU originally.

White TS is almost bone stock, only has the fuel quality plug set to 97RON (I later found out when stripping the interior. It explained why it went so hard compared to my slightly worked red car (I always thought it was the fact it had a second gear synchro that worked!). It also has a fairly growly aftermarket rear muffler which may have little bearing on performance...

I had a little run with a mate from a second gear pull and to make things fair (famously worn synchromesh). Both cars shifted into 3rd at almost the same time. Red car began pulling away in 3rd cleanly, it was more or less was side-by-side throughout second with the white car.

In regards to torque and economy with the Squadra chip. The white car starts up fine, idles smoothly (both with and without the chipped ECU) running on 97RON, open-loop setting (this car has a standard cat and an aftermarket muffler that looks like a Sprinter style muffler, haven't confirmed this). The red car is set to 95RON, closed-loop lambda control (yellow plug) and idles a little rough for a few seconds on cold start, nothing really noticiable. If the red car is changed to 97RON it runs like a pig until it sorts itself out. It will then start okay around 10 minutes after shut-off, leave any longer and it will forget it's parametres and idle rough again for a few minutes (maybe overfueling on tickover IMO, not sure what would cause this?)

Anyway, overall the chip increased economy. I recall getting approx 30-50 more km a tank, but this could be down to a change in driving distance, highway at the time, etc, but I DID NOT notice worsened economy. For performance, the chip greatly increases response throughout the rev-range, especially down low which I put down to better distributed torque curve. The car pulls more linearly in this fashion, and pretty much you don't have to hit 4k each time you want it to go, that said, I do miss the kick in the backside when it 'came on cam'.

I cannot explain the reasons why one car idles fine set to 97RON and the other is rough on cold start. They both don't use the lambda sensor at all on this setting, even at operating temp.. it wouldn't even be used when on tickover for that matter.

Anyway, I recommend the chip and playing with the fuel quality plug first and foremost with the TS, it will give you a noticable amount of power for a fairly small price, considering against exhausts/headwork. Swapping the factory ECU for the chipped ECU is nearly depressing, I can only drive one car with the chipped ECU fitted up to it... otherwise I miss it greatly!

If you do decide on the chip, maybe try 95 closed-loop/97 open-loop setting on your local fuel quality/intake/exhaust setup and see how it idles on tickover (keeping in mind catalytic converter life), either way the chip is well worth it! Hope this has been insightful.

 

[InstiTS]

97RON is pretty close to what you guys call 92 octane, you shouldn't need an octane booster to use the 97RON setting.
Your 98 would be around 102RON over here - I'm not surprised it cost twice as much!

Does the 97RON plug just add ~3 degrees of advance to the ignition?

@InstiTS - My interest comes from having recently converted an alfetta to EFI using mostly GM (holden) parts (see this thread), and having a bit of experience in embedded systems programming/design helps.
When the Motronic discussion started, I was surprised at how little knowledge of the inner workings there was in the community, so thought I'd poke around a bit.
I've been doing some 8051 dissasembly recently, so was familiar enough with the processor's instruction set to make some sense of it all.
I've never seen a TS motor or ECU, so I'm really flying blind here
If you weren't on the other side of the country, I'd be grabbing my test gear and paying you a visit!

I believe the RON setting is a manual adjustment that would otherwise be handled by the CPS/knock sensor... so yes, small increments in advance would sound about right. Reference for those interested here: Craig's Place, Fuel Quality Plug (S30)

Ah yes, thought that might be the case. Embedded Systems is a whole other area, crossing over with Electrical Engineering, etc.
Thanks for the link, I will have a look through that thread tonight. Looks interesting!

Ha, yes. Was good to see a couple of other Aussies on here attempting to understand a 20 year old piece of PCB!

 

[festy]

so yes, small increments in advance would sound about right. Reference for those interested here: Craig's Place, Fuel Quality Plug (S30)

Thanks for that link - a few pieces of the puzzle just fell into place after reading that!

To answer my own question, no - the fuel quality switch doesn't add a fixed amount of timing, it's a seperate map.
The closed/open loop looks like it might really be a leaded/unleaded setting based on the presence of a lambda sensor, rather than specifically controlling closed loop fueling.
For instance, a 75 set to 95RON (unleaded/has lambda) uses a table starting around address 0x4337, or with 91RON setting 0x4301.
Change the setting to 97RON (leaded/no lambda) and the table starts around 0x4229, or 95RON (again leaded/no lambda) and it uses the table starting around 0x425F.
And it looks like if you select the 164 personality over the 75, there's a different set of tables called again - so there's 8 ignition tables so far...