What would be necessary to get ion sensing? (MS3X)

[OliverH]

Ion sensing would be a blast to get it working with MS3/MS3X.
The latest betas support a knock window, which should be the basic to measure on the right crank angle.
If a COP/ CNP supplies a 0-5V signal for ion current values can this be used by MS3 direct or be used with mega log viewer/ Tunerstudio (auto tune) for trimming ignition?

[jsmcortina]

Once you have the ion current as a 0-5V then the fun begins. You have to process the signal to extract the two curves and then find the peak of the second one (PPP curve.) You would need some very tight code to do that efficiently and now slow down the CPU. It might be better handled by a dedicated CPU so the MS3 can focus on running the engine.

James

[Keithg]

If you want a tester, I have ion current in that format and would be willing to test it.

Keith

[OliverH]

@Keith

Is that your Saab DI setup? What's your current setup? DI direct connected to the MS or any conditioner in between?

[Keithg]

Yes, it is my Saab DI. I have 36-1 crank with a a per rev cam sensor. I am triggering with the leds on the v3 board.

[OliverH]

@Keith
This is your current setup?
http://www.msruns.com/viewtopic.php?f=100&t=22579

So you run the DI as a COP Module and don't use the feedbacks from the DI (ion current). And because of that you need the cam sensor. I've got a Trionic 7 DI here as a spare part to my Saab 95 Aero and thougt about using it for the Type4 engine in my VW Camper (DI beside and plug wired to the plugs).

Maybe we need something to measure. There should be flame front (max peak and degree crank) and the second wave which is PPP (max peak and degree crank (estimated at 15-17 degree ATDC (optimized)). Must this be done by a "conditioner" or can MS3X/MS3 handle this. Some code should be around from a finish guy how did some analyzing:
http://vvnet.fi/ville/ion/ion.html
he states that he gives out the code via mail, don't ask me why it's not public.

But can we first discuss what's needed to analyze the ion current and who is responsible to calculate the two curves (find max peak values and crank degree) and how this could be used as:
- a knock sensor or input for a new function (wish not reality) in Tunerstudio to autotune ignition map wit a wizzard
- to identify cyl. number one on a 4 Zylinder engine (Straight/ in line or Boxer).

[jsmcortina]

I took a measurement long ago.
http://www.jsm-net.demon.co.uk/ion/index.html
Analysing the curve in realtime is the tough task.

James

[Keithg]

@Keith
This is your current setup?

No, that is Raul's. We have discussed his implementation and disagree on how best to drive the DI cassette. My experience is that it must be driven sequentially to function at normal engine speeds under boost. Running it in wasted spark mode works fine at idle and during cranking due to the long charge times that affords the capacitor, but at speed and when strong spark is needed (under boost) my experience is that wasted spark does not provide adequate spark.

My old time 'it runs' is here:
http://www.msruns.com/viewtopic.php?f=100&t=8618
I have not updated that in a long time. I need to post something similar on the MS/E web site.

My site is here:
https://sites.google.com/site/ys3al35l/

I have a '91 900 SPG (Aero in europe). It has the Mitsu TE05-12B turbo and I am running the stock Intercooler. It is running a black (T7) Di cassette. When I have run a T5 cassette, I definitely get more power (boost maintained w/o knock) and smoother off idle performance (high throttle angle and low engine speed) from the T7 cassette compared to the T5 cassette.

It is running MS3. I have a 36-1 pattern machined on the back side of my flywheel with a Honeywell GT-1 hall sensor and a home made hall based cam sensor. The DI unit is driven sequentially by 4 2n2222 transistors (on the leds plus the 4th channel I built in the proto area). The sequential fuel drivers are on a board I built and is running 4 VND5N07 fets. I have hacked the APC system so that the MS3 drives the output circuit and the knock sense triggers the MS3 knock input. Idle control is stock PWM valve, A/C idle up is implemented as is the ELCD (Evaporative Emissions valve). I have also implemented gear based boost control by having the stock wheel speed sensor wired in as well. I have run MSnS/E, MS2/E, MS2/E sequential and MS3 on this car at different times. MS3 affords the I/O I need to run it as I have it now.

In short, I have replaced all the functions of the OEM system and added a few that MS3 affords me the capacity to do.

I have asked James about the 'combustion sense' to replace the cam sensor, but that is not super easy as it would require 2 inputs and not many are running a DI cassette, so I am not pressing for it. It would be cool to get rid of the cam sensor, but not really needed as my current solution has worked for 6 years.

Keith

[OliverH]

Hi Keith.

Sorry for mixing up this two setups.
Do you know which kind of signal is on the knock detect pin of the DI? Is this a still interpreted signal?

By the way. The 900 classic 16s wasn't allowed to be called Aero at that time in most countries of Europe, as Opel had a very unknown Kadett convertible called aero years ago and hold the name.

Oliver

[Keithg]

Hi Keith.

Sorry for mixing up this two setups.
Do you know which kind of signal is on the knock detect pin of the DI? Is this a still interpreted signal?

By the way. The 900 classic 16s wasn't allowed to be called Aero at that time in most countries of Europe, as Opel had a very unknown Kadett convertible called aero years ago and hold the name.

Oliver

The knock detect is a 0-5V analog signal as far as I know. I have not logged it, but have gotten this info from a few sources.

T5Suite Manual:

During engine operations the Ignition cassette continuously monitors the ion currents in the cylinders
and sends a signal to the Trionic ECU on pin 44, in the event of knocking. The logic for this function
rests solely in the ignition cassette and is adaptive to be able to handle disturbing fuel additives. The
Trionic ECU is well aware of which cylinder that has ignited and could hence cope with the information
fed through one pin. The signal to pin 44 and ion current in the combustion chamber is related to
each other. When this signal reaches a certain level the ECU interprets this as a knocking event and
firstly lowers the ignition advance by 1.5° on this cylinder. If the knocking is repeated the ignition
advance is lowered further by 1.5 °, up to 12°. In case of the same lowering of the ignition timing
advance in all cylinders the ECU adds a small amount of fuel to all cylinders. If knocking occurs when
the MAP is over 140 kPa the knocking is regulated by switching both fuel injection matrix and ignition
advance matrix. If this is not sufficient the charging pressure is lowered.

In the manual, there is also a plot of ionization current with a scale of 0-5

Lars Erikson's papers give a lot of information:
http://www.vehicular.isy.liu.se/Publica ... 580_LE.pdf

This paraphrases his earlier work with the Saab 4 cylinder and DI in the lab. All of his graphs have ion current listed on 0-5 scale. I wonder if this is reported by the DI cassette as 0-5V?

Keith

[karlo]

James, what circuit did you use in order to read the ion current? I've been wanting to try this myself.

Initially, I think it would be more than sufficient to process the ion current signal off line (eg. in TunerStudio or in a separate application). Doing it real time would be nice, of course, but substantially harder and most of the benefit can be had with an offline solution.

[jsmcortina]

James, what circuit did you use in order to read the ion current? I've been wanting to try this myself.

Initially we had a current mirror and a long chain of diodes. Unfortunately we didn't realise that spark is negative so we immediately zapped the current mirror.

After that I got in the car and the ground of the car was at +400V while the measurements were taken, using just a resistor for V= I.R

James

[bubba2533]

James, what circuit did you use in order to read the ion current? I've been wanting to try this myself.

Initially, I think it would be more than sufficient to process the ion current signal off line (eg. in TunerStudio or in a separate application). Doing it real time would be nice, of course, but substantially harder and most of the benefit can be had with an offline solution.

I like this idea. I would think your right and it would be substantially easier to log this and process the data later. It would need to log with a window and I would think take at least 50 data points. I'm not sure that would even be possible at high rpm.

[jsmcortina]

I'm not sure that would even be possible at high rpm.

Apparently at higher rpms you also find that you cannot ion-sense with inductive ignition because the discharge phase overlaps the PPP. That's why all(?) ion sense systems use CD ignition.

James

[Keithg]

James,

I don't know if that is 100% true. I would have to investigate, but think the T8 Saab/GM uses combustion sense, but it is inductive based wasted spark. I think the key is that they are only looking for knock in a small rpm window. With these engines it is a window with a center at 3100 rpm, the torque peak.

Keith

[bubba2533]

Apparently at higher rpms you also find that you cannot ion-sense with inductive ignition because the discharge phase overlaps the PPP. That's why all(?) ion sense systems use CD ignition.

James

That makes sense.

I brought this website up again to see if they list CD or Inductive and they don't. But, the charge time is 1.4ms which is really short for inductive, but also really long for CDI. So that confuses me, and I wish we could find more information on this.
Delphi Ion Sense Technology
http://delphi.com/manufacturers/auto/po ... s/ionized/

[OliverH]

Would this Italian link help to rebuild the electronic part of the Trionic?
http://www.saabwayclub.it/forum/viewtop ... 14&t=14618

Eng. google translation:
http://translate.google.com/translate?s ... 6t%3D14618

[Keithg]

Would this Italian link help to rebuild the electronic part of the Trionic?
http://www.saabwayclub.it/forum/viewtop ... 14&t=14618

I don't know as I cannot read Italian. Some of it is a discussion
About the different cassettes. Also there are some block diagrams of the circuit showing function.I did not see anything which would help with a diy rebuild.

Keith

[OliverH]

I also saw this abstracted schematics. Maybe someone could define the bits and pieces on this layout or maybe it's possible to read it on the original PCB inside the DI cassette.